Adaptto Midi-E overheating problems?

I just finished upgrading my Motorino XPd scooter with a new motor, new batteries, and a new controller, but I'm not getting the kind of power I was expecting.

My previous setup was 4x nissan leaf modules modified to 16S1P, a Kelly KEB72301X controller, and the stock 500W motorino motor. New setup is 10x leaf modules for 20S2P, an Adaptto Midi-E controller, and a QSMotor 205 V3 50H 3T.

I was expecting going from a 1500w controller to an 8kw one would be a dramatic increase in torque, but it's absolutely not. The kelly actually jumped off the start, I could absolutely not stop it with my feet. This Adaptto is smooth as butter and weak as a kitten. I can floor the throttle with my feet on the ground and I go nowhere. Starting on small hills is a problem with reasonable power settings. So I cranked the amperage to max, shortened the throttle curve, now it starts maybe a little faster than my 500W stock chinesium controller, and has a lot more oomph when opening the throttle and going from 30 to 60, but within a minute of riding the temperature on the controller exceeds 75C and cuts all power.

So far I've run the autodetect process a half a dozen times then tried tuning manually according to doctorbass's guide. The autodetect always seems to want to set my angle correction to +6 degrees or so, but the motor is noticeably noisy at that setting. I find +1-2 degrees sounds smoother. The ind timing gets set around 500 by autodetect, but the motor gets noticeably faster at 550-600. The power timing gets set to +0.51, and changing this to any other value seems to increase heat. The really freaking annoying thing is I can't set it to anything between 0.3 and 0.5. A single press of the button adjusts by 2. Why the hell do they call this fine tuning if you can't possibly tune as fine as the autodetect can? Bah! Curse you, adaptto!

Anyway, I've spent all day on this, trying different settings, and the only way I can keep the controller below 60 degrees is by setting the amperage to such a low setting I need to push off to start, which is simply ridiculous. My kelly controller was able to pull a 200+lb trailer up 6% grade hills from a standstill no problems and the controller never got even slightly warm... I must be doing something wrong here, but I can't imagine what.

Any ideas would be super welcome!

Sounds a lot like a wrong phase/hall wiring combination.

I dont know how the Adappto stuff works, but perhaps if it has an autodetect it isnt able to correctly determine the right combo. If this is a possibility (or if it does not have an autodetect), then perhaps try manually determining the right combo while monitoring no-load current.

There are threads about how to do that, such as How to Determine The Wiring for a Brushless Motor (or something close to that).

Hmm. Wouldn't that cause noise and vibration? Jerkyness? The motor runs incredibly smooth and silent except during the autodetect portion where it's trying all the invalid hall offset angles etc.

The manual for it says to hook up the phase wires arbitrarily, and they're not color coded. As for the halls, I'm not sure I have them in the correct order, since the manual shows a 7-pin conductor and my version of the controller has a 6-pin one. So, I guess it's possible they could be going BCA or CAB instead of ABC. I dunno.

If it actually tells you to hook up phase wires arbitrarily, then it means it should be able to autodetect--but it doesnt mean it is working correctly in this case. Probably is, but its worth a check. And if phase wires are arbitrary, so would the hall wires be, other than power and ground.

Depending on how a controller works, its possible to get a smooth but wrong combo. One Ive seen can spin the motor much faster than it should for the same voltage when unloaded (off ground) but draws much higher current than it should (especially once its loaded), and doesnt have the torque it should.

Another appears to spin smoothly at normal speed but feels different than normal, not exactly rough but different. Draws much higher current than it should.

Most do run rough (if they will spin at all), while drawing the higher current.


Best I can suggest is use an independent wattmeter to monitor power usage while testing phase and hall wire combinations, and find the one with lowest current that also spins smoothly in the forward direction. Note down what you have right now, so you can compare it to what you get when youre done.

Hmm, yeah. I dunno. I think the speed is about correct. On 80v I'm getting around 80kmph. If I increase the OVS (field weakening) setting, it goes up to 105, maybe higher, I don't want to try. This setting gets me more speed at the expense of my motor getting hotter, but it doesn't seem to heat the controller too much. I've seen the motor get up to 80c, but I think if I add some ferrofluid that could be fixed.

I can try putting in a wattmeter, but I'd wager the display on the adaptto is probably correct. I've seen the amperage hit 125A or so, it says my peak wattage is 10.0kw. I suspect if things are getting that hot that quick that number is probably correct.

On max settings it has some get-up-and go, but it doesn't feel more powerful than my Kelly, and it overheats within a few minutes or a big hill. On settings that are similar to my Kelly, I can't start on hills. It might not overheat, but it's not usable like that.

I get the feeling this is a bicycle controller and I shouldn't be using it on a scooter, but on the other hand, 8kw is 8kw, and I'm thinking regardless of the weight, if the ammeter says 100A on a bicycle vs 100A on a scooter, should create the same amount of heat right? And I've read some threads where people are cranking the Midi-E up to 12kw without overheating.

I'm going to take it apart, apply thermal paste, and look into fabricating some kind of heatsink for this thing, but I still wonder if I'm doing something wrong and generating far more heat than is normal.

It maybe down to hall sensors being badly positioned on the motor. Try your spare set which you should have on the qs205. I had similar problem when originally setting up an adaptto with a cromotor and when i tried the other set of halls it stopped overheating. The 'angle corr' setting which fine tunes the hall sensor positions shouldnt really be more than +5 or -5 degrees. You are correct to manual tune as autodetect can be wildly out.


Edit: Try turning down ind timing to about 440 and then try turning up pwr timing. Although ind timing may give you high rpm/speed if turned up it will cause heat issues if too high. After turning it down abit then try putting pwr timimg to about 1.40 This will give you more mid range torque onthe throttle.. The other thing that will keep your peak amps high for more power is turning up 'speed smooth' to its maximum value.

Thanks for the suggestions! Unfortunately I'm already on my second set of hall effect sensors. The "C" channel on the first set is stuck open. The health monitor always shows a "1" for the third column. I think this could be because I didn't know the pinout, guessed that the middle pin on the DIN connector was pin 6 rather than pin 1. I wired it up incorrectly at first, though when I figured out the correct pinout it did work for a few km worth of driving before giving a "!HALLS!" error. I dunno, is it possible to mix-and-match the two different sets of halls? I'd suppose if they're offset at an angle that's an even number of poles away or whatever, that could work?

I'll give those setting suggestions a shot once the weather clears up. I haven't waterproofed the controller or display yet, so I'm not taking my chances.

inedible said:

Thanks for the suggestions! Unfortunately I'm already on my second set of hall effect sensors. The "C" channel on the first set is stuck open. The health monitor always shows a "1" for the third column. I think this could be because I didn't know the pinout, guessed that the middle pin on the DIN connector was pin 6 rather than pin 1. I wired it up incorrectly at first, though when I figured out the correct pinout it did work for a few km worth of driving before giving a "!HALLS!" error. I dunno, is it possible to mix-and-match the two different sets of halls? I'd suppose if they're offset at an angle that's an even number of poles away or whatever, that could work?

I'll give those setting suggestions a shot once the weather clears up. I haven't waterproofed the controller or display yet, so I'm not taking my chances.

Click to expand...

If heat is only your issue then hopefully it can be sorted with some setting adjustments. The other thing that may be worth trying if you are out of ideas is swapping the phase wires around and trying Utodetect again plus some fine tuning. I have read that some phase wire combos give better efficency. Is it just your controller temperature causing heat issues or is it motor as well? And what watt hour per kilometre are you seeing on the stats page.

Gawd- your problem sounds exactly the same as mine except my motor is the one getting hot. Ive got a Midi with a 3000w 50mm motor (which seems decent)
Shit acceleration, poor wh/km, and I can't get to the end of my street before motor temps start rising rapidly. About 5kms is all I get before thermal rollback. Smooth as butter though, runs like it's a Maglev train
But my 500w 9c with cheap trap controller would leave it for dead.
Ive tried every setting combo possible- no luck.
Today I swapped halls I'll let you know if that helps but I really don't think it will. I've pretty much given up....
Think Ive got a dud motor but I'm not going to spend another 400 'trying' a new one.
Sigh... Ebikes, like girls, intoxicating but so many issues :lol:

I ran an Adaptto Mini-E for about 6 months last year and had really similar issues out of the gate. If I remember right I solved the heating issues by looking for Ind Timing and Angle settings used by other members for the motor I had at the time (H3540) and by getting some thermal paste and small heatsinks (Ebay) and attaching to the Mini.
The power at start thing I eventually corrected by cranking the phase amps way up. I noticed early on that I wasn't putting out my amp limit until the bike got up to about 10mph. I did a ride where I ran my usual 18s battery and I had a a 12s battery on seperate breaker switches. I noticed that at 18s I was only putting out 15-20a at start WOT, when I switched over to the 12s battery I had a full 80a off the line. Eventually I cranked the phases up to max and I was able to get the full 80a off the line with 18s WOT. I made sure to watch the controller and wheel temps like a hawk but at that point it stayed really reasonable.

There is also a traction control trick that will get you amazing power off the line. I forget the details but it is on the forum somewhere.

I ended up getting a chance to do some more testing today, the weather was nice enough.

I think I've found a combination of settings that gives me the get-up-and go I want, without overheating the controller too much.

I tired taking the controller apart to reapply thermal compound and maybe silicone the ends of the case for waterproofing, but it seems everything is already siliconed in place. I tried a fair amount of pressure with a flathead and couldn't get the ends off, didn't want to damage anything so I stopped.

I found a 1/4" thick piece of aluminum, maybe 7"x4" or so, drilled two countersunk holes in it, but when I went to remove the mosfet screws in the side, I ended up stripping a couple. It seems they're already siliconed in place, and they seem like very soft screws.

So, I used one screw to hold it in place, and I used as little RTV silicone as I could to adhere the thing in place. I know it's not a very good thermal conductor, but I think it should be removable when I find a better solution. Anyway, this worked pretty good. It didn't prevent the controller from building up heat quickly, but it did wick it away pretty quickly.

As for settings, I'm going to have to go back and check, but I thing I set the hall offset to 2.9, the ind timing to 389 or so, and I forget at the moment which pwr timing ended up generating the least heat. The main thing that helped lower my heat was reducing battery amps to 90 and phase to umm, 220ish? I think. Again, I'm going to have to check, but it's charging right now and I don't want to touch it.

This of course neutered my torque. What I found made the biggest difference in fixing that was: Under traction settings, I had it set to TORQ mode, because hey, I want torque. But it turns out TORQ mode regulates your current at low speeds, making it so your throttle isn't 0-100% amps, but 0-100% acceleration according to the acceleration curve. Changing this made a huge difference. It used to only pull 100A when I was above ~20kmph, and it'd be incredibly slow starting out at 20-40A before getting up to speed.

The other thing was shortening the throttle curve. Under calibration, Throttle Limits, you can set the min and max voltage for your throttle. Mine by default goes from 0.9v to 4.5v or something, but by setting the upper limit to 3.5-4v or so, it's a lot more punchy when you floor it from standstill.

I'm hoping if I get some better heatsinks, use thermal epoxy or something to attach them, force airflow through them with a scoop and duct of some sort, I should be able to get the full output this controller is possible of outputting. Otherwise I'm just going to have to run it on 80% mode. Even still, it's feeling a lot more like my kelly controller, and it's no longer stalling on hills. It can even sorta peel out a bit on gravel. It's still running hot like this, but usually staying under 60c.

Will post an update with more specs.

Alright, since yesterday I've gone and moved my controller to underneath the bike, which hopefully should make it easier to create a duct to force air over it. When I plugged in the phase wires, naturally since they're not marked or color-coded, I must have put them in a different combination. When I tried autodetect it said I had an offset of 240 degrees, I remember this being 120 on my previous phase combo. At 240 it failed to autodetect and just made grumbly noises. I swapped a couple of the phases, now it says 180 degree offset, and when I did the autodetect it seemed to be smoother and as far as I can tell drew less power. It ended up with a hall offset of -0.6 degrees, compared to the +6 I was getting before.

I ended up performing a full reset, since I had updated the firmware without doing so before. I don't think that helped with anything, but it did make it harder to tune in a power setting that gives oomph without overheating again.

My current (no pun intended) settings:
Power profile:
Battery current — 80A
Phase current — 230A
Accel — "--"

Calibration menu:
Throttle limits — 0.9v - 3.8v
Throttle linearity — See notes
Throttle progression — 0

Traction settings:
Throttle mode — SPD

Advanced settings:
Angle corr — -0.6°
Ind timing — 398ms
PWR timing — 1.75 (still playing with this)
OVS timing — 000 (still considering whether this is worth the heat)

So, another thing I've been playing with is the throttle linearity setting. My biggest problem with this controller has been the lack of oomph off the line, and the excess of oomph going from 3/4 throttle to wide open. It's really hard to maintain, say, 45km/h when like a one degree change in throttle at that speed seems to lurch you forward, accelerating to 55 quickly.

I've been trying different throttle curves to fool the controller into doing what I want. I haven't gotten it yet, but I think if I deliberately slow turning the throttle while calibrating, I can make a flat spot where acceleration isn't so crazy in the midrange.

So far with these settings I can drive within the temperature limit so long as I'm consciously going out of my way to pay attention to acceleration bursts. I'm definitely going to need more heatsinks on this thing, and well as ducting air towards it if I want to use it like this though.

Oh and to answer a couple questions, currently my motor doesn't seem to be overheating. The most I've ever seen is 85c when I was riding it really, really hard trying to test the limits of things. Usually under normal load it stays around 60-70c. I think this is within spec, and I'm going to buy some statorade ferrofluid to try and lower this in the future.

My WH/KM is somewhere between 45 and 80, according to the few times I've reset the stats, but I don't think this is representative of real world use. I've only been testing this thing, not actually riding it. So, I've been calibrating it by running it with the wheel off the ground for 10 minutes, then driving up the biggest hill I can a few times, driving it around the block over and over at various specific speeds, etc. Repeat for days. I'm not paying attention to any long-term specs since I think they're probably meaningless at this point.

You sound a lot happier!
The throttle curve is quite fun. Try turning throttle slowly then fast or the other way around and you will get ramp up at different points.

For information.
On some of my Adaptto controllers, accélération was too smooth, even after many calibrations.

I discover that Adaptto changed mosfets thermal sensor at a date, from KT83 to KT84 (from memory).
If you have the wrong thermal sensor on the setup, (after firmware update for exemple), the acceleration is very smooth !

New firmware update sets to KT84 by default, even on old controllers with KT83.
It was corrected on latest firmware, but not on previous.

On adaptto firmware update page :

IMPORTANT: starting from November 2014 our controllers are equipped with KTY84 thermosensors (was KTY83 initially). To check whether the correct thermosensor is set for your controller you need to enter HEALTH MONITOR and take a look at MOSFET temp. If the values are inadequate (negative or too high) please enter CALIBRATION menu and change the themosensor model in Int.TempSens field for KTY84 (or KTY83 if you already got the former). Then also try doing full system RESET. If neither of the measures has helped please contact us ASAP for instructions. The exploitation of controller with broken thermosensor is dangerous and can lead to the total loss of controller.

You sound a lot happier!

Click to expand...

I guess so. The torque is there, which was my biggest complaint. The heat though, is still a problem. I wish I could put 8kw continuous through a 8kw controller, but I guess that's not how it's meant to be used. I'm going to need to buy some heatsinks, thermal adhesive, and I'm thinking maybe some peltier coolers.

I discover that Adaptto changed mosfets thermal sensor at a date, from KT83 to KT84 (from memory).
If you have the wrong thermal sensor on the setup, (after firmware update for exemple), the acceleration is very smooth !

Click to expand...

I did see that. Mine seems to use a KT84. If I set it to KT83 it shows a negative temperature.

I'm inclined to believe the sensor, it should be correct, but on the other hand, when it says the mosfets are 75c, and I feel the side of the controller, it's hot, uncomfortable to keep your hand on it, but at 75c it should be painful to touch. I'm wondering if maybe my sensor is reading higher than actual temperatures. Either way though, I do need to find a way to get rid of this heat. Tweaking the advanced settings doesn't seem to help. I just need to use fewer amps until I can cool this thing adequately.

inedible said:

New setup is 10x leaf modules for 20S2P, an Adaptto Midi-E controller, and a QSMotor 205 V3 50H 3T.

Click to expand...

There lies the Problem.
This low turn count motor is a bit too much for the Midi-E.
It has very low resistance and inductance which leads to high peak currents everywhere inside the controller and creates heat.
A 4T or 5T of this motor would suit the Midi-E much better :wink:

I wonder why you have choosen this motor version?

inedible said:

I'm inclined to believe the sensor, it should be correct, but on the other hand, when it says the mosfets are 75c, and I feel the side of the controller, it's hot, uncomfortable to keep your hand on it, but at 75c it should be painful to touch. I'm wondering if maybe my sensor is reading higher than actual temperatures. Either way though, I do need to find a way to get rid of this heat. Tweaking the advanced settings doesn't seem to help. I just need to use fewer amps until I can cool this thing adequately.

Click to expand...

It is normal that you have temp differences of 20°C or more between outer sheel and heatsink (it is measured near the FETs)

I chose this motor when I was using 16S and a kelly controller. I wanted a bit more speed out of the voltage I already had, and was willing to sacrifice some torque to get it.

I bought a new battery, 20S, and instead of buying a new charger and BMS, I came across a used MIDI-E for $750CAD and figured hey it comes with a charger and BMS integrated into the system, I can charge it directly off a solar panel, and I get a bit more oomph for not a lot more than a bulk charger and BMS would cost.

I picked up some heatsinks and some thermal epoxy, and I've been meaning to get around to installing them. I'll update this thread when I do. Hopefully I'll be able to set the power higher. Right now if it pulls 80A for more than 10sec on a hot day it starts to thermal throttle.

I also bought some TECs (peltiers), and I was thinking of installing them between the FET side of the controller and the heatsink, but after playing around with them for a while it seems their inherent inefficiency means I'll need an even bigger heatsink and probably a fan to keep the heatsink cool. I'm hoping just by covering 3 sides of the controller in 1" thick aluminum heatsinks will be enough to passively cool it. We'll see I guess.

What's the wheel size on this scooter?
You need to be aware that this 3T motor needs alot of amps for a given amount of torque, and it probabaly will never "top out" in terms of RPM so the phase amps will always stay high. It is no wonder the Midi controller overheats.

Did you try the kelly controller on the QS? I think it would blow or overheat even quicker!
What was the kV of the stock motor?

If i were you, i would buy a 4T or even 5T motor (depening on wheelsize).
You will get alot more acceleration and the controller will stay cooler. If top speed isn't enough you could make use of OVS :wink:

The wheel is 10", I believe. That'd be the inner diameter, and this tire is a bit bigger than the stock one, I'd say 2.5-3" thick.

Yeah, you're probably right. I wonder though if I shouldn't just get a MAX-E instead of a different motor though. I'd look at maybe a sabvoton or something, but I feel like I'm invested in the adaptto ecosystem now, and I really do like how integrated it all is, with an BMS that shows the voltages of the individual cells on a LCD on the dash etc. I really like the ability to use cheap used server PSUs to charge it, or a solar panel, or really any sort of source of DC. Hell, if I was really in a bind I could charge the thing off my laptop charger.

My batteries are more than capable of putting out the amps. I think they do 300A constant, 600A peak. It might be a lot more practical and probably a lot more efficient with a different motor winding, but I'd also like the ability to go 100kmph once in a while if I want to. I'd actually like to take this thing out to the drag strip for shits and giggles, even though 98% of the time I'm going to be driving under 40kmph in the city.

inedible said:

The wheel is 10", I believe. That'd be the inner diameter, and this tire is a bit bigger than the stock one, I'd say 2.5-3" thick.

Click to expand...

Could you measure the outside diameter of the wheel, and do you have the 3T (16,3kV) or 3T- (17,7kV) motor?

I wanna calculate the speed based on your 72V battery

Hmm, so I don't have anything to measure with right now.

When I set up the controller I needed the circumference of the wheel, which I think I recall was 1030mm, giving a diameter of 328mm or 12.9 inches, which seems reasonable.

The winding I got is 40*3T, 1110 RPM@72V, 15.42kV. So, by my back of the envelope calculations that'd give 68.59km/h?