The HI-Lebowski: a Lebowski SMD brain running a zombified Honda IMA Inverter: *a HOW-TO guide*

amberwolf

Administrator
Staff member
Joined
Aug 17, 2009
Messages
40,463
Location
Phoenix, AZ, USA, Earth, Sol, Local Bubble, Orion
"Zombified" meaning one that has had it's brain disabled...I suppose lobotomized would apply except that in this case the plan is not to actually *remove* anything from the IMA, just disable the main CPU by holding it in "reset", per this thread:
once I find enough information to locate the reset pin on the IMA's CPU. EDIT: information located, placed in post linked below.

Thread links for what the heck the Lebowski brain board actually is:
The BobC brainboard thread:
The main Lebowski Firmware / chip / manual thread:
The opensource-release / explanation thread:


Some discussion of the project, and what led me here, first. (mostly because I don't yet have the parts to begin the project itself, and much of the upcoming documentation will require having the parts in hand. EDIT: as of last week of April, all the major parts have arrived, and work is in progress, documented below.) If you like, you can skip ahead to the next posts that begin documenting the process (once they're written I'll make an index just below this paragraph).

INDEX:
--Location of Reset Pin on CPU, pinout and other chip documentation
--IMA documentation, Location of Reset Pin on CPU, CPU pinout and other chip documentation (will also include documentation from other IMA projects around the web as I locate them)
--Disassembly, basic dimensions, pics of the insides.
--Brain board wiring and setup, including the 12v DC-DC for running the IMA gate drives, connector pinouts, BobC SMDLebowski board buildup / test instructions & schematic, etc
--enclosure of the brain inside the IMA inverter
--optional mountings, setups, afterthoughts, etc.
--Initial offground testing
--Initial road testing


Base information leading up to the project, and some RAAaannntzzzZZ: :oops:

You can skip all the rest below if you like, and just follow the links above to the details of how to do this yourself. (once they're all written--right now some of those links just go to "reserved" posts).



The plan, when complete, is to run the hubmotors that drive SB Cruiser
with a pair of these, because i have yet to find anything affordable with sufficient power *and* customizability / programmability, that also has all the functions I require. Everything I *have* tried has been disappointing in performance or features at best, and extremely frustrating or even destructive, either in operation, setup, support, or in it's failure, at worst.


The main needs I have are:
-- current (torque) throttle control
-- true sinewave commutation (with FOC if possible, but I'm really after the silence)
-- startup from a complete stop by itself, under high mass (inertia) load.
-- sensorless operation available as well as sensored (for the above)
-- sensorless fallback mode for when sensors fail
-- easy-to-implement analog (proportional) electric braking control with a separate analog input just for this (no wierd setup using the throttle to control braking while holding a brake lever that's just a switch)
-- braking capable of bringing the trike to a complete stop
-- braking that will have the same negative torque capability at high speed that it does at low, basically able to (hopefully, but at least almost) skid the wheel at any speed, when max braking input is applied.
-- easy to use setup program and user interface
-- ability to read out existing settings
-- ability to change hvc, lvc, current limits, other behaviors as the design contains/allows
-- good clear documentation on how to use and setup the controller, and what all it's options are and can do and how to use them
-- good support available for when my brain explodes
-- large enough installed userbase to judge usability/reliability/suitability for my purposes from reported problems
-- low enough cost that I don't have to scrimp on everything else for months (or years) to pay for them (cuz I need two, at minimum)
-- no onboard display required to operate it
-- more stuff I can't remember right now

Other things that would be great, but not required:
-- open source design and firmware
-- ability to actually hold the trike in place with active braking force, without mechanical brakes, for some occasional momentary situations
-- ability to dump enough torque into the wheels to smoke the tires (not that I plan to do this, but having that kind of power could be useful under certain circumstances).






Don't worry, the next paragraph, while long and arduous, can be skipped entirely and you won't miss a thing, unless you, too, are frustrated with (advanced and/or reliable) controller options out there.

Everything I have and *haven't* tried is one of the following (see if you can guess which ones these are, but keep them secret so as not to spoil anyone else's fun): Some beefed-up basically "chinese junk" controller (even if some have better parts, the design is still a clone of a clone of a...of a bad design to start with). Some of *those* are even programmable in a limited way, but the better hardware ones seem not to be, generally, with a few exceptions..but even the programmable ones don't do what i want, and can't in most cases due to their designs. An apparently good (not perfect, but good) hardware design, with decent setup options, not hard to understand, but with the most impossibly complicated dance to set it up (enter values in a spreadsheet...copy hexadecimal from that into a terminal program...connect a serial cable to the controller...send the data...reboot it...and oh, you can't ever read the existing settings so you can never know if it *actually* is setup the way you want....), and a realtime status readout in binary that you have to memorize the codes for.... A probably great hardware design, with lots of variants for every need, and probably every feature you could ever want...but impossible to setup without at least several degrees in subjects I'm not even sure what they are, and if you're not an OEM buying thousands of them, the manufacturer is not going to help you, and you *cannot* use it without doing that setup, specific to each and every different system and motor. Some pretty good designs, very programmable, no degree required, but only available for low to medium power from the one reliable place that sells them, and very expensive for that...the higher power versions are even more expensive, and dont' have nearly as easy to use a setup program. Some opensource firmware projects that look pretty decent for their options and programmability...but only work on what amounts to the chinese junk clone designs. Etc.



The only controller I know of that does almost everything I want, that isn't a zillion dollars, or isn't one of the above, is the Lebowski brain...but it doesn't come as a complete controller. Which sucks. Cuz I'd've just bought a pair if anyone offered them in the right power level, if they weren't exorbitantly priced.

So...the best option I have is to take a good reliable powerstage from a good reliable well-designed controller, and cut it away from it's brain, and then wire in the Lebowski brain to tell it what to do and how to do it, in exactly the way I want it to. It's not the option I wanted, but at least it's not as hard as I thought it would be. ;)



DECISIONS, DECISIONS:

The Lebowski powerstage here:
looks easy enough, but it's cost is rather high (for me), vs using a completely prebuilt (and road-tested, probably for years!) OEM EV powerstage.

I had originally planned to use a single-chip powerstage that was really easy to build, on PCBs designed and provided by Bobc, but didn't have money to buy the parts at the time, and by the time I did, the chips became unobtainable. :( (well, there are still some places that show they can get them, but the prices are outrageous and require buying an MOQ of more than I could use or sell in a lifetime even if i could afford that many--effectively unobtainable).


I considered a number of other powerstages (see some of the discussion here
Lebowski's motor controller IC, schematic & setup, new v2.A1 )
but they are either too large, too heavy, too expensive, or require more DIY than I'm really capable of these days, especially fully assembling them "from scratch" even with a kit.


So, based on the success of Tomdb over at DIYElectricCar,
and with the help of several over here:
to decide which powerstage to use with the Lebowski SMD board designed by Bobc :rip: and built by Kiwifiat,
I went with the Honda IMA Inverter as the best of the available options, for cost and size and ease of use.

These are possibly the lowest-cost OEM EV Inverter that is widely available, at $30-$60 shipped, if you look around hard enough. You can't buy any one of the major parts needed (like the IGBT modules, or the current sensors) for much less than that, if even that. Now, there *are* plenty of people selling them for several times that, up to a few hundred dollars. But unlike the other EV inverters, the IMA inverter does have a few selling them really cheap, and various people on ES, DIYElectricar, etc, have bought and successfully used those without a problem, so they're not duds (so far at least).

The specific version documented in this thread is the
"06-11 Civic Hybrid IMA Power Distribution Module Unit 1B300-RMX-0032"
This link (which will eventually go dead) is for one of these: https://www.ebay.com/itm/06-11-Civi...ution-Module-Unit-1B300-RMX-0032/324436949602

I got mine on ebay from this seller
who gave me a bit more than half-price shipping when i bought two at the same time--not much of a discount, but he didn't have to give any. ;) Total cost including shipping and tax was $130.94. More than I wanted to spend (well, I'd rather not spend *any* money), but well worth it if these are as well designed and easy to convert as they appear. They're supposed to arrive early next week, which should be just in time for me to be able to open them up and document them during my upcoming vacation starting the middle of that week.


WHAT'S THAT THAR INVERTER THANG, ANYWAYZ?

Basically the Honda IMA Inverter is the actual controller used to run the IMA motor. So it's equivalent to the controller that's presently running your ebike, etc. The biggest difference is that it doesnt just use a throttle and ebrake input to tell it when and how fast to move the motor, or to stop it, it requires communication (CANbus, IIRC) with the rest of the Honda computers to do this.

So we have to disable the brain that's in there now, that requires this communication that we can't do (there *are* projects on other forums that *do* use it, just that way, but that's not how we roll here... ;) and that's more work than I'm up for).

There's several ways to do that.

Tomdb cut traces on the board and/or removed parts (see his thread linked previously for that), so the brain couldn't operate the powerstage or interfere with another brain's operation of it.

Kiwifiat removed the brain chip (CPU) entirely.

My plan is to follow Jack Bauer's (from DIYElectricCar and OpenInverter) Tesla Model3 hack, which is to find the Reset pin of the CPU chip, and then ground that to hold it in reset mode, which should tristate the I/O pins, effectively disconnecting the brain with a single wire. That's much less physical work, and less risk of inflicting damage on other parts of the inverter, given my various problems (eyesight, random hand numbness, general klutziness, three giant fuzzball monsters that like to stick their slobbery faces in everything I'm working on, etc.).


This video, starting at 46:01, shows the inverter itself. The rest of the video before that shows disassembly of the IMA unit that the inverter comes from.
[youtube]9Ur_VNd2Lh4[/youtube]
There's plenty of pics of these inverters out there on the web, including internals in Tomdb's DIYEC thread linked above, but I'll post pics of mine when they arrive, with something to show scale. EDIT: As promised, pics, including stuff to show scale (like Yogi, the St Bernard, and the scale itself...Yogi's paws are about big enough to cover the palm of my hand):
file.php

file.php


But since you might not know how big any of those are:
file.php

file.php

file.php


But they're not all that large considering they can handle about 15kw continuously in the Honda Civic hybrid they come from. I don't know what the peak power of those cars is.

As I only need about half that power, and only for a few seconds at startup, at most, I probably won't even need to do anything other than ensure the heatsinks are exposed to airflow under the trike. The only other time I might need more than a few hundred watts from either of them is during braking, if active braking pushes more power thru them than that, and when I'm hauling heavy loads, like the somewhat periodic 300lb-350lb+ grocery trips when I am getting my brother's groceries as well as my own (as his are usually more stuff and heavier stuff, it adds up quickly).



DISADVOLTAGES:

There *is* a disadvantage to using the IMA inverter, for my present system...the IGBTs it uses, vs the FETs the other powerstages I just mentioned, have more voltage drop and thus more wasted heat inside them. For systems over 100v or so, it doens't matter as much because the FETs available for those system have higher resistance and so drop something in the same range, and going even higher the IGBTs will be more efficient than FETs.

However, my system is only 14s2p NMC, so just under 58v fully charged, at present. So there will be more loss in these inverters than the typical controllers used for this system level.

However (again), because the IMA is good for over 150v as used on the Civic itself, I have the option of simply reconfiguring the battery to 28s1p (or seriesing a second 14s2p), without changing anything in the powerstage/brain (except possibly the 12v dc-dc that provides gate drive power, which isn't part of the brain or the powerstage, as it normally came from the Civic's regular 12v system, I believe. In my case a small Vicor brick I have laying around will provide this; only a few hundred mA should be needed as I understand it). I'd have to change other stuff on the trike, like the Cycle Analyst setup, and the onboard charger, but that's all trivial compared to dealing with controller hardware changes of the scale that a doubling of system voltage usually requires. I could also simply add 2p cell groups in series to the existing 14s2p pack to get any voltage desired up to the max of the IMA, if I have to add voltage but not as high as 28s. I'd have to reconfigure the Lebowski brains for the higher voltage max, and the associated LVC, but that's also trivial.


Another disadvantage is they are not light or small, compared to typical generic controllers: About 10.65lbs each, before adding the brain board and any cables or connectors to get throttle, brake, power, serial, phase, hall, etc., I/O thru the casing. And they're around 9" x 8" x 4", roughly. For my trike, not a huge deal, as it's already big and heavy so it can carry big and heavy dogs and cargo. But for many bikes, etc., that might be a dealbreaker.

The Lebowski brain board itself is very small, about a couple of inches across and not much more than a quarter inch tall, if that. It might weigh more than a candy bar wrapper...ok, a couple sticks of gum.

Cables and wires, well, those depend on your usage for it. If you end up using 10g - 6g battery or phase wires, or bigger, it's going to be several pounds more, but then, you would probably already have those wires on your system.



BRAAAAINZZZzzzzz:

For the brain boards...I have a pair of the thru-hole Whereswally606 PCBs, and had begun collecting salvaged parts to build them from, but never got very far, and never assembled any of them, as again my budget at the time prevented purchase of most of the stuff needed, and despite my too-large collection of disused and inoperative electronics :oops: I didn't have very many of the parts required available to salvage. Now I have a bit more money available for a variety of reasons, but the cost of just the parts to build these is higher than buying the SMD versions from Kiwifiat, and I am not sure i trust my assembly skills enough these days anyway, for so complex a project. So these are to be sent on to someone else that might wish to build their own "from scratch" as I thought I was going to. I'll just check with Whereswally606 first to see if he has a preferred list of people that would want them.

So...I bought a pair of Bobc SMD Lebowski brains from Kiwifiat, which came to a bit over $200 including shipping / etc., and they're now in-process and will be here in due time (EDIT last week of April, they have arrived). At his recommendation, becuase I"m using htem with a system that wont' be using the 15v SMPS on them, that section is unpopulated so it wont' waste power/make heat, and so it can't be there to have any failures that could impact other components, or for me to make mistakes with. :oops:


And for the moment, that's it. :)

Note that this post will be edited a fair bit as I either remember details I forgot to add, or links to other sections of the thread as they are created, etc.
 
Last edited:
reserved for IMA disassembly documentation, etc.

EDIT:

The inverters arrived today (4-18-20) after I got home from work (couple days early--apparently Fedex does now deliver residentially on Saturdays, pretty late).

I didn't get to do much with them yet other than take pics (posted in the next few posts), and open one up for internals pics and to find the CPU p/n and such.

More details on getting to the p/n later, but it is a Renesas (formerly Hitachi) CPU 64F7047F40V, the closest page still extant is this:
https://www.renesas.com/us/en/products/microcontrollers-microprocessors/superh/sh7047/sh7047/device/HD64F7047FJ40V.html
which has a link to a page that leads to this hardware manual
https://www.renesas.com/us/en/doc/products/mpumcu/001/rej09b0020_sh7047.pdf?key=4c84846a39e1c995037e853eb84c77da
attached as a PDF to this post, in case it goes away on their site.
View attachment rej09b0020_sh7047.pdf

Page 1.3 has a pinout, showing RES (active low reset) as pin 87, near the center (of course, since that's much harder for me to access on this tiny thumbnail sized 100-pin QFP chip :roll: ) of the fourth side of the chip.
64f7047F pinout.png


With luck, there will be a trace from this pin accessible to me somewhere else on the board, so I don't risk destroying the other things connected to the lines in this part of the CPU or something nearby trying to solder to this pin. I haven't checked the functions of other pins near it, but it might be "safe" to ground all of those pins together, since the CPU would be held in reset anyway.

There's no handy LED like on the Tesla Model 3 controller that JackBauer zombified this way, so I'll only know if the zombification on this is working by checking lines that do have data signals when the unit is simply powered on normally, vs when it is held in reset.

The CPU in question:
20200418_205657.jpg

The other large chips on the board:

Can't find this one under any of it's numbers, so other than it's by Oki(data), dunno what it is.
EDIT: found the numbers easier to read in the pics with conformal coat still on (added pic), it's an OKI L9620Z2. I still haven't found a datasheet for it, but similar part numbers come up as CANbus transceivers, which makes perfect sense if this is one too.
20200418_205704.jpg
20200418_204112 another chip, partly readable, conformal coated.jpg


This is a tamagawa au6802n1,
20200418_205708.jpg
According to the manual here:
https://advantechinternational.com/wp-content/uploads/2014/03/Smartcoder%20AU6802N1%20User%20Manual.pdf
it is a
Smartcoder(AU6802N1) is an R/D (Resolver to Digital) conversion IC used with a brushless
Resolver such as Singlsyn, Smartsyn, etc. It converts the electrical information (analog signal)
corresponding to a mechanical rotational angle of the Resolver to the corresponding digital data and
transmits it.

Can't find this one either, but suspect it is an smps management chip or similar, becuase of the stuff that's around it on the board.
20200418_205715.jpg
 
First up when it arrived was weighing them and getting pics. And of course, the customs inspectors had to check them out, too. Note that the first inspector (Kirin) is about head-high to a kitchen countertop, for size comparisons. The second inspector (Yogi) is shorter by a bit but stouter, and has a bigger wider head. The other inspector (Jelly) was slacking off as usual so she's not in any of the pics until I went outside to get the solvent to get conformal coat off the chips so the numbers could be read. ;)
20200418_195858 Kirin inspecting package.jpg

In box, as shipped, they weighed 22.04lbs.
20200418_200534 Package weighs 22.04lbs.jpg

They were well packed, didn't move around in there.
20200418_200558 Well-packed 1.jpg
20200418_200604 Well-packed 2.jpg


The first one weighs 10.64lbs. It has "inspection seals" on both ends from the reseller, Avicore.
20200418_200814 Yogi inspects first inverter.jpg
20200418_200829 Invertere weighs 10.64lbs.jpg
20200418_200927 End view with reseller seal.jpg
20200418_200934 Other end view with reseller seal and factory labels.jpg


The second one is the same weight, not exactly a surprise. ;) But it was surprising that it does not have inspection seals at all, so I chose that as the one to open up first and get pics of.
20200418_201028 Seocnd inverter weighs 10.64lbs.jpg
20200418_201115 Second unit's factory labels.jpg

This compares both sets of factory labels, there are slight differences:
20200418_201147 Comparison of both units' labels.jpg

This is the molded-in markings on the case (whcih is also the main capacitors), and the phase and battery terminals, which will need ring lug terminals to connect to them. Thankfully I already have some 10g or 8g cabling that has ring lugs crimped to it professionally, salvaged from some powerchair stuff, IIRC, which will mean simply making some of them into lug-to-Anderson converter cables to connect to the existing trike wiring. SB-50s for the battery, and PP75 (same contacts as SB50) for the phases.
20200418_200849 battery input terminals P and N.jpg
20200418_200854 Brand markings molded in cover.jpg
20200418_200914 Capacitance markings molded in cover.jpg
20200418_200922 Phase terminals.jpg
20200418_201046 Better view of Capacitance data.jpg


Basic measurements; nothing detailed. About 9" x 8" x 4". -ish. My ancient leatherbound hand-reeled measureing tape starts at the far outer edge of the ring you hold it with (so you can hang it on a nail/etc at the start end of your measurement), so everything is really about an inch off if you start from the tape's visible end.
20200418_203312 about 9 inches long.jpg
20200418_203333 about 8 inches wide.jpg
20200418_203352 about 4 inches thick.jpg
View attachment 35



At this point, Inspector Yogi was getting bored, and Kirin had already gone to bed, tired of waiting for me. So some disassembly excitement was started, to at least open Yogi's eyes. Didn't last long, without any noms to hold his attention.
20200418_202944 Inspector is bored.jpg


Disassembly is very easy. I followed what is shown in the video linked in the first post. There are two plastic screws, phillips head, that easily come out of the capacitor terminal cover. Then that cover simply lifts off to expose the capacitor terminal bolts.
20200418_202759 Removing plastic screws to access capacitor bolts.jpg
20200418_202803 Removing plastic screws to access capacitor bolts.jpg

20200418_203456 capacitor bolts (shorter than cover bolts).jpg

To actually remove the bolts, you're going to need a long wrench (or socket handle) to break the tension securing them. I did not remember on this one to dig out my torque wrench to see what that tension level is. :( I just used a long 10mm wrench (8 or 9 inches long, I think, by snap-on).
20200418_202831 breaker-arm required.jpg
Once the tension is broken, they will easily come out with a 10mm nut driver; there's no loctite/etc.
20200418_202858 then can remove with nut driver.jpg

20200418_203058 10mm tools needed.jpg

All the bolts are 10mm hex heads, both for the main cover and the capacitors, but the cap bolts are around half the length of the cover bolts. Don't mix them up when you put them back. ;)
20200418_210137 case vs capacitor bolts.jpg


20200418_203439 capacitor bolt cover and tools.jpg
20200418_203447 inside of capacitor bolt cover.jpg
20200418_203806 capacitor bolts removed.jpg
20200418_203814 capacitor bolts out, just terminals.jpg


Once those are removed, you can remove the four main cover bolts, two at the battery-terminal-side corners, and two between the phase terminals
20200418_202815 Main cover bolts 1.jpg
20200418_202817 Main cover bolts 2.jpg
20200418_202822 Main cover bolts 3 and 4.jpg


While the cover is off, you can see under it that the caps are molded into the case (or rather, placed there then potted in place).

20200418_203830 the casing is the capacitors, three large and one small.jpg
20200418_203843 the small cap terminals.jpg


This hole for the data/signal io connector, in particular, along with the capacitor bolt cover, are going to require some sealing-up before this can be mounted on the trike, as they will go under the deck, where they'll drown in the rare but deep flash floods we occasionally have.
20200418_203857 casing is not even water resistant with this connector hole.jpg


The cover / cap module is about a quarter of the weight:

while the heatsink is probably most of the other section's weight:
20200418_210305 case-capacitors 2.56lbs.jpg
20200418_210320 heatsink and boards weighs 7.42lbs.jpg


The rest is pics of the insides, the filenames tell what they are. More details on those as I figure them out.
 

Attachments

  • 20200418_203410 heavy heatsink is about 7 inches wide with main plate about half inch thick wi...jpg
    20200418_203410 heavy heatsink is about 7 inches wide with main plate about half inch thick wi...jpg
    48.8 KB · Views: 7,258
  • 20200418_203957 mainboard overview lowres.jpg
    20200418_203957 mainboard overview lowres.jpg
    85.4 KB · Views: 7,258
  • 20200418_204039 main cpu unreadable, conformal coated.jpg
    20200418_204039 main cpu unreadable, conformal coated.jpg
    85 KB · Views: 7,258
  • 20200418_204047 main cpu  unreadable, conformal coated.jpg
    20200418_204047 main cpu unreadable, conformal coated.jpg
    75.3 KB · Views: 7,258
  • 20200418_204100 main cpu  unreadable, conformal coated.jpg
    20200418_204100 main cpu unreadable, conformal coated.jpg
    80 KB · Views: 7,258
  • 20200418_204112 another chip, partly readable, conformal coated.jpg
    20200418_204112 another chip, partly readable, conformal coated.jpg
    80.7 KB · Views: 7,258
  • 20200418_204121 another chip,  unreadable, conformal coated.jpg
    20200418_204121 another chip, unreadable, conformal coated.jpg
    76.8 KB · Views: 7,258
  • 20200418_204128  another chip, unreadable, conformal coated.jpg
    20200418_204128 another chip, unreadable, conformal coated.jpg
    78.7 KB · Views: 7,258
  • 20200418_204136 another chip  unreadable, conformal coated.jpg
    20200418_204136 another chip unreadable, conformal coated.jpg
    74 KB · Views: 7,258
  • 20200418_204146 another chip, partly  unreadable, conformal coated.jpg
    20200418_204146 another chip, partly unreadable, conformal coated.jpg
    72.4 KB · Views: 7,258
  • 20200418_204221  another chip, partly unreadable, conformal coated.jpg
    20200418_204221 another chip, partly unreadable, conformal coated.jpg
    78.6 KB · Views: 7,258
  • 20200418_204234 maybe switching fets for smps gate drive.jpg
    20200418_204234 maybe switching fets for smps gate drive.jpg
    63.1 KB · Views: 7,259
  • 20200418_204247 probably input spike clamp protection diode.jpg
    20200418_204247 probably input spike clamp protection diode.jpg
    49.5 KB · Views: 7,258
  • 20200418_204300 probably input spike clamp protection diode.jpg
    20200418_204300 probably input spike clamp protection diode.jpg
    55 KB · Views: 7,259
  • 20200418_204732 phase current sensor, one of three, rightside.jpg
    20200418_204732 phase current sensor, one of three, rightside.jpg
    39.3 KB · Views: 7,258
  • 20200418_204754 phase current sensor, one of three, leftside.jpg
    20200418_204754 phase current sensor, one of three, leftside.jpg
    38.1 KB · Views: 7,259
  • 20200418_204800 phase current sensors.jpg
    20200418_204800 phase current sensors.jpg
    54 KB · Views: 7,261
  • 20200418_205503 conformal coat remover.jpg
    20200418_205503 conformal coat remover.jpg
    35.4 KB · Views: 7,260
  • 20200418_205524 conformal coat remover.jpg
    20200418_205524 conformal coat remover.jpg
    50.7 KB · Views: 7,260
  • 20200418_205544 inspector jelly, sleeping on the job.jpg
    20200418_205544 inspector jelly, sleeping on the job.jpg
    27.3 KB · Views: 7,260
  • 20200418_205553 inspector jelly, sleeping on the jobr.jpg
    20200418_205553 inspector jelly, sleeping on the jobr.jpg
    36.1 KB · Views: 7,259
  • 20200418_205657 cpu, reset pin 87 marked.jpg
    20200418_205657 cpu, reset pin 87 marked.jpg
    67.6 KB · Views: 7,258
  • 20200418_205704 oki L9620Z2.jpg
    20200418_205704 oki L9620Z2.jpg
    81.2 KB · Views: 7,258
  • 20200418_205708 tamagawa au6802n1 .jpg
    20200418_205708 tamagawa au6802n1 .jpg
    75.9 KB · Views: 7,258
  • 20200418_205715 not found yet.jpg
    20200418_205715 not found yet.jpg
    79.4 KB · Views: 7,258
reserved for Brain board wiring and setup, including the 12v DC-DC for running the IMA gate drives

Will also contain all documentation I can find or create about the BobC SMD BrainBoard, as there is no single complete reference to it that I can find. (separate documents and posts contain pieces but no one thing has all of them). Anyone that has information, please post it, and it will be added if it isn't already here.


[strike]Additionally, anyone that wishes to help update the SMDcontroller updated 050620.zip (RTF) file to make it an easily-followed, greatly-informational document is welcome to download it, edit it, then reupload it in their own post so that I can incorporate usable material into the final version. (note that this RTF is too large to upload so you must put it in a zip file to upload it). [/strike] Kiwifiat says this is unnecessary, that a new version is already in the works for an updated version of the brainboard, but if you feel like it you can still do it.

Pinouts of the connectors I'm using, with wire colors of the specific cables i have, for my reference. You can use whatever connectors you need to. Mine will probably change later, and be edited here then.

Overall wiring diagram, including specific wire colors I used (for my reference):
brain board  wiring diagram HIL1.png


and a blank version for you to download and markup
brain board   blank for wiring diagram  .png

DB-9 Female Serial (on HIL1) pigtail:
Pin 2 Tx RED
Pin 3 Rx BROWN
Pin 4 "DSR" used for SET jumper/switch WHITE
Pin 7 "DTR" used for RES jumper/switch ORANGE
db9 female colors hil1.png
DB-9 Female Serial (on HIL2) pigtail:
Pin 2 Tx BROWN
Pin 3 Rx RED
Pin 4 "DSR" used for SET jumper/switch ORANGE
Pin 7 "DTR" used for RES jumper/switch WHITE
db9 female colors hil2.png
BobC SMD Lebowski brain board pinout for serial comms port
serial comms.PNG

General DB-9 serial pinout
DB9-Male-and-Female-Pinouts hil1.png

Halls HIL1 Jst 5pin female pigtail:
+5V output WHITE
Ground SHIELD and BROWN
Yellow hall ORANGE
Blue hall BLUE
Green hall GREEN
Halls JST female.png
halls bobc connector.png

Throttle input 3.5mm phono plug (black housing blue heatshrink)
Tip signal black
Ring 5v (unused) brown
Sleeve Ground shield

Brake input 3.5mm phono plug (green housing green heatshrink)
Tip signal red
Ring 5v (unused) white
Sleeve Ground shield

Throttles bobc connector.png




BobC SMD Lebowski brain board schematic and parts layout
View attachment smdlebowskischematic.pdf
View attachment SMDlebowskiLayout buildmap12.pdf

BobC SMD Lebowski brain board buildup / explanation document (updated to fix typos, add warnings, update graphics to match current board layout)
View attachment SMDcontroller updated 050620.zip


Voltage measurement resistor, for BAT pin (which is the pin that measures battery voltage, and is NOT the battery power (BPOW) input pin!!!). This is calculated as

R_meas_battery = 500 x (highest battery voltage - 5)

and then round up, to get the value in ohms of this resistor (which can be a tiny 1/8w or less resistor), to the nearest standard resistance value.

So if like me you have a 14s battery, and if it fully charges at say 57.8v, then you would use

R = 500 x (57.8 -5)
R = 500 x 52.8
R = 26400
R = 27kohm resistor.

This resistor simply goes in series with battery voltage into the BAT pin.
Voltage measurement resistor.PNG
 
Looks like it will fit just fine inside the inverter casing, but the drive state LEDs won't be visible, for troubleshooting/etc purposes.

So, some thoughts about that:

First, I wish there was a connector to plug in external LEDs to for the bank of four drive-mode indicators, so I could put them somewhere I could see when it's mounted on the trike, for troubleshooting purposes later on if necessary.

For now I am planning to add a "window" to the IMA inverter casing just over their position, and put four clear plastic rods down to them.

Haven't decided exactly how to mechanically do it, but probably silicone the rods to the actual board/LEDs, then have them go up to *just* inside the casing surface, and then have a clear window of plastic glued to the casing that the rods can be seen thru without quite crawling under the trike.

Also considered leaving the rods long enough to extend out of the case, and pass thru four drilled holes in the plastic, and sand the ends of the rods into domes so they act as indicators that can be seen from the side if the ambient light is not too bright under the trike. Then I coudl see them a little easier without an inspection mirror on a long handle that I'd be sticking under the trike to look thru.


EDIT: Another option, basic idea suggested by Kiwifiat:

--add fiberoptic strands for each of the status LEDs on the brain board into a "cable" that extends far out of the case (a few feet) so I can see the LEDs while testing and operating the controllers on the trike. This means making a little "cable" of FO strands for each brain, consisting of one strand for each LED, and then going out to a little mounting box to put where i can see it (on the handlebars if the light will go that far (probably not), on the side of the trike near the controllers pointing upwards so it can at least be seen looking down and back if that's all I can get. The mounting box would need a marking on it to show which strand is for which LED, and maybe a lens-like magnifier to spread the light out beyond the tiny dot the FO strands make.

Could also use "optical pickups" at the end of the FO to detect if there's light and then wire their outputs to separate LEDs at the handlebars.
 
reserved for......did you ever notice what a weird word "reserved" is. Is that "re served"? Or is it "res erved". Maybe it's "reserve d". Or "reser ved"? Hmmm.... ;)
 
I'm excited to see this! Kiwi and I keep saying "we" are working on it because my company is funding his work, but make no mistake, he is the electronics brain here. I just buy the stuff, tear it apart, and send it his way to be modified and iterated-on!
 
coleasterling said:
I'm excited to see this! Kiwi and I keep saying "we" are working on it because my company is funding his work, but make no mistake, he is the electronics brain here. I just buy the stuff, tear it apart, and send it his way to be modified and iterated-on!

Sure but you and Tomdb were working on the IMA inverters way back so credit to you guys for identifying the potential of the IMA inverter and importantly none of this would be possible without the genius of the Lebowski controller that does such an awesome job of turning cheap scrap yard junk into upcycled EV tractive power.
 
The inverters arrived today after I got home from work (couple days early--apparently Fedex does now deliver residentially on Saturdays, pretty late), so I've added info to the posts above for disassembly, dimensions, etc.

and lots of pics.
 
Kiwifiat has sent the brainzzz on their way here. After they arrive he'll go thru the connections with me, and the setup/programming menus.

I have a few bits to locate around here that will be needed for it, like serial port connector and USB-serial adapter. If it will work, I'll likely just use the GrinTech adapter, and since it uses a 3.5mm phono plug I'd just use that style jack on the brainboards (I already have both panelmount and cabled styles, from various old audio equipment and wiring). I have some others around here too, including one with a DB-9 type serial end, just have to locate them. Whatever I use has to be compatible with Windows Vista, because the old laptop that came with that on it is the only computer I have available to take outside to the trike for whatever is needed.


I've got some other stuff on the way too, like ring lugs to make cables to bolt to the inverters for battery power and motor phases, and some new 75A SB50 Anderson contacts, and a better crimper than what I have now (not much better, probably, but I don't have the budget for the good ones yet, so some $28 swapable-jaw ratcheting types will have to do for now).

At present the plan is to make a battery cable that goes from my Anderson SB50 from the 8AWG (?--might be only 10?) main battery wires to both of the inverters. I have to decide the physical layout of the inverters under the trike before I make that cable, as I want it to be a "perfect fit" kind of length, so I don't have stuff hanging down that must be tied up in loops, etc. I will use bolt-on fuses between each inverter and the battery positive cable lug for that inverter. I don't know if I have anything bigger than 125A laying around. (and I know they're not the right voltage...but they'll have to do for now). IIRC I have the same fuse type at the actual battery positive terminal itself, too. (and a circuit breaker a short ways past the negative terminal, that is really there as a secondary security/safety switch, but also to be a breaker in case the fuse fails).

I've never pushed the system hard enough with any of the previous controllers to pop anything, but these controllers will (hopefully) be able to push it a lot harder than before. Still be nice to not pop anything...but if something goes awry, better the fuses/etc than the wires. ;)

The phase wires from the motors will simply have the lugs directly crimped onto them, to directly bolt them to the inverter outputs. On one motor I have 10awg phase wires, but the other only has 12awg, which is also the largest I can do on another motor I need to finish building into a wheel to use for testing on this trike.

I am also considering making lug-to-PP75/SB50-contact adapters, to be able to just unplug the "HI-Lebowski" (Hmmm, maybe that's what we'll call these.... ;) ) controllers from the motors and reconnect the generics, for either comparison testing or in case of failure, etc., as the phases already ahve the PP75/SB50 connectors on them. But I could just put the ring terminals on the phases anyway, *and* make the adapters, so the inverters still get the direct-bolt-on connection that's lower resistance, etc., and just use the adapters to go to the old controllers if needed. But I'm pretty sure I have some existing cables that already have ring terminals on one end, and the PP/SB contacts on the other--they're just kinda long, having been battery-to-controller cables on powerchairs in their former lives, so I don't actually need to make adapters for the latter purpose.


I also went ahead and spent a large part of my budget on an LCR meter, the DE-5000, to do motor parameter testing, as well as qualifying various parts (like capacitors) for use in repairs and upgrades of assorted projects I have around here. It will also let me test and characterize unmarked inductors salvaged from old electronics for the rare occasion I would like to build a project that needs an inductor, so that with luck I wont' have to buy one (since usually I already ahve many or all of the other parts). And test old capacitors, see if they're still useful, etc. And verify shunt resistances of generic controllers and whatnot, for testing with cycle analysts or other wattmeters.

I'd been hoping to get input over on this thread about the meter choice:
https://endless-sphere.com/forums/viewtopic.php?f=2&t=105637
but Kiwifiat's was the only response (via email along with our other discussions). I went with that since it agreed with varoius reviews/lists I'd already found out there, and wanted to get the meter here to use during my time off (which starts now, and ends when May starts). The meter arrived last night, but I haven't been able to do mroe than open it, put the battery in, and turn it on; the printed manual is in Japanese, though I have a PDF found on the web of the English version, so I can learn how to use it properly. :)


The rest of my budget at the moment has gone to a new soldering station, because my old cheap junk Velleman vtss5u has been acting up (sometimes not heating up, sometimes overheating, etc., and I would need to buy new tip(s) for it regardless. The Velleman doesn't have actual temperature control, just rather crude "power" control knob that doesnt' really tell you anything about what it's really doing at any moment. I would really really like actual temperature control, so that I mgiht be able to not destroy things (as much) on the very few occasions I am able to deal with SMT components (which I suspect i'll be increasing the amount of rapidly one I start he real work on this poject.

So after spending a few hours on the web reading reviews and comparisons, and failing to find any of the model I really wanted for a reasonable price (considering age and used condition), the Weller WESD51, which I think is the same (or really close to) the ones we used at Honeywell CFSG back in the early 1990s (even late '80s) for SMT and PTH rework, so I know that one well (even if it's been more than a couple decades since I used one)... I went with the Weller WE1010NA plus a pack of different tip types, as it is Weller's own intended replacement model for the one I wanted. It's more complex, which means it's probably more prone to failure, etc., but it also has more features (like presets, and a password to lock the settings).

(EDIT: apparently I dozed off while posting, and somehow submitted it anyway, i don't remember the rest of what I was going to post above so I just trimmed off the repeated keys and ended the sentence. :oops:)
 
I located my little 12v DC-DC brick; I'd thought it was a Vicor, but actually it's C&D Technologies, VKP60MT312, aka VKP-60xT.
View attachment VKP-60L.pdf
20200425_121421.jpg

It's around 2.5" x 2.5" x 1/2", intended to bolt to a heatsink, as it can output up to 60W continous. Since I won't be using it for anything close to it's capabilities, I'll probably bolt it to one of the end fins of one of the inverter heatsinks (since they also won't be used to their capabilities). It'll take anything from 33v-75vdc on the input.

I'd forgotten this, but it has two independent isolated 12v 2.5A outputs, and a 3.3v 18A output. So I can actually power each inverter's brain and gates independently. (isolated to 500vdc between outputs, and 1500vdc from input to output)

Testing using the old Luna 13s4p pack at 51.8v on the VKP's input, I get 12.16v on the V2 output, and 12.17v on the V3 output, and 3.30v on the V1 output.
20200425_122110.jpg20200425_122126.jpg20200425_122255.jpg



Most likely I will use two standard barrel connectors for the 12VDC outputs, so it can be unplugged from either or both brains if necessary. They're not waterproof, but I can put silicone around each one to seal them shut, but still allow cuttng the seal to unplug if necessary. The input power will come from a small-gauge wire pair run to the battery bolt-on terminals of the inverter it's mounted on.


I dug out an old test-equipment cable with a bunch of color + color/stripe wire combinations in about 22awg, stripped it back, and twisted up some pairs and triplets to use for the various signals the brains must pass to/from the inverter, like the gate drives and current sensors.
20200425_194221.jpg20200425_194226.jpg



I also dug out my heavy-gauge wires, most of which are fairly short, though I have around 15-20 feet of 4awg cable. :) Probably not going to use that for this; not really a need.

I have a few feet of 10g and some 8g, from old powerchair stuff, most of it from here:
https://endless-sphere.com/forums/viewtopic.php?f=2&t=67833&p=1263618&hilit=125A#p1263618
Most of them have PP75/SB50 contacts on one end and ring terminals of various sizes on the other, so may already be directly usable for the phase and/or battery connections. Some also have heatshrink-covered fuses bolted inline. If those are like the fuses in similar cables I'm using on my batteries, they're 125A, though they are only 32V.
file.php


The main breaker on the trike is a 63A, up to 230vac (not actually DC-rated :( ), and I've never drawn enough current for long enough to pop it, though I've peaked at around 180-200A numerous times for a few seconds testing some previous controller setups.
file.php

View attachment SGB100H-data-sheet.pdf

So I don't actually have correct protection on the trike's system if something serious did happen and cause a short on the battery lines...but most likely at least one of the disconnects (125A fuse on positive battery terminal, 63A breaker on negative battery cable) would succeed in breaking the circuit.

I have some other bolt-on-terminal fuses, that are actually rated 58v. Two are up to 125A, and four are only 75A. They won't directly fit on the inverter terminals as those are too close to the casing, but they have mounting tab/posts that could be adapted to the purpose, originally intended to bolt them to the SLA batteries in powerchairs.
images.jpg
View attachment Littelfuse_CF_Fuse_Datasheet.pdf


Anyway...it comes down to: I can put a bolt on fuse on each Inverter's battery positive terminal, as a precaution against the unlikely but serious event of an inverter failure that causes a battery short. :)




For the traction battery connection, I will use ring terminals instead of the SB50 presently used for the controllers. I will use some post-mount terminals I have to create a new distribution point under the trike, removing the original battery-output SB50 that is there, and the mess where the Cycle Analyst shunt is, and instead running the battery output cables to those posts, and then ring terminals on each end of a set of cable that runs to each inverter. I'll leave an SB50 on there as well, for connecting legacy controllers in case of a problem with one of the HI-Lebowski units, etc.
20200425_194242.jpg

The shunt will mount between the negative post and the battery negative output cable. At present it is still a single "45A" Grin 1.000milliohm shunt, because there are bugs in the CA v3.14 software preventing my use of the high-range mode, which is necessary for any shunt lower than 0.750milliohm. I haven't found a bolt-on high-current shunt that is 1.000milliohm; so far they're all at most about half of that, usually around a quarter of it.


The built-in charger will also connect to these main posts, via a smaller pair of wires that only have to handle the 12A charge rate, though it will still have a quick disconnect via the PP45 andersons it uses now, mostly so that if it fails I can easily unplug it and hook up a different one, or use the Cycle Satiator, etc., or even use it as a temporary power tap for testing things with battery voltage.


I'd do the wiring stuff in advance, but I'm still waiting for a new crimper, ring terminals, and contacts for the andersons, to arrive. Not expected for at least another few days, and possibly not until mid-May.

I might be able to do the distribution-post stuff; depends on whether I feel like splicing the ringterminal end off of one cable onto the existing battery cable.
 
The brainboards are due today or maybe tomorrow, but I'll still be waiting on a new soldering station to arrive (weller 1010) to be able to do any wiring, and that eta is sometime in the next couple weeks. (the ancient Velleman's power control died so it's stuck on full output, and isn't temperature controlled anyway, so I don't trust it for this kind of potentially delicate work, even when my eyes and hands are working well enough).

So not much completely-on-topic news, other than how these will mount to the SB Cruiser trike once they're done. I would've had something to post before this, but the last couple days I've been tied loosely to the toilet. :( That, along with a massive headache, has kept me mostly in bed not doing anything but resting in the dark (cuz even the light hurt, much less looking at a monitor).

All I really got done was taking some pics, and unscrewing the existing generic controller/grinfineon pair from under the deck between the wheels, and moved them to the bottom left side under the cargo / seatbox.
20200426_142404.jpg


This is to make room for the HI-Lebowski units, as they are so large they'll take up that entire space, and I want to leave the others on there during initial testing so if something fails I can easily just plug them back in and go, to get home, or to work, or whatever.
20200426_142433.jpg



Today was much better, so I got a few things done.

First up was to come up with a way to actually bolt the HI-Ls to the frame of the trike, as they are too heavy to bolt to the thin lightweight deck planks. (well, they'd probably be fine, but I'll feel better this way)

So I need a subframe to bolt the HILs' heatsinks to, that itself will then bolt to the trike's frame. It doesn't need to be very heavy duty, so 1/2" square tubing works fine, with thick tabs welded to that that bolt to the HILs.

The HILs are too large to fit completely within the frame between the dropouts, as the battery input terminals on them will be too close together for bolts to be put in them if I do that. So they'll need to be spaced apart a bit, and offset from each other (one a little forward, one a little back) so the battery terminals are in between each other, so not quite as much space is required as otherwise, and I can still actually unbolt the cables from the terminals without taking the inverters off, if necessary.

The first step was to make the tabs, because the only easy way to do this is to have tapped holes on the tabs that bolts will thread right into, and the tabs go "above" the heatsinks, and the bolts go thru the heatsinks' existing corner mounting holes and up into the tabs to secure the inverters. Because of the tapped holes, alignment will be fairly critical, so if I bolt the tabs to the inverters and then tackweld the tabs to the frame bars, then tackweld the frame bars together, I can then remove the frame for full welding but it will already be aligned.

This then lets me drill and tap *all* the tabs in one operation (excepting the first one done as test article to make sure the process would work), saving a lot of time in clamping/unclamping/etc.

It took a couple of hours to make them, from cutting them off the old retail fixtures they came from (to save time cutting out 8 uniform pieces), to cleaning them up so they could be all clamped in a stack, to shaping one corner on each to fit the rounded mounting spot on the inverter heatsinks, to drilling them and then tapping them and cleaning any flashing up.
20200427_133457.jpg
20200427_133512.jpg


Then test bolted them to the inverters, and clamped a piece of tubing to them to get some pics for concept:
20200427_133620.jpg
20200427_133624.jpg
20200427_133628.jpg
20200427_142847.jpg
20200427_142859.jpg

I'll need to find ten M6x1 bolts much much shorter than the ones pictured above, though, for all the phase and battery connections. Would've been nice if those had been included with the inverters, but no such luck; the ones pictured are the only ones of that size available in the HarborFright bolt/nut/washer kits I have here. I haven't looked thru my bolt bins yet, but I probably have some that are better than these, though likely all mismatched, saved from other stuff.

At that point it was too hot for me (of course we had to have a heatwave about ten to fifteen degrees above normal for this time of year, so it's been 102-105F highs, and 70s for lows just before dawn) so I had to come in and see if I could eat something that will stay down, and type this up. Hopefully cool enough by the time I'm done to go back out and finish the subframe at least.
 
The brainboards arrived, and even though I knew the dimensions, I didn't really have it in my head how small that really is. :oops: They would both fit in the palm of my hand next to each other, and still be able to see some of my palm. Pics of them in the shipping box in their bags, then sitting on the inverters (also still in their bags, to save handling until necessary).

20200427_183212.jpg
20200427_183235.jpg

So they will easily fit inside the casing. I was thinking about using a header connector on the brain board, with the mating plug siliconed to the inverter board itself. The wires to the inverter would be soldered to the plug, then the brain just plugged into the header. Then a bumper made of silicone, with nuts embedded in it, also glued to the inverter board at the other end, for screws to go in to hold the other end in place (like a clamp) using plastic washers or something. Or just glue that end of the brain to the inverter board with a "glob" of silicone at the two corners. It's very small so won't take much to hold it in place even against rough bumpiness from the roads.


I also took a pic of the 12v DC-DC brick sitting on the inverter heatsink fin as I was mentioning, but somehow didn't recall the fin height and edge shape of the inverter heatsink, so it wont' fit. Instead, I'll probably just take a bit of 1/4" thick aluminum sheet I have, and bolt that to the subframe like the inverters, and mount the DC-DC to that. Or I can modify an old CPU heatsink, take off the fins at one end, and use that end to bolt to the subframe, then the DC-DC to the heatsink. Etc.
20200427_183335.jpg


While I was looking at the thicker wire I have available for this, I pulled out the thinnest of the jumper cables I've collected over the years, which look like they ought to be about 8awg or maybe even 6, but are very flexy and floppy and not very heavy. Well, this is why:
20200427_183039.jpg
That *might* be 18g wire in the middle of that super-thick insulation. I suppose if you had a 600v car battery to jumper, they'd be good for that, but I wouldn't ever want to jump a regular 12v car battery with this thing...it'd probably catch fire. :lol: On the plus side, it does appear to be actual copper wire. But I can't use it for this project, so back to the salvaged powerchair wiring on that front.


The subframe is now half-finished. Both of the main rails are now welded to the tabs, and then bolted to the inverters' heatsinks, awaiting tomorrow when I can figure out a way to hold them up to the trike (this assembly is a bit heavy) while I determine the best position and method for the verticals that will bolt these rails to the trike itself (or just plain weld them on).
20200427_220553.jpg
20200427_220559.jpg
20200427_222806.jpg




I also found some short-enough bolts, though they're all allen-head rather than hex head, making it need more spacing between the inverters than otherwise necessary just to fit the allen-head wrench in there if it's ever necessary. The bolts are six short stem-clamp bolts from BMX bicycles on the phases, and the four battery bolts came from the steering-folding plate of an old legpowered Razor kickscooter. They're tougher than the HF bolts, and save me the trouble of cutting the HF bolts to length and then chasing the threads at the cut with a die.
20200427_220604.jpg
20200427_220608.jpg
 
Today I got the subframe finished. I'd considered bolting it onto the trike, so that I could take it off to access wheel axle/dropout stuff for roadside tire repairs but ended up just putting the whole thing several inches further back instead, and welding the frame directly to the trike frame (clamping it on first, then unbolting the controllers so they don't get weld-stuff on or in them. ;) ).
20200428_151625.jpg
20200428_151636.jpg

This moves the trike COG back some, because now they're no longer just behind the axle, but several inches behind, with the heavy heatsinks/etc.

A future version, should they be successful, might see them moved to holes cutout in the bottom of the cargo/seatbox, bolted thru the base of the box so the heatsinks are all on the outside, and the rest of them is inside under an access panel. Would reduce the height of the inside of the box by an inch or more, but if they're on the outboard ends, the middle would still be the same size.
20200428_152032.jpg


Kirin gives supervisory approval,
20200428_152101.jpg
20200428_152111.jpg
so for now, that's done, and the HILs're easy to mount to the frame, just slip into the spot and then bolt in place. The friction-fit of the tabs against the corners holds them in place while manipulating bolts and tool, so I can use both hands for that.


Phase wires will bolt direclty to the outboard terminals with ring connectors in the final setup, but for now I'll use PP75/SB50 to ring terminal adapters on each one so that I can still easily plug in the old controllers if necessary during testing stages.

The battery wires shown aren't permanent, just to get an idea if they're long enough (should be) to do an SB50 Y-connection from the main power to each one of those. (rather than the post-mount ringterminal idea, at least for now).
20200428_153351.jpg
20200428_153333.jpg

Now Kirin is happy I'd finished, so she can go in (instead of staying out there to make sure I don't leave without her knowing).
20200428_153417(0).jpg


Between the heat and the flies, it took several hours to do the above simple stuff, so I didn't get any of the other planned things done yet.

--mount 12v DC-DC to a heatsink and mount that to the subframe.
--wire it wtih two 12v plugs to go to each HIL unit.
--mount jacks for those on the HILs, and run the cable thru the plastic case and silicone in place.
--figure out a 3-wire throttle input jack (other than JST) to each HIL, so I can connect the Y-output from the CAv3 throttle out to them. (that uses JST, but I won't be using that on the HILs in the final setup, so want to put the final jack style on them now, and use adapter from that to JST until I can be sure of operation and take off the old controllers.) The most likely because I have them is 3.5mm phono jacks, in both male and female, from various audio stuff. Use a pigtail version, secured to the HIL exterior and cable run thru the casing just like the 12v plugs, and heatshrink or silicone for water resistance once connected up.
--some other stuff I've now forgotten that I'll remember later. ;)
 
The 12v DC-DC is now wired with pigtail jacks for each independent 12v output, and the input is wired to a pair of PP45s to connect to the system battery.

It's screwed to an old CPU heatsink; drilled and tapped holes for it. Had to clear the tap threads a bunch of times from the aluminum gumming them up, and still messed up one of the holes.

Cut the fins off one end of the heatsink taht sticks out beyond the DC-DC, and then drilled holes for bolts. Drilled holes for rivnuts in the subframe just left/outboard of the leftside HIL, on the front rail. Installed the rivnuts and bolted the DC-DC/heatsink to that, and then siliconed all the wiring and pins on the DC-DC (didn't do it till mounted so it wouldn't get all over my hands).

After lunch I'll be making up the external-connection parts of the wiring that goes from outside to inside the HILs:
--2-wire 12v
--3-wire serial (db-9)
--5-wire halls
--3-wire throttle
--3-wire ebrake

The DB-9 will be mounted to the casing itself, siliconed and screwed in place.

The 12v will be pigtails hanging out of the case, wire glued thru the casing with zipties on either side to prevent pulling thru.

The halls will also be a pigtail (mounted like the above) with JST to fit the JST on the motor itself. Only the left motor has halls right now (wiring not fixed on the right motor yet).

The throttle and ebrake will each be a pigtail (mounted like the above) with 3.5mm stereo plugs. Then matching jacks with JST-adapters to the existing throttle connections, and a Y-adapter jack to the ebrake-COT unit up on the tiller.
 
amberwolf said:
After lunch I'll be making up the external-connection parts of the wiring that goes from outside to inside the HILs:
--2-wire 12v
--3-wire serial (db-9)
--5-wire halls
--3-wire throttle
--3-wire ebrake

Skip the "2-wire 12V " there are pins on the grey male socket for 12V supply that feed the isolated gate supply circuitry.
 
kiwifiat said:
Skip the "2-wire 12V " there are pins on the grey male socket for 12V supply that feed the isolated gate supply circuitry.
I thought about that, but it may be easier to waterproof that socket by just gluing a plate over it, and running the 12v cable thru it's own cable-sized hole in the plastic nearby (along with each of the others thru it's own hole), with strain relief zipties on either side and a silicone seal at the hole. Then it would connect to the pins on the *inside* of the connector, where it goes to the mainboard.

If I run the 12v cable to the pins on that socket, I'm not sure if there's enough space for the interior ziptie to keep from pulling the cable out by accident.

FWIW, if I could physically do it, and had a mating connector for that gray socket, I'd cut the pins (other than 12v/ground) internally, just above the PCB, and use *that* for all the external I/O for the brain. Then work out a seal around that.
 
Because IGBT's are slow in comparison to mosfets we need to set deadtime to 2000ns and pwm frequency to 6kHz.
The following image shows how to wire a BobC brain board into an IMA inverter:

BobC_plug.PNG
BobC_to_IMA2.jpeg
 
Thanks! I'm not yet at the point of wiring in the brainz, but I'm close. I probably would've gotten there today, but I have to go back to work tomorrow, so had to spend the last little while clean up all the stuff scattered around that I've been working with, so the dogs dont' have a party with it while I'm at work. ;)

But since the Weller 1010 soldering station arrived, I cleared off the project table in the back room, and set it up as a workstation for this project. (Eventually this will be moved to an old sewing cabinet, with a fold-down workstation door and storage above and below, but that's gonna be a while). The station includes:
--Miniature "portable" Hitachi oscilloscope
--Sorenson DCS55-55 to power things safely (current limited, etc), rewired to connect to my welder's extension power cable from the dryer outlet. ATM it just has an SB50 output connector on a cable, but at some point I'll add a panel with several different options for the most common things I use, and some banana jacks.
--Weller 1010 soldering station and an assortment of tips and a couple different solder sizes for different jobs
--Weller 80W iron for the huge stuff
--old Gateway Vista laptop for looking up /displaying manuals, pinouts, etc etc.
--old Gateway monitor that has VGA, DVI, component, composite, and most importantly S-video inputs, because:
--an old "student cam" video camera on flexible arm with extreme closeup focusable lens, hacked to output to standard S-video. (so I can see closeup stuff, like working on electronics, and get pics by taking photos of the screen instead of trying to focus a camera directly, which hardly ever works).
--a couple of halogen desk/work lamps
--a 4-foot 2-bulb "shoplight" with extra-brigth 6500k LED bulbs directly overhead, so I can almost see.
20200430_155749.jpg

20200430_164110.jpg

20200430_155738.jpg

Oh, and it's setup with an ESD-discharge strap, whcih I use on my leg so I don't keep catching teh coily cord on stuff with my arm. ;)
20200430_164328.jpg


Anyway, back to the topic:
I cleaned the conformal coat off of the seven areas that will have wires from the brainz: The 12v, the three current sensors, and the three gate drives.
20200430_163019(0).jpg20200430_163024(0).jpg20200430_163031.jpg20200430_163036.jpg20200430_163039.jpg20200430_163043.jpg20200430_163048.jpg

FWIW, I considered using the 12v input pins from the gray connector, but I can't get to them from either inside or outside the connector, to be able to solder them in. (I might be able to fit bare contacts from another plug type onto the outside pins, but I'd have to glue them in place to make sure they don't come loose, and there's no keying, so I could accidentally hook it up backwards, and knowing me, if it's possible, I'd do that). But if anyone else wants to try, these are the +12v pins, on the "inside / bottom" row, the two connected together on the PCB:
20200430_164104(0).jpg


With the new station, I got some soldering done including getting the 12v connector attached to the board (on those empty capacitor pads near the big zener, because there is a hole in the PCB that made it easy to tie down the cable there).
20200430_173306.jpg

I tinned the tips of the wires for the throttle and ebrake cables, and passed them thru cable-sized holes in the plastic casing, where they will also pass thru that same space as above, just inboard of the gray built-in connector. They're not fixed in place yet, as most of hte wire I'll pull back thru out of the case as I close it up. The 5v wire on the throttle and ebrake cables will not be hooked up to the brain board, because both of the brains will be wired in parallel to the same source on each of those, and I don't want slight differences in 5v to cause problems between them, or to have either one be backfed by (or backfeed to) the Cycle Analyst that is actually providing the throttle signal. (I could just put little schottky diodes in there...but I can always open up the controllers if I change it to require the 5v from them to actually run the throttle or brake sensors.)
20200430_174725.jpg20200430_174738.jpg



I haven't done the serial yet, as I wanted to use a DB-9 actually mounted thru a hole in the plastic casing. But I have yet to find any of the bunch I know I have around here. I did find a bunch of my old serial cables, both extensions and regular cables, and "laplink" crossover cables. So I can use those and make pigtails just like with all the other cabling, and is probably what I'll do.


I forgot about the motor halls; I still have to make up a wiring bundle for that. Most likely it'll just be more twisted-pair wire sets, that all run out to a JST connector to match what's presently on the motor. (normally I just directly connect that kind of wiring, but until I'm certain it's all working, I have to leave the option to switch to the old controllers).



I did the "reset pin" mod to the CPU, though not quite the way I'd planned. I wanted to trace the circuit from the reset pin to somewhere else on the board I could get to with a soldering iron, but it goes right to a thruhole to another layer. I might've been able to scrape the green protective coating off and solder to that short trace, but I wasnt' sure I could do it withtout damaging other nearby stuff.

So first I cleaned the conformal coat off teh entire area around the CPU, then used the oscilloscope and multimeter to try to find signals of any kind from the CPU. But I couldn't find antyhing at all, just steadystate 5v or 0v with the meter, and just a low-level AC 60hz noise signal with the oscope. The scope lead works, as it can see the scope calibration signal, etc. I didn't check the SMPS area with it, just thought of that as a test, will do that next chance I get. (have to disconnect the ground on the reset pin).

Just to note that I was powering the inverter with 56.0V from the Sorenson, and that was also powering the 12V DC-DC, which was sending 12V to the inverter as well. I had jumpers clipped from the inverter terminals to the capacitors in the casing, including the ground that goes to the heatsink, since the casing and caps can't be bolted to the inverter and still access the PCB inside.

So I went ahead and did the reset pin mod, by lifting the pin with a dental pick while heating it with the thinnest tip I got for the Weller; it was far easier than I expected. The first pic points to teh trace, and the seocnd to the lifted pin:
20200430_155635.jpg20200430_155652.jpg


Then I soldered a fine wirewrap wire piece from that pin to the ground side of the VCC/VSS caps for the CPU.
20200430_164432.jpg20200430_164534.jpg

I remeasured the area, and found no change in the states of the stuff I could remember (didnt' write any of it down :( ). Still no signals from it on teh scope.


So the inverter itself is "ready" for the brainz to get wired in, and I just need the serial port added to it, then can wire all the external I/O in at the same time.


I haven't opened up the second one yet at all, so if I really screw something up on the first one, I'll be able to figure it out and do it right on the second one to verify it, then maybe fix the first one if i have to. Hpefully it'll all just go ok. :)

Side note: I couldn't see this until I used the camera setup, but there are a number of components whose silkscreened designations are highlighted with a blue or purple marker:
20200430_164922.jpg


Then I was informed by management that it was doggie dinnertime, and after that I didn't get anything else done besides cleanup of the stuff just laying around.
20200430_164936.jpg
20200430_164943.jpg20200430_164945.jpg
 
This is some serious pirate work. Well done. :twisted:

Did you consider using a fault signal to put Lebowski into reset if something bad happens?

Also 2us for automotive power section seems small. I use 3.1us with Volt inverter. I broke two Volt inverters using 1.8uS because i mistakenly calculated my deadtime. Better have some holes in PWM signal than in IGBTs i say.

Also Bas told me i could put both throttles to GND to stop PWM in case i wanted to shift with original transmission. Just FYI.

I enjoy reading about this.
 
arber333 said:
This is some serious pirate work. Well done. :twisted:
Thanks. Nowhere near done yet, though.

Did you consider using a fault signal to put Lebowski into reset if something bad happens?
How would I do that? I mean, what would generate the fault signal?

Also 2us for automotive power section seems small. I use 3.1us with Volt inverter. I broke two Volt inverters using 1.8uS because i mistakenly calculated my deadtime. Better have some holes in PWM signal than in IGBTs i say.

If you're referring to Kiwifiat's post up above:
kiwifiat said:
Because IGBT's are slow in comparison to mosfets we need to set deadtime to 2000ns and pwm frequency to 6kHz.
I am presuming his is based on experience/usage with this same system (though I don't know with what kind of motor)...but I'm all for safer rather than sorrier. ;)


Also Bas told me i could put both throttles to GND to stop PWM in case i wanted to shift with original transmission.
Mine are hubmotors, so no transmission, but it's still good to know that there's a way to "halt" the controller, essentially. I could wire a switch to both the ebrake COT and the CA's throttle output that would disconnect them and switch the inputs to ground, if I find a need for it.


I enjoy reading about this.
I've been trying to follow yours, too, but there's not enough detail about what's going on, decisions, project scope, etc., for me to figure out much.

I've been trying to give as much detail as I can possibly manage, in case someone needs to do something different, they'll see why i chose whatever path I did. :)
 
The crimpers and whatnot arrived, so I should be able to do some more on this Wed/Thursday, if nothing comes up to get in the way; I think I have everything at hand to do the rest of the work now. Since the temperatures are expected to be somewhere over 105F for both my days off, I'll probably be wanting to do inside work anyway (of course, it's in the mid-90s or lower the last few days and will again be after my days off).


So this is the list of stuff I'd been waiting for, some of which arrived earlier (as noted in previous posts):
Crimper: https://www.amazon.com/gp/product/B00Y7M7SG6
Anderson PP75/SB50 contacts https://www.amazon.com/gp/product/B07HB9SF7L
Butt splice crimp assortment: https://www.amazon.com/gp/product/B076H3VHGN
Ring lugs: https://www.amazon.com/gp/product/B07WZRKR7T
Heatshrink with solder bands: https://www.amazon.com/gp/product/B07R4Z13P3
Weller 1010 station: https://www.amazon.com/gp/product/B077JDGY1J
Weller tips: https://www.amazon.com/gp/product/B07P1B2535

I'd kinda needed all the above stuff (except the ring terminals) for several other things, but I only really decided to buy them because of this project, so it will have to take the blame / budget hit for them. ;)
 
Back
Top