Zero torque after "warm start"

[japperrrr]

First of all, let me introduce myself to you guys.

My name is Jasper and I am 21 year old. For the last three years I've been studying Architecture at the Eindhoven University of Technology (Netherlands). Recently, I also started with a bachelor Automotive (Faculty of Electrical Engineering) at the same university. Almost 3 years ago, I bought an electric scooter. The public transportation here is not that fast and I need to travel 15 km to get to the university, so I drive my scooter on a daily basis. I really enjoy driving it and I have never had major issues with it. Until yesterday... Hopefully some of you guys can recognise my problem and help me to solve it.

Also, sorry for the huge amount of text. I made a summary at the end, but I think it's necessary to give this much information as often the solution is to be found in the details.

I'm driving a Novox C50 scooter with a 3500W hubmotor and a matching controller. I replaced the lead batteries with Winston(a.k.a. Thunder Sky) LiFe(Y)PO4 cells. It runs on 51.2V nominal, 16 cells. So far, so good.

About two weeks ago, I noticed that when I drive away from a standstill, the scooter sometimes "shocks" softly one time before accelerating. First I thought that this was not a problem, until recently when the scooter suddenly had less torque (it felt like 1/3rd of torque was gone). It didn't make any strange noises, it simply had less torque. After turning it off and on a couple of times and rotating the motor by hand, the problem disappreared.

For about a week I have been driving it daily without any problem: I drove it to the university every morning and back home at least a few hours later.

Yesterday I drove my scooter to the supermarket where I was inside for approximately than 5 minutes. When I wanted to drive home, the scooter turned on but had no (zero!) torque at all. Only when I turned it on and gave throttle, it would give one really light shock. I took the wheel with the motor off the ground and tried the same things I had tried before when I experienced low torque, but this time turning the scooter off/on and rotating the wheel by hand did not fix the problem.

Then, suddenly, the wheel started spinning when there was no load on it. But when I tried to drive the scooter, the same problem reoccurred. I had no other choise than to push it back to my home.

About one hour later I tried to drive the scooter, and quite surprisingly: I could drive without any problems at all. I drove it for about 300 meters and then I turned the scooter off and back on. Guess what?? Same problem, zero torque again.

So, summarized:
- Used to drive perfectly for months
- Started to give tiny shocks when driving away from stand still
- Lost about 1/3rd of torque for half an hour
- Lost all torque for half an hour

- The scooter drives from a "cold start", but when I turn it off and back on, it does not drive anymore.

- I don't have any idea on how I could determine the problem, as it only appears after a "warm start"


Extra information:

A few months ago the controller short-circuited (strong resistance on the motor), and I received a new controller from the manufacturer. The old controller had been damaged by water entering the controller through the cable harness (design flaw). When I installed the new controller, I changed the position of the carble harness, such that water can not get inside anymore. Together with the fact that there is no unusual resistance on the motor, I assume that the current problem is not caused by water damage.

The throttle is controlled by a hall sensor. Also, the scooter has a built in break switch and side stand switch. But again, I don't see how any of this could cause the problem I am experiencing.

I replaced the standard halogen light bulb with a xenon system, which seems to influence at least the speedometer when I on my lights. The speedometer and xenon light both work on 12V (from a seperate transformer), so I don't think this affects the controller in any way.

If needed, I might be able to buy second hand original parts to replace possibly broken ones instead of repairing them.

If any extra information is required, please ask!

 

[Lebowski]

Man, TU Eindhoven, I studied Electronics there back in the early 90-ies i remember the machine hall off of the main building where we had to do electromagnet practicum, is this where the automotive part is ? I remember a subgroup of Thor was for the power guys, what was it, Odin ? Does the Walhalla still exist ? I studied analog IC electronics which was on the 9th floor then. A few former collegues are professors there now, E. Cantatore and whats-his-name, the huge Status Que fan...

Anyways, as an electronics engineer you should ditch the controller and build a new one.

Looking back at your post i have the feeling something is wrong with the power supply of the gate drivers. Your symptoms sound like FETs are not turning on. But sometimes everything is fine, and since FETs are typically broken or not it sounds more like the gate drivers or the gate drive power supply.
Dependent on howthe supply is build it can be it powers itself, but for this to work it needs a startup circuit. Your symptoms sound a bit like the suppy is weak and barely able to sustain itself, and refusing to start when warm.

 

[japperrrr]

Thanks for the reply.

I'm sorry to say so, but the main building is being demolished as we speak
The Faculty of Electrical Engineering has moved into a brand new building called Flux. A lot of buildings are being renovated (Vertigo finished few years ago) and also the whole public space on the campus is being changed.

Building a new controller myself is not going to be that easy, as I only started my bachelor Automotive this month.

Indeed, what you say makes sense. The motor is probably fine, so if it doesn't have any torque it must be caused by the FETs not turning on. Also, if they would be broken I think there would be a resistance on the motor, correct? Then, to find out why the FETs are not turning on we must probably look deeper inside the controller.

Would it be possible to identify the gate drivers or the gate drive power supply if I open the controller and take pictures of it? The controller is probably high quality as the company that sold these scooters was very proud of it. Therefore, it might be my best bet to try to fix it.

Too bad that Novox went out of business after a fire destroyed their workshop. No support from them anymore.

[edit]
By coincidence I stumbled upon a forum post (in Dutch) where someone turned out to have a broken 5V supply in the controller, and it caused the FETs to work improperly. Could this perhaps be the same problem?

 

[Lebowski]

Just looking at pictures is difficult, especially when it's all SMD components. If it's all commercial IC's you have a chance, if it's discretes then you really need the schematic.

Building a controller is not that hard:
https://endless-sphere.com/forums/viewtopic.php?f=30&t=57877

That was a nice 13 story building, what was wrong with it ? University all of a sudden got a lot of money ? I was there from 1991 to 1995... Still remember the rivalry between the electronics and mechanical engineering students, but I guess EV's put an end to that rivalry

 

[japperrrr]

Haha, looks rather difficult to me! But I might give it a try if I don't have other options. At least then I know how to repair it if it would break again.

Anyway, I will take the controller off the scooter again and open it up. The controller is said to be designed by the company themselves and they did not sell that many scooters, so I might have luck on this one. It's quite a lot of work to remove the controller from the scooter, but I might just have time for that today.

Oh and I forgot to say, I think it has a 6-phase hub motor. At least it has 6 thick wires coming out of it.

Rivalry is still there, but in Eindhoven electrical engineering is winning. The bachelor Automotive is even part of the Department of Electrical Engineering, so that is indeed a huge difference compared to other universities. As far as I know it's even still unique in the world. Though, many courses are still part of the Department of Mechanical Engineering. It's just that the main focus has shifted from mechanical to electrical engineering.

New Metaforum building (mainly the library):
http://www.architectuur.nl/wp-content/uploads/2012/10/tueind2.jpg
http://www.ectorhoogstad.com/sites/www.ectorhoogstad.com/files/project/beeld/copyr_joep_jacobs-bewerkt-eha_klein_0.jpg
http://www.ectorhoogstad.com/sites/www.ectorhoogstad.com/files/blog/beeld/1416e-int-0003.jpg
https://static.tue.nl/uploads/pics/TU_MetaForum-4093--_NvO_02.jpg

New Flux building (Electrical Engineering):
http://www.bouwwereld.nl/wp-content/uploads/2014/11/Flux-TUe.jpg
http://www.nlingenieurs.nl/wp-content/uploads/2015/09/Flux-overview_BartvanOverbeeke-1024x714.jpg
http://www.royalhaskoningdhv.com/nederland/~/media/royalhaskoningdhvcorporate/images/projects/buildings/tu%20eindhoven_thumb.jpg
https://pbs.twimg.com/profile_images/494465717480079360/tMXzk4-y.jpeg

 

[Lebowski]

Oh man you can just tell the artist renderings are fake, I mean, I've never seen any women in the Electronics building but it definitely looks very nice !

Are the 6 wires wired as 2 independent motors or do you get both ends from 3 coils ?

 

[japperrrr]

Opened it up and fuckk, water inside again. The only way it could have gotten inside the controller is by running down on the outside of the cable harness and then entering the controller through the connector... (see red arrow)
The end of the cable harness had been positioned so that water could not get inside. But now I realise, this was not where the water really came from.

Anyway, now we know this, I must say that I am surprised that the scooter would even drive at all! The reason it didn't "warm start" was probably the water being thrown around all over the controller.
It's good to know that it did still "cold start", so that means that there is no major damage done by the water. Otherwise it would simply not drive at all. I'll clean up this mess and let it dry for a couple of days before installing it again. And of course, I will make sure to replace the cable harness to not let this happen again.

How long do you think I should let it dry?

And also, the one photo with my hand on it shows the cables that are coming from the motor. Six thick cables and 6+4 thin cables (sensors?).

 

[Lebowski]

typical, not waterproof so a mechanicals engineers fault

 

[japperrrr]

Yes, indeed. Not really a electrical engineers fault haha. But it still sucks...

To see how the hole for the cable harness had been sealed, I opened up the connector. There was some kind of inverted conical rubber to which the end of the tube should be pushed against. Didn't really work, though

Luckily my creative mind didn't let me down today. Found an old inner tube in the barn, from which I cut off a few cm. Somehow I managed to get all of the connectors through the inner tube and I pulled it all the way down till the other end of the cable harness. Now it is being pushed against the conical rubber and it makes a nice squeezy sound when I try to rotate the cable harness. Guess that means that it's properly sealed now.

Also, I noticed that some sections of the PCB have been damaged by the water. Water has gotten underneath the isolation, and as seen in the picture, a lot of the tracks are now without isolation. Should I worry about this and is there a fix for it, or should I just close the controller and never look at it again?

 

[Lebowski]

If it runs, it runs.

I typically use this rubbery hot melt glue to make things waterproof. Especially on metal ypu can use it as a gasket and it comes off quite easely when doing repairs. I wouls add a bead around where the wires enter the controller. Maybe it is also an idea to drill a small 2mm hole at the lowest point of the controller to let the water out.

 

[japperrrr]

The inner tube should in theory do the trick, at least for not letting water get inside. I don't see how else it could have got there, since the cable harness itself has been completely filled with Polymax adhesive and its opening is orientated away from the rear wheel. Still, the hole doesn't sound like a bad idea.

Last time when I had the same water damage I did think of it, but I thought I had solved the problem already and therefore it could only do harm. Knowing what I know now, I think it's indeed a sufficient way to prevent this in the future. I don't think I will have to worry about moisture getting inside, since it would still drive with even a 100 ml of water inside haha :lol:

[edit]
I measured the resistance of all mosfets and compared them to eachother, and I made sure that there was no short circuit over the main power plug. Everything seemed fine, connected it to the scooter again, and it works like a charm. No shocks or any other problems.

Do problems with water getting inside the electronics occur often?

 

[BahamasBiker]

Very informative thread, thanks everybody.

Don't forget contact cleaner on all connectors. Needed every two years, usually. More often, if in salty weather.

 

[wijnand]

Great to see that you recovered your controller. I have now an issue with mine.. Due to water a lot of aluminium corrosion was developed inside and shorted somehow some Fet's.

I'm now in the proces of replacement and have a question about the taking apart of the controller. I managed to remove the backplate and see the PCB and the Fet's underneath. But how do I separate the case with outside cooling ribs from the massive aluminium strip inside? There seems te be no screws?? See picture below.

Japperrrr was able to manage the taking apart as he showed this in picture nr.3 in his previous post.. How did you do it??
The 4 bolds are loose now, but I think they don't hold the big aluminium strip behind. HELP!

 

[wijnand]

In the meantime I managed to open the controller more and replace the Fet's. Now the second problem appears: no 5V voltage on the controller. It makes it onboard from the main power. It uses a small Hf transformer to lower the voltage from 65V to a lower acceptable voltage for two on board 7805 Voltage regulators.
On the primary site of the transformer there's this high voltage available but there isn't any Voltage on the secondary site. The HF transformer seems not to be switched.
Does anybody know how this works, I don't have an oscilloscope so can't see if there's anything running..

 

[flippy]

you can just grab 12v from the inverter that drives the lights and grabd a 7805 and replace the 5v feed from the controller in order to drive the hall sensors. just put some heat shrink around it, it does not get hot at all.