Motor cutting out

[pajtaz]

Hello,

I joined the forum specifically to ask for your help with my problem. For the last few months I have been reading various forums to help me upgrade my existing bicycle to an ebike. I have read many forum posts on Endless Sphere and they have helped me a lot. Now I have a problem for which I cannot find a solution on this or other forums so I figure I ask what you think.

I have a problem with motor randomly cutting out for a split second or maximum two seconds. I will explain below.

Two weeks ago I completed installation of ebike kit. My setup is:

• Q128C 48V 328RPM 500W rear hub motor from BMS Battery (bought it with the wheel)
  KT36/48SVPR-20A sine wave controller from PSWPower, nominal current 10A, max current 20A +/- 1A
  KT LCD3
  48V 12.8Ah bottle battery with maximum discharge current of 30A
  I have only PAS, 12 magnet, right hand side, magnet KM-DISC17 and King-Meter T281 Double Hall Sensor
  No throttle, no cruise control
  Two hidden wire brake senors

LCD3 settings are:
Max speed = 45 km/h; Wheel size = 700c;
P1 = 212; P2 = 1; P3 = 1; P4 = 1; P5 = 15;
C1 = 7; C2 = 0; C3 = 0; C4 = 0; C5 = 8; C6 = 7; C7 = 0; C8 = 1; C9 = 0; C10 = n; C11 = 0; C12 = 4; C13 = 0; C14 = 2.

Overall, motor is quiet, runs smoothly, motor/controller react quickly to starting and stopping. Problem is that motor randomly cuts out for a second or less than a second. This happens at any assist level, at any wattage, seemingly at random. I changed battery voltage monitoring to P5 = 0 but I saw no voltage drop that would cause the cut outs. Either way it is happening at 150 W or at 250 W or at 400 W at various assist levels. It happened at PAS level 2, 3, 4, and 5 (although I never ride on 5, was just testing it). It only happens for a second or less, and could happen several times within 1 minute or it could happen once every 5 minutes. I am riding on completely flat ground with occasional bump.

When I am talking about cut out, I mean that motor stops delivering power for less than a second, maybe maximum of two seconds, no error codes, no power loss on LCD, just motor not delivering power. It even feels like it is braking at that moment (maybe that is air resistance, can't really tell). Motor power on LCD starts going down and depending on the length of the cut out, it will either drop 50-100 W or all the way to 0.

I ruled out brake sensors because I disconnected them and problem is exactly the same. I ruled out the battery because it happens at various charge levels and I never see a sudden drop in voltage on the LCD when motor cuts out. I have tested various C1 settings for the PAS (5, 6, 7) but 5 and 6 make it worse so I am sticking with 7. I have not yet completely ruled out PAS but I can't figure out why it would be causing this issue. Maybe it's the controller or maybe the motor. It seems to be correlated to demands in power, or at least I assume that is.

For example, on the way to work (17.5 km one way), it was fairly good but occasional cut out of the motor. There was wind in my back. However on the way back the wind was against me, it was very strong, probably the worst wind in the last 6 months, supposedly gusts of 44 km/h. So the motor was cutting out all the time, on all PAS levels, mostly on 2, 3, and 4. There is clearly a connection between strong gusts and motor cutting out. Maybe the motor wants more current when it cuts out and controller can't deliver it or is somehow limited in delivering it. Maybe this is due to settings in LCD3 or due to build of controller. Thing is, it is a 20A max current controller with nominal current of 10A, designed for 36V and 48V (PSWPower KT controller). Besides, the cut outs occur at motor power of way less than 500W. They happen at PAS level 2 (150W), level 3 (270W), and level 4 (380W). It occurred at least 15 times yesterday at level 3 and motor power is limited to around 270W in that level. I cannot see how it can be the motor (unless its faulty) and the controller.

Again, motor otherwise runs so smoothly. It is quiet, no hickups, no grinding, quick start, quick stop, just runs beautifully besides these cut outs. It just doesn't make sense.

Only thing I can say about PAS is that I just now read on the website where I purchased it that the sensor should be installed UNDER the derailleur cables, on the frame below the BB, under the cable guide. I have it now installed over the cable guide and maybe that is the reason for problems. See this photo of the cable guide: https://en.wikipedia.org/wiki/Cable_guide#/media/File:Cable_guide_2.JPG Otherwise the sensor is around 1 mm from the magnet ring but maybe the sensor is just at the edge of detection and it should be closer to the frame. Damn, it would be silly if it was something like this. I will say that some of the motor cut outs were connected to shifting to a higher gear, not always but if I could predict when it will happen, it would be when I am upshifting (even if I am keeping the same pedaling frequency).

Otherwise, if it is not PAS, what do the people think it is? I know motor is not getting hot because I stopped a few times when it was cutting out frequently and touched the motor. It wasn't hot. I can't touch the controller since it is in a case but few days ago I opened the case and saw no signs of excessive heat. Any help is appreciated. Thank you ahead of time.

 

[E-HP]

Welcome to the forum!

Sounds like you did a fair amount of trouble shooting so far, and the fact that you have the problem on flat ground and low wattage, as well as under load rules out a lot of things.

Can we see a picture of the actual PAS installation? That's an interesting sensor you're using, and the mounting seems problematic since it appears to make an assumption about your frame's bottom bracket shell diameter, since it's fixed to the frame with no apparent adjustments. Many are adjustable or are indexed to the bottom bracket axle (set at a known distance/radius) to ensure proper alignment.

Could be if the alignment is marginal and there is any flex in the frame or bumps in the road are enough so the sensor is not picking up the signal from the magnets. I would start there with the current sensor, or purchase a different sensor that allows you to adjust the alignment with the disc.

 

[pajtaz]

Well, it was the PAS. You were right, alignment was the problem but more so the distance between the sensor and the magnet ring, not so much the angle or distance from bottom bracket. I was so sure it was something else but once I took a look at it today by flipping the bike over and testing, I knew it was the PAS.

First thing I noticed is that I installed the sensor incorrectly. Magnet is fine but sensor should have been installed against the frame but I had it over the bottom bracket cable guide. The sensor comes with its own cable guide and a cover. So I had to first remove the cable guide, put the sensor against the frame where cable guide was, put sensor cable guide over it, run the cables for derailleurs then install the cover and tighten everything. It should have still worked anyway because the sensor was still at an ok level with respect to magnets but I should have installed it properly. I obviously had a brain fart when I installed it first because I didn't get why the sensor came with two parts and one of the looked like the cable guide. Duh! But hey, it happens.

Then I started testing. This sensor lights up all the time unless it is sensing a magnet. Once the magnet passes, the light goes off. I noticed that even though the sensor was fairly close to the magnet, the light was flickering a lot. It was obvious it was not sensing the magnets well enough. Once I pushed it so it's basically touching the magnet ring and then tested it, it finally had nice clean evenly spaced signals. I counted all the signals and watched them to make sure they are clean. I finally had clean 12 signals.

I tightened everything back up, had to re-adjust the rear derailleur because of the new cable guide and took it out for a quick test ride. I rode maybe 2km but had no cut outs. The real test will be tomorrow riding to work but I think I fixed it. The only thing I can think that can cause problems is if the magnet ring moves. I am not sure if this happens when strong wind blows or when I go over bumps but the magnet seems to be strongly attached to the chainring with magnetic force. I have now ridden over 250 km with the magnet ring attached like this and when I looked at it today, it seemed to be in the exact same place as when I put it. I was hoping to attach it more rigidly but if it hasn't fallen off so far, it must be adequate.

Man, I am so relieved. I had worked myself up worrying that it is the controller or the motor but at the end you were correct, it was the PAS. It's very sensitive and it must be super close to the magnet ring. If it all goes well tomorrow, I will make a new post and show some photos in case other people run into the same problem.

As for the PAS and the magnet ring, it is these two items I got because I didn't want to remove the crankset.
Magnet ring: https://www.atmparts.eu/p24_eu/eshop/en/2W/KM-DISC17
PAS: https://www.atmparts.eu/p24_eu/eshop/en/2W/KM-T281-R

This magnet ring and PAS are designed to work with each other in terms of diameter of the magnet ring and how the PAS is installed. I am assuming that this was intended to work with Hollowtech II bottom brackets since the magnet ring magnets are at 64 mm diameter and Hollowtech II BBs are installed in an appropriately sized axle. The sensor is being presented as a solution to have a PAS without having to remove the crankset and without having a place to install the sensor on the crankset compatible with Hollowtech II. So I guess King Meter figured that whoever buys this sensor is a person with Hollowtech II BB and therefore the axle size will be compatible. I can tell you that is EXACTLY why I bought it, I didn't want to remove the crankset to install the magnet ring and/or the sensor.

 

[E-HP]

I'm glad it was something simple!!! :thumb: Your detailed description of the problem and your trouble shooting will definitely help the next person having similar issues that finds your post.

 

[pajtaz]

As promised, here are a few photos of how the magnet ring and PAS sensor are installed. You can see the sensor with the small red LED at the bottom of the photo, next to the magnet ring. Magnet ring is attached to the smallest third chainring by magnets.

In order to install the King Meter T281 sensor, first I had to remove the original rubber cable guide. I put the rear gear into the highest and front gear into the smallest chainring. This gives the least amount of tension on the derailleur cables. Then I removed the mounting bolt holding the cable guide and removed the cable guide. Then the sensor goes under the derailleur cables, then a rubber cable guide provided with the sensor, then the derailleur cables over the rubber cable guide, and finally the cover for the cables. The mounting bolt is inserted through the cover, cable guide and the sensor. Before inserting the bolt, you make sure the sensor is close enough to and flat enough against the magnet ring and you test it to make sure the sensor is sensing magnets evenly and correctly. Then you tighten the mounting bolt. Voila! Now you have the sensor against the frame, tight, derailleur cables going over and protected. It's actually neat, nice idea, keeps the sensor in place and somewhat protected from the elements and dirt.