In memoriam LM1875T -- we had so lttle time

[Omega_Void]

I built me two LM1875 amps which have been going fine for a couple of weeks.

Today I built a third. Hooked him up to my junk-grade test speaker, switched on. Works, sounds fine, no problem.

So I hook him up to the woofer on a three-way speaker I made, with the other two amps on the other two speakers. Crossover is being provided by my PC. All three amps on the same PSU.

Sounds cool, everything's fine for an hour or so.

Suddenly -- the woofer jumps forwards and is straining to pop out the cabinet! That'll be DC on the woofer, no?

Naturally I switched off straight away. A visual check of the amp showed no signs of any shorted wires, no visible problem. The other two amps are fine.

What could cause this to occur? The only component difference in the latest amp is the electrolytics on the power pins. Could bad caps have murdered my amp like this?

[Sebastiaan]

Originally Posted by Omega_Void

I built me two LM1875 amps which have been going fine for a couple of weeks.

Today I built a third. Hooked him up to my junk-grade test speaker, switched on. Works, sounds fine, no problem.

So I hook him up to the woofer on a three-way speaker I made, with the other two amps on the other two speakers. Crossover is being provided by my PC. All three amps on the same PSU.

Sounds cool, everything's fine for an hour or so.

Suddenly -- the woofer jumps forwards and is straining to pop out the cabinet! That'll be DC on the woofer, no?

Naturally I switched off straight away. A visual check of the amp showed no signs of any shorted wires, no visible problem. The other two amps are fine.

What could cause this to occur? The only component difference in the latest amp is the electrolytics on the power pins. Could bad caps have murdered my amp like this?

My sincere condolence...

Can you provide a schematic of your amplifier?

1: Do you have an output zobel?
2: Do you use a output inductor/resistor in series?

Did you check the temperature from the chip time to time? And how hot was it?
Do you have a oscilloscope?


Big chance the cause of this sudden suicide is oscillation caused by the capacitive or inductive load from the cables in combination with the voice coil.

With kind regards,
Bas

[Tubbe]

Probably oscillation as Sebastiaan has pointed out.
We will have to wait for the autopsy .

Tubbe.

[Omega_Void]

Originally Posted by Sebastiaan

My sincere condolence...

Thank you -snif!- it's a difficult time...

Originally Posted by Sebastiaan

Can you provide a schematic of your amplifier?

Pretty much like the standard application in the datasheet.

2M2 across the input, 2u2 input coupling cap, 22k input resistor, 22k feedback resistor, 1k gain-setting resistor, 100uF electro decoupling the feedback network from ground.

100uF electro plus 100nF film cap on the supply pins.

Supply is ±27.5 volts off-load.

Originally Posted by Sebastiaan

1: Do you have an output zobel?

Yup, 1ohm 220nF in series, as per the datasheet

Originally Posted by Sebastiaan

2: Do you use a output inductor/resistor in series?

No.

Originally Posted by Sebastiaan

Did you check the temperature from the chip time to time? And how hot was it?
Do you have a oscilloscope?

Chip was not very hot at all. Lukewarm, really. I didn't get crank it up, partly 'cos the power transformer is a bit small for three LM1875 at once.

I don't have a scope.

Originally Posted by Sebastiaan

Big chance the cause of this sudden suicide is oscillation caused by the capacitive or inductive load from the cables in combination with the voice coil.

Now I wish I'd put the output RL on there. Would that kinda thing normally defeat the inbuilt protection?

Thanks

[Omega_Void]

Have checked the input cap, and that's not shorted.

[Omega_Void]

Good news is the woofer's not dead.

[Omega_Void]

Bad news is if the amp was salvageable before, it ain't now.

Decided to check what the output looked like without the speaker connected.

Instantly on switch-on there was the smelly smoke. Also a little bit of dark-coloured gunk on the +ve supply pin, right next to the package. Not quite sure where that got spat from, but wasn't there before.



I kinda figured if the original problem was oscillation, the amp would be okay off-load. Did I figure wrong?

[Sebastiaan]

Originally Posted by Omega_Void

Yup, 1ohm 220nF in series, as per the datasheet

Not that I want to argue with the masters from National semiconductors, but 1Ohm is a rather low value. What is the electrical serial resistance of your woofer voice coil? One of the benefit's of going active like you do, is that you can optimize the Zobel network for each driver.

So if u have the T/S parameters from each driver you can give a optimised Zobel for each driver.

The Zobel (Rz) resistor should be equal to the serial resistance of the voice-coil. Revc)

RZ = Revc

The (Cz) capacitor you can calculate by:

CZ = Levc / Revc2


Currently with only 1Ohm resistor the Zobel you have now has hardly any effect.

Second I think you need to isolate the amplifier with in series 10 ohm, and a inductor of around 12 turns (or around 0.7uH).

Next time, check the chips for heating up carefully, if so there is something wrong. Better is to use a scope and monitor the output.

With kind regards,
Bas

[Sebastiaan]

Originally Posted by Omega_Void

Bad news is if the amp was salvageable before, it ain't now.

Decided to check what the output looked like without the speaker connected.

Instantly on switch-on there was the smelly smoke. Also a little bit of dark-coloured gunk on the +ve supply pin, right next to the package. Not quite sure where that got spat from, but wasn't there before.



I kinda figured if the original problem was oscillation, the amp would be okay off-load. Did I figure wrong?

you mean after you replace the dead chip you power it on again? But did you check the other parts for damage? If not, big chance if the original problem was oscillation, the 220nF capacitor and/or the 1Ohm resistor are damaged, and provide now a short-cut to the output.

With kind regards,
Bas

[infinia]

Sorry about your loss

One simple thing to check that can fail sometimes, is the insulation from chip to heatsink. You can leave HS floating like I do, or grounded with insulator and grease.