Small valve mixing console

[C.B. - Boudio]

Having ANY DC current through transformer windings changes the game as well. Studer* recommend 'demagnetising' input transformers periodically to maintain the quoted distortion characteristics.
* And others, for which I vaguely remember a discussion I had with Mr Lundahl many years ago about the use of 'line level' transformers breing used as a mic input for galvanic isolation purposes (not specifically the sound).

Air gaps and such, isn't it? I'll try to avoid such things (in my schematic, you can see that there are capacitors insulating the transformers), so I don't really have to take those things into account.

 

[C.B. - Boudio]

I finally got to scan the updated schematic in.
I apologize for the not-too-great quality and writing.
View attachment 1713691488852.jpeg
View attachment 1713691532277.jpeg
Now I'm thinking... maybe I should replace the ECC88 in the output stage with two EF80s, which have quite a lot of current.
View attachment 1713691618842.jpeg
And the control layout.

I also designed some other things.

View attachment 1713691775821.jpeg
An 'universal' preamp - the anode resistors of the EF80 still have to be changed (suggestions?). An EC92 is half an ECC81/12AT7, for those who don't know the valve.

View attachment 1713691821470.jpeg
This is a SRPP amplifier, for headphones as well as monitor speakers or to drive heavy loads. It's in fact the output stage of any old Philips television's amplifier. It should put out up to 4-5W with very minimal distortion. The PL84 is the 300mA heater version of the EL86
View attachment 1713692028971.jpeg
View attachment 1713692548969.jpeg
This is my equalizer, just a simple passive parametric design, with more or less the mixer preamp for makeup gain. I am not totally sure about capacitor and pot and resistor values.
So the bands that don't get affected by the EQ, go through a 33k resistor, the cut frequencies through a 33k pot (with one end to ground) and a 33k resistor, and the bands that get 'boosted', go just through a 33k pot to the grid of the triode.
Can it work this way?

This should all be very modular, sort of Danner-style-ish modules in full (EQ and mixer modules) or 1/2 size height (pre and driver), with for the full sized ones a parallel printer port on the top and bottom sides to daisychain everything and the half-sized ones one at the back. The depth of the full-sized modules should then be the same as the height of the half-sized ones (+/- 11cm). This to be able to form a little 'train', for example an 'universal' preamp, a mixer channel, an EQ and the driver amplifier.

I haven't got a compressor yet, but I admire DaveP's one and maybe I'll clone it as a mono version in a full Boudio-module. Or are there still other suggestions for transformerless/SE vari-mu's?

I'm really looking forward to the feedback to improve these designs!

 

[C.B. - Boudio]

I have finally started on building the mixer! Most metalwork is done, as are the wooden side panels. It just needs to get painted.

I am now working on the innards of the first channel strip. The design has changed; I am still using a PF86, but after that, there is a PC92 (basically 1/2 ECC81) and then an EL95, both cathode followers. The PC92 cathode follower, Rk are 330 ohm and 8,2 Kohm, Rg is 2 Mohm bootstrapped, in order to drive the Aux Send and the EQ. The PL95, a small output penthode, uses the same resistor values, but the 8,2 K is replaced by 1,5 Kohm. The master fader that follows it, is 47Kohm
A line signal sounds quite great through this, with to me quite acceptable losses; I only need to figure out the right capacitor values for the EQ.
My aim is now to squeeze as much gain as possible to the PF86 (EF86 but with a 4,5V 300mA heater for use in TV sets).
Ra is now 560K; Rg2 is 2M2; Cg2 0,1uF; Rk 2K2; Ck is 100uF; and Rg1 is switchable between 15K and 10M, the latter to eliminate the need for a DI box.
The coupling capacitor is 22nF, and the gain pot, between said capacitor and Rg of the PC92, is 2M2.
The B+ supply is a little under 350V I think.
This still doesn't get enough out of the mic I'm testing with, an old Shure 565.

Is there something more I can do?

If needed, I'll upload a schematic and some pictures.

 

[ruffrecords]

I also designed some other things.

View attachment 127538
This is a SRPP amplifier, for headphones as well as monitor speakers or to drive heavy loads. It's in fact the output stage of any old Philips television's amplifier. It should put out up to 4-5W with very minimal distortion. The PL84 is the 300mA heater version of the EL86


I'm really looking forward to the feedback to improve these designs!

That is the first SRPP design I have seen using pentodes. But there is something funny going on around the screen grid of the top pentode which does not make sense to me. First off the 2K7 from the HT supply seems to be connected directly to 0V - is there supposed to be a capacitor in there? Also the output seems to come from both the grid and the screen grid of the top tube. I don't understand that. Is there an error in the schematic or is something very clever going on?

I once designed a version of my normal ECC88 based SRPP output stage using a pair EL84 tubes wired as triodes but I have never seen an SRPP using pentode wired pentodes. Do you have a link to a schematic of a typical Phillips TV that uses this as an audio output stage?

Cheers

Ian

 

[C.B. - Boudio]

That is the first SRPP design I have seen using pentodes. But there is something funny going on around the screen grid of the top pentode which does not make sense to me. First off the 2K7 from the HT supply seems to be connected directly to 0V - is there supposed to be a capacitor in there?

I must have made a mistake there copying the original (I redraw such schematics myself in order to better understand them). I've got to dig up the original source somewhere.

Also the output seems to come from both the grid and the screen grid of the top tube. I don't understand that. Is there an error in the schematic or is something very clever going on?

That isn't happening, is it? From the cathode and the screen grid, yes.

I once designed a version of my normal ECC88 based SRPP output stage using a pair EL84 tubes wired as triodes but I have never seen an SRPP using pentode wired pentodes. Do you have a link to a schematic of a typical Phillips TV that uses this as an audio output stage?

It was really quite common, in their radio's and early HiFi amplifiers as well, but getting deeper into the sixties, they started to 'cheat' and use a 10:1 OPT with a single EL84, whilst still using 800 ohm speakers, especially in lesser equipment.

https://elektrotanya.com/philips_19tx410a.pdf/download.html#dl

Here you can find another variation on it; it's not 100% the same I think.

https://elektrotanya.com/philips_17...onstruhe_michelangelo_sm.pdf/download.html#dl

This is an even more elaborate version, with a phase inverter.

There were also radio's that used it, with an EL84 and a UL84 or EL86s. Their top-of-the-range B8X75A, for example.


An update on the PF86 situation: the doubled Rk didn't get more gain out of it.
I measured a bit around and got this: the cathode is now at a bit over 3V, B+ is, with only these three tubes in place, a whopping 400V, the voltage drop over Ra of 560K is only about 50V, and Ug2 is a little over 100V, with Rg2 being 2M2.

 

[My3gger]

Is there something more I can do?

Yes, compare how EF86 sounds with NFB and without. In early 2000 i built several circuits out of the Mullard circuits for audio amplifiers and IO Audio's design with naked EF86 and ECC82, running EF86 at high gain. They were pretty good, although designs with NFB or triode wiring as seen on many American schematics sound more useful to me. You could also play with with EF86's Ua and G2 voltages to set required sound.
One thing to look at is 1:5 input tx ratio, isn't 1:7-10 better suited for tube inputs? The last time i checked UTM and Edcor didn't have "modern" 150:10k mic input, OEP had one in mumetal can with good performance (for ~100eur in the EU!).

 

[C.B. - Boudio]

Yes, compare how EF86 sounds with NFB and without. In early 2000 i built several circuits out of the Mullard circuits for audio amplifiers and IO Audio's design with naked EF86 and ECC82, running EF86 at high gain. They were pretty good, although designs with NFB or triode wiring as seen on many American schematics sound more useful to me. You could also play with with EF86's Ua and G2 voltages to set required sound.
One thing to look at is 1:5 input tx ratio, isn't 1:7-10 better suited for tube inputs? The last time i checked UTM and Edcor didn't have "modern" 150:10k mic input, OEP had one in mumetal can with good performance (for ~100eur in the EU!).

At the moment, I don't have any transformers (yet), nor will I in the near future. And I did away with the NFB, too.
My goal is now to squeeze as much gain as possible out of that PF86; there will be some amplification in the master sections (which aren't built yet), but I'd like to be able to use each channel as an independent preamp.

I tried now kind of a starved plate topology (Ua about 20V under operation) with Ra 1M2, and Rg2 4M7, and it was an improvement output-wise, but a bit too distorted. I'll be continuing in that direction, I think.

 

[ruffrecords]

I must have made a mistake there copying the original (I redraw such schematics myself in order to better understand them). I've got to dig up the original source somewhere.

Can you post the correct schematic or a link to it?

That isn't happening, is it? From the cathode and the screen grid, yes.

Oops, I meant screen grid. Still seems odd to me but there must be a good reason for it.

It was really quite common, in their radio's and early HiFi amplifiers as well, but getting deeper into the sixties, they started to 'cheat' and use a 10:1 OPT with a single EL84, whilst still using 800 ohm speakers, especially in lesser equipment.

https://elektrotanya.com/philips_19tx410a.pdf/download.html#dl

Here you can find another variation on it; it's not 100% the same I think.

https://elektrotanya.com/philips_17...onstruhe_michelangelo_sm.pdf/download.html#dl

This is an even more elaborate version, with a phase inverter.

Thanks for the links. very useful. There is definitely something special happening with the screen grid bias especially as in some case it involves feeding dc though the loudspeakers. Those TV guys were certainly innovative. I need to look at this in more detail. What a way so save on iron

There were also radio's that used it, with an EL84 and a UL84 or EL86s. Their top-of-the-range B8X75A, for example.

Thanks. I will look that up.

Cheers

Ian

 

[C.B. - Boudio]

Can you post the correct schematic or a link to it?

http://server.idemdito.org/blog/histo/radio/tube/tube-amp3d.htm

I think was in here, the second schematic.

http://server.idemdito.org/blog/histo/radio/tube/tube-amp3d0.htm

Here are the original MBLE schematics, where it was developed. All discussed very thorougly in Dutch, so I learned quite a bit from it; for you, the schematics will speak for themselves I think.

 

[C.B. - Boudio]

There is definitely something special happening with the screen grid bias especially as in some case it involves feeding dc though the loudspeakers.

According to the second link in my last message, it's only about 5mA, which they apparently could get away with. This would however help with the efficiency of the amplifier.

 

[ruffrecords]

http://server.idemdito.org/blog/histo/radio/tube/tube-amp3d.htm

That one is very interesting, The top pentode is triode wired.

I think was in here, the second schematic.

http://server.idemdito.org/blog/histo/radio/tube/tube-amp3d0.htm

Here are the original MBLE schematics, where it was developed. All discussed very thorougly in Dutch, so I learned quite a bit from it; for you, the schematics will speak for themselves I think.

Excellent stuff. Thank you very much.

Cheers

Ian

 

[C.B. - Boudio]

I think I will change the tube layout a bit: I will change the PC92 into a PCF802 (I happen to have one laying around), to have an extra stage of amplification.
The plan is now to use the two penthodes (PF86 and P(C)F802) in series, followed by a cathode follower (PC(F)802) driving aux send and the tone stack, and for output still the PL95 in cathode follower. Or is this new idea not recommendable at all?



Also, I play with the idea of making some kind of quadrant fader; I want to mod a high resistance pot, say 1M, to have a tap at 50k or so to decrease the rotating angle. I still have to measure, but it seems like the rotating angle can get small enough to simply connect a bar to the shaft and ultimately get a nice curve like the original. (... if this made any sense)

(edit: I added the schematic)

 

[C.B. - Boudio]

Any thoughts?

 

[ruffrecords]

Quadrant faders have been discussed here several times. Try searching.

Cheers

Ian

 

[My3gger]

Any thoughts?

Yes, it might need some NFB between stage 1 and 2, pentodes without it can have way too much distortion. That would require moving volume control somewhere else, i'm not sure what are two connections for 1M and 100nF right before the stage 3. The rest looks like creative use of existing tubes, good job.

 

[C.B. - Boudio]

Yes, it might need some NFB between stage 1 and 2, pentodes without it can have way too much distortion. That would require moving volume control somewhere else

Good idea! I'll see what I can do.

, i'm not sure what are two connections for 1M and 100nF right before the stage 3.

It's a simple line in, after the mic preamp; it doesn't get amplified (I may do away with it and just use a pad before the mic in, if there's a line level source).
That 100nF is just the g2 cap of the P(C)F802, so that has nothing to do really with that connection.

 

[My3gger]

It's a simple line in, after the mic preamp; it doesn't get amplified (I may do away with it and just use a pad before the mic in, if there's a line level source).
That 100nF is just the g2 cap of the P(C)F802, so that has nothing to do really with that connection.

Moving line in to the input like Ian and Gyraf did for some of their projects is probably better because of additional gain, some sound, etc. I believe you could move volume control before stage 3, this way it doesn't interfere with NFB. One more thing; aren't V2 and V3 PCF802, not P(C)F802? Maybe you tried to show this way it is pentode/triode in on bottle.

 

[C.B. - Boudio]

Moving line in to the input like Ian and Gyraf did for some of their projects is probably better because of additional gain, some sound, etc. I believe you could move volume control before stage 3, this way it doesn't interfere with NFB.

That's probably the best option, indeed.

One more thing; aren't V2 and V3 PCF802, not P(C)F802? Maybe you tried to show this way it is pentode/triode in on bottle.

Yes, the brackets are to make the distinction between the pentode on the one side, and the triode on the other.

 

[C.B. - Boudio]

Which method of NFB would be best?
Per valve from the anode to g1, or something more of a global NFB?

 

[ruffrecords]

Which method of NFB would be best?
Per valve from the anode to g1, or something more of a global NFB?me in both cases

Depends on what you are trying to achieve with it. Global NFB could allow you to vary the gain over a wide range whilst maintaining headroom but ensuring it is always stable is not easy. Local feedback, on the other hand, is relatively easy to stabilise but only gives you limited control.

The improvements in frequency response, distortion, PSSR etc are the same in both cases and only depend on the amount of NFB.

Cheers

Ian