Instant Start 18 fet Infineon Boards are here...

[methods]

Methods, I am about 90% done assembling the controller. I try to avoid as much mistakes as possible. I appreciate if you can answer the following questions:

Sure, shoot ->

1. How do you configure the resistors network in 48V, 72V, 100V? I have read Knucles threads and he mentioned the LM317 voltage regular should be between 12V & 40V. Also, The battery voltage drop about 6.5V for every 100 ohms. Is that true? If so, the value to use would be 300 ohms for 48V, 600 ohms for 72V, 1K for 100V? Since I received 3 x 1.5K, 3 x 1.8K, and 1 X 680 ohm resistors, should I parallel a set of 3 to get the actual value? Ex. 3 x 1.5K parallel = 500 ohm?

Actually, this is a different regulator.
Here are the suggested values:

1.8 || 1.8 || 1.8
input 70v~ 105v

1.5 || 1.5 || 1.5
input 60v~ 86v

1.8 || 1.8 || 0.68
input 50v~64v

1.5 || 1.5 || 0.68
input 30v~54v

2. What is the 6th lead wire use for on the throttle connector?

Hall – 6P
RED +5V
YEL GND
GRN A
BRN B
GRY C
BLK N/C

Throttle – 3S
RED THR
YEL GND
BLK +5V

Brake – 3P
RED ___
YEL ___
BLK ___

CA – 6S short
BLK VBAT
GRY GND
BRN SHUNT-
GRN SHUNT+
YEL SPEED
RED THR_INH

If you analyze the way I set up those pins it will be come apparent that I did 2 things:

1) Maintained compatibilty with the Crystalyte Throttle and Brake as well as the CA
2) Set it up so that plugging the connectors together in an incorrect manner will not cause damage.

Does this answer the question?

3. On the CA connector, do you wire the Ebrake to the EB+ or EB- ?

There are 2 ways to do it:

Brakes Only: EB-
(EB+ would be for a 12V input)

Variable Throttle Control by CA
(Follow instructions for CA, attach to throttle line)

4. On the CA connector, where to wire the Hall to the amplified hall circuitry on the board?

Attach it to any of the three hall signals, A, B, or C.
Normally you would hook it to the SM pin next to the 5V regulator but it is missing a BJT.

5. Is the ON/OFF switch use for enable the controller circuitry? If so, should I connect the VB2+ & Vcc-L?

Connect the switch between V+ (anywhere) and VCC-L
I suggest drilling a hole on the left side cover between the top of the caps and the fets.
This will lock in the switch and allow you to tighten the nut while still allowing easy assembly and access.

Remember to power the CA from VCC-L so that when you turn the switch off the CA turns off

Lastly, can you take a picture of your fully assembled unit top & buttom especially the wiring details?

Look back in the thread, I have already done this.
All I have is a Cell Phone camera, but I can go back through right now and take pictures of how I did it.
Give me 20 minutes as I have some people here now.

Please - If you plan to run 100V replace the base resistors with a 2W 3K.
Also, you need to find a diode for D32 if you want the brakes to work.

-methods


NOTE TO PEOPLE READING THIS IN THE FUTURE !!!!
THE COLORS HAVE CHANGED IN THE CA PIGTAIL SO DO NOT USE THE ABOVE CHART
THIS ONLY APPLIES TO EARLY KITS

 

[methods]

It shouldn't be hard to translate the translation using a hex editor, as I can actually recognize and type in the characters.

I think I may cross-reference a past post, and translate it for you guys. Or, I'll try anyways.

Yep, they pop up right in the hex editor.
Did with the Chinese too but it was hell to search

-methods

 

[swbluto]

It shouldn't be hard to translate the translation using a hex editor, as I can actually recognize and type in the characters.

I think I may cross-reference a past post, and translate it for you guys. Or, I'll try anyways.

Nevermind. I'm on vacation in Missouri and do not have windows XP installation data available to proceed with installing the program on my netbook. So, someone else can do it without feeling like their effort is redundant.

 

[swbluto]

It shouldn't be hard to translate the translation using a hex editor, as I can actually recognize and type in the characters.

I think I may cross-reference a past post, and translate it for you guys. Or, I'll try anyways.

Yep, they pop up right in the hex editor.
Did with the Chinese too but it was hell to search

-methods

I think the most efficient way for someone who didn't know chinese probably would've involved using a Chinese OCR program, like possibly Dragon's naturally speaking. But, even then... that'd be hell. A chinese hex editor probably would've also been difficult to use.

 

[methods]

Please note:

This is by no means the definitive way to populate these boards.
This is just how I make my own when I am in a hurry
I don't bother with the brake lines, I cheese the CA to the ebrake line since I dont use that feature, I hook right into the hall sensors, I glue the caps with QuickGrip, I slather the fets with conformal coating where they are close, I build up the bottom with solder wick only. I will use the board like this for >150A peak and long pulls at 100A cont. 24S Lipo is what I am running (100.8V -> 88.8V)

-methods

 

[methods]

Not bad for cell phone pictures eh?

I also conformal coat the entire bottom edge where it fits into the groove of the controller.

-methods

 

[geoff57]

hi
if SM is the same as on the 12 fet boards i have been dealing with then forget about using it to get a speed reading for the CA on the 12 fet boards it someho was multiplyed by 6 and so the signal is unuseable using a hall signal is much better i would stick to that.

Geoff

 

[dnmun]

methods, you do pretty work. it is not easy to get the mound of solder to be so uniform.

why not just buy a spool of bus wire and use that rather than the solder wick. i saw it in spools at norco.

very pretty.

 

[methods]

methods, you do pretty work. it is not easy to get the mound of solder to be so uniform.

why not just buy a spool of bus wire and use that rather than the solder wick. i saw it in spools at norco.

very pretty.

Really? I thought that solder looked like hell. Now that you said something I will have to try to make my next board perfect.
I did it at only 650F since last time when I did it at 800F I got the board hotter than I intended to.
If you can get the iron up over 700F and use high-flux solder it is easy
I bought some special solder at Fry's that has so much flux that it just oozes off the board.

Bus bar - yea... you are probably right. One part of me wants to use stranded wire but the reality is that we are just trying to create bulk.
I guess it would make a lot more sense to just run big slabs of copper.

yea... You are right. I am going to get away from the wick and go to something larger and easier to work with.
More copper and less solder (the conductivity of solder sucks anyway :?)

Thanks for the idea.

-methods

 

[liveforphysics]

Solder wick is very easy to work with. It's very fluxy, and generally uses good O2 free copper. I've been laying it on traces on boards since my highschool days of being nuts about beefing up amplifiers. I've never had a board failure on something I had beefed up with solder wick.

I think Methy has an unlimited free supply of it as well

-Luke

 

[methods]

hi
if SM is the same as on the 12 fet boards i have been dealing with then forget about using it to get a speed reading for the CA on the 12 fet boards it someho was multiplyed by 6 and so the signal is unuseable using a hall signal is much better i would stick to that.

Geoff

Will do.
Direct shot it is then.

If that is the case then I need to devise a "tap" that is cleaner than running a wire under the board.
Always bugs me when I have to do that...

I tried running 2 wires in one hole and they technically fit but that was the time that one single stray strand poked over and shorted B-C. Took me all day to figure out what was wrong. A pair of hall sensors intermittently shorted together is an awful bug to have... Can create all sorts of odd behavior.

-methods

 

[methods]

I think Methy has an unlimited free supply of it as well

I wish
Unfortunately pilfering is unacceptable where I work so I have to buy mine at fry's.
I did buy the big-papa size so it was relativity cheap.

I am jealous of guys that work in private industry where they have a dumpster that you can just rummage through
A friend of mine is actually encouraged to dig through the parts bin at his company because it costs them so much to throw it away!!!
(environmental hazards etc.)
At my work you are not allowed to even pull anything out of the trash

You WOULD NOT BELIEVE the stuff I see thrown away... (gets me all agitated just thinking about it )

-methods

 

[liveforphysics]

That's uber-ghey they don't let you dumpster dive at your work. I pull out miles of wire from my work dumpster, and so much beautiful aluminum I-Beam and sheet, I'm thinking I'm going to build a huge shop out of it.

We also discard between 50-500 servers daily. When you have 350,000 servers in your datacenter, you are in a constant state of de-com'ing outdated servers and installing new. They used to be thrown away with all the HD's in the servers, but now they are required to pull the HD's out, and smash them with this hydrolic press that punches a 1/2" whole right through the stack of plates. There is a guy who's job is to just remove RAM and snap the sticks, and drive the punch through HD's all day.

It all used to be donated to schools and things, but the new rules for handleng secure data now forbid that any HD or memory storage device of any kind must be destroyed before it can leave the facility.

 

[dnmun]

i noticed there already was bus wire on the phases too and didn't know if you had added it, but you mentioned it was only solder wick that you added. i think the bus wire at norco is already pretinned, and it would be straight so just lay it on top of the trace and heat until it would settle in the solder already there so it makes like a guide rail a fixed distance out from the surface so you could draw the bead out pretty fast and it would be uniform and less tendency to spill out and short to the other leg of the FET, you have more heat than my little 30W could do, that is huge mass, but i think the bus wire is in smaller spools or loops, just unwind and it is essentially straight so it would lay right on top of the bus. and yes, the copper is cheaper conductivity than the solder. 16G? i can't put that much on the leg of the FET, my hands shake a lot too.

 

[methods]

The traces actually come with bus wire soldered in.
Looks to be good for maybe 50 - 70A
I solder the wick over the top.

-methods

 

[steveo]

Hey Methods

I know you posted the Wiring setup above but it doesn't make sence .. I have the 2009/03/28 Board

For Phase wires, I copied docs wire setup becuase i know he uses x-lite x5

a - Blue
b - Green
c - Yellow

For hall sensors, Does it go like this

+5v - 5v Hall
Sa - Blue
Sb - Green
Sc - Yellow
Gnd-1 - Ground

Throttle?
Programer Plug?
5v E-brake for c/a?

LED

Led
Ground

Is there second led like doc did? Where?

C/a Plug I know how..

 

[methods]

Hey Methods

I know you posted the Wiring setup above but it doesn't make sence .. I have the 2009/03/28 Board

Why is that?
You know the colors I listed map to the pigtails I sent right ?
Not the industry standard color pattern of Yellow, Green, Blue.

For Phase wires, I copied docs wire setup becuase i know he uses x-lite x5

a - Blue
b - Green
c - Yellow

Ok... For myself, I would stick to the standard (Yellow, Green, Blue) and switch them at the external Anderson connector but that is fine.

For hall sensors, Does it go like this

+5v - 5v Hall
Sa - Blue
Sb - Green
Sc - Yellow
Gnd-1 - Ground

Again, I would stick to the standard: A=Yellow, B=Green, C=Blue. You can see this in the Technical Reference area. In my opinion it is better to have everybody solder them in the same then change the wiring at the external connector. I guess it does not really matter though, you can assign whatever colors you want.

Now - don't confuse the "standard" colors (i.e. Red, Black, Yellow, Green, Blue) with the physical "Pigtail Colors". The pigtails come in a funny set of colors and I did not want to put the wear and tear on them switching pins around. The colors I listed in the instructions are the pigtail colors. The pin order I selected was to protect the circuit in the situation where the user plugs the Hall connector into the CA connector and the brake connector into the throttle connector. Please note: The CA connector has 100V on it so if you use the hall connection that i sent but in a different pin order you run the risk of blowing your board if you accidentally plug the two connectors together. Does this make sense?

This is the point where someone asks: "Why did you put to connectors on the board that can be plugged together?". It boils down to cost and availability. The CA connectors only ship in pairs. If I did not use the mate to the CA I would just have to throw all of those away.

Throttle?

I listed the colors and where they plug in. (Throttle is the SP hole)
I guess i had assumed that everyone was familiar with the Infineon boards.
Please check out Knuckles threads in the Technical Section, everything is clearly laid out.

Knuckles thread in the Technical section

(Most everything still applies. The regulator is different, the programming hole is a little different, some locations are different. All of the "normal stuff" is still the same though.

Color scheme matches the pigtails sent.
Black is +5, Yellow is GND, Red is signal.
Feel free to move the pins around in the connector if you like.

Programer Plug?

The programmer shipped with a mating header connector.
Solder that into the row of holes right above the date stamp.
Orientate it to the left. You can confirm you have it right by seeing that the red/orange go to +5 and the black goes to ground.

5v E-brake for c/a?

The ebrake is EBS-
Just apply your active low to this pin.
Dont forget to solder the diodes I gave you into D32.
You can run the CA there or you can run the CA to the Throttle line via a diode and a resistor per the CA Instructions

LED

Led
Ground

To me it just looks like Doc hooked an LED into the VCC-L circuit
Button is on, LED is lit
Button is off, LED has no power.

Is there second led like doc did? Where?

See above.

C/a Plug I know how..

OK.


Please let me know if you have any more questions.
As you can see, It can even take a while for experienced controller modifiers to figure out all the details and this is why I am holding back kits from the nOObs.
Eventually I can put all my notes together into an instruction manual but I just dont have time for that now.
I was able to figure it all out on my own by reviewing the past Infineon threads so I figure crafty fellows such as yourself will do ok.

Dont be shy, ask as many questions as you like.
We are all still learning about this -> I am learning about what areas need better documentation, you are picking through the details you need to finish, others are (hopefully) taking notes so that when it is their turn to build they will have it all laid out for them. So questions are always helpful to everyone.



-methods

 

[methods]

P.S. I would not bother with all the trouble of a Status LED. The CA turns on and off with the board and that is a much better indicator of status in my opinion. The diode just makes for a lot more work, more current draw on the regulator, and one more thing that can go wrong. You have to drill a hole for it, glue it in, calculate a current limiting resistor, bah!

-methods

 

[methods]

So I had a chance to test the software out.
Here is a screen shot after a successful burn:

Unfortunatly I cant upload a *.ASV file

Create a file named XXXX.asv
Past this inside:

5:EB218
70
30.0
20.1
1.0
99
0:Switch Mode
100
80
60
1.0
15.0
1:Only Fake Indicate
1:Comm GND
2:
35
75
0:Low
0:NO
0:Fast
0:Cruising Indicate
0:120 Degree


For those who have already had the old software installed you can just run the EXE that Geoff posted.
All I did was run the EXE, choose all my parameters, then save them.
The saved result is what you see above.

I came up with those values by first using common sense, then looking over the values used on the old software, then analysing the choices in the new software.
This is at least a reasonable place to start.

If in doubt set all speeds to 100%.

Remember that the current is a relative value, not an absolute value. If you double it from 30 to 60 you will get double the current. This all depends on your shunt value. They REALLY should have a field to enter your shunt value :wink:

-methods

 

[geoff57]

hi
methods try this I have copied your settings into a .ASV file and saved it as test_1.asv then altered phase to 90 and batt to 40 and saved as test_2.asv, BOTH are readable
both me and knuckles had this problem the answer is NO spaces in the filename i have put both files at the bottom of the post.

I have had to zip the files up to get them on ES.

Geoff

 

[methods]

Cool, thanks Geoff.

On another note:


WARNING - WARNING

This is a Ghetto Service Announcement

Please, do not try this at home kids :wink:
I will void your warranty quicker than you can say supercalafragelisticexpialadocious!

Those of you who have been following the Infineon threads are aware of the fact that geniuses who designed the firmware set it it up so that you have to load the settings right on power-up. Usually this means that you have to implement a power switch on the +5V supply of the TTL programmer.

Programmer hack
(This mod does not appear to work on this chip. I think the chip is powering off of the RX/TX lines)

--- This means that you can not just solder in the USB cable and be ready to go. You have to hack it in some way.

I made a work-around for that on a different controller where I bypassed the +5V from the TTL programmer and used the power switch on the controller to toggle the programming.

Well... Neither of those are working with this chip.
The only thing that I have found so far (Soooooo Faaaaar... i.e. I am sure someone will come up with a more elegant solution) is to lift the ground.

Yea, yea... I know... Never lift the ground
Since I have an MSEE from Stanford and I am a self proclaimed firmware expert the rules dont apply to me

Sooooooo... I made this abomination: *(See pictures at bottom of post)

All I did was pop open the USB plug (it is just press fit)
I de-soldered the ground and soldered a piece of wire to the pad on the board.
I then drilled a hole in the plastic cover with an exacto knife.

Poke the two wires through and BLAM!
Instant switch 8)

The USB adapter can now be hardwired to the board and left in.
When I want to program the controller I plug the USB cord into my computer, start the software, load the settings, press START, then press my "Ghetto Button" (c).

I do not advocate that you do this -> I am just saying that this is how I roll so.... You can do it if you want to but dont come crying if you smoke your chip.

-methods

 

[geoff57]

hi
are you saying that the 5v power supply from the programming lead is not required?
what connections are required on the board?
my standard programming lead terminates with a 4 way connector that I always mate to a connector soldered to the board programming port.
when I set up my switch i put it on the 5v and the ground wires using a DP DT switch as an on of button.
I will contact keywin and find out how he programed the controller boards he makes up himself this may take a few days as it is the week end.

Geoff

 

[methods]

No.

I am saying that with the older boards you can just tie RX, TX, and GND from the programmer to the board.
You can then supply +5 directly from the 5V buss on the board.
When you flip the power switch on the controller (or apply the battery) the code will flash.

What I am saying is that this does not work on this board.
This board programs just like the other board in that "poking" the programmer in for a moment will flash the chip.
What it will not do is work like Philf has his working..
Just putting a switch on +5V is not enough.

I have found that ground must be broken.
BUT -> I played with this for about 5 minutes.
My goal was not to find the best solution, but only to find -A- solution.

Speaking of which... I just got done testing my newest controller :wink:

Fully programmable
100V 150A
Method spec

So... It sounds like your solution of using a DPDT will work just fine because it breaks the GND line.

-methods

 

[steveo]

Hey Everyone,

Some Eye candy for you

my infinion 18 fet board in the process..

8AWG Traces

Is this that part of the board that should be coated? or modded so it wont make contact with the case?

I will be running 79.2v of a123 .. possibly more down the road ... so i went with 3 x 1.8kohms ...

Methods why do you short that mini resistor on the right when adding the last resistor?

p.s. .. I haven't ever followed this new boards .. the last board i worked on similar to this was the original e-crazyman !! .. so bare with me !

About that gap on the "c" phase where the fets are really close .. Methods .. Use a very small washer in between the tight gap.. then solder the fets in place .. it will give you a nice gap that you need!

-steveo

 

[geoff57]

No.

I am saying that with the older boards you can just tie RX, TX, and GND from the programmer to the board.
You can then supply +5 directly from the 5V buss on the board.
When you flip the power switch on the controller (or apply the battery) the code will flash.

What I am saying is that this does not work on this board.
This board programs just like the other board in that "poking" the programmer in for a moment will flash the chip.
What it will not do is work like Philf has his working..
Just putting a switch on +5V is not enough.

I have found that ground must be broken.
BUT -> I played with this for about 5 minutes.
My goal was not to find the best solution, but only to find -A- solution.

Speaking of which... I just got done testing my newest controller :wink:

Fully programmable
100V 150A
Method spec

So... It sounds like your solution of using a DPDT will work just fine because it breaks the GND line.

-methods

hi
from what you are saying you powered up the board for the program transfer from the main bike battery and switched on the controller once you had hit the transmit button on your PC and with the new boards this does not work. If that is how you got power to the chip for the reprogramming I am suprised it worked as when in normal operation the TX and RX lines to the chip do other things the pins are not waisted on just reprogramming the chip. both Philf and my type of programming connector powers up the board from the connector itself useing the 5v that the usb port puts out to power small things like hubs and the TTL converter itself.
So when we program a infineon the controller is switched OFF best to disconnect from the batt but not needed, four connections are made between the computer and the controller board programming port usb ground, usb 5v, TTL tx and TTL rx the data we found would not transfer unless the 5v had a switch in it so the chip could be dead untoll it was powered up at which point the data would be trasmitted, I have no boards yet to check with mine come in about a week keywin just sent them both 18 and 12 fet boards with the new chip in them I will test them but dont see why a USB powered programing cable with a switch in the 5v line would not work, most of the other infineon threads have used this method to program the controller mainly becaus they were trying to find a cheap alternative to the original programming cable that keywin sold that fitted a RS232 port converted it to TTL then sent out the signels down the Tx and Rx wires it also required a usb input to power the chip on the controller this time the RS232 would not work on its own, the cost of this cable was very expensive as well, philf found a cheap usb all in one soloution and things progressed from there till we have the version that most use in other threads, this is the first time I have found an altenative way of wirering up the programming cable.

Geoff