Stuttering on acceleration and regen (problem found)

[zombiess]

Motor, Croatian Hubzilla
Controller, Lyen 18 FET 4115 controller

If I accelerate gently, the motor has some issue for the for 3-5mph, then gets smooth as long as I'm gentle with the throttle. If I try to accelerate like normal I get a
bunch of popping and grinding like the motor is out of sync.

I have a test program and it will sound like it's missing a phase below 3-5 mph sometimes, like it's not firing. As soon as I speed up it clears up. The problem is much worse when the motor has a load on it but clears up at high speed. The last time I had a motor do this to me it was a loose phase wire, but I've already checked them and they are fine.

I've put my scope on all controller outputs and hall wires. The only issue I really see is that when the controller is running the motor the signal coming back from the halls is super noisy. It's about 7V peak square wave signal and the top 3.5V of it is full of PWM. As soon as I let off the throttle, the hall sensors turn back into perfect square waves. I'm wondering if this is the issue and is causing false triggering.

I've tried all different variations of batter amp/phase amp/block time/sensor angle settings (120 is what this seems to want) and they all behave the same. Motor spins freely when hooked up to the controller so there is no short on the phase.

Ideas? I'd really like to get this bike working properly. I guess I can try some different hall/phase combos again tomorrow to see if there is some oddity with this going on, but with my setup all the colors match up 1:1 phase and hall. I'm running out of ideas. This controller worked perfect on my 9c 2806.

 

[neptronix]

Interesting.... could be similar to the problem the MAC and BMC motors have due to their low resistance?
A change of controller could be worth a try.
Too bad you are so far away. I have a pair of EB3xx controllers that sense current a bit different.

But you may want to try some other controllers other than lyen's, if you have some lying around.

 

[Tench]

If the motor has never ran well, the next step i would take is to change the halls for ones you know wont give you problems, its an easy way to rule them out.

 

[zombiess]

Interesting.... could be similar to the problem the MAC and BMC motors have due to their low resistance?
A change of controller could be worth a try.
Too bad you are so far away. I have a pair of EB3xx controllers that sense current a bit different.

But you may want to try some other controllers other than lyen's, if you have some lying around.

Do you think it could be tripping some sort of over current protection? the more phase current I give the Motor the worse the problem gets.

I'm double checking the the hall/phase wiring today, for ever phase combo there is a hall combo that makes the motor run forward/backwards, try one of the other forward combinations.

 

[zombiess]

I definitely have one of the 3 good running forward wiring combinations. Motor runs smooth low amp draw, but seems to lose sync on quick throttle changes when there is a high amp demand, then finds it again. It's the only way I can think of to describe the problem.

First I'm going to drop the voltage down to a 18S setup and see if that helps at all. If that doesn't, I guess I'm going to go cut open my nicely made wiring loom and separate the hall wires away from the phase to see if it reduces the cross talk on them. This thing is driving me nuts. At least I have a scope to look at the signals.

Anyone know if it's normal to see so much noise on a hall sensor when running. I'm gonna go check my 9c 8x halls now and see what the signal looks like on them.

Any ideas on how to fix this?

 

[zombiess]

Just checked hall signals on my 9c 8x8 on 24s/95V When the motor is running they show the same or even more noise than the Croatian motor. That's now off the list.

Going to try 24S on the Croatian motor setup next, see if that improves things at all since I already have a 24S harness ready to go.

 

[zombiess]

Just tried 24S, same problem, going to experiment with settings now, maybe try programming it as different FET model controllers to see if this is some sort of current limit issue since it only occurs during high load demands and it's either losing sync or the controller cutting power for a brief moment.

I should probably rig my scope up to the bike and watch the phase wires to see if they are being interrupted, but that's a little tricky to do safely.

 

[zombiess]

Ok, just did a bunch of testing on 90-100V 24S2P, programming the controller to different settings and here is what I got...Jack shit! Motor will run OK without much issue once it's above 20mph, even switching to high power mode and getting on the throttle it's OK, but at low speed anything but the smallest throttle input causes it to cut out until it reaches speed. I can still cause it to happen in 100% throttle mode around 20mph if I just go WOT instead of rolling it on, but it only happens for a split second and then it's OK.

I tried programming it as a 9 fet controller and reducing all amp settings by 1/4 to get the same as I was as an 18 FET controller same issue, but at first I only reduced the setting by half and got a taste of what 101A of battery current (as seen on my cycle analyst) and 200+ phase amps can do... 30mph power wheelie when switching power levels with the switch... jeez, when I get this thing sorted I can have a motor cycle.

I've tried upping the block time to 5 seconds, setting block time to 0, tried setting battery amps to 65 and phase at 90, battery amps at 20, phase at 90. Basically any setting that makes it behave like a normal ebike causes it to shudder before finally making clearing up at higher speeds.

I also learned that even with the bike sitting on the ground without me on it if I hit the throttle hard it will shudder and stutter not really going anywhere, but with no load at all the wheel accelerates just fine.

I did almost all my testing with the throttle limited to 40%. I just did a quick test voltage was 92.5V, 40% throttle, that gave me a top speed of about 19mph on 24S and at speed the current draw was about 10A (seems high), no load speed and current draw is 24mph, 1.3A (also seems high). This CA is calibrated to the shunt in my controller.

I guess next up is trying different phase combos, maybe the one I have just seems to work OK but is really not right. Gonna go do that now. I really hope this is the issue. If not, I guess the it's the controller and the inductance of the motor is too low causing it to act strange.

 

[dbaker]

Did you try it with the CA unplugged? I was having a similar problem running my 18fet ecrazyman with the Cro motor on 24s2p. Unplugged the CA and it ran pretty strong. I have not had any time to play with it other than just around the block. Cool how the wiring colors matched for me too.

 

[zombiess]

Did you try it with the CA unplugged? I was having a similar problem running my 18fet ecrazyman with the Cro motor on 24s2p. Unplugged the CA and it ran pretty strong. I have not had any time to play with it other than just around the block. Cool how the wiring colors matched for me too.

Yes I did try it with the CA unplugged, still had the same issue. I'm going to go try it again now that you mentioned it just for a sanity check (This thing is driving me insane!) Could you do me a favor and look inside you e-crazyman controller and tell me what model controller it is? If it's a news version it should be an EB318-XXXXX. There are several different versions of these boards available and they are all slightly different. I'd like to know what yours is.

I'm glad to hear that your sanity checking my hall/phase combo as good. I just did a some another quick check by swapping two phases (Blue and Yellow) and going through the hall sequence and was able to make it run backwards smoothly. In the phase B-B,G-G,Y-Y Hall B-B,G-G,Y-Y sequence I did find one odd ball controller to motor hall combo (B-G, G-Y, Y-B) that gave me 45mph @ 3.0A running backwards. Sounded a bit weird and accelerated slower but was smooth at speed. After swapping the phases it didn't exist so I'm just chalking it up to a strange timing sequence that got caused by the halls.

During all combination testing I have the motor set to 36A Phase and 12A battery and can actually hold stall the wheel by hand. Figured this was a safer way to test.

I'm going to go wire everything back up and unplug the CA to see if the issue goes away right now.

 

[zombiess]

Just reprogrammed it back to 35A battery 90A phase, unplugged the CA, still has the same problem, just doesn't want to work right at low speed.

What's going on with this motor/controller? Is there some sort of sanity check in the controller that's causing this? This motor should be pretty low inductance right?

At 0 rpm just trying to start off is the hardest thing the controller can do so I wonder if it's sensing that this motor is a dead short and trying to protect itself. Once it's finally spinning fast it's generating BEMF and is easier on the controller, the problem starts to go away the faster you go. I wonder if reducing the shunt value will resolve the problem?

Also of note is my speeds, I don't think this is a 13KV motor, it's more like a really big 9C 2807 motor so far unless I have something wrong.

I've also put my oscilloscope on everything during operation with no load, everything looks normal, no dropped signals, looks just like my 9C 8x8 under the same conditions.

Thoughts? Anyone? Assistance needed!

 

[John in CR]

I don't believe that controller can survive that motor to begin with. It's more than double your 9C with far lower resistance and inductance. I wouldn't even try to run it hard with an 18fet with 4110's. I'm running one of my slightly lesser motors with an 18fet controller with 4310's, and we have the battery current limited to only 65A and the controller is barely surviving, and that's with a blower pushing fresh air through the controller, and it still gets much hotter than I like. The 4310 isn't as good as the 4110 in terms of real world power handling due to double the Rds on, but it's a lot better than the 4115. Without stepping up to about a $1k controller, I don't think there are any controllers that can do the motor justice above 100V.

Even with a 36fet controller I'd run conservatively, below 100A, with 4115's, so as far as I'm concerned you've just been lucky if you haven't already messed up the controller. FWIW, never connect any accessory devices (CA, 3spd switch, etc) until after you have the bike running properly. They can only cause frustrating issues.

Please post a screen shot of your program settings.

John

 

[zombiess]

I don't believe that controller can survive that motor to begin with. It's more than double your 9C with far lower resistance and inductance. I wouldn't even try to run it hard with an 18fet with 4110's. I'm running one of my slightly lesser motors with an 18fet controller with 4310's, and we have the battery current limited to only 65A and the controller is barely surviving, and that's with a blower pushing fresh air through the controller, and it still gets much hotter than I like. The 4310 isn't as good as the 4110 in terms of real world power handling due to double the Rds on, but it's a lot better than the 4115. Without stepping up to about a $1k controller, I don't think there are any controllers that can do the motor justice above 100V.

Even with a 36fet controller I'd run conservatively, below 100A, with 4115's, so as far as I'm concerned you've just been lucky if you haven't already messed up the controller. FWIW, never connect any accessory devices (CA, 3spd switch, etc) until after you have the bike running properly. They can only cause frustrating issues.

Please post a screen shot of your program settings.

John

John, don't worry about my 18 FET IRFB4115 controller, if I blow it up, then I rebuild it, got a spare board ready because I'm testing mods. I've run it to over 85A continuous current battery draw which should have been around 200A phase and it was fine, my 2806 on the other hand was unable to take it for very long because it's just to much, but the bike would do 25mph power wheelies at 105% throttle if the motor was cold on a fresh charge. I'm not the only person who has done these mods, Markcycle is the guy I copied and he can run massive currents just like me and everything is OK + I monitor the temps of the FETs. Didn't log it this time but I was only messing around for 5 mins at high power because the motor wouldn't take it for long. Controller was only slightly warm. You should seriously try the mods I've done that Markcycle came up with, you'll be amazed. Thanks for your concern, but I'm not worried about the controller on this motor. I honestly wouldn't be surprised if I could burst 12KW out of it for a minute, but to test that I need to get this problem solved. I burst at least that high today for a few seconds when was going around 30mph and flipped the switch to high power and the front end came up about a foot and the cycle analyst showed 101A being drawn. Played with it for about 30 seconds at that power level because it was so much fun, but then went back to trouble shooting.

Also, this motor is not that KV that was posted, it appears to be something like a double wide 2807 based off what I'm seeing, just way way more torque.

Now, onto the problem at hand, here are my controller settings as you requested. I'm running 24S2P right now and these settings are a lot lower than my normal ones, but since I'm testing I'm keeping them lower than normal.

I've tried all the setting I mentioned in the posts above including lots of block time to zero block time, still the same issues. More block time seemed to make the problem worse.

Something else I just tested after Farfle and I had a conversation. He suspected that the commutation noise present on the hall signal was causing false triggering, just as I did in one of my above posts, so he suggest a few things to try. I add 0.01uF bypass caps from all the hall wires to the ground wire to clean up the commutation noise that was present on the hall signals. Went for a test ride, motor/controller still has the same problems, but 0.01uF caps cleaned up the noise nicely, 0.022uF was even better with no noticeable rounding of the leading edge. I hope they do this inside the controller, if not I'm going to start doing it myself. I was seeing over 2V of noise at the top of the signal. Farfle suggested braiding the hall/phase wire sets and separating them from each other as this solved a similar issue he had at one point. Did that and the noise dropped from 2V down to 1V. Then I added the caps and the noise dropped down to about 0.2V. 0.022uF works even better and doesn't cause and rounding off of the signal edge either. Tried 0.1uF and I got a noticeable amount of the front edge rounding.

So now I'm back to suspecting that the issue is controller related and some sort of protection is kicking in due to the current draw. I'm going to solder up the shunts next and try it unless someone else has any other ideas.

 

[andreym]

So now I'm back to suspecting that the issue is controller related and some sort of protection is kicking in due to the current draw. I'm going to solder up the shunts next and try it unless someone else has any other ideas.

I think it`s really good idea. Just do not make it too much.
What is your current r-shunt value?

 

[John in CR]

Use BigMoose's partial bypass shunt mod instead of directly modifying the shunt resistance. It is most accurate and easily changeable. Best of all with multiple resistor choices on switches or a low value pot in line with the bypass resistor, you can have variable current current limits without plugging to your computer.

Also, like I said before, some motor/controller combos don't like the phase/batt current limit ratio too low, and the phase current starvation is noticeable at low speed high load.

Zombies, big motors with lesser controllers never works well except at very low current limits. I'm just trying to save you time and money, but it sounds like you need a stack of burnt controllers for it to sink in. At least it's good repair practice. You've got a big monster motor at significant cost, and you won't be happy with performance until you have a controller up to the task. I assure you that an 18fet 4115 controller will never be up to that task. Maybe on some flat road you might get it up to a high speed a couple of times, but it can't prove durable and will never provide the zippy and fun strong acceleration performance that motor can deliver.

Last, post repair you might want to check the controller using your 9C to verify that it's actually working. The symptoms sound like one repaired controller I have. It works fine no load, but at low speed high load it makes an awful racket. I never could find anything actually wrong with it, nor could the electronic repair guy locally. It went in the broken controller box, because it's simply not worth the time, especially after conformal coating.

Good Luck with it.

John

 

[patrickza]

Do you guys have an opinion on how a lyen 18 4110 fet would handle this motor? 65A battery current ok?

 

[zombiess]

OK, just tried out a 12S2P pack I reconfigured my harness for. Same issue. Same amount of shudder. The only thing that effects how much and how long it shudders is the phase current. My settings were 12S (49V) 35A battery, 100 phase, had the exact same issue. Tried 150A phase, problem got much worse. Dropped down to 70A phase, problem isn't as bad.

Once again, I'm thinking amp protection which I've run into before on the EB212 controllers where I had to back off the throttle to get them to work again. Setting up that 9C back into this bike to test is going to be hours of work because so much needs to be changed, the bike was heavily modified to work with this motor.

Going to see if I can find BigMoose's partial bypass shunt mod that you mentioned. Do you have a link, searching for stuff like this on the forum is sometimes hard and I haven't found it yet.

Sigh, maybe I'll take the time to install the old 9c tomorrow, then I can eliminate the controller as the issue. There goes 3-4 hours of my life, but at least the bike should run again.

 

[itchynackers]

Are you sure you don't have a phase short somewhere? Maybe in the axle. Do a continuity test to be sure they aren't at least grounding to other wires. Maybe others can help diagnose with a multimeter. I know it is a new motor but sometimes shit happens in shipping.

 

[John in CR]

Here it is, but instead of the pot there, since infinite is possible, put the pot only only in line with R1 so resistors set your max gain and the pot varies from programmed settings to whatever ratio you create with the resistors. I'm going to start doing this for any controller I open and mount a 3 way switch on the end. Cut R1 for programmed limits, 50% increase, and double the programmed limits for my 3 hardwired settings. http://endless-sphere.com/forums/viewtopic.php?f=2&t=31643&p=494761&hilit=shunt+bigmoose#p494927

I thought maybe the 9C was easy, either on another bike or able to clamp on the rack to test the controller. Obviously not worth a full swap out. You don't have a vise to clamp the 9C into?

I think this one is the controller, though one time I did have some kind of weird hall issue where the motor did fine no load, but not under load. That stator is set aside for diagnosis once I get some halls in, so I don't know if it's a sensor or a short that rears it's ugly head only when lots of current flows through the phases. They test fine, but something is wrong, and it's definitely not a phase short.

 

[Alan B]

Dropping all the way to 12S should have dropped the phase current a lot. If there was no change then perhaps something else is wrong. If this motor is like a double 9C then phase current on 12S is like a 9C on 24S.

Is it certain that all phase combos have been checked? Sometimes there is one that seems okay but isn't actually the correct combo. Need to verify all six phase lead combos for each of the two hall settings.

Wonder if one hall device is not working right. Scope each one? Could be bad controller input on one hall as well.

Good luck. Frustrating to be so close.

 

[zombiess]

All halls from the motor have been scoped, I even went so far as to make sure they were aligned by using both traces and counting how many mS they were overlapping at a given RPM, they look fine.

I'm not sure which route I want to go first today, spend the time rebuilding the harness for the 9c an mounting it again, or just pull the controller off, take a shunt off another controller and tack it across shunt pad for a 4th one (EB318 controllers had space left for 4 shunts but they only put 3 holes, solder space is still there) blank and try that. (This shunt mod would be easily removable if it's not the problem.

 

[zombiess]

Alright, I just got back from a 30S2P ride on the same controller but I installed my 9C 2806 and it drove well on 12S, but the first few times went WOT from a dead stop it felt weak almost similar to what was happening before but not the same. I then flashed it my normal high power program and no problems at all! Even did a few power wheelies from a dead stop to make sure it didn't miss fire at all, a few 0-50mph pulls/etc and the normal abuse routine i put the bike through. Motor was smelling a little so it was at least 120C. So the controller being bad is now eliminated.

I'm now down to two things, either there is some issue with my motor, or there is a problem with the current sensing in the controller. I'm going to install the hubzilla right now with the exact same settings I just rode at and see if it still acts weird. I read these controller have some sort of auto learning ability but I have no idea what that means or if it goes beyond the hall sensor setting which I have set at 120 degrees.

If it still stutters at low speed then I'm going to shunt mod my controller.

Good news is now I have a 2806 I can easily swap into my bike at any point.

I'm also having Lyen overnight me another 18 FET 4115 controller so I'll have one for comparison.

 

[zombiess]

Hubzilla installed again, same problem. Can make it almost go completely away by running low phase amps, problem gets much worse with higher phase amp settings.
I can tell this motor is going to be awesome once I get this problem sorted out.

Same thing with regen. Low phase amps, smooth stopping, high phase amps, cuts out intermittently until it slows down. Just took the controller off the bike to shunt mod it. Will report back after modding to see if it solves the problem.

 

[zombiess]

Modded the shunt by adding a piece of 12 gauge copper next to one of the shunts and turned the battery and phase currents way down and then some fun settings. Same problem so the issue is most likely with my motor. Probably caused by when I spun the axle wires in the drop out. The hall wires were not harmed, but I'm going to tear everything apart and see if I can find a problem. Will probably replace the hall sensors with SS41s while I have it open.

Dammit, why can't things just work right for me. I still have another controller coming on Tuesday which I'll be able to test with as well.

Guess that's all for now, time to open this beast up and see if I screwed up an internal connection when I redid the wiring.

 

[zombiess]

Gonna go out on a limb and say I found the problem finally

Looks like I have a bad hall. Even though it scopes out OK it must get wonky under load. This hall looks like it exploded and has pits in it. Not quite sure how it got fried unless it happened when I was testing phase combinations. When the wires twisted 1/2 to 1 turn around the axle none of the hall wires were harmed and were all fully in tact. I even pulled them all out and inspected them, none of the insulation was hurt.

Maybe it's just a dud from the factory, what ever it is, it's getting replaced with a Honeywell SS41

Guess I have the joy of replacing hall sensors now... yay. I just hope that fixes this so I can enjoy it. Gonna have a spare 18 FET controller to mod too for just $300 overnighted. I'll probably mod it just like my current one and sell it on here for someone that wants a compact 126V 60-80A controller or just hold onto it as a spare.