Poor Man's Tube Gain Make Up Stage

[MrZpliff]

yes how... is how you say it in swedish, so a little bit of swenglish 

Anyway, the question is more how and where to tap the signal without affecting the quality of it going back to the daw/recorder.

I suppose you need a buffer of some kind and would the best place be at the output or maybe after the input transformer where the signal is unbalanced ?

 

[ruffrecords]

yes how... is how you say it in swedish, so a little bit of swenglish 

Anyway, the question is more how and where to tap the signal without affecting the quality of it going back to the daw/recorder.

I suppose you need a buffer of some kind and would the best place be at the output or maybe after the input transformer where the signal is unbalanced ?

At the output is the normal place to take the feed and yes you do need a buffer so as not to load the output. I tend to take it from the output that feeds the output transformer but a Neve it used to be taken balanced across the secondary of the output transformer, into a 10K bridging transformer to the monitoring controls.

Cheers

Ian

 

[ruffrecords]

Some time ago someone asked about using an ECC88/6DJ8 instead of a 6CG7. As I think I said last time, the 6DJ8 and friends have a lower maximum plate voltage than the 6CG7 so you need to be careful with the HT supply.

On the other hand, the industrial strength version, the 6922 ,has a much higher maximum plate voltage and can be used without problems. For another project I recently designed a 'power mu follower' that would drive +28dBu into a 5K load rather than the 10K load the 6CG7 can manage. It uses a 6922. All you need to do is change the cathode resistors to 220 ohms and the 10K resistor between the tubes to 4K7, plug in a 6922 and you are good to go. Distortion is a low or lower than the 6CG7 version and the gain is just over 29dB.

Cheers

Ian

 

[MrZpliff]

Hmm. The output transformer has two sec.windings which are now connected in series.
If i take one of the sec.windings for the soundcard and the other for the monitor section. Would that make the impedance too low ? Transformers and impedance i find kinda hard to grasp  :-[

 

[ruffrecords]

Hmm. The output transformer has two sec.windings which are now connected in series.
If i take one of the sec.windings for the soundcard and the other for the monitor section. Would that make the impedance too low ? Transformers and impedance i find kinda hard to grasp  :-[

The impedance would be lower but that is no problem. What would be a problem is that the signal level would be halved (6dB lower).

Impedance  is less of an issue these days. The basic rule is that your output impedance must be less than the input impedance it is connected to. The specific rule is that is should be at least one tenth of the input impedance it drives.

So a 10K bridging input impedance should be driven from a source impedance of 1K or less (so 600 ohms is fine).

Here's a quick primer on transformer impedances and levels. Assume we have a transformer with a 1:1 ratio so what you put into the primary comes out the same on the secondary. Now assume we split the secondary in half into two identical windings. If we connect them in series, everything is as before. If we look across just one of the windings we have changed the turns ratio from 1:1 to 1:0.5.  A transformer transforms voltages in proportion to the turns ratio so instead of the full primary voltage, we now see half of it across the single secondary winding.

The thing that confuses people is that impedance is transformed differently. Impedance is transformed in proportion to the square of the turns ratio. When our ratio is 1:1 this is no problem as 1 squared is 1. When we use half the secondary winding and our ratio changes to 1:0.5 the impedance is transformed by 0.5 squared = 0.25 which is different to the voltage which is transformed by 0.5.

Cheers

Ian

 

[MrZpliff]

Thank you for the explanation !
I think I get it 

 

[dirtyhanfri]

Hi

I'm thinking in build a mic pre based on this Gain stage.

I've read some post about it in pages 9 & 10 of this thread.

But I got some questions; ¿The mic input goes direct to the tube stage? ¿Or should I make it unbalanced with a transformer or IC?

The input pot in the tube stage would be Gain? Could I replace it for a stepped switch? I've read something about 150K rather than 470K, and, If I'd like to add a Output Trim pot, it should be placed between the output of the tube stage and the output transformer?

Maybe I'll start working on this soon, I'd like to get some new color in my recordings, and always I record close-miked (AKG414) rap voices, so no big gain is needed. And with this situation in Spain, the Poor Mans philosophy is just perfect! ;D

 

[ruffrecords]

Hi

I'm thinking in build a mic pre based on this Gain stage.

I've read some post about it in pages 9 & 10 of this thread.

But I got some questions; ¿The mic input goes direct to the tube stage? ¿Or should I make it unbalanced with a transformer or IC?

For decent noise performance you really need to go through a mic transformer with a reasonably high step up ratio like 1:7 or 1:10. Sowter, Jensen. Cinemag and others make suitable transformers.

The input pot in the tube stage would be Gain? Could I replace it for a stepped switch? I've read something about 150K rather than 470K, and, If I'd like to add a Output Trim pot, it should be placed between the output of the tube stage and the output transformer?

For a simple single tube gain stage the input pot can be the gain control. Yes, you can replace it with a stepped switch.

The 150K was mentioned in relation to ensuring the microphone sees the correct load. Most 150ohm mikes like to be loaded with 10 times that resistance i.e about 1500 ohms. With a 1:10 transformer, if you place a 150K resistor across the transformer secondary, it will 'look like' 1500 ohms at the primary. If you use a 1:7 transformer you need to use a resistor around 75K. You can easily design a switched pot that looks like 150K or 75K.

Maybe I'll start working on this soon, I'd like to get some new color in my recordings, and always I record close-miked (AKG414) rap voices, so no big gain is needed. And with this situation in Spain, the Poor Mans philosophy is just perfect! ;D

With a close miked AKG414 you will have plenty of level and a single stage will have enough gain. If you want more colour you could change the tube from a 6CG7 to a 6922 (6DJ8 type) and tweak the components so it operates as an SRPP. An SRPP has less gain than a mu follower but since a 6922 has a higher mu than a 6CG7, the result is that a 6922 SRPP has almost exactly the same gain as the 6CG7 mu follower. However, the SRPP will produce a lot more colour than the mu follower (which is designed NOT to colour).You can use the poor man's tube gain make up PCB to build either.

To change the 6CG7 mu follower into a 6922 SRPP stage you need to:

1. Change all three 1K resistors t 680 ohms
2. Leave out the 100nF capacitor.
3. Replace the 10K and 220K resistors with shorting links.

That's all there is to it.

Cheers

Ian

 

[dirtyhanfri]

Thanks for the answer Ian

So, here's an idea, how about replacing the 10K resistor between the tubes with a 10K pot and you can smoothly move from a very good mu follower to a SRPP in one control. The pot will need to be able to take the 5mA and I don't know if it will sound scratchy as you turn it with that dc flowing through it but it is worth a try. The tubes should be fine since only about 50V is dropped across the 10K resistor so the tubes plate voltage only change by 25V each. As designed the circuit is symmetrical so each tube should have the same plate volts.

To avoid the scratchiness of a pot with dc you could retain the 10K and place a pot (or a switched resistor) with a series cap across it to alter the ac conditions of the circuit. Sounds like quite a fun experiment to me.

I think anyone tried it before, maybe I'll do, I'd like to use a 6cg7 tube I've got lying around, but the SRPP idea looks great to me. If I understand it right, if it works, it could be some kind of Clean/Dirty control? that would be awesome.

Anyway, if it doesn't work as expected, I think just a "clean" version with a mu follower would be enough, I think it will be a bit smoother than the ssl9k I'm using right now and wich I'm a bit tired of

 

[ruffrecords]

Thanks for the answer Ian

So, here's an idea, how about replacing the 10K resistor between the tubes with a 10K pot and you can smoothly move from a very good mu follower to a SRPP in one control. The pot will need to be able to take the 5mA and I don't know if it will sound scratchy as you turn it with that dc flowing through it but it is worth a try. The tubes should be fine since only about 50V is dropped across the 10K resistor so the tubes plate voltage only change by 25V each. As designed the circuit is symmetrical so each tube should have the same plate volts.

To avoid the scratchiness of a pot with dc you could retain the 10K and place a pot (or a switched resistor) with a series cap across it to alter the ac conditions of the circuit. Sounds like quite a fun experiment to me.

I think anyone tried it before, maybe I'll do, I'd like to use a 6cg7 tube I've got lying around, but the SRPP idea looks great to me. If I understand it right, if it works, it could be some kind of Clean/Dirty control? that would be awesome.

As far as I know, no one has done it yet. I have just not had the time to try it so I think you will be the first. Good luck!

Cheers

Ian

 

[dirtyhanfri]

Then, I think I'm gonna be the first trying it...

Any special considerations about the pot? Just a normal carbon pot? maybe a stepped switch?

I'm looking for transformers, Ludhal LL1578 (1:10 Mic input) for the input and LL1524 (1:1 600ohm) for the output should be fine or it would be better a LL1517 in the output? (it says ballanced or unballanced line)

For the Gain stepped switch, should it be lin? For 12 position switch, just 12x 12K resistors with a 6K in series?

I'm gonna start ordering parts this week, this looks like a straightforward project...

Thanks

 

[ChrioN]

audio level is not linear. and a 6922 srpp can't drive 600ohm.

 

[dirtyhanfri]

audio level is not linear. and a 6922 srpp can't drive 600ohm.

Thanks, I will calculate the resistors for the Gain switch.

And about the 6922 srpp, I will keep with a 6cg7 tube, I've got a pair of them lying around.

Thanks for the info ChrioN

 

[ruffrecords]

Then, I think I'm gonna be the first trying it...

Any special considerations about the pot? Just a normal carbon pot? maybe a stepped switch?

I'm looking for transformers, Ludhal LL1578 (1:10 Mic input) for the input and LL1524 (1:1 600ohm) for the output should be fine or it would be better a LL1517 in the output? (it says ballanced or unballanced line)

For the Gain stepped switch, should it be lin? For 12 position switch, just 12x 12K resistors with a 6K in series?

I'm gonna start ordering parts this week, this looks like a straightforward project...

Thanks

I am getting confused a bit here where there are comments without quoting what they are commenting on, so here's my take on the above post.:

I guess we are talking about the pot to use to try out the idea of smoothly changing from a mu follower to an SRPP topology. Yes a regular carbon pot will be fine for the experiment. If you are using my standard 6CG7 circuit then a 10K pot is what you need. If you use a stepped switch instead then you should add resistors between adjacent switch positions to avoid  or at least reduce clicks. In fact at this stage I think it would be better not to use a stepped switch. Best try the experiment first with a pot.

As for the input transformer then it depends how much gain you want. With a 1:10 input transformer and a 8922 you would have a pretty good basic mic pre with up to 50dB gain (46dB with a 6CG7).

For the output transformer a 600 ohm 1:1 is fine BUT as I have said before, the mu follower (or the SRPP) will not drive a 600 ohm load but it will happily drive a normal 10K bridging load through such a transformer.


I am not sure what you mean by the gain stepped switch.

Cheers

Ian

 

[dirtyhanfri]

I am getting confused a bit here where there are comments without quoting what they are commenting on, so here's my take on the above post.:

I guess we are talking about the pot to use to try out the idea of smoothly changing from a mu follower to an SRPP topology. Yes a regular carbon pot will be fine for the experiment. If you are using my standard 6CG7 circuit then a 10K pot is what you need. If you use a stepped switch instead then you should add resistors between adjacent switch positions to avoid  or at least reduce clicks. In fact at this stage I think it would be better not to use a stepped switch. Best try the experiment first with a pot.

First of all, sorry about creating confusion, sometimes I have a hard time trying to explain myself in english.

Yes, I was talking about the mu follower-srpp pot, and thats what I will use, a normal pot

For the output transformer a 600 ohm 1:1 is fine BUT as I have said before, the mu follower (or the SRPP) will not drive a 600 ohm load but it will happily drive a normal 10K bridging load through such a transformer.

Ok, thanks for the tip.

 

[ej_whyte]

I haven't had a chance to try it out yet but I'm still interested to see the results of this so please keep us updated on how it goes

 

[dirtyhanfri]

Well, after some time trying to get Lundhall transformers locally I just think the local distributor doesn't work for just a pair of trannies, or something, they just don't mail me back...

So I'm gonna try to build it with OEP's, way easier to get, I was thinking if a A262A3 (1+1:6.45+6.45) wired in series would do the job for the mic input, it would be 1:13?

Also they have the X187B, with 1:9 turns ratio, lower than adviced, but specified for mic input, maybe this would be better?

 

[ruffrecords]

Well, after some time trying to get Lundhall transformers locally I just think the local distributor doesn't work for just a pair of trannies, or something, they just don't mail me back...

So I'm gonna try to build it with OEP's, way easier to get, I was thinking if a A262A3 (1+1:6.45+6.45) wired in series would do the job for the mic input, it would be 1:13?

With both primary and secondary wired in series you get 1:6.45 which is about 16dB gain. You would get 1:13 if you wired the primary windings in parallel but I suspect the reduction in primary inductance would lead to unsatisfactory LF response.

Also they have the X187B, with 1:9 turns ratio, lower than adviced, but specified for mic input, maybe this would be better?

This looks to be a better quality transformer and 1:9 is a gain of 19dB. But if you are going to pay that much for it you might as we  get a Sowter 9610.

Cheers

Ian

 

[dirtyhanfri]

Well I just started working on this, I'm drawing a basic diagram. I'll start just with the pre and phantom, then I'll add variable pad and HPF.

I got a pairs of OEP for trying. Looking to upgrade them to lundhals (Finally I got a quote from their spanish distributor)

I have the PSU working and the tube board fully stuffed. Just a fast question:

I just noticed I have no pot for the output trim, what value should I use? Maybe I'll try different values to see their effect...

Please could somebody check if everything in the diagram looks fine or I'm making some kind of mistake?, I spotted the points where I will implement Pad and HPF, Maybe I should add caps or something between phantom and Input Transformer?

I think I'll start a new thread in the preamp section when I got this working


EDIT: I forgot to draw the Output XLR after the Output transformer...

 

[ruffrecords]

The pad is in the correct position as is the phantom power. You can also add a phase change switch here if you want. Jensen has an app note that shows how to do these three at the input:

http://www.jensentransformers.com/as/as016.pdf

The gain control will work where you have it. It is not ideal but in a one tube mic pre there is not much  alternative. You just need to make sure the leads to and from it are as short as possible ans screened. This is a high impedance part of the circuit and is susceptible to interference. It does need to be wired as a pot so a connection to ground is needed from the bottom of the pot.

The 'character' pot is just a series variable resistor so it does not need a connection to ground.

You can have an output pot on the output of a mu follower stage like this but normally only if it is feeding another tube stage. It will not drive a low impedance pot and a following transformer. If you want an output pot so you can drive the tube harder with the gain pot and reduce the output with the output pot then I would recommend using a 5K output pot. The 6922 SRPP/mu follower will drive this to quite high levels producing some nice smooth tube distortion. The output impedance of this pot should be no more than about 1.5K. If you want a transformer balanced output I would recommend using a 10K:10K type.

Good luck.

Cheers

Ian