New Lebowski disciple!

Beez65

1 mW
Joined
Sep 21, 2014
Messages
13
Location
NorCal
Hi Folks, newbie here! Been lurking and reading until my eyeballs bleed! :shock:

I have assembled my first EV, however small, a Catrike speed with a DaVinci drive, Astro 3220, and Castle Creations controller.
50V at 15Ah, it will go over 30 miles with over 1500 feet of climbing ( my commute, one way ) between charges with a little margin if I am easy on the throttle.

The thing that brought me here is that the control at partial throttle. I liken it to a two stroke motorcycle before it "comes on the pipe". It really only works well at high throttle openings. That's fun for a while but I was trying to build a ( semi ) practicle vehicle. Being able to run smoothly at all throttle openings will make this vehicle much more usable. In addition, the controller limits me to 50V, but the motor will handle a lot more so it would be nice to not have that limitation.

So, I found the amazing Lebowski chip, read a few threads with zillions of posts, some of which I actually understood! :mrgreen: I have ordered a couple. Have not received them yet but I have lots of questions about the build process. I have build electronic devices over many years I feel confident that I can handle this project, however, I have never had PC boards made and certainly never designed any. I have put things together on breadboards, and perf-boards, and built lots of kits of varying degrees of complexity. So, here are some questions:

How much will it cost me to have a small lot of boards printed? Should I use the Lebowski design or Arlo's or Zombiess'? Something else? I especially like the form factor that Zombiess has proposed in his "mini" controller proposal. Where should I look for quotes? :?:

I really want to do something that is modular and expandable to larger and larger motors as I build larger EVs. I want to build a car but it seems like building up from a smaller vehicle, like my trike, would be ideal. I think my second motor would be a Honda civic IMA motor, or a stack of them, to go on a light motorcycle. I like the relatively lower voltages that they run at as opposed to the 500V prius motors.

I guess my main questions are where to get boards printed, and will this avenue work well with sensorless motors such as my existing Astro 3220, and a civic IMA motor? :?:

Thanks in advance for your help, and thanks for the awesome forum!
Chris
 
If you have some made up, I'm sure there will be several builders who will "buy and try". That being said, I wouldn't expect there to be a rush of orders until there are several reports from users who state "The chips from this supplier work great, and improves X, Y, and Z". Best of luck, though...

"Lebowski's motor controller IC, schematic and setup manual"
http://endless-sphere.com/forums/viewtopic.php?f=30&t=36602
 
OK, without wanting to sound too much like the old Dutch commercial (dutchies will recognize this):'we from toilet-duck recommend toilet-duck...

I would start with the PCB I posted in the 'build your own' thread, as this PCB (with external output stage) will give you a working controller. This you can then use to later debug whatever modular system you want to have, it gives you something to compare with. If you start with Zombiess mini brain board you're building something without actually knowing how a properly build controller should behave, you might think all is OK while it is not, or you'll be chasing down errors as you have nothing to compare your measured values/waves with...

13 WC eend act blue.jpg
 
I wish to convert all my EV's to lebowski-based motor control as well. You were wise to have seen the appeal to his system so soon.

I would love to see some well done timing advance options though Lebowski my friend. :)
 
If you do decide to start with the regular Lebowski controller and get a couple of boards made I'd happily buy a chip and a bare board off of you. :)
 
I got a couple chips, thanks Bas!

I'll be looking to print a small batch of boards for my own use now, but I guess I'll get a few extras and pass them along to folks who are interested.

I did also order a dsPIC eval board for $30ish as suggested by Arlo so I have a place to stick a chip and talk to it just to see how things are arranged internally. I'm hoping also to be able to scope some of the PWM outputs from this platform until I get my Lebowski brain board printed and populated.

I'll put up some pictures and videos as things progress.

Peace!
Chris
:mrgreen:
 
Ok,
Status update, I have chips, and boards!
I've downloaded Kicad, but it seems like I have a different version than Bas is using.
It doesn't have the check box for "one part per line" like he shows in the screen shot of creating the bom.
I get a very funny looking bom compared to the one he shows in the screen shots, it has two columns only, just shy of 300 parts, only a few part numbers.
Has anyone successfully created a bom for the latest version of the schematic?


I did print extra boards so I will have a few for sale, they will have a ( very ) slight markup to cover my time and energy to mail them out.
I'll post details as soon as I figure it out. No international shipping.
 
Beez65 said:
Ok,
Status update, I have chips, and boards!
I've downloaded Kicad, but it seems like I have a different version than Bas is using.
It doesn't have the check box for "one part per line" like he shows in the screen shot of creating the bom.
I get a very funny looking bom compared to the one he shows in the screen shots, it has two columns only, just shy of 300 parts, only a few part numbers.
Has anyone successfully created a bom for the latest version of the schematic?


I did print extra boards so I will have a few for sale, they will have a ( very ) slight markup to cover my time and energy to mail them out.
I'll post details as soon as I figure it out. No international shipping.

Awesome; just out of curiosity, what country are you in?
 
Oops! I'm in the US of Hey!
I thought that showed up....
I must have a flag set wrong somewhere...
Chris
 
I have similar problems as you Beez, but I'm lost even more because I have no experience with electronics at the level needed for for the Lebowski controller...
KiCad is installed (Build: 2013-07-07 BZR 4022 - stable) and shows Lebowskis genious 2.30 work. But I was not able to extract a reasonnably looking parts list.

-> Maybe somebody nice and capable in the forum (if not the dude himself...) could provide the service to elaborate a parts list "ready to order"?
If this would be in a .csv or .xls format the installation of KiCad wouldn't be needed if other nice people would offer the boards.
The best case of course would be to place this file at the beginning of Lebowskis thread http://endless-sphere.com/forums/viewtopic.php?f=30&t=36602 .
I'm quite sure that a handful of people would appreciate this and the hurdle for potentially new Lebowski disciples would be lowered.
 
emmgee said:
I have similar problems as you Beez, but I'm lost even more because I have no experience with electronics at the level needed for for the Lebowski controller...
KiCad is installed (Build: 2013-07-07 BZR 4022 - stable) and shows Lebowskis genious 2.30 work. But I was not able to extract a reasonnably looking parts list.

-> Maybe somebody nice and capable in the forum (if not the dude himself...) could provide the service to elaborate a parts list "ready to order"?
If this would be in a .csv or .xls format the installation of KiCad wouldn't be needed if other nice people would offer the boards.
The best case of course would be to place this file at the beginning of Lebowskis thread http://endless-sphere.com/forums/viewtopic.php?f=30&t=36602 .
I'm quite sure that a handful of people would appreciate this and the hurdle for potentially new Lebowski disciples would be lowered.
Where in CH are you ?

What is the problem with getting the .csv file ?

 
Thanks Lebowski! I'm in St. Gallen, so there wouldn't be a problem to come to Adliswil for picking up the chips directly at the source.
The problem is that I'm a total EE noob, never ordered a board or made a parts order at Digikey/Distrelec/Mouser, no scope or whatever. But I recognise the inherent beauty of your work and could handle the through-hole soldering.
I was able to extract different .csv files from KiCad (playing with the options), but they really look different from the one you placed. Yours will certainly bring me forward with the help of some colleagues and the Sphere.
 
Thanks Bas!
I don't know why I dont get email when I have replies to my own thread, probably a problem on my end though. :oops:

I'll be ordering parts today and populating board(s) soon!

Chris
 
Ok, maybe I won't be ordering parts today. I still have a few questions on the bill of materials. :shock:
I was hoping to figure them out myself, but I'm a mechanical guy, and electronics is not my forte.
So, What should I use for Q1, and Q2? Digikey has lots of options for the BC547 specified.
What about C51, C21, etc?
Finally, what about R84? Is this just a jumper?

Ok, that's it for the basic BOM, now I am trying to learn about the power transistors.
My initial controller is for 5000W continuous, 10kW peak. 100V.

The biggest MOSFETs I see that are reasonable are 500W, 150V. Do I parallel ten to get the 5kW continuous? Or 20 to get the peak rating?
So then I need 30 to 60 of them?!?!?

What kind of measurements do I need to make to "match" them and how do I make those measurements?

Thanks! :mrgreen:

Chris
 
good luck. this is a fun journey :) I feel I now know between 75 and 90% of what I need to know.
 
Beez65 said:
Ok, that's it for the basic BOM, now I am trying to learn about the power transistors.
My initial controller is for 5000W continuous, 10kW peak. 100V.

The biggest MOSFETs I see that are reasonable are 500W, 150V. Do I parallel ten to get the 5kW continuous? Or 20 to get the peak rating?
So then I need 30 to 60 of them?!?!?

What kind of measurements do I need to make to "match" them and how do I make those measurements?

Thanks! :mrgreen:

Chris
The 500W you found specify how much they can dissipate, not how much they switch.
e.g. An IRFB4115 is rated for about 100A at your 100V it would mean one switches 10kW DC.
Simplified: As this is AC, you need 2-3 per phase and side, so 12 or 18 FETs.
 
Beez65 said:
The biggest MOSFETs I see that are reasonable are 500W, 150V. Do I parallel ten to get the 5kW continuous? Or 20 to get the peak rating?
So then I need 30 to 60 of them?!?!?

What kind of measurements do I need to make to "match" them and how do I make those measurements?

Thanks! :mrgreen:

Chris
There is to many variables to get a direct answer.
I will tell you I have seen 250 DC amps from a 24 mosfet controller with irfb4110 mosfets in it.

To "Match them" You turn them on part way so apply a low voltage to the gates and measure the drain to source resistance. I did that in this video and this board did not have the drains connected so it was easy to use. I now have a lambda power supply so its a lot easier to get the 2-5v commonly needed. But if you get it in the sweet spot where its ~1/2 way on you will find the most variance between them and you can organize them accordingly. you will want to put the closest fets into groups to run them in parallel this way they will turn on at the same time. I miss read the data sheet it was 10v at the gates for 171 amps[youtube]mYcjsMmaWzc[/youtube]
 
Sooo... it's time to declare myself as another disciple (together with some friends).

- I ordered 10 boards at http://www.quick-teck.co.uk (a tip from marcexec, one board and controller will go to him) and already got them here for ~CHF 25.-/unit (equals ~USD 25.-/unit). For the order I only had to send the original .zip file (30F_pcb_v2p30.zip) from here: http://endless-sphere.com/forums/viewtopic.php?f=30&t=57877 over the quick-teck web interface. Then they checked the file and declared it as feasible. The boards look good, just for some reason the part numbers were not printed.
- 10 Controllers ordered from Lebowski incl. the programmed 12F617 for DC/DC control. Lebowski (thanks!) adjusts 4 of them for low-voltage use (he tested them down to 32V). 6 chips are for 60..150V use.
- LOTS of parts (around 300 per controller!) ordered at digikey for about CHF/USD 1'300.-, enough for 9 total controllers and 15 of the Lebowski-style 6-FET powerstages (using IRFB4110PBF-ND)
As I am unexperienced with electronics I really hope to have selected the right parts... I took everything I could directly from Lebowski's BOM .csv (the important things) and completed the non-specified items as good as possible.
Probably for some of you the complete Excel-list including part numbers and prices will be very helpful, e.g. digikey allows ordering with an .xls. Overall, the total cost per controller incl. powerstage is around CHF/USD 200.- as an estimation for other future disciples.

Cheers, Marcel
 
Time for me to join to join this group of pioneers :).
I already have a board from emmgee's batch above + chips from The Dude (thanks to both).

I will use TO247 MOSFETs, IRFP4568 (150V, 121A@100°C), which raises 3 questions:
  • will the IRS2186PBF gate drivers be able to drive the IRFP4586s? I found FAN7390N which can deliver 4.5A, but they are obsolete
  • do I need to add capacitance? Marcel, I saw that you already increased C2, from 100 to 220µF - I am thinking of a Nichicon UCS2D331MHD, which is 18mm as 16mm with 200V only goes up to 1.75A ripple currents
    for C1 will the EPCOS/TDK one you selected still be enough?
  • what current sensors should I use (I want to cater for at least 2 power "blocks", breaching the (+/-)200A limit for the Allegro ACS758ECB-200B-PFF-T?
    I checked the LEM HASS 500-S & the Tamura L01Z600S05 as examples - the sensitivity obviously drops, in case of the LEM HASS to a little more than 1mv/A, but they all supply 2.5V +/- the offset for the measured currents, right?
 
marcexec said:
I will use TO247 MOSFETs, IRFP4568 (150V, 121A@100°C), which raises 3 questions:
  • will the IRS2186PBF gate drivers be able to drive the IRFP4586s? I found FAN7390N which can deliver 4.5A, but they are obsolete
  • do I need to add capacitance? Marcel, I saw that you already increased C2, from 100 to 220µF - I am thinking of a Nichicon UCS2D331MHD, which is 18mm as 16mm with 200V only goes up to 1.75A ripple currents
    for C1 will the EPCOS/TDK one you selected still be enough?
  • what current sensors should I use (I want to cater for at least 2 power "blocks", breaching the (+/-)200A limit for the Allegro ACS758ECB-200B-PFF-T?
    I checked the LEM HASS 500-S & the Tamura L01Z600S05 as examples - the sensitivity obviously drops, in case of the LEM HASS to a little more than 1mv/A, but they all supply 2.5V +/- the offset for the measured currents, right?

I've run 4568's using the 2186 in 'the big lebowski' controller. You will need to use different values for the gate drive resistors as the 4568 have double the gate charge of a 4115. i would have the resistance values. The peak current for the 2186 is when it switches off the 4568, the current will be 15V/off_resistance.

What do you mean with 2 powerblocks ? You mean two groups of 6 FETs ? What you can do is use ACS758's in each powerblock, so 3 current sensors for each 6 FET stage. You should then give each ACS758 its own resistor in the output line to the controller IC (for R48, R49 and R50). These resistors should be increase such that the parallel resistance is 1 kOhm (so when you use 2 blocks of 6 FETs, you have 2 ACS758's connected to pin 2 (or 3, 4) of the controller IC via 2 resistors of 2 kOhm each). For the 2kOhm you can use the more common 1.8k or 2.2 k . For the controller IC you then have to use 5 mV/A, half the original 10mV/A for 200A sensors.

About capacitors, I would use kind of 'private' caps for each 6 FET stage, and connect the supply lines together at the battery using equal length wiring.

Zombiess uses similar current sensors as you propose in his latest airplane video, maybe you can ask him which type he uses ?
 
I'll go with the ACS758 for starters - change as little as possible and then work my way up. They are not that expensive in comparision. The Tamuras don't look too bad in regard to sensitivity, so I might try them once I actually have two working blocks :)
What resistance values did you use for yours? BTW will the IRS2186PBFs handle two blocks or would you use a 2nd trio for the 2nd block?
 
marcexec said:
I'll go with the ACS758 for starters - change as little as possible and then work my way up. They are not that expensive in comparision. The Tamuras don't look too bad in regard to sensitivity, so I might try them once I actually have two working blocks :)
What resistance values did you use for yours? BTW will the IRS2186PBFs handle two blocks or would you use a 2nd trio for the 2nd block?
I use 10 Ohm and 33 Ohm, so for the bigger FETs I would use half those values (4.7Ohm and 15Ohm). The main current spike for the 2186 comes when you discharge the gate (15V) through the turn off resistance (4.7Ohm). This would be a 3A spike, well within the 4A limit of the 2186. Good for one FET, but maybe risky for two, so I would use a second trio for the second block of FETs.
 
Lebowski said:
marcexec said:
I'll go with the ACS758 for starters - change as little as possible and then work my way up. They are not that expensive in comparision. The Tamuras don't look too bad in regard to sensitivity, so I might try them once I actually have two working blocks :)
What resistance values did you use for yours? BTW will the IRS2186PBFs handle two blocks or would you use a 2nd trio for the 2nd block?
I use 10 Ohm and 33 Ohm, so for the bigger FETs I would use half those values (4.7Ohm and 15Ohm). The main current spike for the 2186 comes when you discharge the gate (15V) through the turn off resistance (4.7Ohm). This would be a 3A spike, well within the 4A limit of the 2186. Good for one FET, but maybe risky for two, so I would use a second trio for the second block of FETs.

Thanks, that was a little challenge for me, so for those who want to follow the thoughts here, have a look at the following from the BYO Lebowski thread first:
gate driver.jpg
  • only the "inner" driver connections (P1/P2 in the example) are used - 6FET setup
  • 10Ω becomes 4.7Ω, 33Ω becomes 15Ω
  • for 6x IRFP4568 Powerstage R98, R100, R21, R23, R10 & R14 will be 15Ω; R99, R101,R22, R24, R11 & R15 will be 4.7Ω
  • briefly thought about going to 1/2W resistors, but since we're halving the value of them, the 2 effects cancel each other out :pancake:

I guess C67-C70 can be left alone even for the IRFP4568s, since you're not mentioning them.

Marcel (and potential disciples lurking), I have attached my latest working version of the spreadsheet including calculations for 96V on the last tab.
How are your controllers working out anyway?

Edit: Spreadsheet updated (can now export to Digikey BOM, some duplicates are created, though) & ordered myself as well :)
399-7736-ND CAP FILM 5600PF 5% 630VDC RADIAL "C43 C65" substituted due to inavailability of the smaller capacity type.
 

Attachments

  • controller_lebowski_230_digikey.ods
    4.6 MB · Views: 102
Hi Marc,
Unfortunately I can't mention any reasonable progress in the last few month and now I'm even more blocked with a broken hand.
The good news: We have 3 brainboards ready (and waiting...), mine is fully equipped with all the drivers for 4 x 6 FETs. Lebowski sent me the ultracool V2.40 of the chip and I modified the board a little bit to connect the temperature sensors. I'm sure that the new findings of Lebowski will solve the hickup problems we had in our test setup and the temperature control option is great.
The bad news: I'm struggling a lot with the power stages, obviously I'm not capable to solder the lebowski-style 6-FET stages together (fat wires, missing equipment and skills). Too much distraction from job, family and the hand problem. I have a provisional power stage, but it looks so horrible that I fear to damage the brainboard when I connect it.
Sooo: A bit blocked, but still very motivated. And to raise motivation: I fried the Alltrax-controller of my Quantya, this cries for another motor coupled with the Lebowski controller.

Nice that you want to try your setup with the 4568. With a nicely made power stage (keeping everything very very closely together to avoid inductance problems as recommended by Lebowski) this should give a good result.
Cheers, Marcel
 
Sorry about your hand.
When you're better, please post the changes for the temperature sensors. Anything else I'm missing in 2.30?

Parts from Digikey just arrived. I don't think I'll be ordering from them again - they are only cheaper because they don't charge VAT and send from the US so in the EU (despite ordering from the local page - .ie in my case and paying in €) you are liable for VAT & duty charges: 91€ order, 25€ extra (@23% in Ireland, 20.95€ of that is the actual VAT, so not too bad).
Anyway, some busy evenings ahead :)
 
Back
Top