The Gotway Gods Destroyed My ACM

[meepmeepmayer]

2 hours ago, Maximus said:

As I wait for my new ACM to arrive, I can't help but wonder what settings to pay attention to while riding up hills.

Ignore temperature, if it gets too high the ACM will beep and stop you anyways.

Watch the current. Per spec, roughly, maximum sustainable current is 20A (old ACM and msuper V3 motors with the 16AWG wires, dunno what the Monster uses) and 30A (new motors with the 14 AWG wires) so that's a guideline.

Problem is the heat not being dissipating fast enough, so the wheel gets hotter and hotter. Both cables and mosfets won't (just) blow from a higher current, but from overheating. So the key is to make breaks accordingly, to allow cooling down.

If you're in the spec, you can probably go on forever. After all, fast riding (in the flat) can produce 20A and nobody reported any meltdowns just due to fast riding in the flat.

If you're above spec, just make more and more breaks the higher the sustained (average, peaks don't matter) current gets.

Here's some numbers that worked for me with my ACM(I have the old 20A wires):

Around 15% incline gives me (80kg and with my riding style) between over 30A average (sometimes up to 40, some higher peaks, but average matters for heat), and this can be done for several minutes (3,4, even 5) without any visible heat damage (opened the wheel several times to check). Then a few minutes break, then again. More or less 1:1 riding/break time (this may be overcautious, but since there is no other way to see what is too much without stressing the wheel too much = heat damage starting, I'd rather be cautious).

So even with the old wires, 30A for 4 minutes is ok. With the new wires, 40 should be good (no guarantee! Maybe the heat will collect faster in the thicker wires? Just added the 10A difference).

The hill I melted my first ACM's 20A cables on (exactly like Marty) produced probably something like 50A sustainably (guessed number!), if not more. Even then, it was over 5 minutes (closer to 10) before the meltdown happened, maybe 1 or 2 minutes before that you'd have heat damage. So even 50A for a minute or two should easily be fine. I will go back to the hill and see what current it produces, as I know 2+ minutes can be done without damage and so 30s or 1 minute is enough to safely read the current there. More numbers then.

So the (sad) summary is, you can only check currents on hills and see if your wheel survives. It's a bit stressful and can ruin your ride's enjoyment, but if you want to do hills, some experience which "feeling of being elevated" produces which current is needed so you can go on rides (with according breaks) later without having to always watch your current numbers religiously.

Also, the steeper the hill or the heavier the rider (higher current, which is all that matters for the cabling), the faster everything happens. I don't think Marty had 5 minutes on his 25% incline. So either the ACM was pre-heated from the lesser inclines, or he had like 80A current that quickly melted everything down.

TLDR: watch current, slowly build towards testing what you can safely do, and in the end, EUCs aren't ready for real mountain riding (no breaks, no second thoughts, just go) yet. Or at least the Gotways are not, maybe Kingsong would be better.

[DaveThomasPilot]

I agree with everything lizardmech said in this post, including this statement:

 

As to more voltage reducing current it depends on motor variables, if you do not change the motor windings increasing voltage will not reduce the amount of current needed to produce enough torque, the top speed of the motor will increase proportionately. If you rewind the motor with more turns you can achieve more torque with less current

That's why a lower Kv motor can deliver more power at a lower speed.

If you go up a hill at specific speed with a low battery voltage versus a high battery voltage, more average current will be drawn.  The duty cycle of the FET on time (if not 100%) will increase to compensate for  the lower input voltage.

Do you agree?

[Chriull]

29 minutes ago, DaveThomasPilot said:

Ok, I was trying to keep it simple.  These motors are actually driven by multiple phases and aren't DC.  But, the same principals apply.  For a given output power, you need at least as much input power.  Higher input oltage means less current (and vice versa).

...

 

16 minutes ago, DaveThomasPilot said:

You don't really need to understand the specifics of how the motor driver works, back emf, etc to know that lower input voltage means higher current. 

Just draw a black box with an input voltage and input current.  Input power equals input current * input voltage.  So, to deliver the minimum power at a lower voltage, higher input current is required.

Your consideration is true for "static" systems - one motor at one battery. Once with high voltage and once with low voltage - both give different speeds and powers...

But with an "regulated" motor in a EUC i see it as follows:

The motor generates a back emf directly proportional to the speed.

To drive at a constant speed some constant torque is needed (to just overcome the resistive forces like friction, etc). As torque is directly proportional to the motor current and also (more or less) direct proportional to the averaged pwmed battery voltage (Ubatt * duty_cycle) minus back emf one has at the motor side a power of  Ubatt * duty_cycly * I motor. At the battery one has more or less Ubatt. So the battery current has to be 1/duty_cycle times higher than the motor current.

So, as @lizardmechconfirmed this in the meantime here (http://forum.electricunicycle.org/topic/7520-the-gotway-gods-destroyed-my-acm/?do=findComment&comment=99837) it is now quite clear why there is such a "tragic" current (and so cable) prob at low speeds!

[Maximus]

8 minutes ago, meepmeepmayer said:

Ignore temperature, if it gets too high the ACM will beep and stop you anyways.

Watch the current. Per spec, roughly, maximum sustainable current is 20A (old ACM and msuper V3 motors with the 16AWG wires, dunno what the Monster uses) and 30A (new motors with the 14 AWG wires) so that's a guideline.

Problem is the heat not being dissipating fast enough, so the wheel gets hotter and hotter. Both cables and mosfets won't (just) blow from a higher current, but from overheating. So the key is to make breaks accordingly, to allow cooling down.

If you're in the spec, you can probably go on forever. After all, fast riding (in the flat) can produce 20A and nobody reported any meltdowns just due to fast riding in the flat.

If you're above spec, just make more and more breaks the higher the sustained (average, peaks don't matter) current gets.

Here's some numbers that worked for me with my ACM(I have the old 20A wires):

Around 15% incline gives me (80kg and with my riding style) between over 30A average (sometimes up to 40, some higher peaks, but average matters for heat), and this can be done for several minutes (3,4, even 5) without any visible heat damage (opened the wheel several times to check). Then a few minutes break, then again. More or less 1:1 riding/break time (this may be overcautious, but since there is no other way to see what is too much without stressing the wheel too much = heat damage starting, I'd rather be cautious).

So even with the old wires, 30A for 4 minutes is ok. With the new wires, 40 should be good (no guarantee! Maybe the heat will collect faster in the thicker wires? Just added the 10A difference).

The hill I melted my first ACM's 20A cables on (exactly like Marty) produced probably something like 50A sustainably (guessed number!), if not more. Even then, it was over 5 minutes (closer to 10) before the meltdown happened, maybe 1 or 2 minutes before that you'd have heat damage. So even 50A for a minute or two should easily be fine. I will go back to the hill and see what current it produces, as I know 2+ minutes can be done without damage and so 30s or 1 minute is enough to safely read the current there. More numbers then.

So the (sad) summary is, you can only check currents on hills and see if your wheel survives. It's a bit stressful and can ruin your ride's enjoyment, but if you want to do hills, some experience which "feeling of being elevated" produces which current is needed so you can go on rides (with according breaks) later without having to always watch your current numbers religiously.

Also, the steeper the hill or the heavier the rider (higher current, which is all that matters for the cabling), the faster everything happens. I don't think Marty had 5 minutes on his 25% incline. So either the ACM was pre-heated from the lesser inclines, or he had like 80A current that quickly melted everything down.

TLDR: watch current, slowly build towards testing what you can safely do, and in the end, EUCs aren't ready for real mountain riding (no breaks, no second thoughts, just go) yet. Or at least the Gotways are not, maybe Kingsong would be better.

Awesome, I'll do some testing when mine arrives...looks like actual delivery for me is about a week, so I've still got a little time 

[DaveThomasPilot]

also (more or less) direct proportional to the averaged pwmed battery voltage (Ubatt * duty_cycle) - back emf one has at the motor side a power of  Ubatt * duty_cycly * I motor. At the battery one has more or less Ubatt. So the battery current has to be 1/duty_cycle times higher than the motor current.

Absolutely!  But, while the battery current is 1/duty cycle time higher than the motor current (assuming 100% efficiency), it's only being drawn "duty cycle" period of time. 

Simple example, 50% duty cycle if motor current is 50 amps, battery current when fet is on is 100 amps, when FET is off, it's 0 amps.  Average is 50.

Temperature rise (and power dissipation) need to include the duty cycle multiplication.

But, again, it doesn't matter what's doing "the magic" inside the box.  Just use energy conservation to analyze.

[DaveThomasPilot]

Watch the current. Per spec, roughly, maximum sustainable current is 20A (old ACM and msuper V3 motors with the 16AWG wires, dunno what the Monster uses) and 30A (new motors with the 14 AWG wires) so that's a guideline.

Shouldn't the control board do this, if it's required to prevent damaging the board?

If it's already measuring the current, it would be trivial to do a longer term average and initiate tiltback or some other kind of warning.  This would be the "maximum sustainable power" specification.

 

[Chriull]

15 hours ago, Marty Backe said:

The app was running but I was looking at the temps. I'm sure you're right that the ACM was running hot, just not hot enough to kick in the tilt-back. It would be nice if you could adjust the temperature where tilt-back kicks in. Then a person could choose to only run their wheels at cooler temperatures. That would have saved me.

The temperature reported by the app is the temperature measured at some place on the mainboard - could be that the cable heating up just contribute a low portion to this measurement. At least they only contribute with a huge delay - imho just when it's already too late.

Better strategy seems to be as @meepmeepmayer posted here http://forum.electricunicycle.org/topic/7520-the-gotway-gods-destroyed-my-acm/?do=findComment&comment=99839 to look mainly at the current reported together with speed and incline and make brakes according to this numbers...

As you wrote before - at steep inclines which only can be handled with very low speeds it is better to walk the wheel uphill! Because, as mentioned around here at low speeds currents get transformed upwards! For example with an back emf somewhere near 10V (v = v_no_load / 67,2V * 10V)  and roughly 800W being 13A at the battery side this would result in a current of 78A at the motor side. This are just very very rough numbers - but one sees the tendency. So once going slowly up steep inclines its just a matter of time until the cables cannot stand it anymore ;( I fear the same will be true for the new models with thicker cables - it will just take a little bit more time...

Edit:

It seems that the motor current stays quite lower, since the coil resisitance could play an important part here - there is/could be a not neglectible voltage drop which lessens the voltage ratio and by this the current ratio...

[Duf]

56 minutes ago, Marty Backe said:

The app was running but I was looking at the temps. I'm sure you're right that the ACM was running hot, just not hot enough to kick in the tilt-back. It would be nice if you could adjust the temperature where tilt-back kicks in. Then a person could choose to only run their wheels at cooler temperatures. That would have saved me.

Either that or a second temperature sensor that reads ambient temperature inside the case. Or.......just over engineer the damn things to prevent pumping massive amps through thin wire, or.............. how about a simple inline fuses that are user replaceable!

 

[Marty Backe]

5 minutes ago, Duf said:

Either that or a second temperature sensor that reads ambient temperature inside the case. Or.......just over engineer the damn things to prevent pumping massive amps through thin wire, or.............. how about a simple inline fuses that are user replaceable!

 

All good suggestions. Too bad Gotway isn't involved in our English speaking community 

[FULspeed]

This kind of thing would not happen if we had a proper information about what happens in our wheel.

In each of my normal means of transport (car, or motorbike, or plane) I have tools telling me if the engine works correctly (oil pressure, water temperature, voltmeter ...)

why not on my wheel?

I would like to remedy to this lack of information (however I have already installed a small digital voltmeter to monitor more precisely the battery status), so now I would also like to control the temperatures (especially now in the summer) to avoid any risk of cable fusion, as happened to Marty.

I tried to find a simple solution, and this would be great for me:

... but I don’t thinks it would work  because 125 degrees Celsius is the maximum detected temperature.

if someone technically more qualified than me have already found a solution, I would be very grateful.

Thanks

 

[meepmeepmayer]

2 hours ago, Maximus said:

Awesome, I'll do some testing when mine arrives...looks like actual delivery for me is about a week, so I've still got a little time 

Do NOT start with looking at currents. You'll see 20A or 30A pop up in no time, and then you're needlessly worried (we don't even know what the real critical sustainable A numbers are, probably quite a bit higher than that, these are just some spec from a table). It will seriously ruin your mood if you restrict yourself because you think some current is too high.

Instead, get to know your exciting new ACM, play around with it, do a long and fast ride, enjoy! (Just don't start right away on 20% loooooong inclines, but shorter inclines are fine) Then, after at least 2 weeks, you have some experience with your wheel, and if you feel like it you can start looking at currents, in the flat, on inclines, whenever. It's important that you have some trust what the wheel can do before you start current watching, otherwise you'll be over-cautious and it really does spoil the entire thing. I did things before looking at the currents that worked wonderfully, had I looked at the currents right from the start I would have been worried. You do underestimate what the wheel can do by looking. Only when you have some experience and go for really notable (and long) inclines, it becomes a helpful way of estimating what you can and can't do. In doubt, do the current watching rather later than early. Look how long Marty happily did mountain rides, everything worked, never worried about currents, and only the one extra steep incline was too much. Maybe only 30 seconds or 1 minute less, and absolutely everything would have been fine.

[meepmeepmayer]

2 hours ago, DaveThomasPilot said:

Shouldn't the control board do this, if it's required to prevent damaging the board?

If it's already measuring the current, it would be trivial to do a longer term average and initiate tiltback or some other kind of warning.  This would be the "maximum sustainable power" specification.

In principle, you're totally right, the wheel should monitor itself and stop when it becomes too much. EUC design 101. Obvious.

But:

This would mean, only regarding the cables, Gotway would have to know which current for how long is critical, e.g. "40A can be sustained for x minutes y seconds before it gets bad". I don't think they have that information, and if they had, it would be needlessly complicated to implement. Because you're effectively emulating how heat collects and distributes, which is complicated. Repeat that idea for every other part. Not really feasible.

Or you get the Kingsong problem: implement conservative limits, and people complain and buy GW.

Afaik, the only safe way forward is, see what maximum current the battery pack can deliver. Then, every cable, mosfet, component must be able to deal with that current for unlimited time. The battery alarm (which apparently is easiest to do, since it is already implemented as a warning) would then be the overall system overstress alarm. And done, nothing else to worry about.

Problem is, for the 100A or 150A (or more?) you could probably get in extreme cases, you need huge expensive components (mosfets, wires,...), redesign the motor (no more e-bike parts) etc.

KS seems on a good way, their new mosfets are up to 350A or so (not sure if peak or sustained, probably peak), if the cabling matches that, they basically solved this problem.

[Chriull]

26 minutes ago, FULspeed said:

 

I don’t thinks it would work  because 125 degrees Celsius is the maximum detected temperature.

 

Imho this could work out nicely! Once the cable insulation gets 125 degrees celcius it should be a good decision to immediatly stop and take break. Plastics/silicone is a good thermal insulatior and the copper should be screaming hot by that time!

26 minutes ago, FULspeed said:

if someone technically more qualified than me have already found a solution, I would be very grateful.

 

Pfff.... qualified .... solution .... pffff 

[meepmeepmayer]

@FULspeed Problem with temperature measurement is, it may not be quick enough.

A really quick and high current spike will blow a mosfet which can be virtually instant. Temp warning won't help then.

A lower but sustained too high current will either blow the mosfet (slow overheating) or, with the ACM apparently, melt the cables first (also slow overheating).

I don't know, but I would guess, the final temperature spike and melting happens quite fast. So you don't just slowly and constantly heat some part up, but it heats, this means higher resistance, even faster heating, etc and the temperature blows up (not literally, mathematically) and you get more heat really fast, and only that does the melting. So maybe it only takes a few seconds seconds from a fine cable insulation to it melting down (in my case, looking at it, almost seems the actual metal started melting right with the insulation, so it had to be fast, assuming the metal needs higher temperature to melt). Who knows. So temp warning might just not be fast enough for this kind of problem.

[Maximus]

17 minutes ago, meepmeepmayer said:

Instead, get to know your exciting new ACM, play around with it, do a long and fast ride, enjoy! (Just don't start right away on 20% loooooong inclines, but shorter inclines are fine) Then, after at least 2 weeks, you have some experience with your wheel, and if you feel like it you can start looking at currents, in the flat, on inclines, whenever. It's important that you have some trust what the wheel can do before you start current watching, otherwise you'll be over-cautious and it really does spoil the entire thing.

Good advice.  I've been slowly working my confidence back up in my KS-14C, trying to avoid hard accelerations and being really cautious about speed.  I certainly don't need more reasons treat my new EUC with "kid gloves"... 

[FULspeed]

19 minutes ago, meepmeepmayer said:

A really quick and high current spike will blow a mosfet which can be virtually instant

Thank you meepmeepmayer, now I'm scared....

[meepmeepmayer]

You mean just the insta-blow? That seems just general knowledge right? At least with older wheels or bad luck. But it needs to be a huuuuge current spike (doing essentially the last second heat explosion immediately). So as long as you don't lock your wheel against a vertical wall, nothing to worry about. Didn't happen very often I think, even with older wheels, just something that can happen in theory.

If you mean the entire thing, this is why you should not look at currents needlessly, you'll grossly underestimate what you can do. So many people doing mountain rides with no problems, Marty is the best example. Just a really steep and long incline was too much. Enjoy your wheel and don't push it too much.

[US69]

3 minutes ago, FULspeed said:

23 minutes ago, meepmeepmayer said:

A really quick and high current spike will blow a mosfet which can be virtually instant

Thank you meepmeepmayer, now I'm scared....

Naaaah, normally, there is a treshold of max current, which for the GW's is a around 120 Amps....as the Mosfets are able to do 180Amp, that "normally" wont happen....from my view that can only be if something other goes wrong, like a short from melting cables..AND THEN you have such a high spike that it blows your Mosfet....

but because of the short before that would be irrelevant....

my 2 cents

[meepmeepmayer]

@KingSong69 is right, the insta-blows may have been just from shorts somewhere else. So forget this, no worries. I'm no expert

[HermanTheGerman]

Brrr, just back from vacation, what do I have to read ?

Thanks God Marty was not hurt !

And thanks for the detailed infos, I think I will remove Gotway from my personal hitlist, so only Kingsong and Inmotion remain.

BTW, I drove up a steep gravelroad to a location which is called "Schweizeben" (1006m) from Bruck (491m), this is 515 meters (~1500 feet) ascending on a distance of 2,5 kilometers, so the average incline is around 20% (11 degrees), which means there must be a lot steeper parts there, because there are also some more flat passages.

Next time when I ride it I'll take a GPS with me to calculate the parameters in more detail.

The KS16B didn't complain.

It would be interesting to know the track parameters (average incline and distance) of your drive until the accident happened.

[Radislav]

Wow! Marty Backe baked his ACM live

The Gotway must be proud that their wheel appeared to be such a die hard while even the Monster with the running fans surrendered 5 sec earlier 

After what happened to @meepmeepmayer's ACM, I never dare to take such a long climbs despite the fact that my ACM motor wires have individual protective sleeves  

[DaveThomasPilot]

Afaik, the only safe way forward is, see what maximum current the battery pack can deliver. Then, every cable, mosfet, component must be able to deal with that current for unlimited time. The battery alarm (which apparently is easiest to do, since it is already implemented as a warning) would then be the overall system overstress alarm. And done, nothing else to worry about.

Agreed, but it doesn't have to be the battery.  Whatever the weakest link is, be it mosfet, cable, connector, or battery.   Advertise a maximum sustainable power based on that maximum current.

If the battery isn't the "weakest length", and a relatively simple design change like using a lower gauge  interconnect wire, or connectors that are rated for higher current might be implemented.  If "maximum continous power" were a key and visible EUC spec, then the manufacturers  would be motivated to make changes

Having worked in power conversion design (DC/DC convertors mostly) for many years, I can tell you that thermal considerations are really a huge part of the design.  Thermal management, heat sink design, airflow.   Analysis and test, test, test.

Power supplies are a good example.  There's very little interest in how much a power supply could deliver if the maximum current wasn't limited so it protected itself. 

I can remember when audio stereo amplifiers advertised peak instantaneous power.  Total worthless spec.  Even maximum continous power specification is meaningless without a distortion spec to go with it.  But, it took many years before an average consumer wasn't influenced by a maximum instantaneous power specification that was so big!   Several hundred watts, oh my!

But, if the EUC market doesn't recognize the value of a spec like maximum continous power,  then EUC manufactures will just throw out something like a motor power spec, even  if the rest of the wheel electronics can't deliver that much power, (or at least not for a prolonged period of time).

I think the only way to develop market pressure to get improved designs is customer education through a forum like this.  Or, a forward thinking EUC manufacturer that not only implements a superior design, but also educates customers on why it's better for the target customer set.

That's why I think its worth posting some technical details about what goes on with our fun toys.

 

 

[Marty Backe]

50 minutes ago, HermanTheGerman said:

Brrr, just back from vacation, what do I have to read ?

Thanks God Marty was not hurt !

And thanks for the detailed infos, I think I will remove Gotway from my personal hitlist, so only Kingsong and Inmotion remain.

BTW, I drove up a steep gravelroad to a location which is called "Schweizeben" (1006m) from Bruck (491m), this is 515 meters (~1500 feet) ascending on a distance of 2,5 kilometers, so the average incline is around 20% (11 degrees), which means there must be a lot steeper parts there, because there are also some more flat passages.

Next time when I ride it I'll take a GPS with me to calculate the parameters in more detail.

The KS16B didn't complain.

It would be interesting to know the track parameters (average incline and distance) of your drive until the accident happened.

I posted my Google Earth track somewhere in Page 2 of this thread, I think. You can view the entire elevation profile of my relatively short ride.

[HermanTheGerman]

1 minute ago, Marty Backe said:

I posted my Google Earth track somewhere in Page 2 of this thread, I think. You can view the entire elevation profile of my relatively short ride.

Thanks, I'll check it.

[Marty Backe]

32 minutes ago, Radislav said:

Wow! Marty Backe baked his ACM live

The Gotway must be proud that their wheel appeared to be such a die hard while even the Monster with the running fans surrendered 5 sec earlier 

After what happened to @meepmeepmayer's ACM, I never dare to take such a long climbs despite the fact that my ACM motor wires have individual protective sleeves  

I've learned my lesson the hard way. I'll still drive my Gotway wheels hard, but not continuously up steep hills. Hopefully I be able to fix it within the next month or so.