Class D guitar amp - SMPS resistor overheating

[Terrymack]

Working on a Peavey Nashville 1000 solid state guitar amp.  A schematic is not available online and Peavey will not provide it.  I have traced the circuit by hand (see attached photo).  I’m guessing it is a 2-switch (DC-DC) Forward Converter.  It gets input from the guitar amps DC Power Supply.  The 50 ohm 5 watt resistor is overheating discoloring the circuit board.  Replaced both MOSFETs, the 50 ohm resistor, and the 270 pf capacitor.  All other caps check good using a Fluke 179 multimeter.

Any ideas or advice is greatly appreciated!  Thanks!

[Vicus]

Are you shure about the schematic?

[Terrymack]

While I can’t be 100% sure I have triple checked it visually and with an ohmmeter and I am reasonable sure it is correct.

[TimFox]

With the 270 pF and 50 ohm resistors both disconnected, is there any unwanted continuity on the board where the 270 pF is located?

[Terrymack]

I removed the 50 ohm resistor and the 270 of cap. There was no unwanted continuity.   All continuity made sense according to the schematic.   Thanks for the suggestion!

Maybe the DC input is the problem -  possibly too high?  I measured 157 VDC but without Peavey’s schematic I’m driving blind.   The power supply +/-15VDC is good so I am assuming the DC supply is working properly.

[Renate]

You're missing the source connection on the "Right" MOSFET to the minus supply.

[Terrymack]

Thanks!  Good catch!  Drawing updated.

[Renate]

What's on the secondary?
If you've got a blown diode and only harvesting half the energy it would go into the snubber.
Wouldn't it?

Check the junction of the two 1µF for about 1/2 157V = 78V, not more than 10V ripple.

[TheMG]

Do you know for sure that there is a problem? Have you compared to a known state and confirmed the resistor isn't supposed to get that hot normally?

It is quite common to find in many products some resistors that run rather hot, and over time will discolor the circuit board from the heat, but is actually normal operation.

How hot is hot? Have you measured the actual temperature of the resistor?

[Terrymack]

The secondary feeds a power diode rectifier circuit.  I’ll put it back together and check.  Thanks!

[Terrymack]

The resistor gets very hot.  I turned it off when it reached 120 deg F.   You could smell it.   I am going to assume that is not normal.   You have a valid point though.  Thanks for the comment.

[andy3055]

Did you check the 1uF caps?

[Terrymack]

Yes - with a Fluke 179 multimeter they both showed 1 uf .  I don’t have an ESR meter.  I may just go ahead and replace them in case one of them is leaking or failing under high voltage conditions.  Thanks for the comment!

[srb1954]

The resistor gets very hot.  I turned it off when it reached 120 deg F.   You could smell it.   I am going to assume that is not normal.   You have a valid point though.  Thanks for the comment.

120F (49C) is not excessively hot for a 5W power resistor. Wire-wound power resistors can commonly get up to 250C case temperature when operated at their rated dissipation. Even lower power resistors are specified at around 150C case temperature at their rated power.

You might want to check that your power FETs don't have spurious HF oscillations. This would cause excess current to flow through the 270pF cap thus dissipating extra power in the snubber resistor. The gate drive waveforms seem to have some HF oscillation on them so it would pay to have a closer look at that as well as the drain waveforms of the FETs.

Make sure your scope probe is good for the signal levels present on the FET drains. The maximum input voltage to the scope probe typically has to be derated for higher frequency signals to avoid damage to the probe and the oscilloscope input circuitry.

[Terrymack]

I found a bad crimp on the transformer secondary center connection.   Would that cause excessive current through the snubber?

[Terrymack]

Repaired suspect crimp on transformer secondary and powered back up. 50 ohm resistor got to 160 deg f (71 c) and rising when I turned it off.   

Measured with a scope at both 1 uf caps junction at the low side of transformer primary (See photo).  Not sure how to interpret it.   If you ignore the spike with HF oscillation the sine wave looks about 70 volts.

[Terrymack]

I think the transformer secondary bad crimp caused the snubber resistor to overheat and burn the board.  The resistor was mounted against the board.   Fixed the crimp and mounted the resistor up off the board.  50ohm resistor temp settled out at 110 c.   The amp still has no output so I need to keep investigating.

Thanks for all the comments and suggestions so far!

[TMM]

The "right" (lowside) MOSFET gate should not have 150vpp on it. I would guess it should have 20-30Vpp on it at most, centered on the negative power supply rail. Scope out your negative supply rail as well - probably there is a lot of ripple on it causing the source of the MOSFET to be swinging around wildly.

The "left" (highside) MOSFET is more difficult to measure since the MOSFET rides on the output node driving the output transformer, which itself may well be 150vpp at the switching frequency under normal operation. You really need a differential probe to measure Vgs of the highside MOSFET since neither gate or source are DC with respect to ground.

That said, apart from a possibly badly decoupled power supply I think it might be working normally. When you drive audio in you should see jitter on the waveforms. If you see nothing but a stable switching then probably the fault is in the input stage that pulse width modulates the switching signals

[Terrymack]

I’m really nervous about putting scope probes on these voltages.   You are right - I need a differential scope probe which I don’t have.

UPDATE:   I found a bad connection on connector from pre amp output to power board.  Cleaned with deoxit and reseated it and now I have my 1khz sine wave to the speaker!   And the snubber resistor is steady at 115 c.    Yay!   See photo of bad crimped wire on transformer!

Thanks to all for the help!!   Now I can practice my pedal steel guitar!

[MathWizard]

Assuming the 5W res. gets pretty hot anyways, what about just unsoldering it and mounting it on some metal bracket screwed inside the case somewhere? Or get 2x25R or 2x100R ?

[Terrymack]

Good suggestion.  I’m still not comfortable with that amount of heat but it is within spec.   I raised it off the board so it has air under it now.  I’ll keep an eye on it.

[TimFox]

Can you replace that device with a leaded power resistor that can be mounted above the board?  Heathkit used to specify that a power resistor should be temporarily spaced above the board by a small piece of cardboard from the shipping box while soldering, and then removed.  Ohmite used to sell higher-power WW resistors on a hollow core, with two mounting brackets that went into that core and allowed at least 0.5 in of clearance below the resistor.

[Terrymack]

I’ll look into that.  I mounted the resistor .25 in. above the board and ran it for an hour with no problem.   See photo.

[tooki]

I’m really nervous about putting scope probes on these voltages.   You are right - I need a differential scope probe which I don’t have.

UPDATE:   I found a bad connection on connector from pre amp output to power board.  Cleaned with deoxit and reseated it and now I have my 1khz sine wave to the speaker!   And the snubber resistor is steady at 115 c.    Yay!   See photo of bad crimped wire on transformer!

Thanks to all for the help!!   Now I can practice my pedal steel guitar!

Do my eyes deceive me, or are those transformer wires tinned and then crimped? I've always been told that's a big no-no, since the solder can creep with time. (And if they crimped it extra hard to try and prevent this, they could have caused a weak spot that would lead to the failure you discovered!)

[Terrymack]

Yes they are and I did not know that and what you say makes sense.  All the wires on this transformer and another inductor are all tinned and crimped during factory build!   I need to fix that.  Thanks!