Audio Compressor Question

[Aub]

Hi all,
Not sure which section to put this in so please move as you see fit, mods.

I've built the compressor, the diagram of which is shown in the photo. It is connected between the microphone amplifier and my Amateur Radio transmitter.

The controls don't seem to do much, so I'm wondering about the circuit. The second triode of the first ECC83 has nothing connected to the grid. Do you think this is a mistake, or is the valve acting as a diode?

Many thanks

Aub

[peterpiper]

Hi Aub, an odd circuit indeed. I wonder if the pin 2 grid should be connected to the pin 7 grid, Dont think it would cause any damage to try it. It wont act like a diode with the OC grid, maybe with grid and anode connected, Pins 1 and 7.

Peter

[Aub]

Thanks Peter. I did wonder if the two grids should be connected together. That also makes me wonder if the two grids of the second ECC83 should also be connected together. Time to get the soldering iron hot.

Cheers
Aub

[Electronpusher0]

I would say the the 2nd half of the first valve was intended to be a diode.
The passive circuit above the valve one uses diodes.

The first half of the first valve is a cathode follower, the (ac signal) output is rectified by the second half and a dc voltage developed on the 1M pot and 0.1uF cap. The greater the signal the lower the voltage.
This dc voltage is connected to the grid of the first half of the second valve which is also a cathode follower and acts as an impedance changer.
The dc voltage on the cathode is then applied to the grid of the second half of the second valve which acts as a variable resistor.
The signal is thus attenuated by a potential divider formed of a 470K (the one above the ECC83 label of the first valve, in series with the signal) and the anode resistor of the second half of the second valve (another 470K) in parallel with the variable resistance of the valve.

I would guess that the 1M pot sets the response time and the 25K pot the compression level.

Peter

[Aub]

Thanks Peter, that's very helpful. You are correct about the functions of the pots.

Cheers

Aub

[m0cemdave]

I would have expected the grid to be connected to the anode for best results as a signal diode.

[thermionic]

It indicates that both the grids are connected together, pins 2 & 7 should be linked.

Cheers. SimonT.

[Electronpusher0]

Originally Posted by Electronpusher0

The greater the signal the lower the voltage.


Peter

I meant of course the greater the voltage, I mixed up electron flow and current - oops.

Peter

[Herald1360]

Originally Posted by thermionic

It indicates that both the grids are connected together, pins 2 & 7 should be linked.

Cheers. SimonT.

Does it, though? The symbol for the other double triode shows the same "continuous" grid across the middle but has explicit connections to the different pins shown.

I wonder if the SPICE models for valves can cope with niceties such as shorting electrodes to convert triodes to diodes and tetrodes/pentodes to triodes?

[thermionic]

Originally Posted by Herald1360

Does it, though? The symbol for the other double triode shows the same "continuous" grid across the middle but has explicit connections to the different pins shown.......

The other double triode also has 2 PIN number identifiers (2&7) whereas the first double triode has a single identifier.

I’ve seen this drawn on other circuit diagrams like this in the past, a spice simulation would be interesting 👍

[Radio Wrangler]

My reading...

The first triode is a cathode follower driven from the input signal divided by two by that pair of 470k resistors.

The second triode is just used as a detector diode with its grid floating and just flapping in the breeze of electrons going past. More usually the grid would be strapped to the anode.

THe .01uF and the 1 meg pot adjust the decay timeconstant for compression.

The third triode is another cathode follower, buffering the compression timeconstant from the fourth triode which is going to be driven into a little grid current.

The fourth triode is where the fun happens.

It's anode gets some positive bias through a 470k resistor
It's grid gets a little but variable positive bias from the third triode.
The cathode is hard-grounded.

So now look at the ECC83 triode characteristics.

Va versus Ia is sloped (a lot more than a pentode would be) because of the field from the anode adding to the field from the grid (there is no screening of the anode) So it behaves as a resistor. THe slope of the curve at the anode voltage chosen tells you the resistance.

Now look at the curves for different grid voltagesm, and you see the slope differs quite a bit.

So the grid volts on the fourth triode controls that triode as a variable resistor - seen at its anode. So the anode impedance acts with the signal source impedance (the horizontal 470k in parallel with the 4th anode bias 470k) an a simple attenuator.

Will it work?

Yes. But it does need quite large signals.

Is it a good circuit?

Yes, not bad but.... It depends on the impedance it is loaded by. As a general circuit, an extra stage of yet another cathode follower would buffer it from the load.


David

[Aub]

Thanks again for all the suggestions and explanations. I'll have another look at it. I may have the wrong type of pots, i.e. log instead of lin, which might be making things seem to be not working correctly.

The microphone pre-amp, which feeds into the compressor, needs to be turned up to max to drive it, compared with when the compressor is not in circuit, so that bears out what David was saying. The compressor feeds into the microphone input on my homebrew AM 80 metre transmitter. I was hoping to get that "punchy" sound that i've heard compressors produce, thus improving the talk power, but the controls don't seem to do much at present.

All the best

Aub

[Radio Wrangler]

THe best punchy sound comes from what's called an RF clipper.

Audio is balanced modulated to make DSB, then filtered to SSB. Then it gets clipped and filtered again, finally it gets demodulated with the same carrier it got modulated with.

Sounds complex, but the compression happens and the harmonics it generates fall outside the bandwidth of the second filter.

So you get punch and a fairly clean sounding signal and you can drive your transmitter to high levels of modulation knowing the audio does the limiting before you overmod.

The SSB need only be at a few tens of kHz. I've seen one elegant design using a phasing type SSB generator using a polyphase network as the audio phase shifter.

A number of HF amateur radio rigs had RF clippers built in and used their SSB system as part of it even when you were eventually producing AM or FM.

David

[Aub]

Ok, thanks David. I always wondered why they are called RF clippers, when they are in the audio chain 🙂. I remember the Datong ones were popular in the seventies.

Cheers

Aub

[Herald1360]

Shades of the CD "Loudness Wars" here

[Radio Wrangler]

It's been going on on radio long before the CD.
Shouting for DX on the amateur bands and pop compressed to within 0.5 of a dB of its life.

It's a brutal world out there, and shouty people seem to rise to the top.

David

[Aub]

I have quite a dull and quiet voice, so need all the help I can get 😄😄

Aub