Solid State Power Amp Idle Current Creep - Hafler XL600

[firerunner]

Hi friends,

I have a Hafler XL600 that was purchased used and is exhibiting some strange behavior.

Initially the first issue was the fan kicks on immediately when powered on, even with no load. A few different owners have mentioned the fan should be quiet and relative to the listening volume, and this unit's fan seems to be running close to full speed, it's blowing the leaves on a plant across the room and quite loud compared to the listening volume.

I figured the idle current was possibly changed by the previous owner and triggering the fan prematurely, but upon inspection the bias pots are set at the factory (I assume) and glued in place.
When I put a meter on a rail of one of the audio boards, it starts around 350ma and creeps up over 400ma (factory spec) and continues to rise. I let it get to 450ma before powering down.

My first thought was to replace the transistors in the fan control circuit, but since these seem to be separate from the audio section, I feel there may be another issue that is causing the idle to increase.

The power transistors are mosfets, could they be exhibiting some kind of thermal runaway? The amp sounds good, no audible distortion or issues.

Attached is the schematic. Has anyone experienced an issue like this in their power amp? Any help or insight would be appreciated, thanks!

 

[JohnRoberts]

That schematic does not show the actual amplifier section.

The important metric is the power output devices class A current. IIRC (Lateral?) Mosfets are often biased on with more class A current than bipolar output devices. IIRC the lateral mosfets have a positive temperature coefficient starting around 100mA so don't typically suffer from thermal runaway.

Look for a small resistance resistor in series with each power device. Measure the voltage drop across that resistance (carefully) to compute class A bias current.

JR

 

[firerunner]

That schematic does not show the actual amplifier section.

The important metric is the power output devices class A current. IIRC (Lateral?) Mosfets are often biased on with more class A current than bipolar output devices. IIRC the lateral mosfets have a positive temperature coefficient starting around 100mA so don't typically suffer from thermal runaway.

Look for a small resistance resistor in series with each power device. Measure the voltage drop across that resistance (carefully) to compute class A bias current.

JR

Hey John! Thanks for the reply. It looks like the last page has the amp schematic, Q101-Q108 are the mosfets. I don't see any resistors on the drains or source but the fuse I pulled is in series between the power rail and the drain of four corresponding transistors in parallel, not sure if that is appropriate.

I was assuming the fan circuit is monitoring the power draw in some fashion and is turning on early due to some extra current draw, however if you think the mosfets wouldn't exhibit that behavior there could be something else going on.

I appreciate the help!

 

[Tubetec]

I have an XL600 also ,
The fan should start off at very low RPM initially and only speed up when the music is loud enough to mask it ,
If everything sounds good and power rails measure right I would avoid poking around with probes in the audio circuit .

I havent looked at the circuit in a while but I think a thermistor on the main heatsink controls the fans speed .

Thats the protection and fan drive circuitry below , R109 and R110 are 22k NTC thermistors , FM101 is the fan ,

 

[firerunner]

I have an XL600 also ,
The fan should start off at very low RPM initially and only speed up when the music is loud enough to mask it ,
If everything sounds good and power rails measure right I would avoid poking around with probes in the audio circuit .

I havent looked at the circuit in a while but I think a thermistor on the main heatsink controls the fans speed .

Thats the protection and fan drive circuitry below , R109 and R110 are 22k NTC thermistors , FM101 is the fan ,
View attachment 134272

Hey Tubetec that's great you have an XL600 to confirm the fan should be quiet, thanks for the info!

I'll see about replacing those thermistors first, I think I can kind of see what's going on there, looks like they trigger the base on Q203

I appreciate it!!

If the rails are to spec and the sound is good should not be too worried about the idle current? I've used the amp for a good while when I first got it and didn't experience any issues.

 

[JohnRoberts]

I found the power amp section. It looks like they do not use source degeneration resistors so forget about measuring current there.

There is a class A bias adjustment pot P1. I do not like the design topology, if the trimpot wiper fails open circuit, the amplifier goes to maximum class a bias current (not good). But let's not redesign the amp today. Adjust P1 for less bias current and see what happens.

The fan running fast may be correct if the amp is getting hot. I ASSume it is measuring temperature of the heatsink.

JR

 

[firerunner]

I found the power amp section. It looks like they do not use source degeneration resistors so forget about measuring current there.

There is a class A bias adjustment pot P1. I do not like the design topology, if the trimpot wiper fails open circuit, the amplifier goes to maximum class a bias current (not good). But let's not redesign the amp today. Adjust P1 for less bias current and see what happens.

The fan running fast may be correct if the amp is getting hot. I ASSume it is measuring temperature of the heatsink.

JR

Thanks for checking it out John. That's interesting about the possible issue with the trim pot! Sadly P1 on both boards have been epoxy'd and are not adjustable. The same red compound is on the various transistor heatsinks so I figure that was a factory move.

Oddly the fan hits high speed immediately with powering on the amp, even with no load on the outputs.
I'll track those thermistors down and see where they are mounted before I ASSume anything haha thanks!

 

[JohnRoberts]

Thanks for checking it out John. That's interesting about the possible issue with the trim pot! Sadly P1 on both boards have been epoxy'd and are not adjustable. The same red compound is on the various transistor heatsinks so I figure that was a factory move.

glyptal (?) or finger nail polish does not mean the trim pots are OK.

Oddly the fan hits high speed immediately with powering on the amp, even with no load on the outputs.

that is not a good symptom. If the temp sensing was accurate it would start up slowly.

I'll track those thermistors down and see where they are mounted before I ASSume anything haha thanks!

From a quick glance it looks like the thermistor's resistance gets lower as they heat up starving the base drive at Q203. When Q203 turns off, Q201, Q202 voltage rises and drives the fan harder.

If you want to suss it out carefully measure the voltages of all the sundry parts in that fan control circuit.

JR

 

[JohnRoberts]

FWIW that fan control circuit is a great candidate to use to develop troubleshooting basic discrete circuitry skills. Its made up of simple passive and active components with well understood operating characteristics. By probing around and looking at the circuit node voltages you can probably identify why the fan is coming on too soon. It could be something as simple as a bad solder joint.

JR

 

[Matt Syson]

I think I would expect the fan(s) to run fast initially when powered up as fans fed with 'limited' voltage/ current may fail to start at all after they have not been used for a while. The 'kick' at startup would be for a few seconds only and then drop to a pretty low speed which would then be controlled by the thermistors on the heatsinks. The important factor is HOW HOT the transistors are getting as running too hot will reduce lifespan. Are ALL transistors tightly bolted to the heatsink and are the thermistors properly mounted so receiving the 'accurate' temperature of the transistors. The thermal drift may be due to the bias setting transistor arrangement not correctly following the power transistor temperature but again what current fdoes it get to after say 20 minutes? Is the quiescent current correct after this time?

 

[firerunner]

glyptal (?) or finger nail polish does not mean the trim pots are OK.

that is not a good symptom. If the temp sensing was accurate it would start up slowly.

From a quick glance it looks like the thermistor's resistance gets lower as they heat up starving the base drive at Q203. When Q203 turns off, Q201, Q202 voltage rises and drives the fan harder.

If you want to suss it out carefully measure the voltages of all the sundry parts in that fan control circuit.

JR

Copy that, I may look to replace the trimmers as well so I can adjust in the future if necessary.

I checked the thermistor resistance in circuit and they are 18k~ and 19k~. Since the fan comes on immediately I figure these would be much lower if they were faulty. Will let it run for a little and see what they read.

FWIW that fan control circuit is a great candidate to use to develop troubleshooting basic discrete circuitry skills. Its made up of simple passive and active components with well understood operating characteristics. By probing around and looking at the circuit node voltages you can probably identify why the fan is coming on too soon. It could be something as simple as a bad solder joint.

JR

Sounds great, will check out all the nodes. I appreciate the analysis as well, seems like Q203 could be turning off prematurely.

I think I would expect the fan(s) to run fast initially when powered up as fans fed with 'limited' voltage/ current may fail to start at all after they have not been used for a while. The 'kick' at startup would be for a few seconds only and then drop to a pretty low speed which would then be controlled by the thermistors on the heatsinks. The important factor is HOW HOT the transistors are getting as running too hot will reduce lifespan. Are ALL transistors tightly bolted to the heatsink and are the thermistors properly mounted so receiving the 'accurate' temperature of the transistors. The thermal drift may be due to the bias setting transistor arrangement not correctly following the power transistor temperature but again what current fdoes it get to after say 20 minutes? Is the quiescent current correct after this time?

Hey Matt thanks for the info. The fan goes on high when the amp is powered on and stays high, even with no signal or load on the outputs. The current increase made me worry as well about the heat in the power transistors. They look bolted on quite well but I will recheck everything, good call on that. I will let the amp warm up and see if the current settles back down to a nominal level. Thank you!