Another germanium preamplifier of the past - the SR Hh63

[Conviction]

Hi everyone!

I hope you're doing well in these weird and uncertain times.

I recently had some time to go through my storage and found this little piece:

It's a germanium preamplifier (of sorts), made in-house at the Swedish Radio in the early 1960’s. The output is designed to go into another preamplifier (the manual glued inside the box says the Lyrec AR2 and some others), so the sole purpose of the unit, which goes by the name Hh63, is to extend the number of available microphone inputs. Max output is 30dB into 200 ohms according to a data sheet I found.

I wonder if I can make this into something useful, say, four stand-alone germanium preamplifiers. The input iron is custom, but probably of great quality. The designer - I always forget his name - made a lot of great sounding pieces back then.

Thing is - without making any claims since I don't fully understanding the circuit - I doubt there’s gain enough. It seems, but correct me if I’m wrong, that the unit merely acts as a “buffer” for the microphones, which explains the low output? I tried to take the signal from after the 10k 50k pot, and it works, but requires quite the gain in the DAW.

I then tried a Lundahl 1:1,14 output transformer, and it helped a little, but still quite useless.

Any ideas? Here's the schematic. Sorry, it's a crappy scan from god knows where:

(original here: https://imgur.com/cqqcOSg)


... And here's the spec sheet:

 

[Bo Deadly]

I like it! That thing is sweet. Unfortunately it looks like the gain stage has multiple feedback paths so it's not obvious as to where you could tweak one resistor. Maybe if you strap another resistor across R113 you would get some boost. But it seems like C101 is competing for feedback.

Make sure your mic wires are correct. At least the pin numbers are obviously a little different from todays conventions of 2/3 for signal and 1 for ground. It looks like they're using pin 2 for ground but I suspect the connectors aren't XLR anyway and you already know this.

 

[ruffrecords]

It is quite clear from the schematic that the output at each of the 1K pots is -20dBm. These each feed an emitter follower which is coupled to the emitter followers of the other stages. This then presumably goes to an output stage with some more gain which is not shown on the schematic you posted.

Cheers

Ian

 

[Conviction]

I like it! That thing is sweet. Unfortunately it looks like the gain stage has multiple feedback paths so it's not obvious as to where you could tweak one resistor. Maybe if you strap another resistor across R113 you would get some boost. But it seems like C101 is competing for feedback.

Make sure your mic wires are correct. At least the pin numbers are obviously a little different from todays conventions of 2/3 for signal and 1 for ground. It looks like they're using pin 2 for ground but I suspect the connectors aren't XLR anyway and you already know this.

Thanks squarewave! I kinda like it too, but as-is it's rather useless (for me anyway). I rarely need a 4 to 1 mic mixer.
Yes, there's several feedback paths, which makes me dizzy. I'll try R113!

A friend of mine modified his unit by strapping output transformers directly after the pots. Not sure of the ratio, but I would guess 1:10. He says the guy who bought it is pleased, and gets "enough gain", but I believe he still needs a 20-30dB boost in the DAW.

I'm aware, but thanks anyway! It looks like the unit originally had those huge Cannon jacks, but was modified to XLR:s. I found a small sticker on it, saying it was "modified and phased".

I'll get back shortly.

 

[Conviction]

It is quite clear from the schematic that the output at each of the 1K pots is -20dBm. These each feed an emitter follower which is coupled to the emitter followers of the other stages. This then presumably goes to an output stage with some more gain which is not shown on the schematic you posted.

Cheers

Ian

Thanks Ian!

Yes, it's quite clear! I however was not clear enough: I'm hoping to modify it, but I really need assistance. Squarewave pointed out that there's several feedback paths and that it might work to squeeze a little more out of each preamplifier section. Adjusting R113 (I tried with a 1k pot) introduced some distortion.

What's happening around R136?, the 1.5k resistor? Care to explain? I'm not a solid state guy, sorry.

Indeed there is an output stage before the 1:1 transformer. Output level is -35dBm, which is not helping.

Thanks!

 

[ruffrecords]

OK, now we have the full picture! The first stage seems to be capable of raising a -65dBm input to -38dBm at its output which is modest 27dB of gain. However, there is no overall NFB around this pair of transistors (it would normally be from Q11 collector to Q10 emitter). The other NFB networks are do do with setting the dc conditions and removing ac from the loop and as squarewave says it is hard to see just what is going on there.

The output stage is similar but not so complex. R134 and R135 in conjunction with R126 set the dc conditions of the output triple. Overall NFB is via R136 in conjunction with R125 and sets the gain at about 16dB. I suspect you could safely increase this to 30dB by changing R136 from 1.5K to 7.5K or 8.2K.

The first stage is faintly reminiscent of the Revax A77 replay head preamp so it might be instructive to look at that for potential modifications.

The only caveat to all the above is I do not know off hand the beta of the transistors used. They may be so low that trying to obtain higher gains will be fruitless.

Cheers

Ian

 

[gyraf]

It's made for connecting multiple microphones to a single mic pre input.

Specs says:

Output: 10mV for microphone input 200 Ohms..

 

[Conviction]

It's made for connecting multiple microphones to a single mic pre input.

Specs says:

Output: 10mV for microphone input 200 Ohms..

Well, yes 

"The output is designed to go into another preamplifier (the manual glued inside the box says the Lyrec AR2 and some others), so the sole purpose of the unit, which goes by the name Hh63, is to extend the number of available microphone inputs."

They were used extensively with Lyrec AR2's, Nagra III's, Maihak MMK6 & 4, and several in-house designs.

The Hh63 also has a sibling, same format but a "real" 2-channel preamplifier with limiter.
I'm looking to increase the gain and make it multi-channel. If that doesn't work, I'll probably bin it or sell it.

 

[mjrippe]

The Hh63 also has a sibling, same format but a "real" 2-channel preamplifier with limiter.
I'm looking to increase the gain and make it multi-channel. If that doesn't work, I'll probably bin it or sell it.

Do you have schematics of the sibling? 

And by the way, never bin anything like this before offering it up here!

 

[Conviction]

Do you have schematics of the sibling? 

And by the way, never bin anything like this before offering it up here!

Of course not, I'll let everyone know! I rarely bin things  Just a little tired watching the scrap heap.

I'll dig up the schematic. The unit is called Hh61:

 

[mjrippe]

Now that's a sexy piece of kit!

 

[Newmarket]

Yes, there's several feedback paths, which makes me dizzy. I'll try R113!

This may sound a 'cheat' for the hardcore discrete people  but if you like the "sound" of the circuit but just find the output level too low then why not simply implement a standard gain block (discrete or op amp based) before the output transformer ? Or  an electronically balanced output either active or impedance balanced. Maybe make it switchable between that and the output transformer for "sonic flexibility" ?

 

[abbey road d enfer]

Gain from mic input to C11 is 45dB; this is quite usable. Actually I find, as well as several others, that most proximity recordings happen in the 30-50 dB gain range.
The output stage, before the pot (at C111) is actually capable of driving a 600 ohm load to about 5Vrms (+15dBu). Only, for actually driving 600 ohms, C111 must be increased significantly; 100uF for adequate LF response but 470-1000uF preferrable for lower THD. Using it with modern (bridging) equipment would relax this constraint. 10uF is adequate for proper frequency response with the usual10k load, and 100uF would give better THD.
The main remaining problem would be gain control. Simplest solution would be to insert the gain pot (R130) between R113 and C103. A 100uF cap must be placed before the pot.

 

[5v333]

tuffa grejer!!

i also think that strapping an output transformer after the c111 is a good idea if you would like to use the mic pres connected individualy to an adda for ex.

doesnt it look like the first stage has both neg and pos feedback. positive with the gate of the first transistor and negative with the lower end of the input transformer?

if so, R116 could be made bigger for gain reduction.
otherwise you could just use the pre attenuation (för-dämpning) for lower gain.

where did you got this by the why?

 

[Conviction]

Gain from mic input to C11 is 45dB; this is quite usable. Actually I find, as well as several others, that most proximity recordings happen in the 30-50 dB gain range.
The output stage, before the pot (at C111) is actually capable of driving a 600 ohm load to about 5Vrms (+15dBu). Only, for actually driving 600 ohms, C111 must be increased significantly; 100uF for adequate LF response but 470-1000uF preferrable for lower THD. Using it with modern (bridging) equipment would relax this constraint. 10uF is adequate for proper frequency response with the usual10k load, and 100uF would give better THD.
The main remaining problem would be gain control. Simplest solution would be to insert the gain pot (R130) between R113 and C103. A 100uF cap must be placed before the pot.

Great, thanks abbey! I'll try your suggestions tonight.

Yesterday I increased R136 to 8,2k and it made a significant increase in gain. I didn't do any measurements more than watching the input levels in the DAW. I would guess it boosted the signal 15-20dB. Simultaneously I noticed some slight distortion but only when R130 was turned to max. I took the signal at R133.

I'll move the pot to R113 and C103, with the added cap.

Since it, according to your calculations, is able to push 5Vrms into 600 ohms, I'll just add a 1:1 transformer for each channel. There's enough room for them in the battery compartment.

I'll report back later!

 

[Conviction]

 

[Conviction]

This may sound a 'cheat' for the hardcore discrete people  but if you like the "sound" of the circuit but just find the output level too low then why not simply implement a standard gain block (discrete or op amp based) before the output transformer ? Or  an electronically balanced output either active or impedance balanced. Maybe make it switchable between that and the output transformer for "sonic flexibility" ?

That's a great idea! Good thing i'm not hardcore. ;D I merely want to turn this into something useful. I actually have a couple of gain blocks from some old 1-to-3 split modules, The only drawback is they run on +/-15V. They should, according to the data sheet, give a 20dB boost into a 600 ohms source. With a 600 ohm 1:1 and some termination it would probably work great.

 

[Conviction]

where did you got this by the why?

Tuffa grejer indeed! But even more so if it was the Hh61.

I found it in a top secret location at the SR Stockholm headquarter  :-X I think I've had it in my storage for about 10 years. But seriously. Got it - like a lot of other stuff I have - from a tech whose boss demanded him to do that final clean-out.

PS: I see you're in Gothenburg too!

 

[Bo Deadly]

Note that if you will always be plugging this directly into a vaguely good ADC, it's perfectly fine to add the gain digitally. Meaning the noise floor of the converter will likely be below the noise floor of the device even at low gain. It's easy to see. Terminate the inputs with 50 ohms, add the correct amount of gain digitally, look at the FFT / spectrum in your DAW, adjust the y-axis so that you can see the noise floor, and see if it goes up / down with the gain controls. If it does, the noise floor of the device is below the noise floor of your converters.

 

[abbey road d enfer]

doesnt it look like the first stage has both neg and pos feedback. positive with the gate of the first transistor and negative with the lower end of the input transformer?

Actually both paths are NFB, one is voltage-to-current, the other is voltage-to-voltage. One decreases the input Z, the other increases it.
This type of circuit was in favour in the late 60's-early 70's. Operators had noticed that solid-state mic preamps with no load connected at the input were quite noisy, so designers endeavoured to minimize this phenomenon, and came with solutions that were invariably based on a NFB path that varied with the actual source impedance.
I believe it may have persisted to this day in ENG and cinema mixers.
A disadvantage is that the frequency response varies with the source impedance. Most microphones have a rather stable impedance; however, some ribbon mics and some curious mics such as the dual-capsule AKG series 200 would act funny.