Need help with a Metcal MX500

[kff]

Trying to fix a Metcal MX500. Q4 dies as soon as the unit is switched on, with all three terminals shorting to each other (2-6 ohm resistance). I've double checked all capacitors, transistors, diodes, and now inductors on the board, and everything looks ok. All capacitors that could have any effect on Q4 have been checked out of circuit. With Q4 removed from the board, the voltages on Q4's gate and drain look ok (checked with an oscilloscope), nothing that would exceed the +/-20V gate/source and 200V drain/source ratings of IRF640N. Some questions:

1) In what scenario would an N-channel MOSFET fail in such a way that all terminals are shorted to each other? Overvoltage on the gate?
2) Is it common for capacitors to test ok at lower voltages (as tested with a multimeter), but fail at higher voltages? What would be a good way to test capacitors that normally operate at ~50V and 13.56Mhz?
3) What other components typically fail when Q4 dies?
4) Any other tips on how I should proceed?

Many thanks!

[zapta]

You do have the schematic and circuit description doc, right? 

Just checking.

[kff]

Yes, this one:

http://www.mikrocontroller.net/attachment/193474/MX-500P-11.pdf

[zapta]

Yes, this one:

http://www.mikrocontroller.net/attachment/193474/MX-500P-11.pdf

Yes, that one.

[BurningTantalum]

Overdissipation?

U4 can cause grief on this circuit, with 56V on the output if it fails (if my memory serves me...)

I also have some recollection of a strange heatsink clamp on this unit (not the metal bar that clamps the aluminium plate but the clamp on the device tabs. One of the transistors has a tin-plate clamp, different to the rest - also are the insulating washers fitted and correct?
Regards, BT

[BurningTantalum]

Just looked at the circuit -
Have you a known good wand connected to the unit?
With Q4 out of the circuit you should still get a green LED to indicate that the wand RF coil is present and intact, after cycling the power switch of course.
Is all OK on the other side of T1 ?

[kff]

Thanks BurningTantalum.

I do have a working handpiece, and the green light is indeed on even with Q4 removed from the circuit.

I am getting a reasonable-looking 13.56mhz sinewave at the gate of Q4. The voltage amplitude is a bit higher than it should be, but that is expected with Q4 removed, as there is no feedback going to U4.

Also, I do see 56V on the output of U4. But again this is what I would expect when there is no Q4 (U4 tries to boost output voltage until it sees 1.23V on the feedback pin, and without Q4, the feedback voltage never increases). I suppose I can desolder U4 and try it out on a breadboard with some resistors.

[kff]

Sorry, forgot to answer, the insulation seems ok everywhere.

[wraper]

Also, I do see 56V on the output of U4. But again this is what I would expect when there is no Q4 (U4 tries to boost output voltage until it sees 1.23V on the feedback pin, and without Q4, the feedback voltage never increases). I suppose I can desolder U4 and try it out on a breadboard with some resistors.

You can bodge one resistor to the schematic and feedback will be there. No need to desolder IC.

[kff]

You can bodge one resistor to the schematic and feedback will be there. No need to desolder IC.

Good advice. With a 4k resistor across pin 4 of U4 and the positive side of C8, I am seeing 2.4V, which looks right. So U4 appears to be ok.

Pretty stumped at this point. Perhaps I can keep the feedback resistor on U4 to limit the output voltage, solder another Q4, and do some probing if it doesn't die right away?

Other troubleshooting tips would be much appreciated!

[Corporate666]

I have an MX-500 with exactly the same problem and manifests exactly the same behavior as yours.

Stumped me too but I didn't have a lot of time to mess around with it.  Hope you will keep us updated if/when you figure it out. 

[kff]

For me, it was a short in T2 -- after desoldering it, the resistance between all 4 terminals is 0.3 ohms. Here is how I figured this out:

1) Made a poor man's signal generator by ordering an AD9850 module on Amazon (http://www.amazon.com/Huhushop-Active-Generator-0-40MHz-Equipment/dp/B00IR5I886/ , $9.99 shipped) and hooking it up to a TI Tiva launchpad board ($12 shipped).  Took only 30 minutes to figure out the protocol and have the board generate a 13.5Mhz square wave with +-2V amplitude.

2) Hooked up the generator across C26. The output at the coax connector was only +-0.3V, so there was leakage somewhere. Desoldered C25 to figure out whether the leakage was to the right or left of it -- it was on the transformer side (i.e. without C25, there was no leakage).

3) Checked capacitors C21 and C22 and they were fine. At this point it could only be T2.

Now, the question is how to replace T2. The schematics specify it as "transformer toroidal 9uH, 10.75 turns each side". Would any other 9uH transformer work here, as long as the number of turns is equal?

[wraper]

Now, the question is how to replace T2. The schematics specify it as "transformer toroidal 9uH, 10.75 turns each side". Would any other 9uH transformer work here, as long as the number of turns is equal?

I don't think so. Better to rewind it.

[bingo600]

Wasnt there some torroid Winding info in the 13.5 MHz metcal psu thread

/Bingo

[N8AUM]

Just a totally wild ass guess: Q4 goes into parasitic oscillation just long enough to wipe itself out ?

Good luck !!!!

[jbryan2]

Trying to fix a Metcal MX500. Q4 dies as soon as the unit is switched on, with all three terminals shorting to each other (2-6 ohm resistance). I've double checked all capacitors, transistors, diodes, and now inductors on the board, and everything looks ok. All capacitors that could have any effect on Q4 have been checked out of circuit. With Q4 removed from the board, the voltages on Q4's gate and drain look ok (checked with an oscilloscope), nothing that would exceed the +/-20V gate/source and 200V drain/source ratings of IRF640N. Some questions:

1) In what scenario would an N-channel MOSFET fail in such a way that all terminals are shorted to each other? Overvoltage on the gate?
2) Is it common for capacitors to test ok at lower voltages (as tested with a multimeter), but fail at higher voltages? What would be a good way to test capacitors that normally operate at ~50V and 13.56Mhz?
3) What other components typically fail when Q4 dies?
4) Any other tips on how I should proceed?

Many thanks!

I would suspect excessive current since it's a 100V FET.  Make sure Q4 is not shorted to the heatsink (look at the notes at the end of the parts list).  Pull T2 and check the inductances and that the windings are isolated.  Work downstream from there.

[kff]

After rewinding T2, there is still no heat produced even though the green light is on, but at least Q4 doesn't blow anymore. Something else must be broken somewhere.

One thing that I noticed is that my unit has IRF640N for Q4 instead of IRF530N. This is a newer model made by Oki, serial number > 120,000. Is it possible that Oki switched to IRF640N recently, or do I have the wrong MOSFET in my circuit? Another difference is that R10 is ~17K instead of 4.7K.

[wraper]

Probably they had reliability issues and switched to beefier mosfet. Unless someone repaired it before you, unlikely that it is wrong.

[jbryan2]

One thing that I noticed is that my unit has IRF640N for Q4 instead of IRF530N.

Big difference in the transfer characteristics.  Three amps versus 20 amps at Vgs=5V.

[kff]

Big difference in the transfer characteristics.  Three amps versus 20 amps at Vgs=5V.

Thanks for noticing this. IRF640N is probably wrong then, I will go ahead and order some IRF530Ns. The unit was repaired by someone else before me, and given that Q4 was blowing due to a short in T2, it is not surprising they decided to go for a higher rated transistor.

[kff]

Also, just curious, why does Q4 need to be double isolated from the heatsink? Is it to reduce capacitance between the gate and ground?

[wraper]

Also, just curious, why does Q4 need to be double isolated from the heatsink? Is it to reduce capacitance between the gate and ground?

Mounting tab is not gate but drain. Considering working frequency likely because of the capacitance.

[jbryan2]

Also, just curious, why does Q4 need to be double isolated from the heatsink? Is it to reduce capacitance between the gate and ground?

Source is tied directly to ground. If the heatsink is tied to ground you had better isolate the tab/drain.

[jbryan2]

Also, just curious, why does Q4 need to be double isolated from the heatsink? Is it to reduce capacitance between the gate and ground?

Oh, and the gate is tied to ground through L10.  If you don't isolate the tab/drain then all three pins are - glory be! - tied to ground.  It's not going to do much that way.

[wraper]

Also, just curious, why does Q4 need to be double isolated from the heatsink? Is it to reduce capacitance between the gate and ground?

Oh, and the gate is tied to ground through L10.  If you don't isolate the tab/drain then all three pins are - glory be! - tied to ground.  It's not going to do much that way.

As I understand he asked about why 2 layers of insulation, not just one.