Axiom: a 100kW+ motor controller

Nice job Marcos looks good! Any release date prevision for the future magic?
Oh, its not an earth-shaking feature, don't worry. The focus now is in the packaging so we can offer more than a populated OSH pcb. No release date, I want it now so we can sell and fund more development, but its not that easy, lots of things can go wrong in the prototyping.
 
ok no problem, I thought at first that you were talking about your FPGA implementation stuff... I was kind of expecting too much :lol:
 
Marcos,

How have you implemented/automated precharge of DC-link on your side with this board? Does the protected GPIO OUT1 & OUT2 from your VESC board are for driving a contactors & a precharge relay? (Power is limited I think...)

I'm asking that because I started to make my own idea on this. I am not planning to use an interface board as you probably know and I have to create a separate PCB for DC link voltage measurement (as already discussed), so I would like to regroup more functions on this board (including precharge). This board would have those functions:

- Automatic precharging the DC-link through an external single IGBT module (automatic Inrush_Limiter) see: https://endless-sphere.com/forums/viewtopic.php?f=3&t=40142&start=150
- Driving the same IGBT module for DC link current & E-STOP (replacing one main contactor)
- Discharging the DC Bus when power is turned off
- Measuring the voltage of the DC-link

I just don't know if my idea hold with the inrush limiter & losses in the IGBT at this scale... I started to work on this idea today:
https://github.com/EnnoidMe/ENNOID-Dual-Gate-Driver VBUS-Sense folder
 
Yes, to get my hands on the FPGA I need either free time, a huge motor or someone to pay for the job, its a ton of effort for 1 person.

The protected GPIOs were added for fan/pump and relays/contactors, so yes, the plan is to use them. DC discharge should not be in control of the microcontroller.
DC link voltage is already measured. I wouldn't use an IGBT as a contactor, most of the times an IGBT fails open, but what if it doesn't? Also the losses are not low.
 
Just went to print this from oh park and looked at the bom. Realized I should ask if you happen to have any board sitting around you would like to sell. I'd be interested in a completed board but even a bare board would be fine (I don't mind doing the pick & place And reflow soldering), bare bone with needed parts would even be better. Basically needing the controller and powerless board. I already have the same igbt's.

I'm at the point where I need to add the motor and controller to a jeep conversion I've been working on. I have a 120kw motor picked out but want to try the controller on a 5kw motor before I purchase the larger motor
 
Marcos what if Youtuber Gerard Morin is right, that these larger drum motors work best with higher voltage + higher frequency rather than higher current? How easy would it be to modify to have lets say an output of between 240-500volts and a frequency range between 15khz - 25khz. In one of his last two videos he is getting what seems to be a tremendous amount of torque from less than 1 amp draw! :shock: Surprisingly I have not found other individuals working on anything like what is described. There is one company (Elmo) that has a controller for high voltage (Drum HV). There is also a company (un-named) making large manned drones in Isreal utilizing such technology. Any thoughts? Can higher volt lower amp mods be made easily on the VESC?

Doug
 
dmann said:
Marcos what if Youtuber Gerard Morin is right, that these larger drum motors work best with higher voltage + higher frequency rather than higher current? How easy would it be to modify to have lets say an output of between 240-500volts and a frequency range between 15khz - 25khz. In one of his last two videos he is getting what seems to be a tremendous amount of torque from less than 1 amp draw! :shock: Surprisingly I have not found other individuals working on anything like what is described. There is one company (Elmo) that has a controller for high voltage (Drum HV). There is also a company (un-named) making large manned drones in Isreal utilizing such technology. Any thoughts? Can higher volt lower amp mods be made easily on the VESC?

Looks like the usual sort of free energy bullshit artist, displaying results with huge emphasis on one variable to appear impressive.

1 amp of draw at 400v is 400w. That power level can send your average ebike climbing the wall and leaving skid marks on the roof. It doesn't matter what the input draw is to the inverter, what's creating torque is phase amps. With a beefy enough controller, that 1a of draw could become 300a through the windings. I could set up a cheap 30mm wide 205mm ebike motor on a stand and run it at only 200 rpm and be completely unable to stop it turning by hand.

He has not invented anything, what he's demonstrating is a sine wave 3 phase motor controller with basic speed control on a cheap outrunner. The talk of resonance and RF frequency is distraction and completely incorrect. He shows input current *while motor is unloaded*, that has zero relevance to efficiency under load. Don't be taken in by this BS artist.
 
First thing I read from that guy: "The purpose of this campaign is to raise money for replicating Gerald's Energy Amplification Experiments.". Can't read past that.

Shaharov, I still don't have any spares left, I was ready to hit the big red "MFG" button and 2 customers resumed testing with >100kw motors and now I want to have their setup tested at full power before making new batch, these are expensive builds and any extra test is very valuable given how scarce test benches are. Last week I found 1 very minor issue in the hardware, and there is an ongoing firmware bug with rotor position algorithm.
 
marcos said:
Shaharov, I still don't have any spares left, I was ready to hit the big red "MFG" button and 2 customers resumed testing with >100kw motors and now I want to have their setup tested at full power before making new batch, these are expensive builds and any extra test is very valuable given how scarce test benches are. Last week I found 1 very minor issue in the hardware, and there is an ongoing firmware bug with rotor position algorithm.

Good to know... Keep me updated. I'm keen to get this started but have a few things to do (including a holiday/trip) before I can start testing with the Controller. Basically, once the controller is tested with a 5kw motor I will be purchasing the motor for the vehicle and there's not much left to prep for that. If I get to the point where I need the controller to continue I'll get the boards made and order the parts... you can then have some more bench testing with a 5kw and >100kw motor... I do like the Vesc and am eager to see how this runs on a larger motor.
 
I Marcos,
First congratulations for your job.
I'm a specialist in electrical motorisation in Switzerland.
I'm always looking for new products to improve the performance and I'm looking for high voltage bldc controller.
Do you plan to sell your board already assembled or a full kit?
The cost of a comparable controller is between 3000 to 6000usd
Do you think your system will be competitive?
If yes, I'm able to buy a first one to do some test in electric car.
 
Hi Marcos,

Was going over the schematic and pcb layout and was wondering the reason for doing a 4 layer board instead of a 2 layer board? I was able to rework the pcb into a 2 layer fairly easily but thought there may be a reason why you decided on 4 layers
 
We should be in that price range for a basic but fully assembled 100kw drive, with a high performance version to come, but I'm lagging behind with my dayjob, I made some upgrades to the schematic that need to be synced into the layout. PM me if you are interested.

About 2 vs 4 layers. First it takes me *much* more time to properly route a 2 layer board, if going 4 layers saves me 2 hours it has already paid for itself. Last but not least, its about current loops, show me a 2 layer board and I'm sure I will find current loops that either pick up EMI or generate EMI. In this case we care about not picking up because the control board is very close to high power switching with no real RF shielding in between. You can very easily have a critical analog signal compromised by EMI, an overcurrent fault (hardwired protections are not sampled in sync with pwm), or an MCU reset. Even an mcu latchup could be a possibility in a high noise envirornment like this.
4 layer board isnt that much expensive,I just paid $7 for a much bigger 4lyr board (ignoring tooling) thats nothing compared to the BOM of a system like this.
 
Is this project still active? I'm evaluate this for a cruising sailboat. A generator and a VESC controlled motor, and some battery. May be it can be efficiency, nonoise... I spoke about a big boat, from 40 to 45 feet, it need a 40 or 50hp motor normally.
 
I believe Marco is still working on it... but real life does get in the way sometime. I know I'm using the current design for what I'm doing, just not quite ready to do the circuit board yet. There hasn't been much time past since the previous responses in this thread.
 
He's right, still high priority, but dayjob and newborn needs me full time for a while. I do have some schematic changes commited but not pushed, I wanted to do the layout of those changes myself.

First of those changes is the comparator IC, I was using them with the common mode voltage out of datasheet spec and could trigger unwanted faults. I informed beta testers but didn't post it here, and I'm not 100% sure about a replacement part#, the comparators with rail to rail input I found are those nanopower stuff with very weak outputs. They work... but I'm not sold.
Second bug is in the resolver sheet, SEO is an input, not an output and should be pulled down.

On the bright side, next version comes with some slight improvements, and I've seen the github board doing solid control of 130kw at 450A. The time I can spend in this I put it in math support and testing, though I'm going to need way bigger motors to level up my testing.
 
Hi,
How much the whole assembly cost with controller
board,resolver interface ,DC link,igbt and gate drivers & heatsinks
Can you connect to me on adityajoshi1904@gmail.com
 
Witaj Marcos
Wspaniała praca .
mam dwa silniki 15KW pracujące na napięcie 30-70 volt i natężenie do 240 Amper KV 140
video1- https://www.youtube.com/watch?v=exJB4CINHs8
video2- https://www.youtube.com/watch?v=o2PyM_OupfE&t=96s
I had two Chinese ESC drivers but they all exploded :)
or smaller motors than 150 kW, for example 15-30 kW, you can use such a driver and whether you would recommend IGBT or mosfets.

I am interested in buying this your system :) !!!!!!!!!!!!!!!!!!!!!!!!!!!!!
 
Hello Marcos
Amazing work .
I have two 15KW motors operating on 30-70 volts and up to 240 Amps KV 140
 
How much the whole assembly cost with controller
board,resolver interface ,DC link,igbt and gate drivers & heatsinks
My BOM says $1251, but it ignores many shipping charges, taxes, custom enclosure, bolts, connectors, cable glands, busbar assembly, manual labor and testing, so its hard to tell at this point.
 
Hello i am so new in this forum, and my goal is to get an open source electric car, the chasis cames from another foro and i am looking for a motor controller and a motor, and my cuestion here is.
If i got one motor in each side in the back line, must i need 2 controllers? one for each motor?
 
lhfever said:
If i got one motor in each side in the back line, must i need 2 controllers? one for each motor?
Short answer: you need one controller per motor. Long answer, 3 phase out of the controller is tightly in sync with rotor position, and I bet your motors will spin independent from each other, so yeah, one per motor.

I'm starting to spend a bit of time on the FPGA part of the board. I'm leaning towards an ICE40 FPGA from Lattice. I'm not fully settled about the FPGA part, my first choice would have been a spartan 3 from xilinx because there are many examples and there are AEC qualified parts, but they don't come in small packages, and honestly, the MCU is not AEC qualified so it would be a long shot to qualify the board right now.

On the other hand, ice40 fpgas come in small packages (QFN48), have a reasonable operating temperature rating, and there is a neat software package to develop the HDL for them which is super clean and easy to install, and also happens to be the only fully open source toolchain. For now the use it is getting is pretty much a CPLD, but a 5000 gate FPGA can do much more and I'm excited about the possibilities, I replaced all the discrete logic using only 0.2% of the available gates.

shoot through elimitation.png

In minutes I set up the logic, validated, sinthesized, and simulated and it appears to work as expected. My next digikey order will include an eval kit to ensure this part is a go. If someone has a better part# I'm all ears, I've been out of of the programmable logic business for years.
 
have a look at the IGLOO nano FPGA's from Microsemi

pro's: (vs ice40)
Flash based (reprogrammable) vs SRAM (so no need for external Flash if you want to be reprogrammable)
Simple SPI ISP programming (FTDI SPI chip is all it takes)

con's: (vs ice 40)
closed source IDE
 
Thanks for the tip. My aerospace friend suggested Microsemi as well, even over xilinx and altera. I've been told they do great under radiation, maybe not the igloo but speaks well about the company.

The main con's vs lattice is that they are not stocked. Where do you order a qfp48 igloo nano in low qty?
 
too bad only arrow stocks the part.. on the other hand.. i think they have free shipping with only a 20 dollar minimum order or something
https://www.arrow.com/en/products/agln010v5-qng48i/microsemi


yea i think they have a whole range of rad hardened FPGA's.

i only have some experience with a smartfusion2 from microsemi. (in a big ass BGA package) not the igloo nano.

the software is not the greatest. but works well enough. only pain in the ass is the licence (free) which you need to renew each year.
edit: just had a look at the software you used.. AWESOME.. clean and simple! libero IDE from microsemi is a big pain in the ass compared to this. (for simple projects)
 
Back
Top