how about post a pic of your burnt thingy. you have photos from the other day. maybe we can reconstruct what it was from before & after shots.

Great idea! See attached. I was surprised to see it in such clarity, but I'm still having problems with the colors. The resistor in question is parallel to the one with 3 orange stripes. The top band color looks maroon to me, but I'm guessing brown then black then brown? I can't find another like it on the board to comp[are, so this photo is probably the best we'll get. What do you think? If it's brn blk brn, would that be 100 ohms?

 

Best I could tell is that it was a 100ohm resistor that I smoked. I carefully removed all the hall sensor leads I had attached and checked the board over to clean it up as close to original as possible. I tried a 100 ohm resistor with fingers crossed, but nothing happened and I couldn't hand start the motor. In studying the circuit closer, I found a little transistor nearby that had a chunk out of the side missing. I replaced the transistor and the burnt resistor and tried again. The new 100 ohm resistor smoked immediately. The resistor smoking problem is now repeatable, but it appears something else is amiss in the circuit.

 

Back at it after a vacation...
I did another look and found that the mosfet I removed is not securely soldered. I'm going to resolder and test and also replace the 100ohm resistor that has smoked twice then try it again.

 

I replaced the mosfet and the 100 ohm resistor and tested. The good news is that the resistor no longer burns up. The bad news is that the motor won't turn even with help. Any ideas/suggestions on what to do next??? I'd like to try to isolate the problem, but not sure where to start.

 

What the board you have pictured above is a Brushless DC Motor Controller (BLDC).

It takes the incoming AC, converts it to DC, and the IC sends out a 3 phase square wave signal to the MOSFETs to energize the correct coils based on the feedback.

Do you have an image of board from the other direction?

With your scope, if it is floating ground, measure the signals at the gate of each MOSFET and compare them, same for drain and source. One will show something not right, so report the results.

Even though it is DC, it is still enough power to kill you, so BE CAREFUL! A meter would be a better choice to check the average voltage on each MOSFET pin.

 

I think you are on to something. Yesterday, I tested the diode function of the mosfets. I got strange results that I attributed to it being in circuit. I found that two of them ( a n and p channel pair) had different results so I replaced them and tried again. The motor didn't work.

I'm now thinking that the phase related to this mosfet pair is to blame somehow and your suggestion to measure voltages of all the mosfets sounds like a promising approach. However, I'm a bit concerned about the floating scope measurement, especially since I am such a newbie and since it is a USB scope and I don't have an isolation transformer. I don't want to fry my computer! Note that my scope does have two channels, so maybe I could do a differential test with the ground leads connected?

I'll first try with the voltmeter while awaiting more suggestions on how to scope it if that is a better way to go.

 

I used my DMM to measure voltage of the mosfet pins when AC supply was connected. I attached the meter's common lead to the base of the surge guard device that leads to the diode bridge after the AC input. Not exactly where this is, but it seems to provide a good reference.

I didn't see any mosfets with different measurements from the others (except for the two channel types).

Here's the data:
Mosfet Channel G D S
1 P -166 -166 -166
2 N -0.1 -166 0
3 P -166 -166 -166
4 N -0.1 -166 0
5 P -166 -166 -166
6 N -0.1 -166 0

 

Questions:
1) Why are all voltages negative?
2) Why are P channel gate pins all at -166VDC? Does that imply a short between the low voltage PWM circuitry and the high voltage motor side? Should I be using a different ground to test the gates?

 

By the way, yesterday, I also replaced the LM324 quad op amp IC. Unfortunately, there was no change in anything afterwards. I haven't been able to find the other IC which is a more likely culprit, the LSI-LS7262 - Brushless Motor Commutator.

 

By the way, yesterday, I also replaced the LM324 quad op amp IC. Unfortunately, there was no change in anything afterwards. I haven't been able to find the other IC which is a more likely culprit, the LSI-LS7262 - Brushless Motor Commutator.

Try here for the LS7262; http://www.anaheimautomation.com/products/ics/lsi-csi-item.php?sID=267&serID=5&pt=i&tID=160&cID=55

And here is data sheet; http://www.lsicsi.com/pdfs/LS7260_LS7262.pdf

 

I have been scratching my head about the high negative voltages, especially on the gates. I tested these with DC since it is after bridge rectification. I just went back and tested the AC voltage the same way and here is what I found:

Mosfet Type G D S
1 P 0.003 0.003 0.003
2 N 0.025 0.022 0.003
3 P 0.003 0.003 0.003
4 N 0.025 0.024 0.003
5 P 0.003 0.003 0.083
6 N 0.028 0.022 0.003

Note the aberration on the source measurement of MOSFET 5. Way higher than the rest. What does that mean?

 

Thanks for the link to Anaheim. I sent them a request last week but haven't heard back. I also ordered a sample from LSI. Free is always good...

 

anyone have an explanation for how DC voltages are so high (especially on the gates) and negative amplitude with low RMS AC voltages? What about the source voltage of one p channel MOSFET being over 25 times higher than the other p channel source pins? I am a bit confused on the AC and DC measurements within rectifier circuitry, so not sure what to do. In any event, seeing the high AC gate voltages suggests a short between the logic and power circuits? Seeing the source voltage higher for one mosfet suggests a problem in the bridge diodes, but they test fine. Any clues appreciated!

 

...One will show something not right, so report the results.

thatoneguy, I'm hoping you can make sense of the voltages I found from your suggested tests. Note that the table of data I provided is kind of hard to read. The first column is the position of each of the six MOSFETs, the second is channel type (P or N) then the voltage for Gate Drain and Source. It's a little jumbled together. I reported DC and then AC. The results really have me stumped...

 

Not quite sure what your doing here? Why would there be an AC voltage on the mosfets? They should be DC. The rectifier changes the AC to DC so anything after the rectifier is/should be DC.

 

Not quite sure what your doing here? Why would there be an AC voltage on the mosfets? They should be DC. The rectifier changes the AC to DC so anything after the rectifier is/should be DC.

My Fluke 111 has true RMS capability and 1% accuracy with 1mV resolution. I assume the AC voltage is a byproduct of the error and perhaps failure in the diode bridge? Since the AC voltage on the source pin of one MOSFET is so different from the others this may be a clue - regardless of the rationale of testing AC voltage following the rectification?Perhaps that circuitry is damaged? The diodes test okay. I'm at a loss as to what to do next? Any ideas appreciated.

 

In thinking more about the AC measurement question, perhaps the AC voltage I'm seeing is ripple? The capacitors should smooth this out, right? Perhaps the caps are bad? There are three each rated at 330uF and 200V (working voltage) wired in parallel, so I'm assuming total should be the sum of the three for 990uF. An in circuit test results in about ~1500uF. That look erroneous so I'll remove them to know more.

 

you're reading 28mV? I don't think that's anything to fret about. even the 83mV seems like it might not be a big deal.

 

you're reading 28mV? I don't think that's anything to fret about. even the 83mV seems like it might not be a big deal.

I totally agree. I was just thinking that the fact the voltages are all equivalent except for one that that could indicate a problem with the winding or circuitry associated with that particular mosfet. I still think it is a clue, but nothing strong enough to draw an obvious conclusion. I think I'll still test the caps because it's relatively easy to pull them and replace them.

 

Before I removed the caps to test, I looked to see if I could find replacements. Unfortunately, they are discontinued. It's a tight fit, so I don't want to risk it unless someone thinks the caps really need testing. I'm thinking this is unlikely given the high capacitance tested in circuit and since they are wired in parallel. In any event, why might an in circuit capacitance test of three 330uM caps in parallel show 1500uM?