A tube tracking console using ITB’s MkIII concepts

[rjuly]

*disclamer* a fair amount of the content of my posting will involve the work and ideas of Ian Thompson Bell. Any errors, distortions, omissions and misrepresentations are soley mine, and reflect in no way on him, nor represent his thoughts on any of these matters.

I’m going to try and build a tube tracking console based on the the work of Ian Thompson Bell (ruffrecords), and the design philosophy he sets out for his Mk III Vacuum Tube Mixer, which he sets out on his website:

Custom Tube Consoles

…and in the Blog for his MKIII tube console project:

http://mark3vtm.blogspot.com
Updates and discussion of the project can be found here on GroupDIY, on this thread:

https://groupdiy.com/threads/mark-3-vacuum-tube-mixer.55613/

I love tubes. I think that, well designed and implemented, they just sound better than more recent analog technology, and not just better, but different. Somehow they add clarity without harshness, sum magnificently and create a very coherent and open soundstage. This possibly due to the harmonic distortion they impart, but whatever the reasons, I knew that, finding the need for an eight track live recording console (which will get paired with a DSD back end), and having looked at Ian’s tube console designs, I had found a tube solution that would fit the bill - if I could manage to cobble it together…

Here are the design points I am borrowing (to put it charitably), from the MKIII Tube Console design:

- 35mm wide pluggable modules for channels and monitor sections.
- The use of aluminium ‘cheeks’ with aluminium extrusions connecting them to form the chassiss frame. Details of this may seen on Ian’s Blog here: You've Been Framed
- The module design of attaching the PCB to a steel panel, and to the control panel, to form a rigid module that also provides shielding between channels. Details of this may seen on Ian’s Blog here: Come Together
- The basic channel mic pre and monitor section designs, which hark back to Ian’s EZ tuber mixer designs, information about which may be found here: eztubemixer.blogspot.com
- The use of a backplane connector design into which channel and monitor modules may be plugged.

What I will be bringing to the party is, in fairness, very little. Essentially just a few ideas of how to adjust the phyiscal and logical topology of the design to suit my needs - which are:

- Eight mic pres with inserts, HPF and direct outs.
- The ability to mix the eight channels down to a stereo line output.
- Master bus inserts.
- Solo PFL with ZERO effect on the direct outputs or 2bus line out.
- No sends or auxes
- Stereo headphone output (provisonal - not sure it will fit)
- Stereo monitor outputs ( useful for basic mixing/ mastering of live material, since DSD can’t be mixed ITB )
- LED VU PPM meters (led because 35mm module width)
- A shallow(ish) chassis suiable for laying on a tabletop and hopefully transportable without a crane.

This is for me a very tall order, particularly since I am barely literate from an electronics perspective, and I certainly don’t understand most of what the designs imply, but that is the beauty of the situation; Ian has done the heavy lifting and is most graciously allow the likes of me to swoop in and take advantage of it, and I have the brains of this forum to pick for anything else. This process will be very educational and exhilarating and I will end up with a very useful and supelative ( if not heart-stopping costly ) piece of audio kit.

The circuit design will be 100% Ian’s, but it will be neccesary to create my own pcbs for the channel, master section and control panels, as I will need the physical topology to be different.

This will be a ‘portable’ console, so the form factor will be flat - meant to sit horizontally on a surface with the controls facing straight up, and we will see how shallow I can get it. Initally I am thinking of a chassis 100mm deep x 315 mm wide (nine 35mm modules, excluding cheeks and any wooden sides) x 300mm tall.

To kick things off, below is my first attempt at a block diagram which is adapted from Ian’s diagram entitled ‘EZTubeMixerSigFlowsm.jpeg’ in the EZTubeMixerDIY section found here:
https://drive.google.com/drive/fold...SN2hXdTg?resourcekey=0-FIqnPwoL8p2m-kLlHc7Cyg


Please feel free to poke large holes in it ;-)

- Richard

 

[ruffrecords]

The basic schematic looks OK to me. The only thing I would say is that BUS has only one S in it

Your biggest challenge I think will be mechanical. A 100mm depth for the chassis is ambitious especially with a horizontal surface. You will need to allow at least 20mm for the motherboard and the wiring behind it which really leaves on 80mm depth for you PCBs. This is only really deep enough for one tube laid horizontally. If you string the three of them out horisitally you will need to take care top mimimise high impedance track lengths between tubes. If you could make the PCB 100mm deep then you could get two tubes side by side. So you could have the two 6922 tubes with their sockets pointing to the back and the have the 7025 tube next to them with its socket pointing the other way. This could considerably simplify and shorten the tracking between the three tubes. But you do have 300mm to play with but does this include the fader????

Very interesting project.

Cheers

Ian

 

[rjuly]

The basic schematic looks OK to me. The only thing I would say is that BUS has only one S in it

Yes indeed - As I mentioned, I am barely literate, apparently not only electronically...

Your biggest challenge I think will be mechanical...

I imagine it will be among my biggest challenges... I also have a huge gap of understanding the underlying concepts to bridge as well. This is just the beginning - the chassis will be as deep as it needs to be, but I will work to make it as shallow as practicable. I would prefer to have the tubes standing upright in a row. I thought I would simply leave a gap in the panel controls where the EQ would normally go to allow the tubes to span from the bottom the pcb up to just under the panel. It would require me to have two pcbs for panel mount switches and knobs.

But you do have 300mm to play with but does this include the fader????

Faders will be (likely ALPS) pots so shouldn't need too much horizontal space. My thinking calls for the faders to control only the levels to the stereo mix outs.

 

[rjuly]

Here is a rough mockup of a possible module layout. I have increased the dimensions to 110mm depth and 311.15mm tall (7RU). Components are much smaller in my imagination!


Not a great deal of space given that I will need to find room for an input and output transformer, and two insert transformers, plus the rest of the circuit components.

- Richard

 

[ruffrecords]

The area of your proposed PCB is about 34 square centimetres. The area of the double Eurocard board I use is about 37 square centimetres so on that basis most of what you need should fit in. My current Mark 3 PCB has three transformers and you are proposing to use four so that will need extra space. However...

Let's have a look at your schematic. At the moment you are using half a TLA for the mic pre, then going into an insert, then through a HPF and into another hall a TLA and then to the fader, and direct out. The thing is the second half TLA is rather underused. It normally provides gain make up for a passive EQ and adds 6dB extra gain to compensate for the 6dB loss in the usual 2:1 step down output transformer/ You do not need any gain make up (the HPF is a cut only EQ) and your schematic shows a 1:1 output transformer so this half TLA basically needs no gain at all. So the second half TLA is just a buffer for the HPF.

So, if we discarded or could move the HPF, the second half TLA is not needed at all. Signal flow would be input, half TLA, output transformer, insert, insert return, direct out, return input transformer, fader, pan buses.

Cheers

Ian

 

[rjuly]

The area of your proposed PCB is about 34 square centimetres. The area of the double Eurocard board I use is about 37 square centimetres so on that basis most of what you need should fit in. My current Mark 3 PCB has three transformers and you are proposing to use four so that will need extra space. However...

Let's have a look at your schematic. At the moment you are using half a TLA for the mic pre, then going into an insert, then through a HPF and into another hall a TLA and then to the fader, and direct out. The thing is the second half TLA is rather underused. It normally provides gain make up for a passive EQ and adds 6dB extra gain to compensate for the 6dB loss in the usual 2:1 step down output transformer/ You do not need any gain make up (the HPF is a cut only EQ) and your schematic shows a 1:1 output transformer so this half TLA basically needs no gain at all. So the second half TLA is just a buffer for the HPF.

So, if we discarded or could move the HPF, the second half TLA is not needed at all. Signal flow would be input, half TLA, output transformer, insert, insert return, direct out, return input transformer, fader, pan buses.

Cheers

Ian

Having done a a rough physical mockup of the chassis (cardboard box) I think I can up the depth comfortably another 10mm up to 120mm.

I do definitely need the HPF, but moving it is no problem. If that opened the way to only needing one TLA per channel, would that mean that I could perhaps design the each module as a 70mm wide pair of mono channels that share the same set of dual triodes?

- Richard

 

[rjuly]

Another thing I am mulling over - my design has no fader for the direct out, only to the mix bus. Therefore my level to 'tape' is controlled only by the mic pre level and associated pad. I was thinking to have an adjustable pad - perhaps allowing me to choose one of three values (-7, -14, -21) this would give me a small amount of granularity in terms of my direct level.

- Richard

 

[ruffrecords]

Having done a a rough physical mockup of the chassis (cardboard box) I think I can up the depth comfortably another 10mm up to 120mm.

I do definitely need the HPF, but moving it is no problem. If that opened the way to only needing one TLA per channel, would that mean that I could perhaps design the each module as a 70mm wide pair of mono channels that share the same set of dual triodes?

- Richard

70mm 2 channel modules is exactly the route I went down in the early days of the Mark 3 design. This was driven mainly by the desire to create a Line In Line Our (LILO) mixer. Input via a transformer went straight to a fader then to a passive EQ, TLA, pan and to the buses. But since you do not need EQ it works also for a mic pre version which is what I do for the poor man's tube mixer.

Regarding a fader on the direct out this is easy to do with a pot across the transformer secondary. A four position switch version with 0,-7. -14 and -21dB should be straightforward.

Cheers

Ian

 

[rjuly]

...So, if we discarded or could move the HPF, the second half TLA is not needed at all. Signal flow would be input, half TLA, output transformer, insert, insert return, direct out, return input transformer, fader, pan buses.

Cheers

Ian

I assume that moving the HPF would mean to before the amp block. Would the overall result look something like this?:




- Richard

P.S. I had a read through youd discussion on insert topology - good stuff!

https://groupdiy.com/threads/balanced-or-unbalanced-insert-switching.69282/post-882534

 

[L´Andratté]

Hi!
A HP filter before the input transformer is not a bad idea, as the xfrmr would most easily overload with LF content, if you know or can find out the inductance of the primary you could possibly even realize a steeper than -6db/oct cutoff.
Sidenote: the polarity switch could be placed after the mic/line switch so you can use it on line signals too (for stuff like M/S-ing, etc.)

A lot of work and happiness awaits you!

 

[rjuly]

These are good tips - many thanks!
- Richard

 

[ruffrecords]

Your schematic is basically correct. As L'Andrette says it might be a good idea to have the phase switch last so it can act on both mic and line inputs. Note that the 2:1 output transformer has a 6dB loss not 0dB.

The 1:10 mic input transformer usually is fitted with a 150K load which is reflected to the primary as 1500 ohms. So for a HPF all you need do is insert the right value capacitors in the hot and cold leads. For a 100Hz turnover, the total capacitance needs to be about 1uF which means you put about 2uF in each leg. 2.2uF in each leg will give a 96Hz -3dB point and 4.7uF in each will give a -3dB point at 45Hz. 11uF in each leg will give a -3dB point at 212Hz.

Cheers

Ian

 

[rjuly]

View attachment 84060

So in looking at this, a couple of questions occur...

- I can see that the signal going to the direct out is balanced, either as it comes back from an insert, or as normalled through the insert. However the leg going to the PFL bus - does that not need to be unbalanced, hence I would connect only the positive signal going to the PFL bus?
- The 10K:10K transoformer - that's only feeding the mix bus. Is it onlyto buffer the signal, since it doesn't change the gain?

- Richard

 

[ruffrecords]

So in looking at this, a couple of questions occur...

- I can see that the signal going to the direct out is balanced, either as it comes back from an insert, or as normalled through the insert. However the leg going to the PFL bus - does that not need to be unbalanced, hence I would connect only the positive signal going to the PFL bus?

Actually your PFL pick off point is in the wrong place. It should be just before the fader which will be already unbalanced. You cannot pick off an unbalanced signal from one side of a balanced output. Contrary to popular beilef, a balanced signal exixts only between the hot and cold connections. Ground/Analogue 0V is not involved in any way. Electronically balanced outputs do happen to refer both the hot and the cold to 0V but this is not a requirement for a balanced system. For example you can connect a couple of transformer balanced output to a transformer balanced input miles away without taking a ground connection along the way.

- The 10K:10K transformer - that's only feeding the mix bus. Is it only to buffer the signal, since it doesn't change the gain?

- Richard

That is correct. It serves to unbalanced the input signal before sending to the fader and buses all of which are unbalanced.

Cheers

Ian

 

[rjuly]

Actually your PFL pick off point is in the wrong place. It should be just before the fader which will be already unbalanced. You cannot pick off an unbalanced signal from one side of a balanced output. Contrary to popular beilef, a balanced signal exixts only between the hot and cold connections. Ground/Analogue 0V is not involved in any way. Electronically balanced outputs do happen to refer both the hot and the cold to 0V but this is not a requirement for a balanced system. For example you can connect a couple of transformer balanced output to a transformer balanced input miles away without taking a ground connection along the way…

Yes that makes sense. Thanks for that information and for taking the trouble to explain it. I will be sure likewise to pay it forward…
- Richard

 

[Trouble Report]

So in looking at this, a couple of questions occur...

- I can see that the signal going to the direct out is balanced, either as it comes back from an insert, or as normalled through the insert. However the leg going to the PFL bus - does that not need to be unbalanced, hence I would connect only the positive signal going to the PFL bus?
- The 10K:10K transoformer - that's only feeding the mix bus. Is it onlyto buffer the signal, since it doesn't change the gain?

- Richard

I would assume that final 10K:10K transformer is unbalancing the signal on its way to the fader. Right at the secondary of that transformer would seem to be the place to pick off the PFL. Don't you think?

 

[rjuly]

I would assume that final 10K:10K transformer is unbalancing the signal on its way to the fader. Right at the secondary of that transformer would seem to be the place to pick off the PFL. Don't you think?

Yes, absolutely. This is much clearer now...
- Richard

 

[ruffrecords]

I would assume that final 10K:10K transformer is unbalancing the signal on its way to the fader. Right at the secondary of that transformer would seem to be the place to pick off the PFL. Don't you think?

That is often the case, especially is there is a channel mute/on switch between it and the fader.

Cheers

Ian

 

[rjuly]

Things are in train. I have order parts from SRS for prototyping the chassis. They should be along in another week or so. The nice thing about the method of chassis construction (extrusions between a pair of cheeks) is that if I need to change the chassis dimensions, it's only the cheeks that need to be redone. Once I have the chassis fabricated, I can start with the placement of the 12 module backplane, and start to get precise measurements for the modules and pcbs.

Here's a picuture of the backplane (which Ian very graciously made available from his stash - many thanks for that! ), with connector's added - I couldn't resist, even though I'm nowhere near being able to plug anything into them...



My plan is to start from the top down with the monitor section, and potentially a headphone amp (if it fits). I will be using Ian's Twin Line amp design as the basis for the main outs, which will be a pair of balanced lines. Since I would like to have inserts across these outputs, my first challemge is to determine where is the best place in the signal path to place them, and where to put the (I imagine at least) one pair of additional transformers.

Very exciting, not least fiscally!

- Richard

 

[ruffrecords]

Inserts on balanced outs are easy. No need for extra transformers you just go:

transformer secondary --> send jack -> half normalled to -> return jack --> output XLR

Cheers

Ian