PDA

View Full Version : Open source 3 Phase Inverter



tomdb
12 April 2014, 0906
I have obtained a opensource inverter kit. The hardware parts, without power electronics. I got my kit from : http://johanneshuebner.com/quickcms/index.html%3Fen_motor-inverter,8.html

The kit looks well layed out.

http://i35.photobucket.com/albums/d187/tomdebree/inverter/DSC_0095_zps0eaa1488.jpg (http://s35.photobucket.com/user/tomdebree/media/inverter/DSC_0095_zps0eaa1488.jpg.html)

Before the hardwork

http://i35.photobucket.com/albums/d187/tomdebree/inverter/DSC_0096_zps5b3f9856.jpg (http://s35.photobucket.com/user/tomdebree/media/inverter/DSC_0096_zps5b3f9856.jpg.html)

Result, the only thing missing from the control hardware is the "brain" an stm32-h103, which is still safe in its box.

http://i35.photobucket.com/albums/d187/tomdebree/inverter/DSC_0097_zps481b18cd.jpg (http://s35.photobucket.com/user/tomdebree/media/inverter/DSC_0097_zps481b18cd.jpg.html)

Mainboard, this houses all the interface for the signals to the stm32 and a powersupply (7-20v). There are a few in and outputs which are software controlable (contactors, precharge, lights) also the inputs are housed on this board (throttle 0-5v brake 0-5v and a few others).

http://i35.photobucket.com/albums/d187/tomdebree/inverter/DSC_0099_zps30f976d6.jpg (http://s35.photobucket.com/user/tomdebree/media/inverter/DSC_0099_zps30f976d6.jpg.html)

This is the driverboard, it converts the pwm 3,3v from the stm32 to 15v-/15v+ to drive the IGBTS

http://i35.photobucket.com/albums/d187/tomdebree/inverter/DSC_0100_zpsb6dfabf1.jpg (http://s35.photobucket.com/user/tomdebree/media/inverter/DSC_0100_zpsb6dfabf1.jpg.html)

This is the sensorboard, it contains two hall current sensors, a temperature sensor and resistor network divider for the HV side.

Besides the kit you of course need some HV goodies. For this i have an broke think city inverter containing 400A/600V IGBTS and capacitors, at my work we also have a few packs i can use as HV supply (350v nom).

This inverter is designed to run opensource software based on FOC code, so its designed to run an AC induction motor with encoder. Luckly at work we just have such a thing laying on the shelf left over from a project.

I started this project because i want to get my hands back on building/designing hardware and coding, currently in my role as engineer im mostly working in cad or having design meetings. Yet i love software and hardware.
The things i want to accomplish/incorporate into the project:

Get the AC motor spinning
Get the STM32 canbus working
Run the STM32 libary for PM motors and run a motor
Design a main board for ACIM and BLDC inputs



I will try to keep this topic upto date on my progress and findings, im open for help/ideas

SplinterOz
12 April 2014, 1638
Nice I am looking forward to seeing how it works out. A guy on youtube is making one of these for his BMW car.

frodus
14 April 2014, 1102
My buddy got that too... He hasn't built it yet but I got him some igbts a few caps... Looks like a decent diy setup.

Just make sure you do thermal calcs and get some decent cooling!

tomdb
14 April 2014, 1200
Today i had time at work to strip the inverter/ dc/dc converter / Charger of all its oem parts (except igbts) and start mounting the kit parts. Tomorrow im going to do some initial test to see if i can get a test setup working before investing in decent busbars and proper capacitors.

As for thermal calcs, none done yet. As the IGBT's are not mine and the same for the motor, i wont put any big loads on them for a while. Im planning on building a smd version of the hardware before i build my final inverter.

tomdb
15 April 2014, 0724
Progress today, i built up the complete setup. Don't mind the mess for now, its an ongoing project.

http://i35.photobucket.com/albums/d187/tomdebree/inverter/DSC_0118_zpsdc5dfdfa.jpg (http://s35.photobucket.com/user/tomdebree/media/inverter/DSC_0118_zpsdc5dfdfa.jpg.html)

The massive slab should have no problem absorbing the heat from the igbt's switching to get the motor to spin up with out load.

http://i35.photobucket.com/albums/d187/tomdebree/inverter/DSC_0120_zpsd553b58c.jpg (http://s35.photobucket.com/user/tomdebree/media/inverter/DSC_0120_zpsd553b58c.jpg.html)

The test beauty, weighs about a kilo or 40. How ever the rotor spins with ease.

Hit a few snags today, my current sensors are reading all over the place, so its time to swap the flatcable for twisted pairs. If this doesnt reduce the issue i will have to check my soldering and the rest of the circuit for faults.

Arlo
31 May 2014, 0652
I love DIY controllers. Any progress lately

tomdb
31 May 2014, 1211
I got an working scope. But the scope is at my parents and the inverter at my work. And were in the middle of switching housing at work so not allot of time to work on it.

tomdb
18 March 2015, 1556
Good news, I switched jobs and got tied up in doing other things. However today i had sometime to spend on my controller.

It works, atleast the complete logic bits, i get a nice squarewave -15 +15 volts.

This means its time to start on building the HV components and casing for further testing.
Parts ordered or on the shelf;
-heat sink aluminium billet (will mount igbt's this weekend)
-plastic for more permanent logic hardware setup.
-Snubbers and capacitor for HV bus
-Brass bus bar material on its way

I now have to get a test motor organized and power supply to continue my testing and experimentation.

tomdb
21 March 2015, 0805
Built the high voltage section for my controller. The heatsink is just a solid chunk of aluminium which I will mill on a later date.

Now i have to find a test motor to make any further testing possible.

According to the igbt specs and cap limit, it should take 400V and 400A (160kw).

I wonder how close i can get to this figure.

Next on the list is mounting the logic hardware over the HV hardware.

SplinterOz
21 March 2015, 1355
Great progress... looking forward to seeing the motor test.

Athlon
21 March 2015, 1425
where are you from , I have a 288VDC induction motor floating around the workshop , is heavvy (60 kg) and not powerful (15-16 kw cont liquid cooled ) but for testing I think can be a good option

tomdb
22 March 2015, 0515
I just got lucky yesterday. Via the dutch Ebay (marktplaats); I found a motor being sold not further then 2Km from me for a very reasonable price.

This means I can start on my motor testing.

@Athlon Im in the Netherlands.

tomdb
23 March 2015, 1301
It goes around and around.


https://www.youtube.com/watch?v=Y7_Zeq-lusg

This is some good news, i just need to get on with mounting my logic hardware so i can easily test it in my garage. Besides that ill need to get some suitable connectors to wire this puppy up.

Biggest issue I am having right now is that my scope freaks out because its plugged into a ungrounded socket.

The other issue im going to run into soon is that i need a capable power source for testing. So i better get looking for a suitable test pack.

Athlon
23 March 2015, 1350
for testing the best is to have 2 motors bolted together one in traction and one in regen , so with smal battery and small charger you can test very high power and also you have a real easy way to measure efficiency , a goo efficiency mean a good tune of the motor/inverter setup

tomdb
23 March 2015, 1435
Yes, this is my eventual goal. However building a simple kart/bike will be the simplest test bench for me, with my limited space.
But im planning on reworking the schematics into new boards using smd components/compacter, so i can have atleast two inverters for testing. Plus i have some of my own idea's for inputs/outputs.

tomdb
23 April 2015, 1414
Made some progress

Working on a simple bms system for test pack i am building. I etched my first own designed pcb. First of many im sure.

Got my hardware in for the encoder setup, got it wired up and it checks out with reasonable values. Maxed it out at about 800 rpm at 31 volts.

Tomorrow im wiring up an "throttle" simple potmeter with a handle for testing sake. When I get my motor responding to the throttle ill look into hooking up a permanent connector and fitting a case for some real load testing.

However in order to do any further testing I will need a suitable power supply, this is why I am doing a little side project on building a spot welder. Once I start to assemble parts of it I will start a little thread to discuss this adventure.

tomdb
24 April 2015, 1412
Good news everyone, it spins in response to my "throttle" input.

Setup an encoder setup, this is quite prone to damage. Now i have proven this works, its time to create a permanent mount with shielding for the wheel and sensor.

I cant get the motor to stop when running from the power supply, I think this has something to do with the low voltage and the in ability to feed current back into it. This means i can get going with building a test rig with batteries for the next stage in development. My next goals is to get a case setup with a proper connector in order to get it running without a laptop input. I will also get going on working out my battery pack buildup, got a lot of old 18650's for laptops.

Now the question becomes what kind of testing rig/vehicle am I going to built, two or fours wheels.

tomdb
16 May 2015, 0703
Alot of head scratching and thinking I figured out what was my issue. After talking to the guy who made the kit and sharing ideas, the conclusion was the most likely the motor is not a four pole (two pole pairs) motor but a two pole (one pole pair).

Now it spins and acts like you would expect.

https://youtu.be/Hkmf01rS6E0
Sorry for the bad video, my phone was acting up.

It took a little longer then i had hoped. However while thinking and pondering the issues I had time to do other things.
- I made an enclosure that will cover the inverter for vehicle testing. I still have to wire up an connector

Most of the time i spent think up and building my spotwelder. The control portion works along with the switching electronics, tested these with a light bulb. Waiting on parts to complete this baby and then ill make a new topic for it.

onegreenev
22 May 2015, 1759
Just wanted to poke in and let you all know I have a running inverter from Johannes Hubener and it is working well. My latest issue was the drive end of the motor shaft getting hot but was told it was because of the high frequency I have been using. With the low voltage I wanted this sucker to spin up fast. Well, it does but in the process heats up the drive end of the motor. So if I really want to the motor to spin fast I need to just at voltage. I will bump the current once I get the load source connected up. Until then you can check out the videos I posted on youtube. I mostly frequent DIY Electric cars but came over here for a peek and saw that a few of you are doing the same inverter. It is a fun project and because its been successful Im moving forward. I also have another inverter to build. I already screwed one up.

Pete :)


https://www.youtube.com/watch?v=NfOHT2Jck90

tomdb
03 June 2015, 1155
Got my self abour 12 kwh of A123 cells to get me started on a test pack, I can vary the voltage from 50 to 500v depending on how many modules I decided to add in series. This will become my test pack.

This weekend will be used to get my inverter built into a case and running as would be in a vehicle.

tomdb
04 June 2015, 1427
got to balance these packs first before going further. I need to source their tiny connectors with roughly 1mm in spacing 2x10. In order to use the integrated temperature probes and voltage sensing.

However if i cant find them ring terminals will have to do.

65556554

Dimensions are 125mm wide 165mm deep and 235mm high excluding the interconnection posts.
I was told these are 16s configuration however i highly doubt it. 18 terminal clamps do not equate 16s, and my voltage measurements do confirm this.
This would mean 56.1 volt nominal with an capacity of roughly 1.1 kwh (17x65) with a weight of below 9 kg's. I could not be happier with this purchase. ;)

tylerwatts
09 June 2015, 1535
Hi tomdb. May I ask how and where you sourced those batteries please? They look great and perfect for a motorcycle conversion.
Regards

Sent from my D5803 using Tapatalk

tomdb
09 June 2015, 2311
I got lucky a student project was selling them off.

tylerwatts
10 June 2015, 0301
Oh nice score

Sent from my D5803 using Tapatalk

tomdb
13 June 2015, 0843
The batteries are going to take some work in adapting their tiny 1.27mm spacing connectors to 2,54mm headers for my cellog 8s "isolation board". This means more pcb design with the possibility of outsourcing due to the size.
My largest question is now to integrate everything on one broad that plugs straight into the module or make an adapter for my existing design. The choices :cool:

Right now i am spending most of my effort on getting the inverter mechanically finished. This means cutting holes in my fiberglass case and painting it. Also milling the base plate with cooling slots.

I had a little delay with this, the webshop was out of stock on the required tap, this was only listed on the invoice with the other parts in the box. This means another week wait before i can finish the base plate.

However progress maybe slow but I am enjoying designing and improving my skills in the field of ; mechanical design, electrical design and software.

65706571

tomdb
03 July 2015, 1213
It is way to hot in the Netherlands to get alot done in the shed. Getting closer to having the inverter "finished". Mechanically that is ready for testing.

Added a 25 pole connector, one the clamps onto a flatband. No more messing with those tiny wires. I will also be mounting a connector to hook up the serial port of the micro controller for ease of read out. Maybe ill just mount the usb to serial converter in the casing for the ease of it.

Once the inverter is back on the heatsink i can make a testing harness for some bench test with all the additional funcions.

All i have to do mechanicaly is mount the back plate to the heatsink and seal it up and she is ready for assembly.

6621

6622

tomdb
13 July 2015, 1334
Some more mechanical progress. This week I can finally try it all out again.

1st order of business was leak testing; passed.
2. clean igbts and heatsink then mount with thermal compound
3. Hack up my nice box so i can actually mount everything, this is the reason for the two interface plates with glands.
4. assemble, this needed some tweaking to get it all to fit. All is in now and snug
5. Connect the 25pin flatband connector to the flatband.

Next thing to do is get an usb port wired up for the serial converter. Then wire the input connector so i can test the inverter and upgrade the firmware.

Besides this I am busy working on my own bms boards, this has taken more time then anticipated. Because I am also researching into charger design, smd and programming.

66356636663766386639

tomdb
13 July 2015, 1336
more pics

66406641

tomdb
15 July 2015, 1344
small deviation. Working on my bms slaves.

Got code running between arduino uno and ltc6804. Proves my wiring schematic works and my 0.5mm soldering worked. Now to finish my pcb design and assemble it.


66446645

tomdb
26 October 2015, 1256
I have been hoarding projects like the world is going to end.

After sending out for my PCB's for the bms slaves again and working out some code. I dove head first into an expansion of this project, reusing existing automotive inverters with this control board.

Started a thread on diyelectriccar. http://www.diyelectriccar.com/forums/showthread.php/honda-ima-163650.html

About reusing honda IMA generation 2 components form the likes of civic hybrids (plenty scrapped here in holland). The power inverter cost me 100 euros secondhand, and the inverter board ran me roughly 60 euros in parts.

With an helpful english fellow, who tore his inverter to bits and took photographs I reverse engineered the gate drive signals. Soldered on some wires, hooked up the inverter board and spun my my three phase induction motor. :p

Taught you guys might like this, since building an power stage from scratch is the hardest and most expensive part. Got to get my batteries ready for a power test, and possibly build a dyno with my two working motors.

6776

6777

tomdb
10 December 2015, 1321
I have not posted about my inverter adventure progress but here is an update; built a semi permanent setup, 17s pack of a123 cells as test pack to start, right now just load testing the IMA hack. If only I had a way of building a dyno

Have been looking at getting more test motors so this a better excuse then ever. This setup will also allow me to test different motors and voltages. A small brushless motor is hooked up, I will run the bldc software I aquired from Damien as a start but will look into getting the STM bldc motor libary running on one of my H103's. Besides those chips I put in an order with olimex for two H405's, porting the code should not be too hard. Plus working in other aspects with the improved hardware should be a nice project.

However a test bench like this will also allow me to test my BMS system, I will run two board (12s each) in series with a simple arduino with a touch screen as BMS master. This way I can easily monitor the pack whilst testing the BMS hardware and software. Guess by the time I get an vehicle most of my testing will be done.

The hack of the inverter was quite easy, I can only recommend getting your hands on some automotive grade high voltage hardware.

tomdb
14 October 2016, 1320
Same story as my BMS, not alot of time and moving to a different country really takes time.
However I got some time to setup my inverter and motor test bench again.

This is not the same brain board as i was using before, this time I used the Lebwoski chip I got off a fellow dutchman on endless-sphere, spun my own pcb to get it nice and small. The test went well, well enough for me to commit to getting the pcb installed inside the OEM housing and then doing some more testing and tuning of this controller. I would believe this to be quite a nice compact and cheap solution for anyone wanting a 200V max system, not sure on what kind of peak loads the Honda IMA inverter can do. Seems like this package however would be very suited to a nice compact bike, lets say 20-30kw peak (the ME1302 would be ideal candidate).

One small problem, I dont have a motorcycle license of right now so I will have to devise another way of testing these inverters. My collection now has a DIY Johannes Heubner Induction motor inverter, Honda IMA based Lewboski inverter and a Rinehart pm100DXR (had to buy it because the price was right).


https://www.youtube.com/watch?v=deUlqjwdXz8

SplinterOz
14 October 2016, 1516
Neat. You are now the 2nd person I know using that Brain. What are you using as the IGBT driver circuit? the OEM one?

tomdb
14 October 2016, 1538
Yes i am just feeding in 5v pwm signals (6, high and low per phase) and tapping off the analogue output from the phase current sensors.

Got a document with all my reverse engineering I did so far, if you are interested I could email it. Quite a neat package that is relatively cheap and available secondhand.

SplinterOz
14 October 2016, 2234
Cool sounds like I need to get me one of these brains! and one from Labowski

tomdb
15 October 2016, 1226
Tested the controller with two packs in series to give me roughly 110V. Oh boy does this motor want to go now, need to ofcourse do some tuning on the motor controller combo once I am happy with everything.

My homemade pcb fits in the oem enclosure after some dremeling away of the reinforcing ribs in the plastic. Looks almost stock except my rats nest sticking out of it. This will be sorted once I get some decent wire, already have a assortment of ampseal connectors which are perfect for this.

Had some small issues with the Honda IMA inverter shutting down, turns out the oem controller got spooked by the pulses it was seeing. So I cut the traces, now this puppy revs even harder. Next tests are at 165v and 200v however these might get into oiling issues.

Think the next real step should be getting the motor mounted in something to start testing.


https://www.youtube.com/watch?v=zTDh4V0WTe4

tomdb
30 October 2016, 0434
Did some more playing with the controller. Got a new motor today, deal was too good to pass up, a Lexus rx350h/rx400h rear motor.

Did some playing with the controller, turns out that I set a parameter wrong and that why it was giving me some issues. Talked to the creator of the chip and he pointed me in the right direction, she spins fine now.

Working on making some busbars for my battery so I can do some higher voltage testing.


https://www.youtube.com/watch?v=exbnxTDsoy8


https://www.youtube.com/watch?v=A5JDapL4EsM

frodus
31 October 2016, 0721
Great work!

tomdb
06 December 2016, 1437
https://www.youtube.com/watch?v=pgPT-Rt6hKk

The Honda IMA stage kick up way to much EMC so looking for a true full metal cased inverter. Have a Rinehart pm100dxr laying around, but that will be kept for when the diy option fails. Love the challenge of building my own solutions.

160v is the max I can get this setup to work at reliably, will probably mount this setup on a small bike or a kart, with a different motor.

tomdb
06 January 2017, 1337
I have bought a new inverter to use as the powerstage this time from a chevy volt / opel ampera.

Two 600amp stages so it should be good enough for what ever motor I currently own. Took me some time to realise I had forgotten to hook up resistors which create a current draw from the current sensor output, this caused alot of issues.

However this is now resolved and she spins nicely. With this part of my diy collection working I can move on to actually getting something moving under electrical power.


https://www.youtube.com/watch?v=7cJJQT425fo

Spoonman
09 January 2017, 1000
That is wicked dude - nice work!