The Workbench

Re: BM1000 FM detector circuit

Paul — Tue, 11 May 2010 09:30:28 GMT

Excellent!Thanks.

I have a scope.I think I will need to get a signal generator soon.I did move the pot as I was cleaning it,but before moving it I measured the resistance so I could return it to original.I will put the headphones on and try adjusting it.

Thanks again

Paul

Re: BM1000 FM detector circuit

tournedos — Tue, 11 May 2010 09:07:00 GMT

The purpose of the detector is to output a voltage that is proportional to the FM modulation, i.e how far the carrier is from the nominal frequency. It can go a certain value up and a certain value down, but outside these limits the signal starts to distort as the detector goes out of its linear operating range (the curve looks a bit like an S). This adjustment is there to set the operating point in the middle of the linear range, otherwise a strong modulation will clip at one end or the other, creating distortion. The effect can be similar to if you just tuned a little bit out.

If you have an FM signal generator - it doesn't need to supply a stereo signal for this adjustment - you could make it supply a strongly modulated signal, look at the detector output with an oscilloscope and adjust so that it doesn't clip at either end of the demodulated audio signal swing (or clips equally much at both ends).

Lacking these instruments, you'll still get a pretty good adjustment just by ear. Mark the original position of the trimpot if you haven't touched it yet. Tune to best possible signal, adjust the pot for best audio quality, maybe repeat a few times

Here's a pretty good writeup on FM detectors.

Re: BM1000 FM detector circuit

Paul — Tue, 11 May 2010 08:42:09 GMT

Thanks Mika

Makes sense.I didn't think of that.If the diodes and coils are of fairly loose tolerance then the pot can help compensate for this and a fixed resistor would be difficult or impossible to match.I assume the many circuits I have seen with fixed resistors are simply a compromise?

In the manual it simply says"adjust for best noise suppression"I assume this is to mean noise,as seen on a scope with proper signal injected,but not nessessarily obvious by ear?

My thoughts on the vastly improved stereo separation with the new caps were along those lines.Those tired,old tantalums were suppressing the 19k pilot tone,and a new pair of caps brought it out.Maybe someone can elaborate on this,my electronics is a bit rusty,haven't studied for over 25 years.

Paul

Re: BM1000 FM detector circuit

tournedos — Tue, 11 May 2010 08:23:21 GMT

I'd say keep the pot and just adjust it properly. I don't believe there's much of a chance of getting a symmetric detection curve with fixed components, no matter how carefully selected. There'll be enough of loose tolerance around, not least the germanium diodes and the coils which you can't help much.

I believe the "noise" that is being mentioned here is actually (phase?) distortion - apparently the faulty tantalum caps messed up the higher frequencies enough to partly kill the stereo MPX signal which rides on the 38 kHz subcarrier. Their purpose in a detector circuit like this is to filter out any remaining RF (IF, actually) energy.

BM1000 FM detector circuit

Paul — Tue, 11 May 2010 07:16:14 GMT

I just finished replacing all the electrolytic caps in my BM1000.The amp is sounding excellent with my CDX but the FM was a bit flat sounding with very little stereo separation.In fact when switching the mono switch,I actually had to listen carefully to hear the difference.That is until I changed the last two caps shown in the diagram.I replaced the two tantalums with carefully matched 4.7uf electrolytics(measured 4.8uf).These two caps made more difference than any others in the entire amp.The stereo separation and soundstage through FM are excellent now.

The question I have is this:What is the technical function of the 2k trimmer called "noise suppression"?I have not seen this in other circuits.Was this used to balance the detector should the components drift over time out of the close tolerance reqiured in this circuit?If so,if I carefully match to within <1% with modern components,can I simply replace the pot,and the 1.5k resistor, with an exact matched pair of 1.5k resistors?Perhaps Metal film or carbon film?

Thanks

Paul