Well, someone else, maybe one of you, posted a superb timestamp summary in the comments on the video a few days ago, beating me to the punch.
Great job.
I'll repost it here for anyone who's curious. He was heavy on the timestamps, it's too many for me to link each one, you'll have to scan manually.
-----------
FOC: Field Oriented Control. HV: High Voltage ~80-200V. IM: Induction Motor.
OVERALL: discussion is about Toyota parts, but is generic for almost any permanent magnet motor which will need FOC software.
TEST EQUIPMENT: <<INEXPENSIVE>> It is assumed that only basic equipment is available such as light bulbs, switches and multi-meters. (12 V wires can be automotive grade, but HV cables MUST BE RATED FOR ~ 300 Volts).
0.51: Connections to VCU for Toyota Gen 2 Inverter.
4:27: Overall schematic of inverter and VCU Gen 2, 12V wiring including external items such as throttles and resolver connections
4:55: Discussion about VCU Gen 2 PCB components: purpose of; location of power supply ccts; pre-charge/charge circuitry etc.
6:21: location of solder jumpers for 'open collector' requirements with Asynchronous (IM) motors with encoder feedback.
6:37: VCU Gen 2 connections to inverter Gen 2 and trans-axle Gen 2.
7:08: Grey connector for throttle, brake, 12V power etc.
7:15: Black connector goes to Gen 2 inverter.
7:24: How to SAFELY apply 12V power to VCU Gen 2 for reducing damage from mistakes such as 12V to wrong pin...
9:46: 55W lamp, car battery, ideally fuse (10A?).
10:30: VCU energized: should be two LED s light steady, and third one flashing (shows STM IC executing software,
).
11:14: Initial hook up to Gen 2 inverter. Bypass inductor, Use 80-160 V DC supply initially. Use ~230V incandescent light bulb.
11:21: Signal harness has been connected at this point in time.
11:41: 2 pole connector is 12V supply to Gen 2 inverter. Watch polarity.
12:45: VCU Gen 2 AND Gen 2 inverter connected to 12V supply. Light bulb behavior slightly different to only supplying VCU Gen2.
13:00: Small glow of lamp due to quiescent steady state current consumption of Inverter Gen 2 electronics (Gate Driver board).
14:01: Bulb glowed as expected, can connect 12V supply directly, bypassing globe. Please use fuse in series with supply to protect wires.
15:04: Trans-axle Gen 2 temporarily wired to VCU Gen 2, HV battery, trans-axle resolver connected to VCU Gen 2.
15:42: Only MG2 of Gen 2 trans-axle connected to MG2 output of Gen 2 inverter
16:13: Trans-axle planet split device has been welded together and the trans-axle bolted back together. No oil in trans-axle.
16:27: MG2 and MG1 now spin at same speed all the time. See other videos for info on welding.
17:00: Stubs for trans-axle differential output were not available, so Damien put tape on the petrol connection for seeing rotation of the axle.
18:20: Discussing running trans-axle in open loop mode first, as part of FOC Tuning process.
18:43: HV battery connection to two bus-bars, (bypassing orange input cable port) to bypass buck/boost inductor.
19:04: Build up commissioning plan using the most simple steps conceivable. Can always add other steps later.
19:17: Bus bar closes to black capacitor square is HV positive, Bus bar closest to aluminum die-cast box is HV negative.
20:05: HV negative wire connect directly to HV battery pack (or equivalent) [Remember to use HV rated cable
]
20:51: Recommendation: use HV battery rated 80-160V (150V?) good value...
21:17: Use ~20 Amp fuse HV rated and fuse holder HV rated, light bulb 230V rated 100W HV rated, bypass switch HV rated.
22:03: Fuse protect wires from shorts (does not protect electronics). Switch open for pre-charging inverter capacitors and seeing inrush current flow into inverter electronics and seeing how current flows into trans-axle during early commissioning phase.
23:03: If you don't understand pre-charging inverter capacitors, please pause watching this video and learn what this means first.
23:41: HV light bulb limits any fault currents that may flow during early commissioning phase.
24:08: HV light bulb protection for limiting fault current for wiring inverter incorrectly, inverter faulty, battery wired wrong polarity, any other unplanned conditions. (Please note, shorts on HV battery circuits can carry thousands of amps, with burning and UV flash-over)
25:36: Negative battery connection crocodile clip attached to HV battery (HV positive connection was already connected). Result, light bulb flashed for <1 second. Means pre-charge current for inverter capacitor Gen 2 flowed, then once caps charged light bulb went out.
26:50: Total system of trans-axle, inverter, VCU, HV battery , LV battery is now ready for overall system commissioning.
27:05: Now go to computer talking to VCU, and set f to 3 Hertz. And Ampnom to 10%, Can hear whine coming from trans-axle.
27:35: Ampnom ramped to 20% then 25%, now get petrol axle jittering and trans-axle rocking back and forth.
27:42: Ampnom set to 25%, and now we see axle rotating slowly in open loop mode. PROVES: inverter/trans-axle/HV/LV Battery/VCU wire connections system WORKS as a whole. We are now READY for the next stage of FOC tuning.
28:20: We observe HV light bulb is glowing dimly, because we are consuming current in the transaxle rotating.
29:09: Ampnom is now set to 0 amps via the laptop. We can now close switch to bypass HV light bulb, and then continue open loop tuning.
30:05: The spinning petrol shaft axle means: inverter/trans-axle/HV/LV Battery/VCU wire connections system WORKS as a whole.
This is a significant step.
30:32: The next step is to commission this permanent magnet trans-axle using the Openinverter forum FOC firmware.
31:19: For those commissioning their permanent magnet trans-axles, please pause what you are doing and carefully watch this entire tutorial
.
32:00: Demonstration of resolver circuit. Sinusoidal excitation signal verified with a tiny second hand loudspeaker.
32:33: Set 'encmode' to 'resolver' on laptop.
32:59: How to lash up 12V supply to VCU on bench to test excitation output signal. Connected pins 4 and 5 to loudspeaker.
32:42: Discussion of how to get the 6 wires of the trans-axle resolver connected to the correct pins of the VCU.
33:58 Resolver discussion applies to practically any permanent magnet motor.
34:20: Need paper and pen, basic multi-meter, wires to connect to trans-axle resolver plug (hopefully you have the mate for the plug)
35:25: Without any resolver wiring diagrams, point to point sets of three pairs of wires. Write down resistances of pairs of wires. Lowest resistance pair is the excitation coil pair. This applies for any make resolver.
38:26: The two high resistance pairs are the sine and cosine pairs of wires. They have almost identical resistance.
39:23: Which pair is the sine or cosine pair does not matter for the Openinverter forum FOC firmware. Therefore you don't need to lose time trying to work this out. There is a video by Johannes that describes the maths for this.
40:01: Need to make chart in the notebook for 4 possible permutations of connections for the sine and cosine pairs. Need to connect the same polarity of the sine and of the cosine coil to the ground return connection.
40:40: CRITICAL: Only 1 of these permutations will work your motor properly. The rest will only make your motor partially work. Even if you have a schematic from the manufacturer, you still need to go through this work.
41:44: Strongly recommend you bring out the resolver wires from the VCU and connect to some connector block with at least 6 tunnels in it. Put exciter coil wires to the right and the rest to the left. To facilitate changing sine and cosine coil polarities.
42:19: Wire the exciter coil of the trans-axle resolver to the exciter output of the VCU. Please note, polarity of exciter does not matter
. Because all you are doing is swapping polarity of BOTH sine and cosine coils, which does'nt matter.
44:32: Now we look at outside world connector of VCU. E.g.: Pin 1 is sin resolver connection.
47:40: With correct resolver connections of the system and tuning of the FOC firmware, end result commissioning is demonstrated by easily changing speed of the trans-axle axle by simply depressing the accelerator pedal.
48:35: Damien will post a parameter file on the Openinverter forum. Please post yours here too so we can build up data for everyone to use.
49:03: Page from notepad showing last configuration was the successful resolver feedback connections for correct sin/cos polarity.
49:34: Huebner web interface. >Load FOC firmware. > hwver>Prius. >Opmode = OFF.
Must make sure resolver signals are being received by STM32 in VCU.
Make sure HV completely turned off.
Put vice grips onto shaft and turn manually for full revolution.
Computer, Go to Commands . Put opmode into Manual / Run.
angle L , R. Can graph from left or right.'
Go to bottom of screen, hit Plot button, and see waveforn on screen. Plot 204 - 207 degrees. This plot includes jitter. We expect some, but not a lot. Jittering >5 degrees, would consider that a problem. For this, check wiring, screens of resolver cable. E.G. jumping over a hundred degrees, STOP problem in resolver wiring or something else, broken resolver winding etc.
Restart Plot, rotate axle by a few turns look at waveform. Should get jumping wave forms that move from >0 to <360 degrees. We don't expect plot to capture 0 degrees or 360 degrees, because sample rate is too slow.
So screen goes 0 -> 360, 0 -> 360, 0 -> 360.
Third test, crank up limit data points and burst length. Start plot and hand rotate axle again.
Expect to now see 0 - > 360 in some of the waveform display.
Now go back and press Stop Inverter button on screen.
55:51: lasterr, Throttle1, if this displayed low resolution instead, micro-controller is telling us amplitude of received signal is too low.
56:02: Remove vice-grips from axle.
56:17: Have to configure current sensor positions for FOC firmware in Gen 2 inverter. Has to be done manually through custom command box.
56:46: Set pinswap 5 Send Custom Command. In messages should see Set OK.
57:31 We expect PWM1 would be measured by IL1, PWM2 would be measured by IL2, PWM3 not measured..
Not so Toyota, they use PWM1 not measured, PWM2 measured by IL1, PWM3 measured by IL2.
This confused the FOC software.
58:37: pinswap = 4 is a bit change command which relabels the current sensors to be as per Toyota arrangement above.
58:38: first bit of pinswap swaps PWMs, second bit of pinswap swaps current input labels.
So now PWM1 sensed by IL1, PWM2 sensed by IL2.
1:00:10: we now need to set some parameters for our motor. Polepairs = 4; Resolverpolepairs = 4;
encodermode to resolver.
syncoffset =0 to start with (Has range 0 - 65535). At this stage of commissioning, please leave at 0.
pwmfrequency to 4.4kHz.
3.1 current gain seems good value for MG2 power stage in Gen 2 inverter.
udcmin = 0, udcmax to some high value.
heatsink temp max to 150
1:02:20: save parameters one more time. Now press refresh button, and we should see condition off.
1:02:46: Reconnect HV battery 160V in this case, through 100W filament lamp.
1:03:06: We now start tuning in our commissioning process. Hit Refresh, confirm we are in manual/run.
1:03:21: Hit manualid box and press 1 amp, expect to hear a whine coming from trans-axle.
1:03:38: Increase to 2 amps, motor is now beginning to turn. Then motor stops and light bulb illuminates. This is perfectly fine, so set opmode to off.
If we did not follow that procedure, the inverter would try to inject maximum power, trans-axle would be jumping around the room and may damage windings. So cheap light bulb is a great asset for current limiting.
1:05:02: Can now switch on our bypass switch.
1:06:36: So when we inject 3 amps we see axle spin freely. We have achieved correct resolver operation AND correct current sensor operation
.
1:07:15: Johannes and Damien then looked at reverse, and a bug was found in the software. So now reverse works perfectly fine, as does forward
.
1:08:40: We will now move over to the Gen 3 system because this trans-axle has oil in it (Gen 2 does not, so don't want to damage the gears without lubricating oil). Does not matter that we flick from Gen 2 to Gen 3 system, procedure is exactly the same. Key thing is resolver and permanent magnet physics is same.
1:09:55: Gen 3 trans-axle has welded planetary gear device in it and using Yaris Gen 3 inverter.
1:12:00: A sketch of syncofs tune is shown, vertical axis is (manualid causes spin on its own [no manualiq]) effectiveness of syncofs value, horizontal axis is syncofs value from 0 to 65535 (Looks like notch filter).
1:12:41: syncofs (Synchronous offset) is the most important parameter we need to tune. This parameter is the most misunderstood.
1:13:43: For all different values of syncofs to left of notch, for all values of manualid, the motor would spin up.
For all different values of syncofs to right of notch, for all values of manualid, motor would spin up.
As we approach the correct value (the notch) from the left, we will notice it will take more and more manualid to achieve spin. this is good. We are onto the correct method of tuning. There may be only 100 points in the notch.
1:18:06: For this set up at 150V battery voltage 10,200 seems to be a very good notch value. But we will set to zero and perform tuning together.
1:19:13: Start inverter in manual mode. Go to Parameters, confirm syncofs is 0.
1:19:14: Now put in 3 Amps for manualid, and check motor spins. Then go back to manualid = 0.
1:21:30: My rule of thumb is to go in 5,000 jumps. So apply this then check manualid= 3 Amps. Now please note, the motor still spins but more slowly. so we are beginning to approach the notch
.
1:22:50: So now change syncofs by half as much again or 2,500 to give us 7,500.
Does not move at 3 Amps!, Try 6 Amps, and we can get movement. Vital Clue!
1:39:15: Now lets try 8,750, (increase by approx half again or 1000). Now we need 12 Amps!
Lets try another 500 points, 12 Amps axle is slower again. Note, if we see shaft rotating the other way, we have found ourselves on the right hand side of the notch. We don't want to be there, we want to be at the minima.
1:26:40: Now go in steps of 200 to 9,200. 6 Amps, nothing, 12 Amps nothing! We have to go up to 100 Amps before we hear some rumbling, but no movement. This is a clear sign we are close to the notch. Stator and rotor magnets are nearly perfectly aligned, but not perfectly because we are still getting rumbling or jitter.
1:27:59: Lets now try 10,000 points (jumps of 100), not at individual values yet.
Lets say we have jumped to say 20,000 points, and wacko, we are spinning in reverse! So we are now on the right hand side of the notch! In other words, if we overshoot our notch syncofs position we go in reverse.
So lets go back to 10,100, and we get a small amount of movement at 12 Amps.
1:31:23: Lets try 10,120 points. No shimmy at 100 Amps!
1:32:18: So next test in our tuning is to try manualid = 0.1 amps, and manualiq = 3 Amps, and motor spins.
1:33:13: If we want to check reverse we just try manualiq = -3 Amps, and it spins backwards!
1:33:25: We are now very very close to the notch point. We now go to Parameters, and save parameters to flash.
1:34:07: Final test, put motor into normal run mode , now use throttle pedal, and as we can see we can spin the motor up using our throttle pedal
.
1:34:50: RECAPPING
Step 1: Use sine firmware and check our motor and inverter work.
Step 2: Put FOC firmware in. Calibrate our current sensors. Set up resolvers. Can now run motor with manualid with syncofs = 0. IF at this time we get resolution error, we have to backtrack and fix before proceeding any further.
Gen 2 and Gen 3 transaxle files will be put onto openinverter.forum, please put yours here too.
-------------
There's not many comments, but I was the first person to thumbs it up, so, if you'd like to re-visit the link and thumb up his points for visibility, more people will see them (or Damien could pin it).
On my first 12v setup, I needed max boost and 95% ampnom to get similar results to Damien on Transaxle... Start with lower ampnom value, and work your way up as you've been doing. I think it's to do with how many Amps you need to get initial spin. The lower the voltage, the higher the amps, and inverse as well. Did not touch any other values.
