Big fire ball today - Ctrlr REPAIRED !

[Hugues]

Got into trouble today with my bike. I'm listing below the chain of events, before i forget them:
- rode the bike this week-end, no problem, inspected the whole battery pack, everything ok
- recharged the pack last night
- bike to work today, no problem, no rain
- tonight, leaving from work, i realised my wheel was a bit stiff to turn
- did not know what it was, switched on the controller, the display does not come on on the CA
- i open the battery compartment to check connections, all seem ok
- remembering some recent posts about stiff wheel, i disconnected the phase wires from the controller
- the wheel is turning freely
- leaving the phase disconnected, i disconnect the battery from the controller, also the precharge circuit
- then reconnect the pre-charge circuit, and when i want to connect the controller to the battery, there is a spark on the connector
- try a second time, spark, but manage top connect them anyway. strange, looks like the precharge resistor is not doing its work
- call my wife, put the bike on the car, back home.
- there i remove the battery pack completely from the bike, measure voltage, 86 V, all looks ok
- so i wanted to reconnect the whole thing. Connect first the precharge circuit
- and when i connect the controller to the battery, a huge fire ball starts around the harness, scared the shxx out of me, thought it was a lipo
- i manage to pull some cables and disconnect the controller
- after the smoke is gone and my heart rate comes back below 200, i can see the precharge resistor has literally melted. Pictures attached

YOu can see the 3 battery on top, connected in series, then you can see the connector for the controller down left, not connected to the battery, below is the precharge resistor, melted, vaporized. Last long blue cable on left is my charging cable.

what the heck is going on ? I'm not sure where to start.
I checked if there is continuity between 3 phases: negative. Continuity between phase and bike: negative.
If i reconnect phases to controller: the wheel is stiff to turn, even though controller is not connected to battery.
Batteries seems fine, no puffing, voltage still 86 V.
Fire ball has left 2 white spots on my bike, paint has melted it seems.

Really need guidance here before i screw things more. Thanks in advance.

:|

 

[liveforphysics]

Your controller had a FET fail shorted. Its why it couldn't precharge, why the wheel turned stiff, and its why the connector flashed.

 

[gwhy!1]

Sounds like you were lucky to get some of the connections apart and not do more damage. From what you have said it sounds like you have a short in the controller ( or wiring ) I assume you do not have a inline fuse from your battery ( if you dont , it maybe a good plan to fit one ). A short in the controller will cause the pre-charge resistor to burn up and also any wire from the resistor to the short. Check with a ohm meter across the power input into the controller ( black and red leads ) this should be a min of around 8kohms, if its a lot less than this the the controller is at fault and could cause the failure.

 

[Drunkskunk]

It was a dead FET, but at this point, possably a dead controller.

No fuse on the battery?

 

[Hugues]

Thanks for the quick feed-back guys, i appreciate.

Measured the resistance between red and black going to the controller, pretty much zero ohm.

2 questions first before i forget:
- what's a ok value for a precharge resistor ?
- what's a ok value for a battery fuse ? where to mount it ?

Now i opened up the controller, pictures below. Visually i cannot see anything damaged. How does a failed FET look like ? Can i measure the resistance for example across a pair of their 3 legs ?

I've got a spare controller, although last time i checked it was not working, so could be a source of spare FET if need be.

thanks for your input.

 

[gwhy!1]

the value of your pre-charge resistor will depend on your battery voltage ( i cant remember what exact values I have used in the past but they are normally around 300ohm and 1 watt (this is for a 50v setup ) but its not that critical to have a exact value . the value of the fuse will also depend on your max current under normal conditions, if the current maxed at 20A then a 30A would be good enough ( as long as your wiring is rated for a higher current than 30A )

 

[itchynackers]

Is it me, or do the legs of that blue resistor look cooked?

 

[Hugues]

Is it me, or do the legs of that blue resistor look cooked?

no, this one is ok. I made a modification when i received the controller , there is a long thread about this on the forum.

I'm searching the net now to learn how to test a FET. I often find the test where we use a digital multimeter capable of putting 3 volts at least on the FET.

Don't think i can do this with my multimeter, i will check.

is there a simpler way ?

 

[AussieJester]

7th fet from the left looks like it is missing the middle leg?

KiM

 

[bigmoose]

Now i opened up the controller, pictures below. Visually i cannot see anything damaged. How does a failed FET look like ? Can i measure the resistance for example across a pair of their 3 legs ?

The trouble is often if a FET shorted from an overvoltage spike and didn't have a current blow up, it looks just like a normal FET.

Here is how I would go forward: Controller disconnected from everything.

When you look at the front of each FET, the leads are 1-Gate 2-Drain 3-Source from Left to Right. All N channel FETs have an intrinsic body diode That has the Cathode on the Drain, and the Anode on the Source.

Use your meter in diode conduction mode. A good FET should look like this:

1) Meter red to pin 2 Drain, Meter black to pin 3 Source: No conduction
2) Reverse the meter leads on the FET
3) Meter black to pin 2 Drain, Meter red to pin 3 Source: Should show body diode conduction of 0.6 Volts.

A blown pair of FETs should show less than 0.6 volts in both directions. Now since the FETs are paralleled, both could be blown, or only one. You will have to unsolder one then do this test again on each individual FET to figure out which one or if both are blown. If one is blown, it is probably prudent to replace both, as the other guy saw the same voltage over stress.

On some H bridge designs it is best to not power the controller without a motor attached.

I read this over about 4 times... I sure hope I didn't reverse the colors or pins in a dyslexic fit!

 

[Drunkskunk]

looks like it to me. And I can't see the burn marks on the blue resisto's legs, but it does have that powdered blue look to the middle like they get when they cook..


Precharge resistor I use is 1K at 1w. Overkill but would be fine. I use Automotive fuses you have a 40A controller, so a 40A fuse.

 

[John in CR]

Identifying and replacing the blown fets is relatively simple. The big question that defies good explanation is how did the controller blow. I have had exactly the same thing happen, that is a working bike running fine, no changes or events, and bike doesn't work next ride. I have had exactly those circumstances occur too.

 

[Njay]

On a FET I've blown up, the resistance between gate and source/drain went down to 4 Ohm (tooking the output transistor of the driver down with it). The resistance should be infinite. If this is the case, it's easy to identify which FET(s) are blown without picking up the soldering iron.

 

[knoxie]

when will folks learn and start fitting fuses? I find it hard to understand why you would fit a pre charge resistor to protect your connectors but dont fit a fuse to protect the system as a whole? Use a std 40A automotive fuse, my harness has 3 fuses, 2 on each parallel string and 1 on the main line, they offer literally no resistance as they are soldered inline but they offer so much peace of mind, unchecked a blown controller or severed battery wires would most likely destroy your battery, bike and what ever is near.

Anyone reading this who hasnt.. FIT A FUSE / FUSES FOR FROCKS SAKE

 

[sn0wchyld]

when will folks learn and start fitting fuses? I find it hard to understand why you would fit a pre charge resistor to protect your connectors but dont fit a fuse to protect the system as a whole? Use a std 40A automotive fuse, my harness has 3 fuses, 2 on each parallel string and 1 on the main line, they offer literally no resistance as they are soldered inline but they offer so much peace of mind, unchecked a blown controller or severed battery wires would most likely destroy your battery, bike and what ever is near.

Anyone reading this who hasnt.. FIT A FUSE / FUSES FOR FROCKS SAKE

i guess some (ie me) didnt think of it, since there's no step by step guide to building ebikes, at least that ive found... though now I know this it'll be the first thing I do when I start holidays next week, and till then Ill never leave the controller pluged in to the batts when not in use. It seems obvious in hindsight, but yea, for those of us on our first build it isnt necessarily the most obvious thing. I do however remember thinking that its strange that people feel comfortable leaving their controller plugged in 24/7... Ive always unplugged it when not in use. Now I know...

perhaps this topic should be stikied (or something similar) as a warning to the noobs out there?

guess its lucky it was a wirefire, not a lipofire...

 

[John in CR]

when will folks learn and start fitting fuses? I find it hard to understand why you would fit a pre charge resistor to protect your connectors but dont fit a fuse to protect the system as a whole? Use a std 40A automotive fuse, my harness has 3 fuses, 2 on each parallel string and 1 on the main line, they offer literally no resistance as they are soldered inline but they offer so much peace of mind, unchecked a blown controller or severed battery wires would most likely destroy your battery, bike and what ever is near.

Anyone reading this who hasnt.. FIT A FUSE / FUSES FOR FROCKS SAKE

+1

Correct me if I'm wrong, but doesn't a precharge resistor circuit also need it's own small fuse, since we don't want anything approaching acceptable mains power going through the thin wires of the precharge circuit such as in Hugues' flash event?

 

[Drunkskunk]

when will folks learn and start fitting fuses? I find it hard to understand why you would fit a pre charge resistor to protect your connectors but dont fit a fuse to protect the system as a whole? Use a std 40A automotive fuse, my harness has 3 fuses, 2 on each parallel string and 1 on the main line, they offer literally no resistance as they are soldered inline but they offer so much peace of mind, unchecked a blown controller or severed battery wires would most likely destroy your battery, bike and what ever is near.

Anyone reading this who hasnt.. FIT A FUSE / FUSES FOR FROCKS SAKE

+1

Correct me if I'm wrong, but doesn't a precharge resistor circuit also need it's own small fuse, since we don't want anything approaching acceptable mains power going through the thin wires of the precharge circuit such as in Hugues' flash event?

If the resister is sized right, even at a dead short between battery leads, it should never pass more than 1 watt
But this thread has me thinking, Maybe I ought to have 1 fuse per parallel string as well. could save some trouble later.

 

[sn0wchyld]

when will folks learn and start fitting fuses? I find it hard to understand why you would fit a pre charge resistor to protect your connectors but dont fit a fuse to protect the system as a whole? Use a std 40A automotive fuse, my harness has 3 fuses, 2 on each parallel string and 1 on the main line, they offer literally no resistance as they are soldered inline but they offer so much peace of mind, unchecked a blown controller or severed battery wires would most likely destroy your battery, bike and what ever is near.

Anyone reading this who hasnt.. FIT A FUSE / FUSES FOR FROCKS SAKE

+1

Correct me if I'm wrong, but doesn't a precharge resistor circuit also need it's own small fuse, since we don't want anything approaching acceptable mains power going through the thin wires of the precharge circuit such as in Hugues' flash event?

If the resister is sized right, even at a dead short between battery leads, it should never pass more than 1 watt
But this thread has me thinking, Maybe I ought to have 1 fuse per parallel string as well. could save some trouble later.

the resistor should effectively be its own 'fuse' shouldnt it? ie so long as your resistor isnt so big that it can disipate the power running though it continuously, it will eventually burn up and sever the link that way. ie a 1000ohm resistor on a 100v system should be smaller than say, 3 watts. That way it should handle the spike of power during normal connections without blowing up (10w for maybe 1/2 a second or so before tapering off to <3w), but if there's a dead short then itll die in less than a minute (and your batterires are only 'putting out' ~10w untill this happens, so no struggle for them). though I dont actually know how long a sub 3w resistor would take to blow at 10w... somthing ill have to experiment with I think .

 

[gtadmin]

Ohm's law states that the current through a conductor between two points is directly proportional to the potential difference across the two points. Introducing the constant of proportionality, the resistance, one arrives at the usual mathematical equation that describes this relationship: I = V ÷ R

For your 100V and 1,000 ohms, I = 100 ÷ 1000 = 0.1A and because W = V × I then 100 × 0.1 = 10W ==> hardly wire melting power and might take a while to burn your resistors (unless you are only using a 1W or worse, ¼W resistor)

 

[sn0wchyld]

Ohm's law states that the current through a conductor between two points is directly proportional to the potential difference across the two points. Introducing the constant of proportionality, the resistance, one arrives at the usual mathematical equation that describes this relationship: I = V ÷ R

For your 100V and 1,000 ohms, I = 100 ÷ 1000 = 0.1A and because W = V × I then 100 × 0.1 = 10W ==> hardly wire melting power and might take a while to burn your resistors (unless you are only using a 1W resistor)

hence why i was saying less than 3w resistor... but does anyone know how long a resistor will take when more than its rated is pushed thorough it? im guessing it varies dep. on resistor?

 

[gtadmin]

...
hence why i was saying less than 3w resistor... but does anyone know how long a resistor will take when more than its rated is pushed thorough it? im guessing it varies dep. on resistor?

Hope you don't have it mounted next to your LiPo then No, I'm sure there is a way of calculating it though! (or try it )

 

[sn0wchyld]

...
hence why i was saying less than 3w resistor... but does anyone know how long a resistor will take when more than its rated is pushed thorough it? im guessing it varies dep. on resistor?

Hope you don't have it mounted next to your LiPo then No, I'm sure there is a way of calculating it though! (or try it )

same story with fuses surely? i know they're more 'contained' than a resistor, but you still dont want motlen metal near your batts ! but yea, maybe a fuse would be better overall... no doubht ill get curious anyway and see what happens with a 1w 1kohm resistor and 100v

 

[knoxie]

Fuse in series with the pre charge resistor isnt a bad idea but most likely not required.

For newbs I wouldnt suggest even using a precharge resistor to be honest its just 1 more thing to get wrong or drain off your pack, fit properly rated connectors to your battery and controller leads and the small sparks caused by hot plugging the controller in and out dont wear the contacts as much as you might think. I dont use pre charge resistors on any of my rigs, I simply put a blob of vaseline in the contacts every 6 months or so to prevent capacitor arching becoming an issue.

Bigger controller is a different matter though and a pre charge is a must, the main thing is the fuses, its the sort of thing you will wish you had done when its too late sadly, :x you could save 1000's of bucks worth of kit with something that costs you just pennies and a few minutes to install, if installed properly it wont affect performance of the system but will offer you some protection, you can still of course have a short circuit condition the other side of the fuse but its a must to have at least one fuse fitted in the discharge leads from the battery, if your controller blows thats bad enough but imagine that and the battery going up in smoke as well when it could have been saved.

I think a fit fuses sticky is a very good idea, folks can post up wiring loom examples with details of the kit and component vendors making it as common a part of the build as the throttle cable

 

[gtadmin]

Fuse in series with the pre charge resistor isnt a bad idea but most likely not required.

How would you know it was blown? If it were and you then connected your main leads, you would then blow your main fuse anyway (given you have a short somewhere in the rest of the circuit).

...
a precharge resistor to be honest its just 1 more thing to get wrong or drain off your pack,
...

Yes, could drain your pack over a week or so if you leave it connected

 

[Hugues]

interesting feed-back guys, thanks.

As often the case when i brake something on my bike, i do learn a lot in return, and probably others too through this forum.

Lunch time here. Last night i ordered a 1'000 ohm resistor to replace the one i've blown up.
I also ordered this "baby" fuse:
https://www.distrelec.com/ishopWebFront/catalog/product.do/para/keywords/is/40ET,40ET,Bussmann,Semiconductor_fuse_40_A_super_fast-blow,High_Performance_Semiconductor_Fuse,_Super_Quick-acting/and/language/is/en/and/shop/is/YY/and/series/is/1/and/id/is/01/and/node/is/DC-49442/and/artView/is/true/and/productNr/is/273449.html


Thanks to Bigmoose especially for the detailed procedure to control my FET. I don't think my multimeter has this function. But i'll check if i can find one. In the meantime i managed to isolate the problem to the blue phase by measuring the continuity on each phase, blue shows zero. I also measured the resistance between the legs of all FET and those on blue phase show zero.

So the next step for me will be tonight to remove those 4 FET and replace them by new one i get from my spare controller (not working for other problem). Then tomorrow when i get my fuse and resistor i can re test the whole thing and hopefully nothing else is damaged on my controller.

Anyone got a good link on how these FET work in general and specifics for our controller. Like: why 4 FET per phase, what each one is doing...I always like to understand what I repair.

thanks a lot