Do EB3XX controllers limit phase current? - Looks like it!

[zombiess]

*Take this post with caution as I have not actually measured the phase current out of the controller, but the way the controller is behaving I strongly believe it is limiting phase amps*

As some of you have seen I have posted about issues I have had with some EB324 controllers. Xie Chang's new style 24 FET controller. What does that have to do with the title of this post? Well I decided to stress the snot out of one that I made a repair on to see if it would blow up. So far I only have 7 very hard miles on it.

On to the testing and the discovery.

First a little background. EB324 controllers can not be programmed as a 24 FET controller since the software that supports them only goes up to a 18 FET controller. Stock the 12 FET controller uses 2 shunts, 18 FET controller uses 3 shunts and the 24 FET controller uses 4 shunts. The few EB324 controllers I have built only have 4 shunts installed into them because I was unsure how changing the shunt effected amp settings.

Since I first got my hands on an EB324 controller about 2 years ago I have strongly suspected that the controller actually limits phase current through some algorithm or measurement scheme that is very different from the EB2XX controllers. I have suspected this because lowering the shunt appears to have no effect on the phase current setting, only the battery current. I have also noticed that the EB2XX controllers are a lot more punchy compared to EB3XX controllers.

The test setup:
Greyborg Warp Frame with Cromotor V2 in 19" wheels with 25.3" tires
18S3P 15AH battery pack, 75V fully charged
CA V3 in pass thru mode with wattage limit set to 20KW and current set to 200A and ramp rates at 10V/s
EB324 IRFB4115 controller programmed as an EB318, block time 0.0s, shunt value of 0.83mOhm, controller expects 1.00mOhm
Multiple dead stop to WOT passes made on each setting to get the highest peak possible.

Test 1:
Set battery amps to 30A in the software, this will allow ~52A to be pulled from the battery.
Phase current programmed to 105A in the software
Test ride CA showed a peak current of ~54A

Test 2:
Set battery amps to 49A in the software, this will allow ~85A to be pulled from the battery.
Phase current programmed to 105A in the software
Test ride CA showed a peak current of ~89A

Test 3:
Set battery amps to 100A in the software, this will allow ~175A to be pulled from the battery.
Phase current programmed to 105A in the software
Test ride CA showed a peak current of ~102A <--- Ah HAH! What's going on here?

This is where I started to scratch my head since I was expecting the peak to overshoot by quite a bit before settling down as the EB2xx controllers would do. I made several pulls just to be sure. Went back and checked my laptop and the settings were just as I have posted.

Test 4:
Set battery amps to 100A in the software, this will allow ~175A to be pulled from the battery.
Phase current programmed to 120A in the software
Test ride CA showed a peak current of ~118A... looks like the battery current is following phase current, as it should.

Test 5:
Set battery amps to 100A in the software, this will allow ~175A to be pulled from the battery.
Phase current programmed to 150A in the software
Test ride CA showed a peak current of ~149A... yup, battery current is following phase current.

Test 6:
Set battery amps to 100A in the software, this will allow ~175A to be pulled from the battery.
Phase current programmed to 175A in the software, Figured why not, I have plenty more of these controllers I'm afraid to sell until I prove reliability
Test ride CA showed a peak current of ~175A... battery current is following phase current.

Test 7:
Set battery amps to 49A in the software, this will allow ~85A to be pulled from the battery.
Phase current programmed to 175A in the software
Test ride CA showed a peak current of ~85A... battery current is now limited by battery current setting x1.73.

So there you have it, the EB3XX controllers appear to limit phase current regardless of what the shunt value is and I suspect it does this at all throttle positions. There is very little if any overshoot noticed, at least on the CA V3. To really detect current overshoot I would need to hook up some current sensors and use at high speed capture loop to record the maximum. A device I eventually plan to make.

So if your settings are as follows (battery amps is actual amps, not just the programmed value) Battery amps are effected by the shunt
If Phase Amps > Battery Amps then Battery Amps = Amp limit
If Phase Amps < Battery Amps then Phase Amps = Amp limit

Wish I had someone shooting a video of the 175/175A ride. It's just enough power (I saw 11KW on the CA) to do a smooth slow rising wheelie if you yank up on the handle bars when giving it throttle. I managed to ride out a few power wheelies like this from really low speed with the longest one being about 25 feet with the front end hanging about 2' off the ground before it settled down.

End result of the ride, motor was warm to the touch, warm enough after sitting for a minute that I decided it was best to take it on a nice low speed ride to help cool it off vs let it sit and bake. In all I probably did somewhere around 40 dead stop/slow WOT pulls to 10-35mph with many being back to back. Controller case was slightly warm, outside temp was about 60F.

Other good news is this controller is taking some hard abuse. I did not it expect it to live at these levels. Only time will tell if it is going to work OK. This is the second controller I have repaired and installed on this bike. The first one kept blowing up the same FET (I gave up on it after 7 repairs ) on the high side of phase B. They are both returns from a customer who had issues with them. I hope they work so I may finish up the rest of these controllers I have.

 

[snellemin]

That is some good "data logging" man. I'll wait and see how this controller of yours does in the long run before I order me one. BTW the 18Fet 4115 controller I got from you works great with the cromotor for the few runs I have done. Super smooth and quiet. Running 21s2p 5ah lipo pack.

Would you know what settings you have programmed into that 18fet of yours? I haven't programmed it yet and want to have a reference before I start to screw around with things.

 

[zombiess]

Based on my observations, how would this controller be monitoring phase current? The old versions estimated it based off of the shunt value, but that is not the case with these since phase current limiting is unaffected by the shunt value. There is a small squiggly trace coming from each phase. I would need to reverse engineer the circuit, but I believe it leads back to the gate signal line eventually.

Not having a known data sheet for this micro make it hard to figure out. I am curious as to how this limiting is being accomplished without phase shunts or sensors.

 

[gwhy!1]

Hi Zombiess,
This post has intrigued me , as I have also found that the EB2 do seem to have a better hit than the EB3 but when I was trying to work out why this was I was getting different results to what you found with the EB3's but like yourself I was not actually measuring the phase current ( only assuming )
all my EB3's and EB2's has a shunt mod and has approx 2.5x adjustmet in programmed currents.

If a programmed battery current = 40A ( 40x2.5 = 100A) and a programmed Phase limit of 90A (90x2.5 = 225A ) max battery current reading when riding = 100A

or if I reduced the programmed phase current to 40A (100A ) max battery current reading was still = 100A when riding ( but acceleration suffered big time as expected )

but if I reduced the programmed Phase current even more e.g 20A ( 50A ) then max battery current ever seen was only 50A when riding ( the max current would climb higher once the motor go upto max rpm ( a very long stright/flat run )

so in my case the phase current was following the modded shunt programmed value :?

I am in the process of collecting some bits and pieces to do some bench test measurements on these 2 models of controllers ( EB2's and EB3's ) as the EB3's have never seemed as good as the EB2's and this post has has fired me up again to try and find out why

Have you came to any more conclusions with your controllers ?

 

[zombiess]

Gwhy, it's hard to follow what you are saying since you are not specifying the controller version you are talking about.

EB3xx doesn't require and adjustment to the programmed phase value when modding the shunt from what I've seen. EB2xx does. EB2xx controllers are known to over shoot the phase currents by quite a bit as seen on dyno plots.

 

[gwhy!1]

Gwhy, it's hard to follow what you are saying since you are not specifying the controller version you are talking about.

EB3xx doesn't require and adjustment to the programmed phase value when modding the shunt from what I've seen. EB2xx does. EB2xx controllers are known to over shoot the phase currents by quite a bit as seen on dyno plots.

Sorry , the above results was with the EB3's and this conflicts with your findings.

 

[zombiess]

Gwhy, it's hard to follow what you are saying since you are not specifying the controller version you are talking about.

EB3xx doesn't require and adjustment to the programmed phase value when modding the shunt from what I've seen. EB2xx does. EB2xx controllers are known to over shoot the phase currents by quite a bit as seen on dyno plots.

Sorry , the above results was with the EB3's and this conflicts with your findings.

Different motors, different results, not surprising to me at all. As I said I have no idea what is different in this version, but I like it (except my issues with the new EB324 controllers). At least I like the simple to use software and the fact the boards are dirt cheap. The power stages on them suck a fat one, but that can be resolved (I am working on something).

Only motors I have tested are 2 different KV Cromotors, the normal 9.3KV and the 18KV I have. Both with the same results after modding the shunt. I have also driven both motors with EB2xx controllers. I lack any hard data because I don't care enough about it, but it looks like all of the anecdotal info points to the EB3xx overshooting the phase current less than the EB2xx.

 

[gwhy!1]

Thanks,
If I find out anything conclusive with the tests I will be doing I will post here.