To add or not to add fuse !

[yuweng]

Spoiler

 

 

When i first opened up this brand new wheel 1 1/2 years ago, i already thought of adding fuse but then again, reported incidents are close to non. Then they announced that they are adding fuse but i can't see what type of fuse or rating they are using. Then last week my worse nightmare has happened 

Just before this happened, i just went up & down to a very steep hill no problem. Then went into nearby MTB trail, there was this steep rock garden down hill that i would normally just glide thru skipping all those rocks but that day i was not confident enough so i swayed to my left to brake. Then i hear a 'pop', quickly i sat the whole EUC to the ground. Then it continue, i can see explosion thru the shell, it was like fire cracker popping inside & smoke for close to 10 seconds non-stop ! My riding buddy was shouting 'battery fire battery fire' from behind, luckily it was just the board or else li-ion fire could have started a forest fire !

Not only that the board was destroyed, most of the wires are melted as well 

[RagingGrandpa]

Benefits of fusing:

1. After a board failure, fuses can help avoid burnup and arcing of the control board (like you experienced).
    
2. After a board failure, fuses can help avoid overcurrent discharge of the EUC battery pack (which can damage cells in ways that are very difficult to inspect).

 

Drawbacks:

1. A fuse will not save your board.
   In your case, your board overheated and failed, and was instantly destroyed. After that, there was arcing and burning... but even if you had used a fuse, the board would require replacement after the failure.
    
2. Choosing a current rating.
   The fuse rating must be chosen carefully. Too low a value could result in nuisance opening (typically at low speeds). Too high a value will give no benefits.
   I suspect one 50A fuse at the input of the MCM5 controller would be appropriate, based on my measurements from similar unicycles.
    
3. Additional complexity and points of failure.
   The fuse must be installed in a robust way. If the connections to the fuse fail, the EUC will cutout and you will crash.
   Soldering to blade-style fuses (like Gotway did!) is not recommended- a fuse holder should be used instead (like Veteran did).
    
4. Specialty fuses.
   The cheap automotive fuses (like what Kingsong, Veteran, and Gotway use) are only rated for reliable operation at 32V.
   If you use those fuses to interrupt a short circuit in your 84V EUC, the plastic case of the fuse will explode, and an arc is produced. Maintaining physical separation between the fuse and any other nearby wires or terminals is important, to ensure the arc does not reach another circuit before dissipating.
   It would be better to use fuses rated for 125VAC or higher (which will not explode when they melt at 84VDC). But such fuses are larger, more expensive, and more difficult to install.
    

If I were you, I'd replace the board and not add anything more.

[HighEndScooters]

Very detailed response thank you @RagingGrandpa

Is it worth retrofitting fuses to machines like gotway Tesla mcm5v2 etc in your opinion? 

[RagingGrandpa]

On 5/1/2021 at 9:02 PM, HighEndScooters said:

Is it worth retrofitting fuses to machines like gotway Tesla mcm5v2 etc in your opinion?

No. You're adding complexity for minimal benefit.

Regarding electrical upgrades: it would be awesome to somehow reduce the risk of pack fires (which are rare)... but output fuses don't help that issue.

[RockyTop]

It would be very difficult to pick the correct fuse. Fast blow vs slow blow. How many amps? It would take a lot of research to get to fuse size and type correct. Kingsong is probably the only company I would trust getting it right. 

[RagingGrandpa]

Apparently the Gotway 120A fuse does open during certain board failures, impressive... I still hate how they soldered to it, though.

https://www.facebook.com/groups/ElectricUnicycle/permalink/3935915353173065/

 

Edited August 17, 2023 by RagingGrandpa
(img url)

[RagingGrandpa]

... and sometimes the Sherman burns its board before its fuses

https://www.facebook.com/groups/ElectricUnicycle/permalink/3941266965971237/

 

[Scottie888]

If its OP's MCM5 we're talking bout here, I believe he's mentioned in other posts its modded with an aftermarket 84v 1300wh Li-ion battery pack & controller from its stock 67v 340wh.

FWIW I've tossed my stock 800wh MCM5v2 innumerable times & have even dunked it sideways in 8"+ puddle of water & slush this past winter/spring, atleast a coupla times. Its a rock then & still a rock now. Its possible the mods may have something to do with this unfortunate event.

The MCM5 (V1/2) has been around for a few yrs now & there's been zero confirmed incidents of fires or weak mobos. Personally I certainly wouldn't go find skeletons where there's none. 

Even if a wheel's climbing an unclimbable & overloads, it would just cut out, wouldn't it.

[yuweng]

My riding buddy who is also a dealer got me a new board quite fast but still haven't got the time to install it back. First thing i notice is that they are using the old TO-220 instead of TO-247 is it bcos TO-247 shorted out will destroy the board ? He also got me a fuse, he said there are only type 120a & 150a. Damn bcos of U.S sanctions all China product prices went up  

Gonna open up the motor to inspect as well in these few days.

15 hours ago, RagingGrandpa said:

... and sometimes the Sherman burns its board before its fuses

https://www.facebook.com/groups/ElectricUnicycle/permalink/3941266965971237/

 

That is what exactly has happened to my wheel, it continues popping non-stop  

12 hours ago, Scottie888 said:

If its OP's MCM5 we're talking bout here, I believe he's mentioned in other posts its modded with an aftermarket 84v 1300wh Li-ion battery pack & controller from its stock 67v 340wh.

FWIW I've tossed my stock 800wh MCM5v2 innumerable times & have even dunked it sideways in 8"+ puddle of water & slush this past winter/spring, atleast a coupla times. Its a rock then & still a rock now. Its possible the mods may have something to do with this unfortunate event.

The MCM5 (V1/2) has been around for a few yrs now & there's been zero confirmed incidents of fires or weak mobos. Personally I certainly wouldn't go find skeletons where there's none. 

Even if a wheel's climbing an unclimbable & overloads, it would just cut out, wouldn't it.

As mentioned, been riding for 1 1/2 years with 2.5K km mileage no problem. Everyone at one point will open up their wheel. The purpose of this thread is to discuss the Pros & Cons of adding fuse to old GotWay/ Begode wheels since all new ones do come with fuse.

As RagingGrandpa has pointed out that even with fuse & it still continues to destroy the board 

[Scottie888]

3 hours ago, yuweng said:

As mentioned, been riding for 1 1/2 years with 2.5K km mileage no problem. Everyone at one point will open up their wheel. The purpose of this thread is to discuss the Pros & Cons of adding fuse to old GotWay/ Begode wheels since all new ones do come with fuse.

As RagingGrandpa has pointed out that even with fuse & it still continues to destroy the board 

Understood & I get it this is a discussion. However in order to have a viable discussion, all relevant facts have to laid out on the table or otherwise, an uninitiated reading this might come to unwarranted & perhaps unfair conclusions.

For the most part, Gotway is the creator of its own destiny & should take the fall for that. However lets not pile on more crap bcos its fashionable & appeals to our biases. Diy mods maybe simple in theory but perhaps not so much in practise & can be destined for a rocky marriage even for the experienced. I also get that its almost always easier & convenient to blame others.

Anyhoos back to discussion. I'm far from being an elec engineer but don't mobos normally fry bcos of shorts? Overloads are not the only source of shorts. We're basically tossing the wheel around every time its ridden (nevermind crashing) so I'm only saying its part of maintenance to regularly open it up & ensure all connectors are attached correctly & everything else are screwed tight in their proper locales as its not difficult to imagine a loose connection frying the board as well.

[RagingGrandpa]

On 5/8/2021 at 11:34 AM, Scottie888 said:

I'm only saying its part of maintenance to regularly open it up

Sure.

Remove and prevent conductive debris.
Secure the wire harness to prevent insulation wear-through and vibration fatigue.
Inspect & remedy loose mounting of PCBs.

But your controller could still burn for many other reasons. More thoughts below:

  

Just now, RagingGrandpa said:

> don't mobos normally fry bcos of shorts?

Well, not really. The typical causes for modern EUC controller "burn"-type failures, based on available information from our community, are:

• Board-temperature-related
  This covers most cases of 'overloading' - the popcorn-sound board burnup during riding, typically while riding up steep grades. It means a "good" controller, which had no defects in components nor manufacturing, experienced failure of components, conductors, or interconnects due to excess temperature.
   
• Manufacturing defect
  This covers the occasional but significant percentage of boards that burn during the first few miles of normal, non-aggressive riding. The likely cause being defective components (bad chips / capacitors / etc) or assembly defects (wrong component / missed component / poor connection / solder debris).
  PCB testing during manufacture can catch many such issues, but not with 100% confidence.

The final result of these problems can be "uncontrolled excessive current flow," which you could refer to as a short-circuit... but that's a result, not a root cause.
 

> its not difficult to imagine a loose connection frying the board

Poor connections are bad, of course. We could consider a few:

1. Weak solderjoint in the high-current path: (PCB defect)
   Result: local heating, desoldering, arcing
   (and potential for arcs or molten solder to short other nearby circuits)
    
2. Weak or intermittent spring-type terminals (any 'connector' on the controller): (Terminal defect or excessive wear)
   There are different ones to consider here:
   Fan- should result in overheat alarm before damage occurrs
   Buttons- only result in failure to respond to user input
   Lights/USB5V/Speaker- only result in failure of those devices to operate
   Hall sensors- typically result in cut-out and error beeps, not board damage
   Motor cables- potential for the same 'heating' issues as #1 above. But, they're the heaviest connections, and are quite unlikely to work loose.
   Some controllers (e.g.: Veteran; Inmotion) use bolted (screw-down) terminals for the motor phases. If loosened, the same effects may occur.
    
3. Wire harness open/short circuits: (Assembly defect, fatigue, or insulation failure)
   The only circuits related directly to a "burned controller" are the motor phase wires, where shorting of phases to eachother can cause excessive current flow in the controller. This did happen frequently in older EUC's due to insulation failure (melting), but has become less and less common with today's modern EUC motors.
   Short circuits in the battery harness will abuse the cells, but not so much the controller.
   Open circuits cause devices not to operate.
    

Finally, all controllers are vulnerable to mechanical damage: which could lead to board burn, but more frequently leads to board failure (without 'burnup').

• Vibration, impact, or other mechanical stresses could cause components or conductors to become damaged and fail.
   
• Conductive debris, such as a loose screw, metal shavings, or wire cuttings in the EUC shell, could cause short circuits. 
  Controllers are especially vulnerable to debris at locations such as through-hole capacitor leads.

 

[Scottie888]

@RagingGrandpaSo correct me if I'm wrong but basically from what I can gather from your notes, adding a fuse btwn the battery & mobo only prevents a sudden power surge which isn't likely to happen anyways.

I am somewhat surprised though that with the amount of years & experience (not to mention the quantity involved) in mfgr'g mobos, we can still manage to make 'weak' boards. Maybe its within the margin for a few to fail. At least I haven't heard of many MCM5 board failures.

[Tawpie]

I always try to remember that a fuse is meant to protect upstream wiring/components from a failure. Same as a circuit breaker (a resettable fuse) is there to protect the house wiring from a short in your blow dryer, a fuse between the battery and controller will protect the battery and wiring from a short in the controller. Which is a good thing, but when a controller short circuits it usually clears the short very quickly by blowing up components. The controller is dead and a fuse can’t save it. The fuse could protect the battery from a long lasting short but I think a long lasting short would be a rare (but always possible) occurrence.

I don’t know that the relatively long time required to melt a fuse is going to save the battery from getting wounded anyway—depends on the battery.

Edited May 10, 2021 by Tawpie

[RagingGrandpa]

What fuse value should I use?

Fuses take an annoyingly long time to open... choosing a value for an EUC should start with a conversation like:

1. How much current do I need to sustain for durations longer than 5 seconds?
   My guess is 2x rated power -> 60A for Sherman.
    
2. How much current is damaging to cells and must be interrupted in less than 1 second?
   My guess is 20A per cell -> 200A for 10p Sherman.

Now consider typical opening times for melting fuses of this type (ANL).
If we need that <1sec opening for 200A, then 200A becomes the "400%" rating, and we arrive at (one) 50A fuse. Or, 25A per pack.

Then to check for safe operation at the maximum sustained load for our fuse selection, it will tolerate 1 hour of operation at 150%, so 75A.
More than the 60A we needed, so that's lovely.

Quote

 

Should I fuse each pack separately? Or put one big fuse in front of the controller? Or both?

If your goal is to protect the cells, put fuses as close to the cells as possible. This ensures that short-circuits in any part of the wire harness will be protected.

If your goal is to prevent arcing failure at the controller, put a fuse at the controller input. This ensures a single current limit for the controller, regardless of how many packs are connected to it.

If you really want both things, you should consider using both.
(Like Inmotion does...)


What if one pack becomes offline?
(Blown fuse, broken wire, open connector, etc)

This gets tricky.

The best answer is "warn the rider" - but Sherman and Gotway controllers have no way of knowing about a pack offline. The EUC will seem to behave normally, just with shorter total range.

In the Sherman example, if we selected 25A fusing per pack, and one pack becomes offline, consider the fuse opening time at 60A sustained: about 250% of the 25A rating, so the fuse should open in about one second. Not great.
If I had a Sherman, I would increase the rating to better tolerate this kind of one-pack operation: 30A seems nice, which should give more than 100 seconds of continuous torture at 60A. Given 100 seconds of such torture, the controller would fail before the fuse opens.

And hey look, Veteran agrees, and is now fusing Sherman at 30A per pack
 

On 6/10/2021 at 4:07 AM, fbhb said:

The fuses when originally added as a protection to the Sherman, following Marty's OHH burned board and eWheels intervention, were indeed 50A up until a recent decrease down to 30A by Veteran.  I know this to be a fact following my 2x control board replacements that were both sent out including fuses to retrofit/replace the existing 50A my wheel already had fitted. 

The first board arrived in January with 50A, the second in April with 30A so the change was put in place sometime between January and April

Edited August 17, 2023 by RagingGrandpa
(img url)

[RagingGrandpa]

Jason shared an interesting example of melting fuse delay:

On 11/9/2021 at 8:22 AM, Jason McNeil said:

1. Keep discharging with 60A current on 24S4P, battery can run about 10 minutes, then fuse blew (8:45 in video below)
2. Current more than 102A, in a few seconds the fuses blew, that means any current more than 100A in a few seconds our fuses could be blown

 

 

In this test, the green mini-AT fuses for each pack were rated 30A.

Frankly, I'm surprised they blew at all, with two packs paralleled and subjected to a 60A load. That's 100% of the fuse rating: it should not melt.

Perhaps the current was not perfectly balanced in the parallel connection... but I was still expecting 125% rating sustained indefinitely. Hmm.
 

 

p.s. the fireball is why these fuses are only rated for use up to 32V... but hey, the circuit opened, fine.

 

Edited August 17, 2023 by RagingGrandpa
(img url)

[yuweng]

According to my dealer friend, begode/ gotway has recently introduce a self-resetable fuse in all of their wheels, lets see if fire will be a thing of the past for them 

Spoiler

 

[thendless]

RagingGrandpa I was just thinking about your comment about warning the rider that a fuse has blown in one of the individual packs. I was originally thinking there must be a way electronically to signal this and and mentioned there is no real way, other than you noticing our reduced range or if you cutout after expecting the same amount of power from your wheel. Perhaps a primitive method for signaling may be all that is needed, IE smell. Maybe there is a material that we can affix to the fuse via shrinkwrap that is heat activated (but not combustible?) that is heat activated and lets out a unique smell.  Have no idea if something exists but maybe something simple like this is all we need.

[RagingGrandpa]

Smellovision?

The practical answer really depends on "which EUC we're talking about" - because they all handle electrical faults differently. Some don't have fuses at all, so this "one pack offline" situation isn't even possible.

Which EUC did you have in mind?

 

Edited November 15, 2021 by RagingGrandpa

[thendless]

Smellovision would be great! I mean there must be a simple way to utilize the heat generated from a blown fuse to trigger some sort of notification.

I was thinking more in terms of retrofitting existing packs. Its should be easy to add fuses in downstream of each battery pack, similar to how gotway added the 120A fuse downstream of the controller (Male XT60 - Fuse - Female XT60).

[RagingGrandpa]

I would look for a DC buzzer that would make noise at 2V and could survive 20V. Connect it "across the fuse" - potential would only be present if the fuse melts. If one pack's fuse opens, and another pack is still connected in parallel, the remaining pack will continue discharging and the voltage seen by the fuse buzzer will slowly increase, alerting you...

[thendless]

After some more thought all these fuses appear to be automotive fuses not rated the higher voltages that EUCs utilize. As a result when the fuse blows a small fireball tends to happen and its more than likely we would be able to smell something.

[yuweng]

Nowadays i get all my answers from here instead of google 

[yoos]

13 hours ago, thendless said:

I was originally thinking there must be a way electronically to signal this and and mentioned there is no real way

There is a simple way for the controller to recognize that a pack went off-line: measuring pack battery resistance. Even more, it could be easily implemented in EUCWorld or any other app that tracks voltage and current.

To elaborate: all wheels exhibit voltage sag: V(t)=V_1(t) - I(t)R where V_1(t) slowly decreases as battery depletes while the R is the internal resistance of the battery and I(t) is current drawn from the battery at time t. From the V(t), I(t) data R can easily be determined. For N parallel packs the resistance is R=R_0/N, where R_0 is the resistance of a single pack. So if one pack goes off-line the resistance will change abruptly. It should be trivial to set an alarm triggered by such an event (e.g. 8% change in R as a threshold for a 10p Sherman or 20% change for a 4p wheel where a single missing pack increases R by 33%)

Now there are two basic algorithms  to implement this:

1) dynamically track R and trigger an alarm if it experiences a high enough jump

2) have a proper value of R stored (it's easy to calculate from battery specs of the EUC, this data could be pulled from some table of EUCs or put in manually if your wheel has a modded battery setup) and check whether it's matched by the R that is derived from V(t),I(t).

Approach 1 works for any wheel, no need to know the battery setup in advance. Approach 2) has the advantage of noticing if a pack goes missing during a power-off/power-on cycle, which is actually quite a common problem.

I would combine the two, triggering on both events, i.e. if R jumps in value and if R does not match the value specified for this EUC. 

 

[Planemo]

54 minutes ago, yoos said:

There is a simple way for the controller to recognize that a pack went off-line: measuring pack battery resistance. Even more, it could be easily implemented in EUCWorld or any other app that tracks voltage and current.

@Seba would this be possible?

Good post though! Although I do wonder if a warning would still save a crash - lose a pack on most wheels whilst riding and I would think it's going down in most instances, unless you happen to be doing some slow chilled riding. Still deffo worth having though, even for the reason of picking up an offline pack on power on/off as you say  

Also be handy for post-crash analysis/cause investigation. I always run EUC logs when riding for this reason. 

[Chriull]

5 hours ago, yoos said:

There is a simple way for the controller to recognize that a pack went off-line: measuring pack battery resistance. Even more, it could be easily implemented in EUCWorld or any other app that tracks voltage and current.

 

4 hours ago, Planemo said:

@Seba would this be possible?

Afaik internal resistance calculation is already in testing/?one of the latest versions? of EUC World. But should only be availabe/make sense for wheels reporting the battery current like Inmotion or Kingsong.

There is/was battery current estimation for Begode/Veteran wheels but as far as i got it maybe too inacurate for such a purpose.