Small valve mixing console

[C.B. - Boudio]

Could it be a viable option to use a 2K2 wire wound pot (connected as a variable resistor) as the cathode resistor of the PF86 as a way of gain control?

 

[C.B. - Boudio]

Otherwise stated: what would be the biggest drawbacks of doing this?

 

[tk@halmi]

You will constantly change the DC operating point with volume changes. This is not desirable because it produces changes to headroom, symmetry and distortion with each setting. It might be a cool effect, but it will not make for a predictable recording channel.

I would err on the lower overall gain side for the channel. Today’s audio interfaces are very high resolution, and we also have lots of inline gain boosting solutions for quiet dynamics and ribbon mics.

 

[C.B. - Boudio]

You will constantly change the DC operating point with volume changes. This is not desirable because it produces changes to headroom, symmetry and distortion with each setting. It might be a cool effect, but it will not make for a predictable recording channel.

I see! Maybe I'll make the g2 decoupling variable, then?

I would err on the lower overall gain side for the channel. Today’s audio interfaces are very high resolution, and we also have lots of inline gain boosting solutions for quiet dynamics and ribbon mics.

The overall concept of this project is a 'prosumer mixer in the early sixties', and I want to keep everything in line with that; I'd even like to mod a Revox x36 to 4-channel (in the distant future). Which is why I really want quite a lot of gain, even at the cost of a pinch more hiss and/or half a percent more distortion (maybe I should have said that earlier...).
It's a bit lo-fi (I do hate that term, though) punk-ish whilst trying to get results as decent as possible, if that makes any sense (well, it probably doesn't and raises some hair and eyebrows with some of you ;-))

 

[tk@halmi]

Another option is to substitute the 15K transformer secondary loading resistor with a 25K audio pot in parallel with a 39K resistor. The capacitor would go to the wiper of the pot and into G1. You will need a 1M resistor or something of your choice to bias the EF86.

Input transformer saturation on transients sounds much nicer to me than oddly squashed tube signals.

 

[C.B. - Boudio]

Another option is to substitute the 15K transformer secondary loading resistor with a 25K audio pot in parallel with a 39K resistor. The capacitor would go to the wiper of the pot and into G1. You will need a 1M resistor or something of your choice to bias the EF86.

Sounds like a good idea!
But I'd like to have the impedance switchable, so 15k/10M (or something a little lower, I'll see what works best in the end), for high-impedance consumer valve equipment, and mainly to eliminate the need for a DI-box and to be able (not nescesserily par default) to get something like a fuzz bass just straight from the mixer. So that's why I am also considering other options.

 

[C.B. - Boudio]

This is the last revision of the design.


It's a bit messy, and the corrections and additions and some hastily erased things don't help it, but you'll get it, I think.

Any thoughts, great or small?

 

[ruffrecords]

The first stage will not work as drawn. The two switched resistors from grid to ground will be largely shorted out by the transformer secondary dc resistance.

Cheers

Ian

 

[C.B. - Boudio]

I thought that the 10M resistor would raise the impedance of the transformer quite a lot (let's say a 10k:10k, that's really just 1:1 but optimised for frequency response etc. at that impedance?), for use with high impedance sources, so that was the thought behind it.
Thanks to point it out!

 

[ruffrecords]

I thought that the 10M resistor would raise the impedance of the transformer quite a lot (let's say a 10k:10k, that's really just 1:1 but optimised for frequency response etc. at that impedance?), for use with high impedance sources, so that was the thought behind it.
Thanks to point it out!

Except it is also part of the negative feedback. The current circuit creates a virtual earth at the grid (not a very good one because of the low open loop gain). Even if the PF86 only has a gain of 30 times, then the 1M feedback resistor looks like 33K from grid to ground.

Cheers

Ian

 

[kags]

I have nothing useful to contribute to this amazing thread, but every time I see the topic, I get this picture of a "small valve mixer" in my head... maybe the valves should replace the faders? Pro Tip - Don't let GPT design your mixer!

 

[Brian Roth]

@kags LOL!

 

[Brian Roth]

@kags FWIW those are "officially" called sillcock valves by pro plumbers.

 

[C.B. - Boudio]

Except it is also part of the negative feedback. The current circuit creates a virtual earth at the grid (not a very good one because of the low open loop gain). Even if the PF86 only has a gain of 30 times, then the 1M feedback resistor looks like 33K from grid to ground.

Cheers

Ian

This made me think... could it be possible to inject the NFB at g3 (not coupled to the cathode of course)?

 

[ruffrecords]

This made me think... could it be possible to inject the NFB at g3 (not coupled to the cathode of course)?

I don't know about G3 but G3 has certainly been done before. Rupert Neve did it in his earliest tube mixers with an EF86 followed by half a 12AX7 cathode follower which fed G2.

Cheers

Ian

 

[C.B. - Boudio]

I don't know about G3 but G3 has certainly been done before. Rupert Neve did it in his earliest tube mixers with an EF86 followed by half a 12AX7 cathode follower which fed G2.

Cheers

Ian

That's interesting! It's in that schematic you posted last month in the BC-6B, isn't it?

I figured, g3 is almost always connected to the cathode, which is the same phase as g1, so injecting the opposite phase there could work. I'll try it.
I want to keep g2 as-is, because I've got that gain control over there, but if it doesn't work with g3, I might try his way.

 

[MidnightArrakis]

@kags FWIW those are "officially" called sillcock valves by pro plumbers.

View attachment 141007

>> This looks like it could be an -- LM19310 -- "regulator". (HINT: This is some rather weak humor be conjured up by what appears to be "LM19310" stamped on the side and referencing that to the old and standard National Semiconductor "LM7XXX" voltage regulators of days-gone-by). Oh, well.....at least I got a chuckle when I thought it up.

/

 

[ruffrecords]

That's interesting! It's in that schematic you posted last month in the BC-6B, isn't it?

I figured, g3 is almost always connected to the cathode, which is the same phase as g1, so injecting the opposite phase there could work. I'll try it.
I want to keep g2 as-is, because I've got that gain control over there, but if it doesn't work with g3, I might try his way.

G3 is what separates from a tetrode from a pentode. Its purpose is to catch any electrons which " bounce" off the plate after being accelerated by G2. If you don't connect it to the cathode or ground it can no longer do this so you may end up with a tetrode style 'kink' in the tube curves. G3 is a very coarse grid so its gm will be very low.

Cheers

Ian

 

[C.B. - Boudio]

Didn't think of that aspect. Now I'm even more curious as to how it'll work out, although there may be even more reasons why it wasn't really a thing.

In the end, I'll may have to just resort to doing away with the cathode bypass cap and inject the NFB over there, but nothing better than a little experimentation!

 

[ruffrecords]

Didn't think of that aspect. Now I'm even more curious as to how it'll work out, although there may be even more reasons why it wasn't really a thing.

In the end, I'll may have to just resort to doing away with the cathode bypass cap and inject the NFB over there, but nothing better than a little experimentation!

Definitely time to get the soldering iron out and see what's what.

Cheers

Ian