Broadcast mixer from 1952

[Backtomad]

My first post here so greetings to you all from Finland!

I have just got my dirty hands on something quite spectacular: A Finnish Broadcasting Company Designed Tube Mixer model 652.



What a beauty! 6ch of pure 50's tube magic! What I do know is that it was built for broadcasting the year 1952 Helsinki olympic games, it has 6ch (2 inputs per channel) and 3 mono outputs. And oh yes, it works 100%.

Anyway, I'm a total noob about electronics, I just record and mix. But I have a good friend that knows a lot about these kinda things but doesn't know much about mic preamps as he has been more into tube guitar and bass amps, old radio's etc.

So here's the thing, I want this mixer to work as a 6ch tube mic preamp with 6 direct outs for tracking. Maybe it could also be expanded somehow into 6:2 and used as a cool stereo summing unit? But that comes later. First thing to do here is the direct out thing.

Here is the schematics that my friend draw. What do you guys think is the best way to do it? Can you help us out here?

 

[emrr]

You'd need to 6 preamp output transformers and connectors, unless you can deal with Hi-Z outputs. Then it's still 6 low gain single tube with no gain control. It is what it is, use it like it is. Build some other preamps from scratch instead. I'm doubting there's room for changes, and you'd essentially destroy it's historic and still totally useful value trying to make it do much more. Pretend it's an Altec 1567A, but a better one.

 

[Script]

Build some other preamps from scratch instead.

+1

That is a truly beautiful piece of gear. Inspiring form factor.

I envy you ! If you don't like it for what it is, please send it over. I promise I'd maintain and keep as is til 2052 and longer...

 

[NOON]

I'd restore it as is, 6:3 is a fairly useful setup.

 

[ruffrecords]

The big problem with retrofitting direct outs to a mixer like this is that the mic pre outputs are quite high impedance and you probably need low impedance ones. This means you need to add some kind of buffer to each output. I can see three ways of doing this.

1. Add a step down transformer. In order to not load the mic pre stage too much this would probably need a ratio of about 10:1 which would lose you 20dB of gain giving you a maximum gain of about 34dB (assumimg the input transformer is a 1:10 step up)
2. Add a high voltage FET follower. This solves the impedance and signal level problems so you get the full 54dB or so of mic pre gain but you have an unbalanced output. You could easily make this impedance balanced or if there is room add a 1:1 output transformer. The problem is you now have semiconductors in the signal path.
3. Re-purpose some of the existing tubes to create cathode follower buffers for the mic pre outputs. This avoids using semiconductors but would lose some of the authenticity of the mixer. You would need to carefully choose which tubes to repurpose. For example, two triodes looks to be used as an oscillator and cathode follower buffer. If you can do without the oscillator function this could provide two direct out buffers. The mixer also appears to have three programme amps with transformer balanced outputs. YOu need two at most so you could consider repurposing the tubes in one of them as direct out buffers.

Edit: When I was working I visited Finland many times, mostly to Helsinki to see a company called Wartsila. I have the greatest respect for Finnish airline pilots who can land a plane safely in unimaginable cross winds.

Cheers

Ian

 

[Backtomad]

You'd need to 6 preamp output transformers and connectors, unless you can deal with Hi-Z outputs. Then it's still 6 low gain single tube with no gain control. It is what it is, use it like it is. Build some other preamps from scratch instead. I'm doubting there's room for changes, and you'd essentially destroy it's historic and still totally useful value trying to make it do much more. Pretend it's an Altec 1567A, but a better one.

+1

That is a truly beautiful piece of gear. Inspiring form factor.

I envy you ! If you don't like it for what it is, please send it over. I promise I'd maintain and keep as is til 2052 and longer...

+1

That is a truly beautiful piece of gear. Inspiring form factor.

I envy you ! If you don't like it for what it is, please send it over. I promise I'd maintain and keep as is til 2052 and longer...

Thank you all for your answers. But don't worry, every change amd mod we are doing to it will be done with utter respect to this great piece of broadcasting history so everything can be reverse engineered back to normal again in the future if we ever want.

 

[Backtomad]

The big problem with retrofitting direct outs to a mixer like this is that the mic pre outputs are quite high impedance and you probably need low impedance ones. This means you need to add some kind of buffer to each output. I can see three ways of doing this.

1. Add a step down transformer. In order to not load the mic pre stage too much this would probably need a ratio of about 10:1 which would lose you 20dB of gain giving you a maximum gain of about 34dB (assumimg the input transformer is a 1:10 step up)
2. Add a high voltage FET follower. This solves the impedance and signal level problems so you get the full 54dB or so of mic pre gain but you have an unbalanced output. You could easily make this impedance balanced or if there is room add a 1:1 output transformer. The problem is you now have semiconductors in the signal path.
3. Re-purpose some of the existing tubes to create cathode follower buffers for the mic pre outputs. This avoids using semiconductors but would lose some of the authenticity of the mixer. You would need to carefully choose which tubes to repurpose. For example, two triodes looks to be used as an oscillator and cathode follower buffer. If you can do without the oscillator function this could provide two direct out buffers. The mixer also appears to have three programme amps with transformer balanced outputs. YOu need two at most so you could consider repurposing the tubes in one of them as direct out buffers.

Edit: When I was working I visited Finland many times, mostly to Helsinki to see a company called Wartsila. I have the greatest respect for Finnish airline pilots who can land a plane safely in unimaginable cross winds.

Cheers

Ian

Thank you for your very detailed answer and respects, Ian Yeah, my sisters husband is a commersial airline pilot and yes those guys are like superhumans

The input Jørgen Schou transformers are indeed 1:10. You may have missed it (or maybe it's not in that drawing? Don't know, I suck at reading them) but there is already -20db pad per channel so that should handle the too much load thing?

We want to keep the signal chain as original and tube only as possible so FET is out of the question if possible.

We also thought that we should maybe use the existing other tubes for the buffering so thank you for your assuring comment as now we know we are on the right track! We were also worried if the one tube per channel would had enough power to drive the whole channel but it seems not to be a problem right?

Oh and we do have 6 Lundahl LL2801 1:1 600ohm output trannies ready for this project so with them we get nice insulation for the channels, right ohm's and balanced output's too.

Do you have any ideas how to make the semi-stereo outputs? I was thinking the easy way could be that if the first 1 to 3 channels would be L and the 4 to 6 into R? Of course adding pan pot into each ch would be more useful thing to have..

 

[Script]

How does the unit sound and perform (measure-wise wrt hum, THD etc) stock?

 

[Backtomad]

How does the unit sound stock?

Very nice indeed. I tested it with 2 different mic's quite briefly but the sound is clean, warm and nice. No humming and gain pots and everything is totally scratch free. Unfortunatelly I wasn't able to record those tests and now the main amp is out already and so on.

But like I told you guys, all the original parts are kept safe so we can always reverse it back to it's original working form if we want to.

Oh and thats not all. I also have another old Finnish Broadcasting Company 4ch tube mixer too. Don't know what we are gonna do with it yet but if the first one turns out to be great then it's next on the table. This one is from the year 1957

 

[Disco Volante]

Nice chickenheads. Great FINd!

 

[ruffrecords]

Thank you for your very detailed answer and respects, Ian Yeah, my sisters husband is a commersial airline pilot and yes those guys are like superhumans

So true. I rememeber landing at Helsinki one February in a raging snowstorm. As we approached, the plane must have been flying at 45 degrees to the runway. At the last moment he whipped it straight and slapped it down on the runway. The entire plane erupted in cheers and applause.

The input Jørgen Schou transformers are indeed 1:10. You may have missed it (or maybe it's not in that drawing? Don't know, I suck at reading them) but there is already -20db pad per channel so that should handle the too much load thing?

Yes, I had noticed the input transformers and guessed they were most likely 1:10 step up types. This transformer gives you 20dB of gain. The first tube, the preamp, has switched gain. In one position it is about 14dB more gain for a total of 34dB. In the other position that gain increases to 34dB for a total of 54dB. If you were to then add a 10:1 output transformer for a direct out you would drop the gain by 20dB (effectively negating the gain from the input transformer) but if you use a cathode follower and 1:1 transformer you do not lose any gain.

We want to keep the signal chain as original and tube only as possible so FET is out of the question if possible.

Understood!

We also thought that we should maybe use the existing other tubes for the buffering so thank you for your assuring comment as now we know we are on the right track! We were also worried if the one tube per channel would had enough power to drive the whole channel but it seems not to be a problem right?

That will depend on the tube characteristics. Unfortunately I cannot see what triodes are used on the schematic you posted. The thing to remember is that the drive capability of a cathode follower is mostly determined by the quiescent plate current which in turn depends mostly on the tube type. If the double triodes in your mixer happen to be ECC40 then they can be run at up to 6mA quiescent which should be enough to drive +10dBu into a 600 ohm load. However, if you can guarantee not to load the outputs with less than 10K we could safely run it at a lower quiescent.

Oh and we do have 6 Lundahl LL2801 1:1 600ohm output trannies ready for this project so with them we get nice insulation for the channels, right ohm's and balanced output's too.

Lundahl does not have any data for the LL2801 other than to say it is a 600 ohm 1:1 line out transformer. Unfortunately this tells us very little about what kind of drive circuit it requires. Fortunately they say the core is radiometal which should give low distortion and high inductance.

Do you have any ideas how to make the semi-stereo outputs? I was thinking the easy way could be that if the first 1 to 3 channels would be L and the 4 to 6 into R? Of course adding pan pot into each ch would be more useful thing to have..

Whichever route you decide to take you will need to create another mix amp (EF40(7) on your schematic. There is one towards the bottom right of your schematic that is already connected to a progam amp. Then there is the question of routing. You could simply set it up as a a pair of 3 into1 mixers which would be relatively straight forward and not alter the load on the mic pre outputs. As soon as you think about adding pan controls you have to consider the additional load on the mic pres. Probably the easiest pan is a simple Left-Centre-Right switch routes the signal to either the left, the right or both buses.

Cheers

Ian

 

[Ike Zimbel]

Very nice indeed. I tested it with 2 different mic's quite briefly but the sound is clean, warm and nice. No humming and gain pots and everything is totally scratch free. Unfortunatelly I wasn't able to record those tests and now the main amp is out already and so on.

But like I told you guys, all the original parts are kept safe so we can always reverse it back to it's original working form if we want to.

Oh and thats not all. I also have another old Finnish Broadcasting Company 4ch tube mixer too. Don't know what we are gonna do with it yet but if the first one turns out to be great then it's next on the table. This one is from the year 1957View attachment 142389

Maybe think about using the 4-channel as the 2-mix stage for the 6-channel...

 

[Backtomad]

So true. I rememeber landing at Helsinki one February in a raging snowstorm. As we approached, the plane must have been flying at 45 degrees to the runway. At the last moment he whipped it straight and slapped it down on the runway. The entire plane erupted in cheers and applause.

Yes, I had noticed the input transformers and guessed they were most likely 1:10 step up types. This transformer gives you 20dB of gain. The first tube, the preamp, has switched gain. In one position it is about 14dB more gain for a total of 34dB. In the other position that gain increases to 34dB for a total of 54dB. If you were to then add a 10:1 output transformer for a direct out you would drop the gain by 20dB (effectively negating the gain from the input transformer) but if you use a cathode follower and 1:1 transformer you do not lose any gain.

Understood!

That will depend on the tube characteristics. Unfortunately I cannot see what triodes are used on the schematic you posted. The thing to remember is that the drive capability of a cathode follower is mostly determined by the quiescent plate current which in turn depends mostly on the tube type. If the double triodes in your mixer happen to be ECC40 then they can be run at up to 6mA quiescent which should be enough to drive +10dBu into a 600 ohm load. However, if you can guarantee not to load the outputs with less than 10K we could safely run it at a lower quiescent.

Lundahl does not have any data for the LL2801 other than to say it is a 600 ohm 1:1 line out transformer. Unfortunately this tells us very little about what kind of drive circuit it requires. Fortunately they say the core is radiometal which should give low distortion and high inductance.

Whichever route you decide to take you will need to create another mix amp (EF40(7) on your schematic. There is one towards the bottom right of your schematic that is already connected to a progam amp. Then there is the question of routing. You could simply set it up as a a pair of 3 into1 mixers which would be relatively straight forward and not alter the load on the mic pre outputs. As soon as you think about adding pan controls you have to consider the additional load on the mic pres. Probably the easiest pan is a simple Left-Centre-Right switch routes the signal to either the left, the right or both buses.

Cheers

Ian

This just in: the tubes in it are all EF40 so they are pentodes.

 

[Backtomad]

Maybe think about using the 4-channel as the 2-mix stage for the 6-channel...

Now thats a great and clever idea actually! Thank you!

 

[emrr]

Looking again, we have 3 of the same output amps. I'd be inclined to make them be 3 independent high gain pre. I'd be running the EF40's as triodes to knock gain back to a more reasonable input stage level. That's the easiest path with least destruction. There's no reasonable path to stereo that doesn't involve drilling a lot of regrettable holes as the starting point.

 

[ruffrecords]

This just in: the tubes in it are all EF40 so they are pentodes.

That is puzzling because the schematic definitely shows some double triode tubes. Can you ask the guy who created the schematic what he thinks they should be?

Cheers

Ian

 

[Somkereki]

I have more than 10 Siemens - Halske microphone preamplifiers from 1963. They even have an input for ribbon microphones. The most amazing thing is their signal-to-noise ratio! 132 dB!!! Today's semiconductor amplifiers do not achieve this. They even thought about the linear phase response, there is a capacitor of a few picofarads, with which you can equalize. And their internal structure is symmetrical up to the output transformer. I plan to reduce the noise even further. The only flaw was the increased leakage current of the capacitors coupling from the anodes to the next grid. I replaced them with excellent polypropylene capacitors. I tested them in my friend's recording studio, they are better than any (most expensive) modern ones. They will have the best place there! (I just listen to music on a (Hungarian) stereo amplifier manufactured in 1963. Its output transformer has a 16 square centimeter iron core cross-section, with primary and secondary windings divided into many layers. It is only 2 x 12 W, but it is enough for me to listen to music. Of course, their distortion spectrum image contains almost exclusively even harmonics. The input transformer has multi-layer shielding of Permalloy - felt - Permalloy. I don't remember how many layers. They knew how to build exceptionally good devices in the past. I wish everyone a happy new year!

 

[ruffrecords]

I have more than 10 Siemens - Halske microphone preamplifiers from 1963. They even have an input for ribbon microphones. The most amazing thing is their signal-to-noise ratio! 132 dB!!!

I find this very hard to believe. 132dB signal to noise without qualification as to the gain, the bandwidth and the source impedance is meaningless. Can you post the actual Siemens spec for these?

I just found the spec for this amplifier on line. Its EIN for Input 1 ( a 50 ohm input, probably the ribbon one you refer to) is -132dBu which is pretty average for a 50 ohm input. On the other hand, the quoted figure for Input 2 (a 600 ohm input) is only -123dB. Bearing in mind this is a 15KHz bandwidth measurement it is pretty mediocre.

And don't forget this is NOT a signal to noise measurement, it is an equivalent input noise. So when the gain is set to 60dB, the noise at the output will be 60dB higher at -63dBu.

Cheers

Ian

 

[Backtomad]

Happy new year to everyone!

That is puzzling because the schematic definitely shows some double triode tubes. Can you ask the guy who created the schematic what he thinks they should be?

Cheers

Ian

Yeah I will ask him.

I did run some tests today. First with 2 different mic's and noticed instantly that yes the overall sound and tone is good but it lacks the lower register. So I hooked up a signal generator and spectrum analyzer and hit it with pink noise. And that confirmed my observations of the lack of the lower end.


So what can be causing this? Can it be some component in that signal chain on input? I also thought that maybe it has something to do with input impedance being too low?

Any ideas?

 

[ruffrecords]

The low frequency drop of could be caused by a number of things, probably a combination of them. The first thing I notice is that the 20nF output capacitor after the first EF40 had a load of about 100K or less. This CR has a turnover frequency of 80Hz so that will be 12dB down at 20Hz. If you change that to 220nF the turnover frequency will drop to well below 10Hz (all the other CRs turnover around 7Hz).

The second most likely culprit is the input transformer. Do you know the output impedance of your signal generator? This will have a large effect on the measured low frequency response. The other thing is we do not know what impedance microphones the input expects. If it is 50 ohms then using a modern 200 ohm mic will definitely loss low frequencies. The only way to fid out for certain is to measure the primary inductance of the input transformer at 100Hz.

Cheers

ian