Strange power break downs with KT36 controller with MAC rear hub motor

[HolgiB]

Hi all,

In the last time I have had strange issues with my MAC hub motor running together with a PAS sensor, KT36 / 20 Amps controller and LCD8. It is kind of hard to explain. So I rather link to my video (please ignore the German text :
https://www.youtube.com/watch?v=5wUuqlKUeLk

I see those break downs only at higher power levels (4-5) which translate to ~ above 400 watts. It seems to me that the three phase wires running parallel to PAS-Sensor cable cause disturbances. I have had similar issues with my old 9C RH205 hub motor which I used prior to the MAC.

I already exchanged the controller to a smaller 14 amps peak KT36 but the issue remains.

My questions:
- Has anyone of you run into similar issues ? I guess the higher phase current, the more prone to influence those sensor wires become.
- Would it make more sense to add shielding to phase wires or the sensor cable ? I was thinking about wrapping it with aluminium foil and use some short wire to ground the improvised shielding.
- Or should I exchange the unshielded sensor cable to shielded cable ? (e.g. microphone cables).

Thanks in advance,
Holger

 

[lionman]

What battery do you have?

Looks like it is sagging quite a bit for such a tiny load.

Its dropping around 4V with only about 10A of load?

I would be checking battery connections. If there is a bad connection it could be resulting in voltage drop. After that, check internally in the battery to make sure its well balanced and that there are no signs of dead cells.

 

[HolgiB]

Thanks for your reply !

Battery pack is a 10s5p build out of Samsung 29E cells which handles my bikes quite well.

I am unsure how you see 4V drop in voltage. Between the first part of the video and the second I was already riding a bit. Thus voltage from the battery drops as expected. IIRC the second part was after quite some demanding uphill rides with high power level.

Just look at the second part of the video where I drive uphill with cruise mode enabled and you see that the battery voltage drops close to nothing.

0:23 min - 37.1V - 350 Watts
0:31 min - 36.5V - 450 Watts before "break down"

0.6V drop in voltage is not really bad given that the Samsung 29E is not exactly a high current cell.

Also the pack powers my other e-bike (28" with PSW Power 350W rear hub motor) quite fine. Not power drops up to 450 watts.

 

[lionman]

Ah I see, I didn't realise that so much time had passed. It appeared as if you where crawling along drawing under 100w and then you accelerated drawing 400w and the voltage dropped.

Do you have a throttle to test it with to rule out issues with the PAS?

 

[HolgiB]

Yes, I have a thumb throttle lying around somewhere.

Do you think the PAS could be cause of the problem ? It seems pretty strange to me that I saw this strange power loss during cruise mode without pedaling.

BTW: Yesterday I changed the routing of my cable a bit and moved the PAS wires away from the three motor phase wires going up the seat tube.
Guess what ? Nothing changed.

Again power drops when riding uphill with more than support level 3 from the controller.

What drives me nuts: This issue somehow seems to be power depending. I had power drops at level 4 when the battery was at full 42V but was running level 4 on my way home for over 90 minutes without issues (but the battery voltage was lower than).

I will try with thumb throttle and report back.

 

[docw009]

My questions:
- Has anyone of you run into similar issues ? I guess the higher phase current, the more prone to influence those sensor wires become.
- Would it make more sense to add shielding to phase wires or the sensor cable ? I was thinking about wrapping it with aluminium foil and use some short wire to ground the improvised shielding.
- Or should I exchange the unshielded sensor cable to shielded cable ? (e.g. microphone cables).

The majority of commercial ebikes use a 8 or 9 wire cable which lays the Hall sensor wires right next to the phase wires. No shielding. I figure the frequency of the Halls and phases is on the order of 100 hz. I would not worry about crosstalk.

Since it happens at high current levels, maybe the MOSFET's are breaking down. Are the screws tight on the heat sinks?

 

[HolgiB]

The majority of commercial ebikes use a 8 or 9 wire cable which lays the Hall sensor wires right next to the phase wires. No shielding. I figure the frequency of the Halls and phases is on the order of 100 hz. I would not worry about crosstalk.

Thanks for that hint ! :thumb:

I was already wondering if there are similar issues with commercial kits and that shops would rather sell shielded cables if things like this happen more often. Separating phase wires and PAS sensor wires by around one inch didn't change anything. Same issue.

Since it happens at high current levels, maybe the MOSFET's are breaking down. Are the screws tight on the heat sinks?

Good hint but it happens with both controllers I tried (KT36 / 14 Amps max & KT36/20A max).

After tinkering around yesterday I can say so far:
- Changing the PAS sensor didn't solve the issue.
- Running with thumb throttle works great. No power drops. No hickups from the hub motor at power levels above 400 Watts.

Unfortunately this doesn't help since riding with PAS sensor is mandatory where I live and also I prefer this riding style.

Open questions to me:
Could this issue be related with unclean commutation of the hub motor ? From what I know the MAC (even a slow model like mine) spins much faster than the average Bafang geared hub or even slower turning DD motors. I have never run into issues with sensorless controllers but may be this would help with the MAC ?

Would a PAS sensor with more signals help here ? E.g. a V12 with 10 magnets in the disc ?

 

[HolgiB]

OK, after tinkering around with various setting for limiting max current (Parameter C5), PAS sensitivity (Parameter C1) and general PAS responsiveness (Parameter C14) I didn't exactly find a solution but a work-around. :lol:

Limiting the maximum current to 12A removed the strange issue with my controller. I can now ride even slower than 6 km/h slowley pedaling at support level 5 without those hickups. I am still not shure what is happening though but since I experienced a similar issue with two different controllers (KT36 / 14 Amps & KT36/20 Amps) as well as with two different hub motors (9C RH205 2705 & MAC 12T) I am pretty confident that I can rule out a hardware issue specific for any of the components.

Eventually the controller simply runs into a peak current it is not able to deal with?
May be having a controller with Hall sensors would help here or may be not ?
May be a faster PAS sensor providing more signals at lower pedal speed could improve things ?

But unless someone has a good explanation what is happening here, I will rather spent money on this topic.

Using a more solid controller such as those EB306 from Grin might be a smart approach but I ended up with a price tag of 430€ when I composed a kit for me yesterday.

Not something my wallet would easily handle !

 

[amberwolf]

If you're going to replace the controller anyway, you could also look up Casainho/Stancecoke/etc's thread(s) about the KT/Kunteng/Kun Teng series of controllers, and their custom firmware for it (and the LCD3 if you have one), and see if that resolves your problems.

If it doesn't, no money lost; if it does, then no money spent.

 

[HolgiB]

Hey amberwolf,

thanks for the hint ! :thumb:

I already know about the custom firmware for KT / S12 controller series from the german "pedelecforum". I guess you are right from a money standpoint but so far I have bolted down four options for me:

1) Buy some EB3XX 9FET / 12FET controller plus Cycle Analyst to solve my issues imediatly. This would mean around 300-400€ plus soldering Hallsensor connection and thermo resistor on my MAC which I currently have not connected.
2) Buy some FOC / Sinus controller plus JLCD-display which will most likely solve the issues. This would mean around 200€ without any modification to the controller necessary.
3) Getting a big KT-controller with something like 30-40 amps max current and limiting it down to something like 22-25 amps in the hope that more FETs (12 compared to 9) are better able to deal with the Mac in high current situations.
4) Buy a preconfigured kit with nice LCD display plus FOC / Sinus controller, which a german shop also sells in combination with the Mac themself which would mean > 350€ in the end.

Here is something I found in the web researching valid controller options for the Mac:

The 500W Mac motor is recommended for use at 36V or 48V with recommended controller limit of 25-30A but this is really
dependent upon the application and the motor winding that is used. For further info on suitable matching of Motor and
Controller please refer to the kits section, but typically the High Spec Infineon Controllers (IRFB4110 or IRFB3077 fets) are
recommended. The slower wound motors such as 12 or 10T will work well with the 6 fet, but the 9 fet is recommended for
the 8T and possibly the 12 fet for the high speed 6 T if they are used with wheels larger than 20”.

Since my Mac is only 12T in a 26" MTB rear wheel and used with a 36V battery, I would expect any 9FET controller with "good" FETs to work with it.

Hall sensors should also only be a nice to have but not a must have. KT controllers (while widely used) are rather average controllers and since the Mac is somewhat of a "hub motor princess" does work below its possibilities. :lol:

The more I look into the possibilities, the more unsure I am what to choose. Especially since the bulletproof solutions are the more expensive ones....as always in life...

 

[HolgiB]

Okay......stuttering with the MAC has come to an end now ! :lol:

I have ordered a Grinfineon C4825 25 Amps Sinus controller plus PAS plus CA V3 and the slim digital aux button satellite.
While my wallet somewhat "bleeds" my heart as e-biker is eagerly awaiting the package.

 

[Jatem]

Did you figure out the cause?

 

[HolgiB]

Good morning

yes, I figured out the cause :lol:

The cause (so to speak) is the MAC itself. While apparently every other hub motor I used seem to work fine with those cheap KT controllers from various vendors, the MAC seems to be somewhat the princess of brushless motors. Even the Grinfineon is barely able to cleanly control this sissy without Hall sensors.

This is even more funny since Grin / ebikes.ca has been selling this as a kit. If I was in the position again I would rather choose a more beefy geared hub motor similar to the 500 Watts Bafang rather than sinking ~ 500€ into a hub motor.

Oh well.....

 

[Jatem]

I was considering a MAC 10t, with a KT s12s and LCD8, for a big dummy cargo build. Maybe I should look for another motor if I'm going to use the s12s controller.

I would really like to keep the cabling and display tidy looking, hence wanting an LCD8 with KT controller, as the CA is very chunky.

 

[HolgiB]

Well, while the MAC itself has a very high reputation for being able to deal with high current under high loads, I still think that another high powered geared hub motor (e.g. Bafang 500W) plays at similar levels if you crank up the current (> 20 amps). I have been running a DD 9C RH205 before together with a KT 12FET controller with 22 amps peak current and it played along nicely.

KT plus LCD8 makes a nice and clean package IMO but does not like the MAC.

There is one big benefit using the CA plus MAC and this is temperature control. The NTC in the hub motor plus CA allows you to fine tune your temperature limit which then lowers the current if the motor gets too hot.

On the other hand: All other e-bikes I have build over the last years have never had issues with a fried hub motor which of course becomes more likely if you crank up current and ride up steep hills with heavy load.

There are simply limits in regard to hill riding capabilities for hub motors which also apply for a MAC. Sometimes a BB drive similar to a Bafang BBS HD might be a more valid solution.

 

[Jatem]

I won't need to climb long hills, and most of the time it'll be <7% grade. I was hoping to limit the current to 15A to save battery life. I could get an ezee motor, but I've read they are fussy about controllers.

 

[HolgiB]

Well, if no long trips through mountains are planed I would simply stick to a big DD like those 750W / 1kW kits you can buy at e-bay. Depending on the vendor they will be most likely sufficient for what you need as well they are robust enough.

Edit: I mean something like this:
https://www.ebay.de/itm/Hinterrad-26-48V-1000W-Elektro-Ebike-Conversion-Kit-Elektrofahrrad-Umbausatz/392252655832

The website is german but I guess the pictures speak for them self. A more slower motor would be better for increased efficiency in hilly terrain.

I can not comment on ezee hub motors but all other geared / non-geared motors I tried worked perfectly with the Kuteng family.

Only the MAC.....
But we already had this story....