HESC V1 150v 600Amp VESC based motor controller

[mxlemming]

Here is a little test in sensorless mode because the motor has no halls.
https://youtube.com/shorts/pLYTusTA7lk?feature=share

About the waterproofing, since I won't use it on a bike I was thinking it would be fairly safe from water inside a lawn tractor.
That was my initial plan anyways.

I think I'll get the CL700 for my car.

In a car you would not want a sensorless system that has equal chances of going backwards as forwards when it begins to get rotor motion. That's a crude sensorless. A good FOC setup will not go backwards for a rotor tooth before it goes forwards.

I don't think this is a fair statement, or it needs clarifying. The Shul controllers run VESC firmware, the exact same firmware as the "state of the art" axiom you've supported, and have access to the same feature set with probably better performance owing to closer replication of the reference VESC hardware.

If you know of any magic feature above and beyond HFI and openloop startup, please drop a hint and I'll implement it in my firmware and offer to VESC.

 

[kiwifiat]

Implement the method described in the pdf attached to the first post: https://endless-sphere.com/forums/viewtopic.php?f=30&t=103025#p1506964
and you will have your magic bullet.

 

[mxlemming]

Implement the method described in the pdf attached to the first post: https://endless-sphere.com/forums/viewtopic.php?f=30&t=103025#p1506964
and you will have your magic bullet.

I love this method, I've read it and understand it completely, but unfortunately it is completely incompatible with VESC hardware since it requires high pass filters implemented with analog opamps.

VESC has HFI methods that are incredibly effective now without additional hardware, they work on the principle of tracking inductance from injection in a single direction in dq axis. They work with the shul controllers.

 

[eee291]

Hey mxlemming, looks like I have you to thank for this.
Using the mxlemming_lamda_comp observer I managed to spin this motor 4x above base speed, 8600 rpm vs 2k rpm.
It was drawing 1kw no load lol
with the 5.03 firmware, I only managed about 4k rpm.
I don't actually plan on using this much fw, it's fun to test though, not to mention it's running sensorless.

Sensorless works fine with more current, in the video I'm demanding a few mA, so it just rotates back and forth, but a bit more throttle, and it springs to life instantly.

 

[mxlemming]

Hey mxlemming, looks like I have you to thank for this.
Using the mxlemming_lamda_comp observer I managed to spin this motor 4x above base speed, 8600 rpm vs 2k rpm.
It was drawing 1kw no load lol
with the 5.03 firmware, I only managed about 4k rpm.
I don't actually plan on using this much fw, it's fun to test though, not to mention it's running sensorless.

Sensorless works fine with more current, in the video I'm demanding a few mA, so it just rotates back and forth, but a bit more throttle, and it springs to life instantly.

This makes me happy. glad I've helped.

Give the HFI a try. HFI45 is normally much better on outrunners but I don't think it's ever been tested on a big inrunner. I'm genuinely intrigued to hear how well it works with the HSG motor.

 

[eee291]

Vss works about the same as regular sensorless, HIFI doesn't seem to work at all. I haven't messed with the HIFI settings much tough.

Edit: got it working now, I set sensorless erpm to below 100, regular HIFI seems about the same as sensorless. HIFI 45 deg v0 starts better.

https://photos.app.goo.gl/noxL3Dx4DLsfcXwu9
When the motor moves back and forth a little, I'm demanding about .2A, it starts reliably at or above .6A (motor current)

 

[DEHiCKA]

Did anyone independent torture any this controllers to full VA specs? All I have seen since are no load or very light load which tells nothing about the controller design quality.

 

[eee291]

[youtube]fiW32WjSGtY[/youtube]
Their official channel

 

[DEHiCKA]

Their official channel

Thanks, I have seen that. It is not independent and it's least powerful CL300. Most problems arise on high currents.

Can anyone open the cover and make a photo of the power and logic board with part numbers visible?
Or just write down the markings on the mosfets and mosfet drivers, so I can make some estimation.
I have checked their website and FB posts, no photos with part numbers visible unfortunately.

Also I can not find their firmware repo or any HW profiles in main VESC repo.
VESC is GPL and for any hardware sold FW source must be available before any sales made.

 

[JRP3]

Thanks, I have seen that. It is not independent.

Did anyone independent torture any this controllers to full VA specs? All I have seen since are no load or very light load which tells nothing about the controller design quality.

I'd say testing beyond rated current to failure is not a "very light load".

 

[eee291]

The cont. Output for the CL350 is about 100A DC, if you stay within that range, I doubt it's ever going to fail.
I could do a few hard runs in a LEV within a week if you'd like.

 

[DEHiCKA]

I'd say testing beyond rated current to failure is not a "very light load".

Very light load from independent party.

The cont. Output for the CL350 is about 100A DC, if you stay within that range, I doubt it's ever going to fail.

I am mostly interested in CL1000 and CL1400. At such high currents lots can go wrong.
At least I need more info before spending $1K on the thing without a warranty.

 

[eee291]

The problem for me is the price, honestly.
With 1.5K€ I can get a Tesla drive unit

 

[JRP3]

The problem for me is the price, honestly.
With 1.5K€ I can get a Tesla drive unit

Link please!

 

[eee291]

If I was in the US you'd see me with 20 drive units in my backyard considering the cheap stuff over there

 

[JRP3]

If I was in the US you'd see me with 20 drive units in my backyard considering the cheap stuff over there

I just took a quick look on Ebay and they have gotten a lot cheaper. Problem is you'd need to machine an output housing for the one end of the motor and an enclosure to relocate the inverter. Has anyone done this for the Model 3/Y small front motor?

 

[eee291]

Why would you do that to a tesla drive unit? Just use it as is.
A more ideal motor for that would be from a Nissan Leaf, a full inverter motor stack is like 1K€.
You need the input shaft from the gearbox to weld a clutch hub onto it, and a metal plate to adapt it to your transmission.

 

[JRP3]

Why would you do that to a tesla drive unit? Just use it as is.

Because I wouldn't be using it in a car and don't want the transaxle, looking for compact power, like a QS motor and 3Shul. Seems as if the Model 3 front motor could have a similar footprint if modified as I suggested.

 

[eee291]

You're out of luck then, unless you can do quite a bit of fabricating, look at a teardown vid of a tesla unit :|

 

[JRP3]

You're out of luck then, unless you can do quite a bit of fabricating, look at a teardown vid of a tesla unit :|

I have looked at teardown vids and I think I can do the needed fabrication, just don't know if I want to do it.

 

[pwd]

From the Electric Motorcycle Builds group on Facebook:
https://www.facebook.com/groups/electricmotorcyclebuilds

Some footage if a CL700:
https://www.youtube.com/watch?v=TiSyLB2NZyA

associated data:
https://metr.at/r/mhybK?fbclid=IwAR0eAdHdwqP-9DzrRcsk9dDlhV91WUm2W3il7zFuHJNLzvptoll39lrcw78

Kevin (Alien Rides) is a 3shul dealer in the USA

 

[JRP3]

From the Electric Motorcycle Builds group on Facebook:
https://www.facebook.com/groups/electricmotorcyclebuilds

Was there supposed to be 3Shul specific posts? Because I didn't see any.

 

[pwd]

From the Electric Motorcycle Builds group on Facebook:
https://www.facebook.com/groups/electricmotorcyclebuilds

Was there supposed to be 3Shul specific posts? Because I didn't see any.

That was a link to the FB group I was referring to, if you search for vesc within the group you'll see some:
https://www.facebook.com/groups/684227311714643/search/?q=vesc

or this:

https://www.facebook.com/hashtag/3shulmotors?__gid__=684227311714643

 

[DEHiCKA]

I have searched that FB group/pile regularly. Unfortunately still nothing on any of Shul ESCs even remotely close to full spec.
Those kawasaki log is just 250A max PhA which is not even half the spec.

 

[j bjork]

I dont see the price as a big problem, it is pretty much in line with other comparable controllers I think?
But the difference between cl1000 ($900] and cl1400 ($1600) is on the large side I think..
I have a qs180 90h with a fardriver 961800 now, but I am curious about these controllers.
Would be nice with higher voltage, variable regen and a better program.

With this price difference I would like to find out if maybe cl1000 would be enough, smaller size would sure help too.
dougf seemed to only use 1000-1100pA on his dyno run, but that was from maybe 2000rpm.

I did send them a email with some questions, and got a reply with some of the answers. I then clarified some were I didnt got the answer I was after, but didnt get any more reply.
I havent seen any wiring diagram for these new controllers, but on the old ones there was no extra inputs for reverse, 3 speed switch etc. From the answer I got it seems like they only have removed some of the data connections that was not needed on the new ones, nothing about added inputs. I didnt get any answer to the question if there were any.

Is that a vesc limit, that there cant be any inputs except throttles and motor?