Inmotion V12 HT board: "More durable MOS"

[supercurio]

11 minutes ago, fbhb said:

Even the much maligned S18 actually uses 12 TO-247 MOSFETS despite it's much lower performance threshold, as can be found on King Song's website:

KS-S18-18 inch series-King Song

Ah yeah, with the description of 4000W+ peak power.

On @Jack ex-KS's climbing V12 video, the peak recorded is 6500W+ with 12 smaller MOSFETs

[Tawpie]

uhm... TO-220 and TO-247 are the package specs and *mostly* define different thermal characteristics. The recommended working voltages of any given part are primarily controlled by the silicon, not so much by the package. TO-247's benefit is that as a package it has a lower thermal resistance from die to heatsink than does the TO-220—so it is possible to cool the silicon die better (which means it should be able to handle more average power before it goes poof). As voltages go up though, you do start to worry about clearance and creepage distances and the lead spacing does come into play. 150V wants 0.6mm (ish) and the TO-220 is getting a little close... the TO-247 has twice the lead spacing so it has more margin with respect to the risk of arcing lead-to-lead. If you mind your board cleanliness and conformally coat the board, you probably don't *need* the extra lead spacing, but more margin is always welcome as long as it doesn't cost too dearly elsewhere (cost, availability, board space, repairability etc. etc.).

How you attach the package to the heatsink, and how the heatsink is cooled are also part of the total power handling equation... the thermal resistance of the package is just one part of the overall thermal budget. Changing to a TO-247 package will get the heat to the edge of the package more efficiently, but if you can't get get it from the edge of the package to the air you haven't improved your overall thermal margin. Survey the CPU overclockers boards for an idea of the many different ways to get rid of heat, there are a lot of options! Absolutely, when the package thermal resistance is taking a lot of your thermal budget, looking for a more thermally efficient package is the place to start. Unless you're really pushing it, the package is usually one of the smaller rocks in the thermal stream.

Make no mistake, if the package allows the silicon to get too hot bad things happen. But the package in this application doesn't affect the breakdown or max working voltages.

[Planemo]

7 hours ago, Tawpie said:

uhm... TO-220 and TO-247 are the package specs and *mostly* define different thermal characteristics.

Agreed, I dont think the V12 issue is one of thermal runaway. There just isnt enough time for heat build up during the stress test given some have failed on the 2nd or 3rd attempt. So I suspect its simply a spiked over current/over voltage situation.

That said, theres something reassuring about knowing theres 12x beefy TO-247's keeping you upright, whether the size is essential or not.

On the flip side, I'm still amazed that Inmotion, with their undoubtably huge knowledge, would make such a big error on component selection. They must have known that the fets were rated for 100v and they were building a 100v wheel. So I'm still not ruling out the possibility of them having had a shady batch of fets. Either that or their EE team needs a serious reshuffle.

[onizukagto]

19 hours ago, Planemo said:

Agreed, I dont think the V12 issue is one of thermal runaway. There just isnt enough time for heat build up during the stress test given some have failed on the 2nd or 3rd attempt. So I suspect its simply a spiked over current/over voltage situation.

That said, theres something reassuring about knowing theres 12x beefy TO-247's keeping you upright, whether the size is essential or not.

On the flip side, I'm still amazed that Inmotion, with their undoubtably huge knowledge, would make such a big error on component selection. They must have known that the fets were rated for 100v and they were building a 100v wheel. So I'm still not ruling out the possibility of them having had a shady batch of fets. Either that or their EE team needs a serious reshuffle.

If it was a design fault,

I wonder why batch 1 hasn't failed yet? 

If its component fault, then its the QC fault. 

I'll be brutally honest here, rarely do Chinese companies invest in through testing procedures. Third party validation is very rare and in-house QC testing is basic at best and just superficial A+B testing from staff's family and friends. 

The type of product QC testing rigors that are shown in promotional videos from say, Dyson or IKEA.Or like Testing and Commission procedures of installation products. 

In 10yrs of working with electronic product design in Asia (China) I've never ever seen it in person.

Maybe it does, but for the international selling stuff like Chinese mobile phones, but for me personally, it's just really uncommon. 

[fbhb]

11 minutes ago, onizukagto said:

I wonder why batch 1 hasn't failed yet?

 @supercurio has done an outstanding amount of work on this subject and he can confirm from his findings that some "batch 1" V12's have failed!

[Tawpie]

While batch 1 wasn't exactly immune to the problem, remember that the first batch was only 300 units. And it was the first batch. Most folks from line workers to QC to management pay closer attention to the first units off the line than they do after putting the same screws in the same holes for the thousandth time. Sure they're better at it, but attention to detail is challenging for everyone to maintain. The number of parts in incoming inspection gets larger, the pressure to meet the daily production goal gets higher, new people are added and trained. Lots of things evolve on the production line, and not all of them are beneficial to product quality.

If you are able to do it properly and a few balls bounce your way, the production rate scales without incident. More often though, you learn a lot about weaknesses in your design and purchasing and manufacturing and acceptance testing... and you fix the issues one by one. Rate production ensures that all of your tolerance stackups will get just outside of their worst case limits all at the same time—if you have enough margin everything is still good. If you don't, you have another problem solving "opportunity".

Edited January 14, 2022 by Tawpie

[Rywokast]

this is why i never buy a launch wheel lol... this is horrible, i feel bad for everyone who shelled out a huge chunk of change and are now probably afraid to ride their wheels, it is is unacceptable and inmotion should be providing replacement boards for everyone free of charge.. very glad i decided to wait until next summer to get the V12

[RagingGrandpa]

10 hours ago, onizukagto said:

If it was a design fault, I wonder why [all V12's don't fail]

 For me, component variation is a rational explanation for the inconsistent failure rate...

On 1/4/2022 at 4:52 PM, RagingGrandpa said:

Bear in mind, there is always component variation...
Semiconductors are made by dissolving chemicals in other chemicals, to create conductive pathways. This chemical process cannot be perfectly repeatable, and so part-to-part variation of some amount is expected. The 'average' part likely exceeds the rated limits by a notable amount; but a few parts may end up right at these limits. Which would help explain a [20%?] failure rate for controllers of identical construction.

For the cutout issue: we should be honest about we know and don't know. There has not been a measurable cause of failure identified yet. The cause could be some additional thing, unrelated to the FET spec. But this out-of-spec part is concerning, in any case.

Edited January 14, 2022 by RagingGrandpa

[digithom]

You cant' simply compare 100V mosfets with 100V battery and say "they are silly"
The driver of the motor is composed by a sort of H-bridge, revisited for the three phase contest, here it is a basic example

Each phase is driven by two mosfets, one on the positive side of power supply ad the other on the negative side, and the motor is on the middle of the bridge.
In this configuration, as they are connected in series to the load (the motor), each mosfet is sharing only half of the power supply, that is 50V, so a 100V mosfet is double the voltage it have to handle.
Without this, a 100V mosfet, running on a 100V power supply, will not wait for the Inmotion test to burn, but it will do very early.
I bet they are fighting with a stock of clone or fake mosfets than with a project issue.

Edited February 1, 2022 by digithom

[Chriull]

3 hours ago, digithom said:

In this configuration, as they are connected in series to the load (the motor), each mosfet is sharing only half of the power supply, that is 50V, so a 100V mosfet is double the voltage it have to handle.

Only if both of the mosfets are off.

The normal state of these 3 half bridges is that in one half bridge both mosfets are off and in the other two half bridges only one top side and one low side mosfet is on.

So at the two half bridges with the conducting mosfets the closed ones have to bear (almost) the full supply voltage.

So your following statement does not fit, too:

3 hours ago, digithom said:

Without this, a 100V mosfet, running on a 100V power supply, will not wait for the Inmotion test to burn, but it will do very early.

As these mosfets "withstand" continously almost 100.8V.

Reasons/design guides can be found in infineons app notes regarding these kind of mosfets and the avalanche energy they can withstand. Depending on the exact series and which app note they can be used in designs to withstand about 1.2-1.5 times the drain source breakdown voltage.

 

3 hours ago, digithom said:

I bet they are fighting with a stock of clone or fake mosfets than with a project issue.

Could be. Or just bad manufacturing if one sees pictures in reports of teardowns with how the mosfets were mounted on the heatsinks...

[RagingGrandpa]

7 hours ago, digithom said:

they are connected in series to the load (the motor), each mosfet is sharing only half of the power supply

Not during use.

A peak-voltage condition is as shown here, when one (or more) FET is active: 

On 12/21/2021 at 2:49 PM, RagingGrandpa said:

it's a function of temperature... if you exceed 100.0V across the FET at room temperature, it starts exceeding 1mA of leakage current, rising hugely with every additional volt.

When the EUC is off, with no FETs driven, no harm- we don't have battery directly across any one FET.
But when the motor is active with a very full battery, it does seem that leakage could occur.

I think the FET will act with a zener behavior, dropping about 100V from the circuit and then passing the excess charge in an unlimited way.
It shouldn't be damaged instantly... just from the thermal overload that could result if nothing else limits that current.

Perhaps the design can tolerate it because:

• motor commutation is a pulsing behavior, so no one FET will suffer the leakage current continuously, and
• the FET (or at least, the silicon die inside the package) warms up quickly, so the first few moments of operation (and even the leakage current itself) may be enough dissipation to warm it up enough to raise the breakdown voltage, and
• if there is any significant (forward) motor torque being driven, resistive losses (including battery sag) reduce the apparent voltage, reducing leakage.

The part is still 'outside of spec,' but practically speaking, a small occasional current leakage might not be destructive.

And even if it seems to work most of the time, I wouldn't want to stop suddenly with a cold, full-battery V12 riding down a hill, with outside-spec parts like that.
Additional margin in the design would be nice... even Gotway uses 125V-rated FETs.

.02

[digithom]

On 2/1/2022 at 5:24 PM, RagingGrandpa said:

Not during use.

A peak-voltage condition is as shown here, when one (or more) FET is active: 

The condition you shown in the picture should never happen, FETs ar always driven in couple in crossing left and right side, most of times, the gate of the two mosfets are driven my a dedicated electronic so that is not possible to switch only one.
But we are supposing everything, we should talk looking at the real schematic diagram of the controller of V12, does anybody have it ?

[Chriull]

1 hour ago, digithom said:

The condition you shown in the picture should never happen, FETs ar always driven in couple in crossing left and right side, most of times, the gate of the two mosfets are driven my a dedicated electronic so that is not possible to switch only one.

This post was not for general discussion about v12s inverter but how it can happen that almost the full 100.8V can drop over one mosfet.

Seemed to happen that only two instead of 3 h bridges were drawn - but that does not matter for this specific topic. One or three h-bridges would show exactly this same point.

1 hour ago, digithom said:

But we are supposing everything, we should talk looking at the real schematic diagram of the controller of V12, does anybody have it ?

Afaik this did not leak and got public. Chances are quite high that they use  three half bridges, 2-3 shunts for current measurement. And as they afaik use 12 mosfets there are always wo in parallel.

[tonto_euc]

Saw a post from a guy who had a V12 do well on the free spin test. He then changed his tired, hit a curb and bam, the controller went out. not sure if he damaged something during the tire install, the tire was too tight on the wheel so had added resistance, or just bound to happen due to a sudden bump.

now i'm considering selling my V12 despite passing the free spin test after 5 times. BTW, i know they said do the free spin test 10 times, but is it 10 times back to back or 10 times total? my back was killing me, so i had to stop at 5. i haven't since tried to do the test again.

so if InMotion has no plans to replace our current V12 controller, i may opt to sell the V12 or buy the V12-HT controller (assuming this works and has better rated MOSFETs with different part number).

Any idea what to expect?

[mrelwood]

19 minutes ago, Tan Ho said:

BTW, i know they said do the free spin test 10 times, but is it 10 times back to back or 10 times total?

Check MADpack’s video where he reads and explains Inmotion’s answers to two dozen V12 mosfets issue questions:

If I understood him correctly, you let the wheel accelerate then yank back. That is one yank. You should do only two yanks in a row, then stop and put the wheel down once the tire stops spinning. Now you have done one serie of two yanks. Repeat this serie 10 times, so that you have done 2x10= 20 yanks in total.

 But more important than the exact amount of yanks is that you must do them correctly! Shoot yourself on a video as you do a few yanks and the deeply committed people here can tell you if you did it correctly.

 Not doing it correctly can also be the reason for the “guy”’s V12 failing even after a successful “test”.

[tonto_euc]

@mrelwood thanks! very helpful video! there were some confusion in regard to the term. so in the video, he said do 10 groups of tests. each group consists of the yank, brake, yank & then brake. i did do that but accelerated one more time to let it cutoff: yank-brake-yank-brake-yank.

sometimes it was difficult to do and i ended up yanking to cutoff sooner than i expected, so did not count that as a successful group. in total, i have done 5 successful groups out of probably a dozen attempts. looks like there is a brief stop between each group of tests.

so since i only successfully did 5 groups of tests, do i do another 5 or should i start over again with 10 groups in a roll?

it's a bit disheartening to hear that InMotion did not commit to swapping out the controller for the new controller with the improved MOSFETs.  

[Rawnei]

Inmotion is updating their app with a function that will show you if you are doing the tests correctly (I don't know when it's out).

[conecones]

18 hours ago, Tan Ho said:

Saw a post from a guy who had a V12 do well on the free spin test. He then changed his tired, hit a curb and bam, the controller went out. not sure if he damaged something during the tire install, the tire was too tight on the wheel so had added resistance, or just bound to happen due to a sudden bump.

now i'm considering selling my V12 despite passing the free spin test after 5 times. BTW, i know they said do the free spin test 10 times, but is it 10 times back to back or 10 times total? my back was killing me, so i had to stop at 5. i haven't since tried to do the test again.

so if InMotion has no plans to replace our current V12 controller, i may opt to sell the V12 or buy the V12-HT controller (assuming this works and has better rated MOSFETs with different part number).

Any idea what to expect?

This is very likely to be happening. There is no official announcement, but word of mouth from dealers has confirmed Inmotion is sending out the next revision of the boards in March (maybe April). Due to chip shortages across the industry, this date is actually pretty good. For comparison, it's going to take longer to receive my Ipad that I just ordered last week.

A recall is likely not happening, which means for owners of the previous gen board, they are out of pocket for the new board or just live with the uncertainty of the old board. If the wheel test through the APP is effective and we don't see evidence of more cutouts after passing the App test, then I can respect this decision of no recall from Inmotion.

Gotway regularly does unannounced updates (boards, bearings), and yet people still not only worship that brand but then go and spend hundreds/ more on aftermarket capacitors, fuses, BMS'/Custom batteries, shells, bumpers, bearing/bearing covers/waterproofing treatment, pedals, pedal hangers,..etc, because the stock wheel has so many damn flaws . You think Gotway did those updates for us? No they did it to make a known problem go away (a big enough problem that actually affected their profits).

[Planemo]

2 minutes ago, conecones said:

Gotway regularly does unannounced updates (boards, bearings), and yet people still not only worship that brand but then go and spend hundreds/ more on aftermarket capacitors, fuses, BMS'/Custom batteries, shells, bumpers, bearing/bearing covers/waterproofing treatment, pedals, pedal hangers,..etc,

Theres a difference though.

I just read of another V12 that passed the stress test, went on for another 700 miles then failed. It's sheer luck that the rider didn't break bones (or worse).

It's the whole uncertainty with the V12 that would make me never step foot on one until theres confirmation exactly what the problem is, or another board is thoroughly tested and supplied.

The difference is that the raft of things you listed above re GW don't generally lead to a faceplant if left untouched (bar the potential 900Wh pack fires).

As I said long ago, riding a V12 is like Russian Roulette. I can't believe people are doing it tbh, but I did read another post (which was replying to the failed one) where the rider said he was thinking of selling but had decided to 'risk it'. Fcuk that, a braver man than I.

I read that Inmotion can't provide serial numbers of which units are affected which I am saddened by - I expected more from them re production line management. Without that crucial information theres no way I would simply rely on the stress test when it has already proven to be inconclusive.

[tonto_euc]

44 minutes ago, conecones said:

This is very likely to be happening. There is no official announcement, but word of mouth from dealers has confirmed Inmotion is sending out the next revision of the boards in March (maybe April). Due to chip shortages across the industry, this date is actually pretty good. For comparison, it's going to take longer to receive my Ipad that I just ordered last week.

A recall is likely not happening, which means for owners of the previous gen board, they are out of pocket for the new board or just live with the uncertainty of the old board. If the wheel test through the APP is effective and we don't see evidence of more cutouts after passing the App test, then I can respect this decision of no recall from Inmotion.

Gotway regularly does unannounced updates (boards, bearings), and yet people still not only worship that brand but then go and spend hundreds/ more on aftermarket capacitors, fuses, BMS'/Custom batteries, shells, bumpers, bearing/bearing covers/waterproofing treatment, pedals, pedal hangers,..etc, because the stock wheel has so many damn flaws . You think Gotway did those updates for us? No they did it to make a known problem go away (a big enough problem that actually affected their profits).

that's great news that a new board is coming in March/April timeframe. but sucks that current V12 owners with working controllers are out of luck. i hope eWheels will allow me to pay a discounted price to send my V12 in for rework with a new controller board.

When will the InMotion app come out to do the app test? Is the test just to let us know if a particular spin group test is valid? So we're going to have to do the same 10 groups of free spin tests again?

Edited February 4, 2022 by Tan Ho

[Bizra6ot]

It seems to be the same for the last batch of V11, probably with the same bad series mosfets as for the V12, this is the third reported case I've seen since the 2022s failed during braking

 https://youtu.be/M68BhtHvJiU?t=657

 

Edited February 4, 2022 by Bizra6ot

[tonto_euc]

32 minutes ago, Planemo said:

Theres a difference though.

I just read of another V12 that passed the stress test, went on for another 700 miles then failed. It's sheer luck that the rider didn't break bones (or worse).

It's the whole uncertainty with the V12 that would make me never step foot on one until theres confirmation exactly what the problem is, or another board is thoroughly tested and supplied.

The difference is that the raft of things you listed above re GW don't generally lead to a faceplant if left untouched (bar the potential 900Wh pack fires).

As I said long ago, riding a V12 is like Russian Roulette. I can't believe people are doing it tbh, but I did read another post (which was replying to the failed one) where the rider said he was thinking of selling but had decided to 'risk it'. Fcuk that, a braver man than I.

I read that Inmotion can't provide serial numbers of which units are affected which I am saddened by - I expected more from them re production line management. Without that crucial information theres no way I would simply rely on the stress test when it has already proven to be inconclusive.

i was on that thread and said i'm thinking of selling my V12. i was surprised to hear the guy i was corresponding say he had considered it but decided to "risk it". that's just nuts. i'm about to be 50 next week and had dislocated my shoulder due to a 7mph pedal clip last year. can't afford to get hurt on avoidable potential hazards, like these flawed V12 controllers, or my wife may just take away my riding privileges. 

i'm going to email eWheels and see what my options are.

[Waulnut]

47 minutes ago, Tan Ho said:

i'm going to email eWheels and see what my options are

Let us know the result if possible. I'm hitting over 400 miles on my wheel so far. I'd pay for half the cost to get a new controller. 

[conecones]

7 hours ago, Planemo said:

 

The difference is that the raft of things you listed above re GW don't generally lead to a faceplant if left untouched (bar the potential 900Wh pack fires).

While some are minor issues, there are also reported instances of "random" faceplants on Gotway wheels. If you frequent the Gotway groups, you'll see examples of (usually new) riders cutting out "randomly" because of riding in the rain. I don't know how they missed the memo but many thought it was waterproof .

One of the early RS19's in our local group cut out at high speed from a fried mosfet. Smoke started coming out the wheel and the batteries were disconnected immediately as a safety precaution. This is a first gen large bearing RS with green board. As a result, everyone in the same group buy got free replacement black boards from Gotway. This is one of those unannounced upgrades Gotway did, and the green board is rumoured (but never officially confirmed) to overheat as a result of allowing too much power. Many riders have also noted less torque and top speed after switching to the black boards (which also come with larger heatsink). None of this is official - why? Because Gotway purposely doesn't tell you when they screw up.

Black Cobra's random cutout on the Mpro - was a cause ever confirmed?

Recently witnessed a local rider on his ExN instant faceplant in front of me on a trail. Wheel turned off but no idea why - there were no bumps in that section and it turned on fine after. Total mystery, been working fine for a few thousand KMs previous.   

This Inmotion problem is only getting more attention because Inmotion officially addressed it. There's many Gotway cutouts if you look around.

7 hours ago, Planemo said:

 

I read that Inmotion can't provide serial numbers of which units are affected which I am saddened by - I expected more from them re production line management. Without that crucial information theres no way I would simply rely on the stress test when it has already proven to be inconclusive.

Agreed 100%. This is something I hope they can get confirm and release to the public. Let's not forget the first batches were sent to many reviewers who absolutely beat the living hell out of their V12's riding them much harder than most people would. They seem to be fine, so it does show the design is solid, even if the electrical engineers on this forum are all up in arms about the 100V mosfet spec.

 

 

 

[supercurio]

@conecones please note that on the public cutout tracker, batch 1 represents 25% of the wheels affected.

@Nick McCutcheon also reported a cutout followed by burned MOSFET on an early production V12 EUCO got as demo wheel.

It's a common misconception, I believe initially driven by Inmotion that only some batch 2 V12 are affected, blaming bad components from Infineon in the process.

However the data which was collectively gathered contradicts this view.

I think you're right mentioning the many other board failures, some of which have been particularly harmful affecting Gotway wheels as well. I can't help but wonder if That Black Cobra's Monster Pro would have survived the type of stress test which were invented to identify the worst V12 units.

I know some will never even contemplate the idea, afraid of destroying a fine wheel but I would recommend anyone getting a new wheel (any manufacturer, any model) to try the same stress tests before riding.