Ampera/Volt inverter with Lebowski brain

[arber333]

Also i tested the compressor some more and i get very good results with:

1. changed max phase amps to 25A

2. Changed acceleration limiter to 1 and 5

3. Changed wiggle settings
f) wiggle range: 90 deg
g) wiggle rate: 180 Hz

Now it will spool off at once as i apply tiny bit of throttle

 

[kiwifiat]

Yes please let me know how it goes. I have great experience with simple Johannes resolver interface.
EDIT: I may put a dual Lebowski board in production. This one i intend to make with Prius resolver chip. @Kiwifiat, how did you find that chip operating in resolver 2 communtation mode on its own? I would provide oscilator and filters on circuit. Do you think it would be good to work with?

I have a layout for the resolver to digital chip off the Ampera/Volt inverter but haven't got around to building it up yet. One of the things that was putting me off was all the expensive peripheral parts you need plus the fact that you need to clock the chip and monitor the error pins. So in the end I built up the resolver to HALL example that I referred Johannes to over on diyelectriccar way back based on an Arduino and tweaked the software a little and tested it on the Sonata HSG motor. It works perfectly with the lebowski controller. I have yet to test it on the Leaf resolver, I expect to have to tweak a few things since all these resolvers have different transfer functions between the exciter and feedback. The only thing is that it is only exciting the primary at 2 kHz so I was thinking of upping that to 4kHz or so. Most of the dedicated chips run the exciter at 10kHz.

I expect the prius Tamagawa chip would also work well if you can get it working, probably a better solution for automotive use in any case.

 

[arber333]

I am really happy! I got automatic AC start e.g. by AC clutch signal.

Like Lebowski stated i made simple voltage divider in front of throttle line. 10K/3K3 and i added 470uF cap across 3K3 resistor.
When i apply 13V to this line it makes 3V4 on divider, but cap charges it in about a second. This signal coupled with acceleration limiter causes AC compressor slowly to spool up and reach its working RPM. Voltage rise is slow enough to get it working every time.
When i remove 13V signal motor spools down slowly and turns to standby.
I will experiment with it more and i think of calibrating throttle so it works up to 75%.

This will work with any AC system in a car, since you only need AC clutch command to start compressor and every car has this, even more elaborate automatic ACs.

 

[arber333]

The car is moving! I installed inverter, calibrated throttle, setup reverse switch and wired my PRD relays. I calibrated RLS encoder to get sensored start. The car was automatic. I have brake boost and EHPAS pump working. I just need to connect my DCDC CAN instruction.

I will need to reinstall front battery box since I forgot to tighten my coolant temp sensor. When I tried to put liquid in it was leaking... DOH!

 

[Lebowski]

Just saw the post on your leafdriveblog about backing out and throttle being twitchy. What you can do is in the throttle menu, change coefficients from1,0,0 (which is linear throttle, so .5 throttle gives you .5 torque) to 0,1,0 which is quadratic throttle (so 0.5 throttle gives 0.25 torque). This should make the throttle easier to operate for manouvering. 0,0,1 is 3rd order throttle, so 0.5 throttle is 0.125 torque.

Its in the manual...

 

[arber333]

You are ahead of me. I wanted to gather some more data before I accuse your firmware here... Well here goes...

I use 1.2,0,0 throttle and -0.4,0,0 regen like Arlo suggests and I like smooth response. I will try your settings too.

Right now at speed (or higher rpm) I feel motor instantly pulling back as if it would apply regen just momentarily. It happens as if shaft would hit something as it rotates. Up to say 2000rpm it works smooth and after that it starts jolting. Could that be connected with new addition in code?

I use "hall assisted sensorless" option from the menu. Is that OK?

Otherwise I feel teh powah. I took the car to 100km/h according to GPS the first time I drove, but I still need to sort speedo out. Drive is smooth, I like 406 coupe suspension.

Any comment on the symptoms? Would that be hall sensor filter? I use 100hz with 0.90 gain. If I increase gain over 1 I get sluggish motor. Should I decrease it? All of sensor points are at confidence 6 since i use RLS magnet encoder in UVW mode.
I didn't zero it out. Maybe I should zero it on one phase applying 12V to one phase and reset position jumper on encoder?

Tnx

A

 

[Lebowski]

I would not use hall assisted sensorless, then try again. It sounds like the halls cant keep up with the higher motor speed or you have too much caps on the hall lines. Sensorless should be fine.

 

[arber333]

I used caps per schematic... 1nf if I recall correctly. Should I use less?
I dont use any 4K7 pullups since sensors provide push pull signal of their own.

I will remove hall assisted setting and test again.

BR

 

[Lebowski]

I used caps per schematic... 1nf if I recall correctly. Should I use less?
I dont use any 4K7 pullups since sensors provide push pull signal of their own.

I will remove hall assisted setting and test again.

BR

I think the advanced chip you use is sensitive to the magnet being exactly centered above it. If not it might miss transitions occasionally, giving the behavior you mentioned.
Indeed you dont need the pullups but you still need the 1k series resistors...

 

[arber333]

Damn! I had loose screws on front wheels!


Also I see signs of magnet case on axle hitting hall speedo sensor. I think this was the source of rubbing. Very disturbing I should think. Of course hall sensor is dead...
More work to do.

 

[arber333]

I still have rubbering shakes when i drive at speed. I tried to rebuild my speedo sensor and it is now out of the way of magnet collar. Feeling is kind of like you would drive on tiles. But when i investigated further it seems this only happens when i have the steering wheel centered or to the right. If i am turning left the shakes vanish. Something in the brakes maybe? I need to investigate this some more... Gopro would be a good solution.

 

[Lebowski]

I still have rubbering shakes when i drive at speed. I tried to rebuild my speedo sensor and it is now out of the way of magnet collar. Feeling is kind of like you would drive on tiles. But when i investigated further it seems this only happens when i have the steering wheel centered or to the right. If i am turning left the shakes vanish. Something in the brakes maybe? I need to investigate this some more... Gopro would be a good solution.

Maybe wheel bearings ? (i have no clue about cars, im a motorbike guy)

 

[amberwolf]

My mom had a big "boat" Oldsmobile from the seventies that did something like what you describe. The tie-rod and linkages were worn, and had to be replaced. I don't think it had any bearings, just the metal-to-metal itself. I didn't do any of the work so I don't know what it actually looked like, but that's what the shop told her and charged her for.

I've also heard of tires themselves having a problem, due to wear, and replacing or rotating the tires (swapping from one wheel to another elsewhere on teh car) evened out the wear (over time) and fixed this.


But like Lebowski, I'm not really a car guy...I prefer bicycle-type creations.

 

[arber333]

Yes I think it is axle related. But I am a bit on edge since I spliced Nissan Leaf and Pug shafts and I think I did a good job. But still...

 

[arber333]

Yes I think it is axle related. But I am a bit on edge since I spliced Nissan Leaf and Pug shafts and I think I did a good job. But still...

Well after a visual inspection of both axles i found out i didnt use the bearing retention plate. I havent got it from my motor supplier or from driveshaft supplier. Leaf manual ferers to it as consumable as it warns against using the original retention plate when assembling the bearing.
I ordered one 39776-JD000 plate from ebay. I will install it and well see.
Meanwhile i noticed that if i use shielded LAN cable vs normal UTP cable i get much worse performance from EVdisplay wattmeter. Then i removed my UTP cable and rout it in a different way. I have seen much better results. So i see routing my comm cable in paralell with DC cables causes EMI buildup in my comm cable. If i use a different route EMi falls rapidly in proportion to square with the distance.

 

[Lebowski]

Cars are Greek to me .... but looks like something a setting in the controller will not fix

 

[arber333]

Let me rephrase this !
https://www.youtube.com/watch?v=wVFfeTLZY84

 

[Lebowski]

Does it now run good, also with the fieldweakening etc ? Did you try at higher speeds ?

 

[arber333]

Well the first time I went fast I got to 100km/h on GPS. But now I am a bit reluctant to try more before I get the retention plate in. Otherwise motor still wanted to pull at that speed. I think fw starts at 4krpm mechanical and works really good. Regen as well.
I had to shield my phase cables. I will shield DC cables as well. Still I have some artifacts in EVdisplay.
I use 1.2,1,0 for throttle and I get quite fast torque response. Though no wheel spins. What if I used 1.5,1,0? Would it start up sooner? Regen I use -0.4,0,0 as Arlo suggested.

 

[arber333]

Retainer is now in and i need to wire the battery and BMS in again. Will report back...

EDIT: The rubbing is still present so the retainer was not the cause of it.
EDIT2: I am thinking it might be the RH driveshaft. It may be oscillating and causing the rubbing.
EDIT3: Or i might have assembled the bearings inside the driveshaft in impropper order and now there is some rubbing where the bearing seat is not completely aligned with bearing. DOH!

Lebowski what do you think. I use 10nF decoupling caps on the hall sensor lines. What would you recommend i change them to? I would like to try if hall sensors are causing motor to vibrate maybe. But i need a reference to which value i focus to.

 

[arber333]

Well i found out what was bothering my Pug.
The left inside CV joint has a halfhsiezed bearing. In by itself i couldnt determine that since the wheels had so little resistance when lifted up. It was only when my friend an i tried to rotate both wheels simultaneously we noticed the LH side was grabbing the differential. I noticed a slight tick in the CV joint when turning outside of allignment. I guess this was something that was declaring itself before failure.
I will need to take the CV apart, inspect the bearings and the outer surface. Maybe it will be enough to just replace one bearing. If not i will need to change the whole Leaf CV bucket.

 

[arber333]

I still havent changed the CV joint, but i changed motor settings.
I used 75uH and 48mR like Arlo mentioned in one of his earlier posts. This setting gave more power from start and ability to spin rubber. However there is still shaking in control when starting from 0 and this causes delay.

I noticed with existing settings if i change setting d)accept direction change in menu d)ERPM limits
to lower value i get faster start or maybe less shaking in start from 0. Any comments on that?

Should i also lower option f) erpm dr2 jump to dr3? Would it help if difference from e) to f) would be less?

 

[arber333]

Well, remember when i said i had siezed LH CV joint tripod bearings? Well that was not true. Rather i made a dumb mistake when copying the length and forgot to pull out on the tripod so it would be extended like in its natural position. Instead i just copied it in the innermost position with wheels down. Naturaly you all know right triangle? Well i forgot the hypotenuse is allways longer than cathetes. All in all i didnt listen at my geometry class!
So my CV axle was pushing at the lip of the tripod casing everytime it came around. In 10000km i wouldnt have an axle...

So i went and cut my LH CV axle again. This time i made many measurements and tried to assemble the joint without grease and rubber. I needed to take off 20mm of the axle! I went to a machine shop with a lathe and they cut the shaft and beveled the edges so it would be easier to weld. Then additionaly they made a pin to fit on the inside of hollow shaft. Next they drilled 2 8mm holes in each shaft section at 90deg and assembled everything. I went to the car with everything and tried one more time. The shaft spun clear. Sadly i hadnt had my phone handy so no pics of the process .
Then i ask my friend with huge MIG/MAG welder to weld the shaft for me. He filled the beveled circumference and the 4 holes. Then i liberaly sprayed the shaft with zinc spray. Now hopefully everything would carry 200kW if needed!

I went for the test ride and everything works, car is silk smooth on the road and i am surprised how good the wheels are tracking. Now what is left is to go for wheel aligment, just in case i made some errors when assembling the front bridge.

 

[kiwifiat]

Great news. Does that mean that those rumbling noises you thought were coming from the leaf motor were mechanical all along and nothing to do with the Lebowski parameters?

 

[arber333]

Great news. Does that mean that those rumbling noises you thought were coming from the leaf motor were mechanical all along and nothing to do with the Lebowski parameters?

Exactly! CV shaft was rubbing on the edge of CV tripod casing. Whenever the three lips of the casing would come near the shaft on the underside they would rub on it. In time this would have destroyed my shaft. As it is now everything now works and i am trying to tune up the transition like Arlo said. Well i figured i want to have transition to sensorless at as low RPM as possible, because whenever i try to start motor shakes a bit and 1/2s goes by without traction.
But higher in the RPM i am able to burn rubber even at 60km/h! I am sure i could get more power out of it, but my poor battery developped bulges at 400A! So calculating 400A at 355Vdc i pushed some 142kW into the motor. That is not "halfh" bad. 8)
Notice the pun at 1/2 bad? Since my plan is to use both sides of Ampera inverter...