FIXED::::I reversed polarity on controller. Please help,

[EdwardNY]

I accidently reversed polarity on my Lyen 18fet controller. I don't get +5 volts on throttle or halls. All I can determine is that I get a small voltage like 1 volt in the +5 throttle when I turn the controller on.

I opened it up and this is what I can determine. It smells a bit burnt inside but I can't see much.

This capacitor in the middle looks a bit burnt but not too bad. The top of it looks fine where the K is. The other tiny capacitor on the bottom left next to the letters DG always looked like that based on looking at other lyen 18fet images. You can see the shunt here which looks fine.


You can see two blue identical things in the middle. The one with the two white wires attached looks burnt. The wires go outside the controller and you connect them for 48 volts, and disconnect them for 49-88 volts. This I assume was to run on multiple voltages, as I requested this from Lyen. At the time of reverse polarity I had the two white wires connected for 12s lipo. I disconnected these wires and ran on 18s lipo but the controller still didn't work.



Just another shot, you can see what looks to be a burnt spot on the middle capacitor. The other two huge capacitors on the right look OK, that white residue I don't think is from them burning. I saw another photo of a lyen 18fet and they had similar white stuff on them.


Just a full shot of the top incase you want to point something out to check.


Just another controller image I found for a lyen 18fet. You can see those large capacitors have white residue on them also.

 

[dnmun]

so after you reversed it at one voltage you increased the voltage another 25V and tried again with it reversed or did you connect polarity correctly the second time?

how many of the capacitors are shorted? did you test the mosfets to see how many burned up and are shorted?

if the two white wires were connected then there should not be any voltage across the second power resistor but it does appear to have overheated too which implies a large current flowed through it which should not be possible if the two wires were connected. use the diode tester on the mosfets to see how close they now are in forward voltage across the body diode. maybe only a few carried the current.

 

[EdwardNY]

The second time I hooked it up correctly in both 48 volts and 74 volts or 12s and 18s lipo. I certainly won't do something like that again.

That is interesting that that second power resistor is black like that considering I had the wires connected.

The resistance in the fets seem to be the same when checking resistance through the phase wires. They are all around 19.

I'll have to look into checking the fets with the diode tester. None of them look burnt.

 

[amberwolf]

The below is speculative:

I can't tell without you tracing the circuits out on the PCB with a meter, but the only way I can imagine the resistors being burned / intact like that is if the white wires actually short across the *un*burned resistor, and that the power feeds in thru the burned one first. (despite where it looks like the white wires physically go to)

Well, actually you could put yoru meter on Ohms setting across each resistor, one at a time, first with the white wires not shorted, and then with them shorted, and see which one shows the low resistance with them shorted (and what their actual resistance is now vs what it should be).

What I suspect is that the burned resistor is now changed in resistance to somethign pretty high, and it's preventing enough of the pack voltage from getting thru into the low-voltage regulation section, so that you don't get enough voltage to run the MCU/etc., and that is why you only see ~1V on a 5V line.


It is also possible that the first regulator (12V, usually) in the LVR system was damaged by the reverse voltage, and that is preventing the rest of the system from getting enough voltage.

Either way, you could simply trace what voltages you get (relative to battery negative) at different ppoints along the low votlage regulation path, first at the "ignition" wire at the PCB, whcih ought tohave pack voltage on it, and then at each side of each of those resistors, and then at the input and output pins of the 78xx regulators (if it has them). I can't see everything on the PCB clearly enough on my little screen with my eyes, so it may be one of the types that doesn't use those, adn has a more efficient (and more complicated) regulator system.

You can also trace backwards from the VCC or 5V pads to the regulators that would be somewhere between the pads and the "ignition" wire and/or battery + input to the system.

 

[dnmun]

if you measured the source/drain resistance and they are all around the same number then the S/D capacitor is shorted to start out imo.

so go through and measure resistance across the caps. if it changes under measurement that implies it is not shorted.

your mosfets may or may not be shorted but if the cap is shorted across the busses then it is impossible to make the measurements and you have to remove it first.

then you can use the diode tester between the phase wires and the red and black power leads.

remember that the mosfets are all arranged with the drain connected to the higher voltage side. B+ on the hiside mosfets and the phase wires on the loside mosfets. the mosfet should not conduct with the red probe of the diode tester on the high side but when you reverse it and put the red probe on the source leg or the phase wires for the hiside, then you should see the forward bias of the body diode. all should be identical in forward bias. identical.

 

[crea2k]

A good start would be to get a multimeter and test all of the burnt looking components, the capacitor values are on the side of the capacitors, so they are easy to test, the resistors, if im feeling lazy I will test the same one on the board if it has several and compare the resistance, or look up online the colour band code to see what it should be, r115 just looks to be a big power resistor. Its hard to tell by the photo the colours, but you can use this to calculate the ohms value it should be :
http://www.digikey.co.uk/en/resources/conversion-calculators/conversion-calculator-resistor-color-code-5-band

 

[crea2k]

If I had to hazard a guess id say the blue resistors are these, bearing in mind I cant properly see the colour band though on the photo.
http://www.sourcingmap.com/200-pcs-100-ohm-1w-watt-axial-lead-metal-film-resistors-p-161686.html

 

[dnmun]

i suspect the power resistor is not 1% and more than 100R too. but i think this guy dropped out already. no replies.

 

[dnmun]

bullshit. people learn how to fix stuff all the time. that is what this place is about. this guy just doesn't wanna fix it himself but for others they use the information we can provide to figure out the problem they are having.

 

[EdwardNY]

Sorry guys, it did seem a bit overwhelming so I decided to wait to see if Edward Lyen would fix it. He fixed this controller already when I first bought it for a good price and noticed he fixed someone else's controller who reversed polarity for a good price.

Anyway, Edward Lyen is not responding to any Emails so it looks like he isn't in the business anymore.

It is really bothering me that I did such a stupid thing to ruin a $200 dollar controller, so I'm going to go ahead and try and fix it. I hope you guys will still help me.

I found out information from another post about that blue resistor. Like crea2k said, I might as well replace what is burnt first. I want to try and get my +5v back before I go ahead an check other things.

This is what was said about the blue resistor.

I talked to lyen and got the correct value of resistor I needed (2W 200 ohm resistor, 72-100v a 300 ohm) but could not find that locally so I bought a 330ohm 5w resistor and replaced that 100v 470uf blown cap, hooked everything up and ....

I assume that I have two resistors as the 2nd resistor gets cancelled out when I connect the two wires for 48 volts. So for 72 volts I would use those two resistors for 200 ohms. Since I won't hook this up with 48 volts, I can just get a 200 ohm resistor.

If anyone is interested here is his post and it looks like he blew the same resistor and also the same cap. It is the same controller but could be a bit different.
http://endless-sphere.com/forums/viewtopic.php?p=574408#p574408

I believe that he replaced that resistor and cap and his controller worked. But he didn't reverse polarity. I also think his +5v worked.

Here is his damaged parts.

 

[dnmun]

why even worry about that? i am looking at the blown up S/D cap that i told you you had to cut out and you act like it doesn't even matter.

 

[EdwardNY]

If I had to hazard a guess id say the blue resistors are these, bearing in mind I cant properly see the colour band though on the photo.
http://www.sourcingmap.com/200-pcs-100-ohm-1w-watt-axial-lead-metal-film-resistors-p-161686.html

The color is Brown, Black, Black, Black, Brown.

I think this is the exact resistor.

If there are two 100 watt resistors at 1% tolerance.

How important is it that I match this?

Lyen mentioned that the resistor was 2W 200 ohm resistor, 72-100v a 300 ohm. But the guy replaced it with a 330ohm 5w resistor and it worked.

I'm just asking because the place I'm going to order the parts from doesn't have exactly the resistor I need. I don't want to spend $10 bucks for a large quantity also.

They do have a metal oxide 100 Ohm 1W resistor but it has 5% tolerance. Is that 5% tolerance important to match in this application?

Thanks.

 

[EdwardNY]

why even worry about that? i am looking at the blown up S/D cap that i told you you had to cut out and you act like it doesn't even matter.

Which cap is the S/D cap eactly?

Sorry, I just want to keep it simple at first and replace the obviously melted resistors/caps.

 

[dnmun]

the cap that has the top blown off there in your picture. you did not show this picture before so we could not see the blown cap where i thought you would find it. that resistor does not matter. if your mosfets have all blown up then it is never gonna work.

 

[EdwardNY]

the cap that has the top blown off there in your picture. you did not show this picture before so we could not see the blown cap where i thought you would find it. that resistor does not matter. if your mosfets have all blown up then it is never gonna work.

Oh, those caps weren't mine. That was from another post. But it has the same resistor melted and my cap on the first post also looks burnt.

I'm going to have to order a capacitor tester to test it.

He over volted his controller and that is why they blew. But his +5 volt worked.

I know checking the fets are tricky unless you remove them. But I'm more concerned about the +5 volts.

I'll try and check the fets resistance

 

[dnmun]

you do not have to order a capacitor tester. use the ohmmeter on your DVM like i said originally. you already said you measured 19 ohms. i asked if that was the resistance across the S/D cap in the middle of the busses. you even talked about a spot on it. where the plates shorted together inside. i explained that you have to remove it to measure the body diode of the mosfets. that only requires some dikes to cut the legs off.

 

[EdwardNY]

dnmun, I just wanted to say that I personally don't believe the fets or caps are shorted. I think the problem is with the +5 volts and maybe the +12 volts.

When I reversed polarity what I did was cut some 12 gauge stranded house wire and at the ends bent 1 single strand of copper. This is very tiny but not as small as hobby king stranded wire which is hair like. I touched the battery wire for a split second with this single tiny strand and saw a big spark and the single strand melted immediately (the single strand was just at the tips of the wire, so most of the wire was fully stranded).

I question how much flow could have went though like this and if it really would have enough power to melt the fets and caps. But I just don't know, maybe it does.

I'm going to do some more reading and checking. Lots to learn here. I checked the blue resistor with the ohm meter, it reads 100 Ohms for the left resistor and the right burnt resistor tests at 200 ohms. Both resistors test at 300 ohms. So they seem to be fine, but I'm not sure if they can be tested on board.

 

[crea2k]

If I had to hazard a guess id say the blue resistors are these, bearing in mind I cant properly see the colour band though on the photo.
http://www.sourcingmap.com/200-pcs-100-ohm-1w-watt-axial-lead-metal-film-resistors-p-161686.html

The color is Brown, Black, Black, Black, Brown.

I think this is the exact resistor.

If there are two 100 watt resistors at 1% tolerance.

How important is it that I match this?

Lyen mentioned that the resistor was 2W 200 ohm resistor, 72-100v a 300 ohm. But the guy replaced it with a 330ohm 5w resistor and it worked.

I'm just asking because the place I'm going to order the parts from doesn't have exactly the resistor I need. I don't want to spend $10 bucks for a large quantity also.

They do have a metal oxide 100 Ohm 1W resistor but it has 5% tolerance. Is that 5% tolerance important to match in this application?

Thanks.

If thats what the colour code is thats the value of the resistor on there, if you stick it into any resistor calculator on the internet it should give you the same value for that colour code, a lot of circuits will allow a certain flexibility with resistors, so very slightly different should be ok. If you unsolder any suspect parts and test them off board with a multimeter thats the most accurate way of doing it, as testing some in circuit can give strange results sometimes. The easiest way to remove big parts like that is just with a $5 hot air gun, just direct it onto the soldered side of the board and then just use some pliers to pull the component out, do this with the board on its side though, as things have a habit of falling off if its upside down. I use that method for harvesting components off boards, apart from I just turn them upside down and shake them and they mostly just fall off. Either that or just get a big soldering iron, Hobbyking sell one with 1/2" tip on it for next to nothing, you'll want something a bit smaller for smaller components though, but a bit iron is best for the components with a lot of solder, as the less time the components are subject to heat the better, and irons like that get solder hot super quick. As for testing the capacitors this can be done with a regular multimeter on the capacitor test mode, even the cheap testers have this usually.The resistors should read the value of the band out of circuit, the last band of the resistor is the percentage band, it gives the +/- allowable variance in the value, so if its a 100ohm resistor and the band is 1% it means the resistor can be 99ohm, 100ohm or 101ohm and is classed as being good, any other value and its on its way out or just damn right bad, if it reads nothing or open circuit its completely dead. The capacitor test is pretty simple, just put the multimeter in capacitor test mode, put the neg on the side of the capacitor with the dash marked on the capacitor and the pos on the other side and the value should match whatever is written on the capacitor, or there abouts depending on the % of the capacitor. When you replace the capacitors make sure that the legs go in the right holes, so the one marked with the dash - , needs to go back in that hole. Most resistors dont have polarity so can be put in any direction.

 

[crea2k]

To be honest I have a Fluke 17B and if I remove the capacitor from the circuit and use the capacitance test on the meter iv never had a problem with it, as it always seems to read the value correctly.

 

[crea2k]

You can always compare the value of a capacitor on the board to another though, so if they are marked up the same, but read completely different values, even if they values are incorrect would indicate a problem with one of them.

 

[dnmun]

why you people are talking about testing the capacitor for capacitance is beyond me. to me it is obvious that the capacitor is shorted. i was trying to get him to measure the short of the capacitor with the ohmmeter so he would realize it has to be removed. it has to be removed to measure the mosfets to see if the body diode of the mosfet is intact.

that voltage dropping power resistor just burned like that in normal use when he raised the voltage on the input because the controller's circuit current is now very high and the resistor was designed for a lower circuit current. the higher circuit current is due to the shorted parts.

it seems obvious if you are analyzing the expected failure sequence when the the polarity is reversed.

 

[EdwardNY]

If I try and measure any capacitor what happens is the resistance starts climbing. Sometimes it is falling. The value doesn't stay the same. But wouldn't I need to remove the capacitors?

The standard FET test through the Negative battery connector to phases are all 9.8. From the Positive battery connector it continues climbing from like 10 and up. This is the same thing that happens to all my good controllers also.

So for checking capacitors and fets I would assume I need to remove them? I don't see how they could be tested connected.

Something else, I decided to see if my controller was programmable. I think Lyen mentioned to check this out to see if the micro processor works and if it is worth fixing. So I hook up the USB connection to do this and it actually worked. I then decided to check the voltage of the throttle while connected to the USB and it was 4.6 volts, or the voltage from the USB connector.

I went to go check the halls voltage and notice it was very low and then the throttle dropped to around 1.3 volts.

Then I wasn't able to program the controller anymore. I also notice that when the USB is connected to my controller it only outputs 1.3 volts instead of the 4.6 volts.

Did I break my controller more? Why was I able to program it at first and the voltage was 4.6 volts, but now it is only 1.3 volts when connected to my controller. The USB adapter is 4.6 volts when disconnected from controller. For some reason the controller now lowers the voltage coming from the USB adapter.

 

[dnmun]

not sure what the standard FET test is. it should be open circuit, not 9.8, when the body diode is reverse biased. it should be about .4V when the body diode is forward biased.

if the resistance changes when you apply the ohmmeter probes, then the capacitor is not shorted.

 

[EdwardNY]

If the fets are not shorted, what would you do as the next step to check?

I located the 5 volt voltage regulator.

 

[dnmun]

measure the circuit current going through the input voltage drop down power resistors. it should be about 50mA.

the FETs are shorted. you already measured it.