leach amp inductor smoking

[Bosium]

hi there

I've just completed a dual mono leach amp, 750VA trans with separate bridge and 10000uf caps per channel. It sounds fantastic. I've been building it for ages and it seems tip top, BUT...

I was testing one channel with a 1khz sine wave input on a 8ohm dummy load and the resistor with the inductor wound around it started to smoke.

I removed power immediately, amp still 'works' and no fuses blew. What's up with that? I'm certain that I've assembled it correctly, both channels are biased at about 110ma and play audio perfectly into both 4 or 8 ohm loads.

Any help would be appreciated.
Gareth

[JonMarsh]

It could be indicative of oscillation, as little current should flow through the resistor except at very high frequency. OTOH, what gauge wire did you use for the inductor- it should be pretty healthy, like AWG14. or better.

[slowhands]

The resistor is what value (ohms, watts)? You need a 1 ohm 5-10 watter here. A more likely candidate to smoke is the Zobel net resistor, but at higher frequencies than 1 kHz. I suspect this inductor is not needed anyway, but people always seem to put them in.

[JonMarsh]

I suspect this inductor is not needed anyway, but people always seem to put them in.

For many/most bipolar output stages, even the classic EF triple T circuit (much less a CFP output stage!) both an RC zobel and a series parallel RL network are necessary to assure stability with reactive loads, inductive and capacitive.

There are ways around it, but I'm not telling....


The inductor does introduce frequency dependent load sensitivity, unfortunately- it often does have some kind of impact on HF sonics, depending on the speaker load. I wouldn't recommend removing the RL network from a standard Leach amp, unless you're just using it for subwoofer duties, for example.

The resistor in parllel is for damping the network; the inductor value will typically be in the range of 2-4 uH.

If you have a scope, you can determine the real problem pretty quickly.

It's brave souls who build and test amplifiers without a scope..

Best regards,

Jon

[slowhands]

Originally posted by JonMarsh


For many/most bipolar output stages, even the classic EF triple T circuit (much less a CFP output stage!) both an RC zobel and a series parallel RL network are necessary to assure stability with reactive loads, inductive and capacitive.

Jon

I am familiar with all the arguements for using the damped inductor to buffer capacitive loads. Most of the expert opinion says it is de riguer, with one notable exception (Pass). I realize suggesting it it not necessary is heretical.

The Leach amp uses feedforward compensation, from driver to inverting input. Furthermore, it has a low-pass filter on the output feedback, which makes it even more immune to hf instability with capacitive loads. In my experience with similar amps, if properly applied those two techniques keep an amp stable into capacitive loads.

Nelson Pass pointed out that the inductor degrades the damping factor at higher frequencies, and so in fact can impact tweeter damping. In light of that, I'd at least experiment with leaving it out.

[wrenchone]

One thing you might want to consider - if the Zobel inductor is the standard type with wire wound around a resistor, make sure it's sonnected well at both ends. Otherwise, you are pumping all the power through the resistor and it will smoke....

[Bosium]

thanks for the replies, wrenchone i think you have the answer, my inductor isn't well attached, i'll try that.

G

[peranders]

The inductor you have wound, what wire area did it have? Could it take the current?

[lucpes]

The thing here is that all the power should go through the inductor which should be quite large to avoid troubles. You may even solder the resistor underneath the PCB, it only provides oscillation dampening for the coil.

Bullet-proof sized inductor compared to the white 5W resistors:

[peranders]

I have made an inductor of 1.5 mm Cu wire, 5 mohms = much current before it glows.
Picture here