Yamaha YZ250F

[j bjork]

Tomorrow I will torture it on the track. I tried a little today, and I think it is lacking "punch" as it is. I suppose missing a few 100:s of phase amps might have that effect

 

[scianiac]

I wonder if the controller issue is the somewhat common overcurrent fault when transitioning from sensored to sensorless? I know this is being worked on with the next VESC versions but I don't know if it's fixed yet. For me the result of this issue is always if I turn the pA up too high and then apply full throttle when it's still in sensored mode it will overcurrent fault in the transition and the realtime log show as much. I was however able to turn the pA up quite a lot by tuning some settings (and not just cranking the absolute max current to the moon).

 

[speedy1984]

Hope it goes better for you fella

 

[j bjork]

I wonder if the controller issue is the somewhat common overcurrent fault when transitioning from sensored to sensorless? I know this is being worked on with the next VESC versions but I don't know if it's fixed yet. For me the result of this issue is always if I turn the pA up too high and then apply full throttle when it's still in sensored mode it will overcurrent fault in the transition and the realtime log show as much. I was however able to turn the pA up quite a lot by tuning some settings (and not just cranking the absolute max current to the moon).

It is quite possible that it was in the transition that it happened, it was at rather low speed.
These logs didnt show anything though:

(This log is from the track today, not where the overcurrent happened)
Even though I knew where in the logs it happened it didnt show anything, no fault or anything.

Can you read the realtime logs after a ride, or do you need to catch it on the computer when it happens?

I wouldent call wanting over 1000pA as max limit on a 1400pA controller "cranking the absolute max current to the moon"
I expect to be able to run 1400pA, not 800..
I actually think it is rather annoying that still in the end of October there dosent seem to be a program to get the full power out of a controller I bought in Mars or something like that? Not that I got it then, but I think that is when I payed for it.

Another disturbing discovery I made when I was building the wiring:

This is the wiring I got with the first controller. There are a bunch of unused holes, with no plugs or anything in.
I didnt think much of it at first, I thought the seal in the connector had a little "skin" as this type of connector with built in seal often has.
(So It is sealed until you push a wire through, after that you have to put in a plug if you take the wire out)
But that is not the case here, it is open for water to go strait in

 

[j bjork]

I can add that the controller worked just fine during today's ride though, to not sound too negative
The logs are very interesting to look at afterwards too

It dosent have the same kick as the fardriver, obviously I have a lot less pA so that is to be expected. But I think it is in the throttle too.
Not necessarily a bad thing, it is easier to control on mountain and that sort of thing. Maybe just a matter of getting used to be harder on the throttle when you want power.

I have modified one of the psu:s with 100v caps instead of the 63v ones, to be able to go higher in voltage.

There are 2 of the bigger ones, and one small.
The problem is that the 100v ones are bigger, so I have problems fitting them:

The little circle with the + is where the small cap is supposed to be, but I cant fit it there.
So I moved it, I hope the slightly changed position and the rather long legs isnt too bad?
It has worked to charge one of the batteries with, and to top up the other one so far.

 

[scianiac]

Can you read the realtime logs after a ride, or do you need to catch it on the computer when it happens?

Unfortunately no, the logs were the issue becomes very obvious at the ones where you tell it to capture a very short bit of data to it's internal memory and then you download it later. You can set the trigger to record right before a fault code, so an overcurrent fault. It's a bit of a pain but the result is you see the perfect sine waves and then a hiccup where they don't match.

Honestly I wouldn't bother with the realtime logs as it's easier to find out if this is the issue by trial and error.

I wouldent call wanting over 1000pA as max limit on a 1400pA controller "cranking the absolute max current to the moon"

Oh not at all, I was just pointing out that that is a solution but not a very good one as setting the absolute max current, which is the one that triggers the overcurrent fault and cuts the power will just tell it to ignore the issue even though it's still happening.

The first thing I would do is to figure out via trial and error if this is indeed the issue. The symptoms are kinda like this, when the motor RPM is over the sensorless transition rpm you can apply full throttle no problem and even if you up the phase current and then just use less than full throttle to get up to speed and pass through that RPM then after that you can use full throttle with no issues. Below that RPM you can also apply full throttle with no issues but as soon as it hits that RPM it generates an overcurrent fault and cuts all power abruptly. You can move the sensorless transition rpm up and down and if that RPM at which the above is happening then it's pretty clear that's your issue. Also from my experience this doesn't happen when unloaded (IE rear wheel in free air).

If you've confirmed that this is the issue you have lots of options to improve it. Updating the VESC version if it's not up to date is a good first step.

Trying different observers I've read is a big one (I actually haven't tried this one)

Changing the observer gain can also help.

Rerunning the motor parameters settings at a high current setting and see if the values it gives you different.

Setting the overcurrent fault slow setting.

Turning on sample in V0 and V7.

Honestly I tried lots of settings and many did improve it but the one the made the biggest improvement in more ways than one is turning on MTPA. I'm not really sure why it works so well on this issue but it did. Now you should have at least some understanding of how MTPA works before you just go turning it on because you're probably going to really want to after I tell you what it does! The guy who wrote the MTPA algorithm for VESC did a pretty good pos about it here: MTPA merged into VESC! | Details | Hackaday.io, he's also around here but his handle escapes me at the moment.

MTPA only really works if you have an IPM motor with a fair amount of saliency. Which you can tell because you will have some Ld-Lq difference value in the motor setup. But most IPM motors do have a fair bit and if I recall you do have an QS IPM motor. But with those motors you get a lot more torque, especially at high RPMs. In my setup it was stupid how much more torque there was at the same pA. While it's a little unclear I think in some circumstances it may be a little less efficient but I think it's very slight or even more efficient in some cases because I don't know that a simple FOC drive is ideal for IPM motors anyway and there are maximum efficiency algorithms that would probably be better. It's kind of a moot point though as if you don't have excessive motor heating (I've had zero issues here) and notice any range issues (also haven't noticed it) then it doesn't really matter since you're after power.

The other thing to have to keep in mind when turning on MTPA is, it will just by the nature of how it operates use all of the phase current you have set as field weakening once you are over the base RPM of the motor. Now this means you will have power way way past that base RPM but in that range it is noticably less efficient and ideally you should only use that occasionally and gear the bike accordingly, like it probably already is. (Using it the top speed of my bike is about double the base RPM of the motor and gearing) And that much field weakening does have a risk that some controller fault when you are running the motor at those RPMs can generate a back voltage spike that will cook a controller. Now I think it's pretty easy to just disable this if you just set the max ERPM limit to what the base rpm of the motor would be with your battery.

 

[j bjork]

My motor is ipm and has rather much saliency, Hackey reacted on that it was a lot more than the qs 138:s.
I am already using MTPA.
At first the transition was rough, Hackey tuned that earlier and also changed to the beta version 6.05 on the first controller.
This one is still on 6.02 though, but the transaction is smooth anyway with the old settings.
I dont want to mess too much with it, if this controller breaks too..
The slow overcurrent is off, but I dont think I want to change that yet.
I will chase Hackey this week and see what he says if I can get in contact with him

 

[scianiac]

Did he suspect the first controller blew up due to the 6.05 beta firmware? Do you happen know which observer it is currently set to?

 

[j bjork]

No, the problem with the first controller started before changing firmware. It started loosing a phase after some riding, then it got more and more. I suppose bad soldering or something like that. I have to check observer, ortega original or something?
I actually got the unlocked phone app yesterday, so now I can get full power as long as I dont connect the computer

 

[j bjork]

No one said my caps change in the psu was a bad idea, with the longer legs and changed position and all.
So I went ahead and did the same thing an the other charger. One thing I didnt do an the first one for some reason was adding this:

I didnt have the right resistors when I made the first, but I should at least have drawn the wires for it.
Now I had to drill out the rivets to open it again

Anyway, It is one 1,2K ohm then two 560 ohm resistors. I have 4 positions now to choose. The difference is then 3v from position 1 to 2, then 1,5v from 2 to 3 and 3 to 4.
Or what it does is I can choose to charge to 4v/cell, 4,1v/cell, 4,15v/cell or 4,20v/cell
Maybe I should have had even bigger difference between position 1 and 2, to get a better storage voltage at maybe 3,7v/cell.
But I dont think I can even go that low with just changing one of the 2 series connected psu:s.
And it is on that voltage a long time anyway, so it is not very hard to just unplug on a reasonable voltage.

I think I will probably be on position 3 (4,15v/cell) 90% of the time..

 

[speedy1984]

Any update on the vesc now you have had time to test and tune fella?

 

[j bjork]

Not much, it has worked fine. I have ran it on 1100pA the last two rides. I will go to 1200 soon, but that will probably be about as high I can go. Hackey said there is a 1400pA limit in the firmware, so I will need some headroom for that. That is a little disappointing as I was expecting the full 1400pA output, not as a safety limit.
I am still waiting on the unlocked computer vesc tool though, so if I connect the computer I have to connect the phone after to change back the limits..

I sometimes have problems connecting the phones Bluetooth, and sometimes I loose connection when I try to log.
On the latest ride I thought I had a nice log, but when I looked at it it had no gps map
Well, I dont think that last part is the controllers fault though

The more precise throttle control sure is nice on slippery rocks, there was some of that on the last ride. And also very steep, rather long climbs on sandy dirt that I had problems getting up last time I was there (a few months ago).
Now the only problem was not falling off or flip over, not traction. Maybe it was manly because it is very wet now, so the dirt sticks together more.

I tried a twist throttle on the left for the brake, but it was too weird so I went back to the thumb throttle:

It worked fine the latest ride, but the time before on super wet sand that got in everywhere it got stuck a few times.
This is the upside of a battery from that ride:

That is how much got in between the seat and battery, that are tight fit.
I drove in rain gear that time:

 

[speedy1984]

Glad you managed to get the controller working,it's been a long time coming.I also liked the throttle when my first vesc was working but did feel less responsive than the Fardriver but that was probably more a tuning issue.I think now just seeing how reliable the vesc will be is the next mission as you ride long and hard and in horrible weather conditions for electronics,so that will be a great torture test.Looking forward to more posts

 

[j bjork]

I havent done much to the throttle, more than set the response to 0.1s. I think there are some settings for rpm/sec that might do something too, but it can be a good idea to not go crazy there. It can be nice if the rear wheel dont go from 20 to 100km/h in 1 sec when you loose traction

I still havent learned 3d cad, but I found some useful drawings from other people:

This was a cover for a small vesc and a wattmeter.
I used it for the wattmeter with shunt and the voltage regulator.
I left as much room as possible if I want to add something more.

This is how it looks now, a lot better than stuff just hanging in the wires..
Could have been better still with something custom designed, but I think this will do.

 

[j bjork]

Oh, I should have used this picture instead:

With a little knob to turn

 

[speedy1984]

Yes there are lots more settings including curve that is a great addition I found and made all the difference with my trampa boards vesc and the 3shul

 

[Geoffers]

I've really enjoyed reading this build thread, great work. I only discovered this forum a little while ago and have spent maybe too long reading up on electric dirt bikes.
I'm considering doing a conversation on my wr450f , probably not until spring but have settled on using a qs180 motor.

I'd like to have maybe 30s battery, would you recommend the 3shul controller? I see you've used a few different ones over the tread.

I don't plan on racing but do like a bit of go in my bikes

 

[j bjork]

The qs180 sure works good, but it is big and heavy. It is also a lot of work to build a good jackshaft system. I think it could be a good idea to look at something like the 10kw rated qs 160 with built in reduction, but I am not sure if it is avalible now.

The 3 shul (vesc) sure is more sofisticated than the fardriver, but I think it should have more inputs and outputs. Also the quality is a bit questionable. You will probably have very quick and direct support if you have simple questions. If it involves some work from Hackeys side, you will probably get a response that he will fix it tomorrow, but weeks later you are still waiting.

 

[Mugenski]

The 3 shul (vesc) sure is more sofisticated than the fardriver, but I think it should have more inputs and outputs. Also the quality is a bit questionable. You will probably have very quick and direct support if you have simple questions. If it involves some work from Hackeys side, you will probably get a response that he will fix it tomorrow, but weeks later you are still waiting.

After seeing the somewhat poor craftsmanship on it's exterior I'm not so sure I would be happy seeing it's internals for the price he's asking. I have got in contact with him but his date to follow up has been left with no response like you menetion. I'm thinking a different controller might be the prudent option considering it doesn't seem like the 1400 is even capable of 1400A without jumping through hoops.

 

[Geoffers]

Thanks, I had a quick hunt for qs160 but not much came up?
I can see there's alot of work in the jack shaft setup, yours is a work of art! Anything I do would be a bit more agricultural, which is what I use the bike for.
I'm not in a hurry at the moment, just seeing what the options are as this is all new to me.

Will you use the bike much now? Would be interesting to see how the controller goes.

If you were doing it again would you still go with the same setup?

Cheers Geoff

 

[Geoffers]

That's interesting, that's the kind of performance I'm after, I'm guessing without jack shaft the top speed is higher?
Have you got some pictures or a build thread on your bike? I'd like to see it and don't think I have!
Cheers Geoff

 

[quelsurprise]

Hi,
Thanks for sharing, how are you finding the jackshaft set up you have, anything you would change?
I’ve got the QS180 and trying to see if a belt pulley could be put on the motor shaft, if not I guess it’ll have to be sprockets and a chain, but would rather a belt for less noise, and less messing around trying to seal the chain housing so oil can stay in there.

 

[j bjork]

Ok, lets see if I can get as much as possible..
The motor I am talking about is this one: QSMOTOR,0.5-12kW Electric Hub Motor & Mid Drive Motor Manufacture China
There seem to only be an expensive, pre production unit available so far though.
The housing seems to brake on the 138:s with high power though, so I guess time will tell what this can take.

There is nothing wrong with my setup, but if I would do it again I would probably try something like that.
It would probably be a lot easier, and it would be at least 10kg lighter I think.
Not as powerful, but probably enough.

If I would change something on my jackshaft setup?
The splines. I used the same as the motor, but it is a rather odd one.
I wont use the same sprockets anywhere else anyway, so there is no real reason to use the same splines.

Sometimes I think I should have a chain tensioner, sometimes not.

More reduction would be nice, but not easy with the chain and sprockets design.
Not sure if I would try to make a gear version.
With more reduction in the primary I could have less in the secondary, bigger front sprocket (less wear) and smaller rear (less damage from rocks)

I think the oil in the primary is a good thing. So far no bearing problems in the motor (no water trough the shaft seal).
No wear on the motors splines. Keeping the primary reduction cool.

For now no riding in the evening, it gets dark early here.
So once a week, 6 laps on a normal about 8km track is the usual.

I would probably go for the same controller unless I find something else, more interesting.

Battery, maybe I would go 28s instead. Just because there are more controllers available, and not a big loss.
I would probably build 2 batteries of 21700:s, maybe molicell p42 (if p45 still is twice the prize) about 6kwh each.

 

[Geoffers]

Thanks for taking the time to reply, that motor does look good, maybe a bit to expensive for me at the moment. As I'm not in a rush I'll wait a bit and see what happens!

I'd be using the bike daily and hadn't really thought about the chain oil bath protecting the motor seal and bearing but that's a really good point and would be important enough for me to go to the extra work of doing it.

A stone or something dragged round the socket would make pickle of it quite quickly.

I would more than likely build my own battery, as long as can get enough capacity to get round seeing livestock at an appropriate rate
Much food for thought

 

[j bjork]

I dont know enough about belt drives to really say, but I guess with the small pulleys you can fit for the primary drive it would probably take a wide belt to take the torque.
And then you would have to worry about small rocks and dirt getting in there.
Would be quieter for a street bike though.