Coil heating problem

[burner]

Hi,
I just finished building my first truepath. After some sticky muting issues it actually worked! At least almost

2nd channel seems to work, but the coil of the 1st channel heats up quite quickly. Almost too hot to be touched after a minute when no input or load. I have tested it briefly with some audio input with a speaker connected and it actually seems to work, but coil keeps heating intolerably hot.

I have tried to track the problem without success. The 2nd channel doesn't have any issues, but I can't find any differences i.e. when comparing voltages and resistances between the 1nd and 2nd channels.
My strongest theory would be that there is an strong high frequency AC voltage in FB1 net (which my multimeter can't measure). Otherwise there can't be current flowing trough the coil. There reason could be the feedback circuit (FBKOUT1/FBKGND1) but there is not any anomalies there either.

I have only a multimeter (no oscilloscope) to test with. 2nd channel has about 80mVac on its output on idle and so has the 1st channel. Is it normal? Has anyone else had a similar problem?

[isomebloke]

One of my early Trupaths had this problem - sounds fine, everything working fine but one inductor too hot too touch. Identical symptoms. My problem was that the litz wire still had enamel on it and the solder joint wasn't good. Cure - remove, soak off solder with solder wick, scrape gently and resolder.

Checking resistances once the components are soldered can give missleading results as you don't really know where you're measuring to and from.

Idle output should be adjusted closer to 0mV using the trim pots near the input end. Short the INPUT, not output and with multimeter across the output turn the trimpot to get as close to 0mV as possible. Repeat for other channel. Getting exactly 0mV will probably be impossible but you should be able to get within a few mV.

G'luck

[V-bro]

There is always some kind of load, even the zobel provides some kind of load.... It is well possible the coil has some malfunction. Even a cracked core can be problematic.

Try to get a replacement through 41hz....

Also do the feedback resistors on that channel get hot? If it is only the coil that pretty much confirms the theory.

[burner]

Thanks for the replies! Unfortunately I still haven't solved the problem
I did managed to tune both channels outputs to around 1-2mv on idle.
I'm not 100% sure (but 90% sure ), but I think first it was only the coil that was getting hot. Then I replaced the coil with standard 2301-V-RC coil, but it was getting hot also, so I don't think the coil is the actual root cause.

Now, after desoldering the fets and soldering then again those get hot too (again, not sure if those got hot also earlier). The R217 gets hot too and that didn't got hot earlier for sure. I was a bit amazed how it was possible, because there must have been current flowing from coil trough the R217 (where else it could go)!
Another thing I found was that the voltage over high side Rs was 2,2mV and 3,2mV on the lower side. heating doesn't occur when the amp is muted. I also found from Audiophools Fault finding guide that this occurs with broken fets.
So my current theory is that the fets were a bit broken in the first place and after desoldering them I destroyed then even more because of excess heat. I'll try to get another set of fets later. I'll report back once I have replaced them.

[isomebloke]

If you're replacing output FETs then you could do worse than;

http://www.maedos.com/Components.html

Not sure what the kit ships with these days but i think IRFP4228 is still best for Trupath.

[V-bro]

41hz only use IRFP4321, which is the same performance as 4228

[burner]

I finally got replacement fets, but the problem still remains. I also tried another truepath IC without luck. So now I'm truly puzzled with this. Any ideas?

[audiophool]

Please post some pics of your coils when fitted to the Truepath. Please also post some pics of the soldering above and below the PCB. Please also describe, and post photos of your general setup, for example, are you using test resistors?

Heat means current. Why not trace the current path? http://www.maedos.com/Guideline.html...on_trace_short

A defective MOSFET will not cause a coil to overheat when there is no load.

How sure are you that one coil is hotter than the other? In fact the coils do run fairly hot during normal operation. Do you have any way of measuring the temperatures? If so please post.

If I were fixing it, the first thing I would do is replace the coil with a nice neat non-litz one like in the top left corner of this pic and then see if this makes any difference.
DSC00842.jpg

I wonder if the hot mosfet is a secondary problem caused by poking around whilst trying to fix the first problem?

[burner]

Hi again,
Sorry for my extra slow progress on my truepath project. But I'm determined to make it to the finish line!

Please post some pics of your coils when fitted to the Truepath. Please also post some pics of the soldering above and below the PCB. Please also describe, and post photos of your general setup, for example, are you using test resistors?

Unfortunately I'll have only a crappy phone to take pictures with. I'll try to take some later on but actually I have made some progress with this problem...

Heat means current. Why not trace the current path? http://www.maedos.com/Guideline.html...on_trace_short

This indeed was a big mystery for me. Where the current goes if none of the other components gets hot. Then I realized that if the fets keep turning on and off at high frequency (>300kHz) creating -48/+48 voltage before the inductor and the inductor filters this out to a steady 0V, no current would flow trough the 20ohm resistor or the caps. The multimeter would read only 0V, because it cannot measure high frequencies. Or what do you think?

A defective MOSFET will not cause a coil to overheat when there is no load.

How sure are you that one coil is hotter than the other? In fact the coils do run fairly hot during normal operation. Do you have any way of measuring the temperatures? If so please post.

I don't have temperature readings, but the other coil got too hot to be touched, probably over 60 degrees.

If I were fixing it, the first thing I would do is replace the coil with a nice neat non-litz one like in the top left corner of this pic and then see if this makes any difference.
DSC00842.jpg

I wonder if the hot mosfet is a secondary problem caused by poking around whilst trying to fix the first problem?

Thanks for the input. It was really helpful. Previously I tried a a some off the shelf inductors ending up with the same problem. But finally I got it work by just unwinding one loop of the truepath inductor. So it turns out that inductor needs to be well in the specifications which was something I didn't think of (sound quality is another issue of course). It seems there was a shortcut in litz wire and unwinding a loop did the trick.

What I have now is the following:
*Fets don't get warm at all
*Inductors are only warm. And this seems to be normal.
*Multimeter reads 128 mVADC (and 0mVDC) in outputs. Normal?
*I have managed to get only 28 loops trough the inductors? And IMO I've done a good job wiring the inductors Is this enough and how many loop there should be? What toroids the trupath ships with nowadays?
*It's impossible to insert the litz wire trough the trough hole pad. I had to remove 4 strands for that. Do I have correct litz wire?
*I need another set of litz wires. Where could I get those?