Redd. 47 preamp + Redd EQ

[emi2345]

Hi all, I have a spare 3U rack case lying around and I've been listening to a lot of Revolver/White album, so it figures I'd have a go at the Redd.47 preamps they used.

I was thinking along the lines of two channels, with Ian Thompson Bell's Redd/RS127 EQ combined with a further ECC88 stage to makeup the gain and drive the output. All switched so it could be used as a stereo unit. Phantom power, 20dB input pad, phase, high-pass filter, hi-z jack input, Sowter 9970 & 9980 transformers, GZ34 power supply with parallel dual 0A2 stabilisers to get me 80mA, VU meters with drivers.

Here's my draft schematic and a bunch of questions, feel free to tear it apart.

1. do I need coupling caps to block phantom power from getting to the input transformer? what value?

2. how should I implement phantom power short protection? a 200mA fuse?

3. how much current does phantom power current draw? 50mA per channel?

4. VU meter current draws 30mA per channel from +48V regulated power supply. So can I pull 160mA from the 50V tap on the transformer rated for 144mA @600V?

5. high-pass filter what values for 100Hz? Brian Sowter kindly gave me the measurements of the 9970 transformer: 23.4R primary; 1635R secondary, 13.06H. Original 'rumble' filter uses 1.8uF and 3K for 30Hz, the original A92 transformer appears to be 18H.

6. Zobel filter values? Sowter 9970 is 13H original A92 transformer appears to be 18H. Brian said the Zobel probably wouldn't be needed. How do I wire the mic transformer with centre tap without zobel?

7. Sowter 9970 has 35% tap, is that for zobel use or for alternative/switchable impedance? the secondary looks wired in parallel in the schematic

8. What happens if we bypass the negative feedback altogether? How much more gain do we get? Do we get lots of tasty distortion?

10. 600 ohm T-pad attenuator OK? It needs to be before the EQ so before the output transformer, so unbalanced.

11. Power supply - the dual 0A2 stabilisers max out at about 40mA, so can I wire two pairs in parallel to give me 80mA? Are my CLCRCR values ok?

All help would be very much appreciated. I've also drawn a schematic of the EQ circuit with the ECC88 gain/buffer stage. I don't want to post it without Ian's permission, but it's pretty self explanatory, I was going to use the same values from the REDD .47, and I was going to use his PCBs for the EQ.

 

[emi2345]

And here's a draft of the power supply...

 

[emi2345]

Fair enough no-one got involved, it's a lot in one post and there are answers to some of these questions out there somewhere. I think I'll go with a regulated PSU based on the G9 instead of the very complex tube one. Here are a couple of questions I couldn't find answers to...

- Has anyone implemented a 100Hz high-pass filter on the primary (as opposed to the 30Hz 'rumble' filter) and what values would work?

- Has anyone tried a no-negative feedback higher-gain setting? What would the results be like in terms of gain and distortion?

 

[TwentyTrees]

I don't have much useful to add (apart from a supportive bump), as I haven't done any tube stuff yet, but this sounds like an awesome project and I'm watching with interest as have been thinking of looking at something similar at some point. Subscribed!

For what it's worth, I understand that phantom power current draw is very low - API VPR spec is for 5mA per module, for example.

Good luck!

 

[emi2345]

Thanks! Yeah I did some more searching and found answers to a few of my own questions. I'm going to drop the Zobel for now, and the VU meters, I don't think I need capacitors to block phantom power. I could tweak the RC values to get a 100Hz high-pass but I'd love to have a good starting point if anyone else has got there, and the same with the gain/feedback settings.

 

[ruffrecords]

There was a long thread about adding EQ to a REDD 47 maybe a couple of years or so ago. Definitely worth searching for it.

Cheers

Ian

 

[emi2345]

Thanks Ian! I was going to buy your pcbs for the eq, with the pots and switches off-board. Do you know if anyone ended up trying it with the ecc88 circuit from the redd .47?

 

[ruffrecords]

Thanks Ian! I was going to buy your pcbs for the eq, with the pots and switches off-board. Do you know if anyone ended up trying it with the ecc88 circuit from the redd .47?

As I said there was a long thread about this and they were talking about using one of my EQs. I do not recall the details but I think the removed the output transformer and fed the unbalanced output into the EQ. Then they added another ECC88 stage to make up the gain lost in the EQ and used it to drive an output transformer (not sure if it was the same one). The thread is on here somewhere.

Cheers

Ian

 

[emi2345]

I'll have a look, so far I found the thread where you described that as an idea Ian, but not one where someone actually tried it out.

 

[ruffrecords]

I'll have a look, so far I found the thread where you described that as an idea Ian, but not one where someone actually tried it out.

OK, I could have sworn that he went ahead an implemented it but it is possible that was another thread. There are so many REDD 47 threads. Can you link to the one where I suggested it?

Cheers

Ian

 

[moamps]

I was thinking along the lines of two channels, with Ian Thompson Bell's Redd/RS127 EQ combined with a further ECC88 stage to makeup the gain and drive the output.

I don't like this idea, but will try to provide some answers.

1. do I need coupling caps to block phantom power from getting to the input transformer? what value?

No.

2. how should I implement phantom power short protection? a 200mA fuse?

You don't need dedicated protection for phantom source because the current is limited with 6k8 resistors anyway. But if you like, for two microphones,  50 or 63mA fuses will do.

3. how much current does phantom power current draw? 50mA per channel?

14mA shorted, max (48V/3k4)

4. VU meter current draws 30mA per channel from +48V regulated power supply. So can I pull 160mA from the 50V tap on the transformer rated for 144mA @600V?

Not really, because this part of secondary wire will most probably overheats loaded with 160mA for 50V tap and two B+ currents.

5. high-pass filter what values for 100Hz? Brian Sowter kindly gave me the measurements of the 9970 transformer: 23.4R primary; 1635R secondary, 13.06H. Original 'rumble' filter uses 1.8uF and 3K for 30Hz, the original A92 transformer appears to be 18H.

It really depends of a transformer, I got about 0.82uF for 125Hz, you are using two caps, so about 1.5uF can be a good start.

6. Zobel filter values? Sowter 9970 is 13H original A92 transformer appears to be 18H. Brian said the Zobel probably wouldn't be needed. How do I wire the mic transformer with centre tap without zobel?

The zobel shouldn't be connected to the secondary's middle tap. That's the mistake from the beginning of copying the original drawing, IMO.
 

8. What happens if we bypass the negative feedback altogether? How much more gain do we get? Do we get lots of tasty distortion?

You will get about 55-60dB (both transformer included)

10. 600 ohm T-pad attenuator OK? It needs to be before the EQ so before the output transformer, so unbalanced.

You can't use 600ohm T-pad there, only after output transformer. The proper attenuator  placed after anode of E88CC should have 600x7x7 ohms. Besides that, there is higher output voltage which is 7x greater than at the output of the output transformer.

11. Power supply - the dual 0A2 stabilisers max out at about 40mA, so can I wire two pairs in parallel to give me 80mA? Are my CLCRCR values ok?

The better way is to design two separate B+ outputs, one for each channel.

 

[emi2345]

OK, I could have sworn that he went ahead an implemented it but it is possible that was another thread. There are so many REDD 47 threads. Can you link to the one where I suggested it?

Cheers

Ian

On page 4 of this thread you mention it and show how to convert one of your pcbs to implement it https://groupdiy.com/index.php?topic=51030.60

 

[ruffrecords]

On page 4 of this thread you mention it and show how to convert one of your pcbs to implement it https://groupdiy.com/index.php?topic=51030.60

Yes, that seems to be the one. Basically you just repeat the RED 47 output stage for use as a gain make up amp. Unfortunately i do not make the poor man's tube gain make up amp any more (I have one left in stock).

Cheers

Ian

 

[emi2345]

Yes, that seems to be the one. Basically you just repeat the RED 47 output stage for use as a gain make up amp. Unfortunately i do not make the poor man's tube gain make up amp any more (I have one left in stock).

Cheers

Ian

Thanks Ian,

That's ok I was going to use tag board/strip.

I'm not quite sure what the signal flow was in the original .51 desks was, would it have gone .47 amplifier - inserts - channel eq/ fader /pan pot - mix bus? And would these stages all have been passive unless another .47 was added to make up for any insertion loss? I ask because perhaps using just the ecc88 part of the .47 circuit might not be a very close approximation to the signal chain in the desk, and also might taking the ecc88 out of its circuit with the ef86 and feedback loop change the performance a lot?

Thanks,

Edmund

 

[emi2345]

"I was thinking along the lines of two channels, with Ian Thompson Bell's Redd/RS127 EQ combined with a further ECC88 stage to makeup the gain and drive the output."
I don't like this idea, but will try to provide some answers.

Would love to hear what you've got to say about this, see the post above.

It really depends of a transformer, I got about 0.82uF for 125Hz, you are using two caps, so about 1.5uF can be a good start.

Great info thanks, is your transformer the Sowter or similar?

  You will get about 55-60dB (both transformer included)

That's great, do you know what feedback resistor values I should use to get 52dB? Then I could go for around 58dB with no feedback. Distortion is sort of the point with these higher gain settings.

You can't use 600ohm T-pad there, only after output transformer. The proper attenuator  placed after anode of E88CC should have 600x7x7 ohms. Besides that, there is higher output voltage which is 7x greater than at the output of the output transformer.

I don't quite follow you here. Since I don't want to saturate the EQ section which is coupled to the E88CC output, I need to have my attenuator there (unbalanced E88CC output, before output TX). Would an L-pad be more suitable?

Thanks for all your help!

 

[ruffrecords]

I'm not quite sure what the signal flow was in the original .51 desks was, would it have gone .47 amplifier - inserts - channel eq/ fader /pan pot - mix bus? And would these stages all have been passive unless another .47 was added to make up for any insertion loss?

That is pretty much it;  mic -> REDD 47 -> EQ -> fader ->pan -> bus -> REDD 47 -> master fader -> REDD47 -> out

Everything except the REDD 47 was passive balanced 200 ohms.

I ask because perhaps using just the ecc88 part of the .47 circuit might not be a very close approximation to the signal chain in the desk, and also might taking the ecc88 out of its circuit with the ef86 and feedback loop change the performance a lot?

You are right it will not be like the signal chain in the original desk. If you want to get close to that you need to add a 200 ohm or 600 ohm EQ followed by another REDD 47.

Cheers

Ian

 

[moamps]

Would love to hear what you've got to say about this, see the post above.

There is no optimal place in REDD47  for inserting an EQ, IMO.

Great info thanks, is your transformer the Sowter or similar?

Two Sowters 9970, slightly different. You should do some experiments to get right value.

That's great, do you know what feedback resistor values I should use to get 52dB? Then I could go for around 58dB with no feedback.

It depends of particular tubes you use (gain, etc.), I can't help you more.

.I don't quite follow you here. Since I don't want to saturate the EQ section which is coupled to the E88CC output, I need to have my attenuator there (unbalanced E88CC output, before output TX). Would an L-pad be more suitable?

You simply can't load anode of output tube in REDD47 with 600 ohms, whichever topology  of pad you use.  Anodes of E88CC are originally loaded with reflected impedance of, let say, 600 ohms  load from secondary to primary side of the output transformer.

 

[emi2345]

There is no optimal place in REDD47  for inserting an EQ, IMO.Two Sowters 9970, slightly different. You should do some experiments to get right value.It depends of particular tubes you use (gain, etc.), I can't help you more. You simply can't load anode of output tube in REDD47 with 600 ohms, whichever topology  of pad you use.  Anodes of E88CC are originally loaded with reflected impedance of, let say, 600 ohms  load from secondary to primary side of the output transformer.

Ah yes for some reason I assumed without really looking that the E88CC was in a low output impedance circuit but now I realise it's just a parallel common cathode gain stage. If I plug the numbers into this calculator https://www.ampbooks.com/mobile/amplifier-calculators/output-impedance/calculator/ (16K plate resistor 400R cathode resistor, 6922 tube) I get 2.27K output impedance per triode. So If I design a bridged t-pad with a matched impedance (5K6) would that present the tube and the transformer the correct load?

 

[ruffrecords]

Ah yes for some reason I assumed without really looking that the E88CC was in a low output impedance circuit but now I realise it's just a parallel common cathode gain stage. If I plug the numbers into this calculator https://www.ampbooks.com/mobile/amplifier-calculators/output-impedance/calculator/ (16K plate resistor 400R cathode resistor, 6922 tube) I get 2.27K output impedance per triode. So If I design a bridged t-pad with a matched impedance (5K6) would that present the tube and the transformer the correct load?

Unfortunately not. The transformer needs to be driven by as low an impedance as possible in order to maintain an adequate low frequency response.

Cheers

Ian

 

[emi2345]

Unfortunately not. The transformer needs to be driven by as low an impedance as possible in order to maintain an adequate low frequency response.

Cheers

Ian

So is there no type of attenuator that would work before the output transformer primary? If so am I going to need another set of transformers to interface the eq properly with the amplifier?