Ferrite inductor puzzle

[DangerMan]

Originally Posted by mole42uk

I am building a valve stereo decoder for FM broadcasts based on the circuit shown here.

Hi Richard,

You've probably already sussed this but something looks a bit odd in that circuit.

The ECH84 heptodes actually have G2 and G4 connected internally whereas the diagram shows an external connection with the 38kHz signals fed in there (and then they get shorted together via the valves).
The circuit as it is drawn actually has two connections to G2/4 and none to G3.

It would make more sense if the 22K/2uF were to feed both valves' G2/4's and the 33pF/470k 38kHz feeds went to each valve's G3.

[mole42uk]

According to the Philips data sheet, G1 is pin 2, G2 & 4 are pin 7, G3 is pin 1 and G5 is on 3 with the cathode.

So, my reading of the circuit is that the 38kHz goes anti-phase to pin 7 on each valve, both G3's are connected to the 22k/2µF via pin 1 and the G1 is the signal input.

It looks as though the draughtsman made a couple of errors - the connection of G3 is one of them, the other is the "180Ω" resistor in the lower 38kHz filter.....

Richard

[Roy]

This circuit featured in Mullard Technical Communications Vol. 6 No. 59, with the same error in the drawing of the connections to the heptode grids. A correction was printed in No. 60.

38kHz switching signal should go to G3, not G2/G4. The attached PDF gives the full story.

Roy

[mole42uk]

Originally Posted by mole42uk

I think that, somewhere in here, lies the answer.

Finally resolved!

I took both cores apart to check, and I had both top halves on the 38kc/s coil, and both bottom halves on the 19kc/s coil. So - a large gap on the one, no gap at all on the other.

I had tuned the large gap one with extra capacitors, which I have now removed.

Both coils are now behaving at something quite close to the correct frequency!

Many thanks to everyone who helped.

Richard

[Radio Wrangler]

That seems to explain the whole thing, and it's a very easy confusion if you're not familiar with all the tricks of pot cores.

Getting two of the unground halves is a quick way to make a gapless transformer core when you need one in a hurry.

Gapless cores, just ground flat are often labelled "oL" where the Al value would have been.

Fitting an adjuster only increases inductance, and the increase depends on the gap size, so you need to dig into individual core data, and you need to back off the turns to get the adjustment range centred on the target value.

David