Smoke. Need help understanding what went wrong.

[curriergroh]

I had my first setback today. One of four drives released its magic smoke during a test power-up of the control box. I don't understand what went wrong. I'm looking for help diagnosing the problem. I want to understand what went wrong before I move forward again.

Driver: G201X
Kelling Power Supply: 6520 (measures at 62.5v with no load)
BOB: PMDX-122

I have been measuring continuity and voltages at every step of the build. Going very slowly. I have been testing with one driver up to this point and have not had problems. The driver that blew is not one I have used yet.

A picture of the control box immediately after the smoke
I pulled the power plug as soon as I heard the pop and saw smoke.
The drive that blew is the one detached from the aluminum.
All of the other drives are fine.


Pulled the drive out and left the terminals connected. Photographed right away so that I could check to make sure connections were right. I was pretty frustrated at the time and didn't want to trust my memory.
Direction, Step, and Common 3 conductor shielded cable.

Power and motor windings.

Popped the cover off to see the damage.

The board has at least 5 spots that show damage. The row right under the 2 screws has most of the damage that I can see. What are these pieces?

I doubt this is related but I noticed that the screw on the left of the image is not tightened, and the connection to the aluminum back plate was not even because of it.


What are common causes of drivers blowing?
The only ones I know of are shorts or improper connections.
I couldn't see any shorts before the problem or afterwards.
Unless I'm blinded by rage I think my connections were in the right places and had solid contact.

What should I do before trying again?
How can I diagnose this without risking another driver?

Still having fun, just sad I lost a gecko. Bye little buddy.

[Mountaincraft]

Is it possible that screw was part of the grounding system, and could cause probs when not tight? Or perhaps the loose screw s allowed the board to swivel just enough that one of the solder puddles touched it?

Did you have any motors or exterior plugged in when it blew?

I'd call Gecko in the AM and get their input.. Who knows, maybe it's just a bad unit, and they'll send you a new one or repair that one for free...

You're making me glad I am springing the extra few bucks for the 203Vs....

What is the purpose of the ferrite rings? Reduce interference? Conductor heating?

What are the PVC adapters for? Air intakes?

[RomanLini]

I'm not an expert on that device but it looks like the power mosfets, their drivers and current sense resistors are badly cooked. The most likely cause would have been a short circuit of the motor wires.

The destruction would probably have been avoided if you had a fuse on the power input to the driver, I always use fuses between the PSU and drivers for exactly this reason.

It's also possible that it might have been caused by connecting the PSU voltage in reverse to that one driver, although that is something you would probably have noticed.

[bill south]

One nice thing about Gecko Drives is that Mariss will most likely fix it for you. Their support is second to none.
Also, he may help identify what was "most likely" wrong with the installation.
Are we having fun yet????
Chin up.
B.

[curriergroh]

4 motors were connected.
The G201x were running open as far as current, so 7A.
The motors are rated for 7A.

I took out all of the drives except one and powered up again. It ran fine. It was not in the same position as the bad drive.

The ferrite rings eat noise. I had some so I put them on for fun. Not sure if they are needed.

The PVC adapters are for other cables not yet configured. The limit switch circuit, estop, machine ground, and parallel cable.

The motor wires are still connected to the terminal. Is there a way to test if there is a short in them? Visually the connection looks clean at the gecko, could the short be somewhere else and do the same thing?

The power is connected correctly, continuity is ok, and there is no continuity between + and - in the cables. Just so I'm clear, if there was continuity between the + and - in the cable, is that what a short is?

If I want to put fuses in are they inline on the +? What amp rating, fast or slow? I have 7A motors.

I found the return form on the gecko site. If I use that will the drive get to Mariss?

Thanks for the help.

[poopy_pants]

I would double and triple check the wiring to the motors. It looks like a short. I had the same issue with one of the drives in my G540. It turned out to be a bad connection in one of the DB9 connectors used on that system. Trying to stuff too many wires and a resistor into the shell without shorts was challenge. I realize your connections are different but I would investigate that thoroughly. By the way, Gecko usually gives you one get out of jail free card so I would contact them even if you believe it could have been self inflicted.

[curriergroh]

Since I cut power so quickly when the problem happened I wasn't sure if the other drives were ok or not.

It was the Y-axis that went up in smoke.
I have reconnected the X, A, and Z drives one at a time and finally all together and they are all working.

I have inspected the 4 motor wires and plug and tested for continuity between wires, there is none, and along wires, they were good.

I'm debating whether or not to connect one of the remaining good drives to the y-axis to see what happens.

[Mariss Freimanis]

Best guess is reversed power supply polarity.

Reason: both H-bridges are blow (burned-open 0.04 Ohm resistors) and all 4 IR2104 half-bridge drivers are blown. Because each bridge operates independently, a common factor (power supply) had to be involved. The only common factor is the power supply. Both brown capacitors show some swelling which clinches the guess.

If you haven't returned a drive before, return the destroyed drive (yes, it's totally and completely destroyed) and mention on an included note "One time, stuff happens deal". We replace one drive for free no matter what was done to it. Run over it with a forklift, spill sulfuric acid over it or connect it to 220VAC and it will still be replaced for free (These things have actually happened). That also includes free return shipping.

Mariss

[bill south]

Mariss;
Without a doubt, the best service in the industry!

[curriergroh]

Wired up another working drive to the same location as the bad one. Works fine.

Seems clear, based on what Mariss said, that it was the power supply, but I still don't know how or what it did. Anyway, all positions are working now, just need to send in the dead gecko.

Thanks for the help.

If you want to follow my progress my build thread is here:
The Prototyper: 4x8, 8020, cncrouterparts

[keebler303]

Is it possible that the terminal blocks were pushed onto the board offset by 1 position? If the power connector was set over by 1 terminal it would connect DC+ to one of the motor terminals and DC- to the DC+ terminal. If that caused power to be routed back through the flyback diodes in the transistors it could cause a "reverse polarity" situation.

Did you take a picture of the terminal block before you pulled it off the drive? If you plugged them in after the drives were mounted, it seems it would be fairly easy to stick the terminal block on in the wrong position with the way you have them mounted.

Matt

[Mariss Freimanis]

A flaky terminal block wouldn't blow the bridges in the drive. It takes a lot of energy to do that.

The "off by one position" scenario has two problems that make it unlikely.
1) The cover of the drive doesn't permit placing the connector if it's off by one position.
2) The resulting catastrophe would be limited to only one bridge.

The MOSFETs are rated at 33A continuous at 100VDC. Peak single-pulse current is 150A rated per MOSFET. Only reverse polarity can blow out both bridges simultaneously because reverse polarity forces the MOSFET intrinsic drain to source diodes into continuous conduction. This over-current then blows the 0.04 Ohm resistors. When they go, the half-bridge driver outputs are no longer ground referenced and the 4 IR2104 drivers blow (witness the blast craters in each SOIC8). This is absolutely characteristic damage caused by reverse polarity.

Mariss