Poor Man's Tube Gain Make Up Stage

[ruffrecords]

Hi Ian,


  wow. followed link. Your tube mixer looks fantastic. All your philosophy rings true to me. Wish I had loads of money . . . Oh well . . .


   I really am incredibly grateful for all your time and dedication to this project. We are not really worthy . .. oh,all right then. we are . . .


      Kindest regards,


       ANdyP

Hey, Andy, no problem, in fact my pleasure. I spent all my working life in two industries. The first professional audio, at Neve in the 70s, where quality came first and cost second. The second part was in consumer electronics where gaining the maximum functionality from the minimum parts without sacrificing reliability was the primary goal.  The poor man's Pultec is an attempt to apply the lessons I learnt in the second part of my working life to the products where my heart lies, and that is pro audio. Project studios do not get the thrashing that pro audio equipment did in the 70s where it was operated 24/7. That means the same sonic qualities should be capable of being realised without the necessity of 'building like a tank' and the associated costs that brings. So you don't really need to use Grayhill switches and 1% capacitors to make a pro sounding EQ.

Cheers

Ian

 

[strangeandbouncy]

. . . Yay . . . .

 

[ej_whyte]

Where do A, B and C relate to with the EQ schematic?

To ensure the overall gain is unity with the controls set flat you need a pot at the input to the gain make up amp. That is the connection to point C.

If you also want to have an EQ bypass switch you need to replace the EQ with a pad in the bypassed position; the input to that pad is point B and its output is point C.

Cheers

Ian

Aah yeh had a look at schemo and think i get it now. Point B would be from the input Xformer via the bypass switch, and point c would either connect to the EQ output or point A depending on the bypass switch.

Looking like a good project

 

[ruffrecords]

Aah yeh had a look at schemo and think i get it now. Point B would be from the input Xformer via the bypass switch, and point c would either connect to the EQ output or point A depending on the bypass switch.

Spot on. I think I'll do a drawing showing that.

Cheers

Ian

 

[ruffrecords]

I Just had a thought. Rather than having a PCB with two channels worth of gain make up and extra holes plus a score line so it can be cut up into two, why not just do a PCB with a single gain make up channel on it and use two for stereo?. That way, in a stereo set up you could separate them a bit (especially in a 1U rack) to reduce crosstalk etc, and it would be very easy to build a single channel or indeed a multi channel unit.

What is the consensus? is this the way to go?

Cheers

Ian

 

[menigu]

great idea, Ian.

 

[strangeandbouncy]

plus one here . . .

 

[chrispbass]

+1

 

[culteousness1]

+1!

 

[ruffrecords]

OK, looks like the 'Ayes' have it. I'll do a single channel layout incorporating the other comments too.

Cheers

Ian

 

[ej_whyte]

Im not too bothered about the stereo/mono issue, but i guess mono would be most flexible. Im in the process of making a complete schematic for the EQ section and your mu follower stage, with stepped pots to replace the cut/boost pots, mainly just to sort it all out in my head, but if anyone is interested i don't mind posting it (it is your project after all ). I've got the lin pots done, nice and easy, but got a little stumped on the log pots. I was using one of the stepped attenuator calculators online, but wasn't sure what value to put in the max attenuation field. I know in theory we would want infinite for this situation, but i have a feeling that would be impractical? Can anyone suggest a suitable value? I've seen 40dB thrown around a couple of times.

 

[ruffrecords]

Im not too bothered about the stereo/mono issue, but i guess mono would be most flexible. Im in the process of making a complete schematic for the EQ section and your mu follower stage, with stepped pots to replace the cut/boost pots, mainly just to sort it all out in my head, but if anyone is interested i don't mind posting it (it is your project after all ). I've got the lin pots done, nice and easy, but got a little stumped on the log pots. I was using one of the stepped attenuator calculators online, but wasn't sure what value to put in the max attenuation field. I know in theory we would want infinite for this situation, but i have a feeling that would be impractical? Can anyone suggest a suitable value? I've seen 40dB thrown around a couple of times.

I have measured a couple of readily available plastic track pots and their off resistance is really excellent - a few ohms at most. We don't need that as you say but the off position of the pots is the flat response position so there is no reason it should not be as good as it can get. 40dB is probably not enough - that's 100 times so the 470K lo cut pot would look like 4K7 which I think is too high. 470 ohms might be OK i.e 60dB but 47 ohms would be better i.e. 80dB.

The other thing is that although log pots are used  the actual law of boost/cut versus pot rotation is not really log. A typical log pot is 20dB down at the centre i.e the 470K lo cut  pot becomes 47K which gives about 6dB attenuation. This means the attenuation changes from about -16dB to -6dB  i.e. by 10dB in half the rotation, then from -6dB to 0dB i.e by 6dB in the second half. If you are going to the trouble of making stepped pots then it might be worth the extra effort in adjusting its law to get a constant dB change per step.

 

[ej_whyte]

Isn't the low cut a lin pot?

 

[ruffrecords]

Isn't the low cut a lin pot?

Ah, I see what you mean. That is my mistake. I just checked with the original Pultec. Both the hi pots are linear and both the lo pots are log.

Cheers

Ian

 

[ej_whyte]

Ok here is a your capacitor spreadsheet with my resistor calcs added in, how do they look?

I am not too bothered about having every step the same attenuation, as the main reason I want stepped is for easy stereo matching.

I used the calculator at http://margo.student.utwente.nl/klaas/audio_step.htm by the way, and for the actual resistor values I tried to use values that mouser had a fair number of choices for.

Thanks

 

[ruffrecords]

Ok here is a your capacitor spreadsheet with my resistor calcs added in, how do they look?

I am not too bothered about having every step the same attenuation, as the main reason I want stepped is for easy stereo matching.

I used the calculator at http://margo.student.utwente.nl/klaas/audio_step.htm by the way, and for the actual resistor values I tried to use values that mouser had a fair number of choices for.

Something is not right. A regular log pot is usually at 10% of its nominal value at the half way point so the 47K pot should be at 4K7 at your position 6. Similarly the 470K should be at 47K at position 6.

So I checked out the link you posted and after a bit of head scratching I think I have it worked out. The program creates resistor values for equal dB steps. So if we want to be -20dB at the middle position, then if the steps are equal we must be at -40dB at the last position. So I ran the program with 11 steps,  total impedance 47000, highest level 0dB and lowest level -40dB. If you then add up the resistor values from the bottom to the mid point it equals 4701 which is exactly what I would have expected.

So the answer is to use this program with highest level set to 0dB and lowest level to 40dB. The maximum attenuation is infinite since in the last position the switch wiper is connected to ground.

Cheers

Ian

 

[ej_whyte]

OK cheers, I did think the graph curves looked a little extreme. I'll redo the calcs tomorrow, bed time now, session starts at 8 :/

 

[ruffrecords]

OK cheers, I did think the graph curves looked a little extreme. I'll redo the calcs tomorrow, bed time now, session starts at 8 :/

LOL, no rest for the wicked and of course you will be watching the Royal Wedding as well 

Cheers

Ian

 

[ej_whyte]

I was just wondering as well, why is the bottom resistor always bigger then the one above it? For example with 47k, 0 to -40dB the bottom 3 resistors are 436, 275 and 470, but you would think as all the other resistors are lower than the previous one, the last resistor would be lower than 275? It just seems strange and I couldn't see a reason :s

P.S sorry to hijack this thread with non-makeup related rambling  :-\ I'll put it in the original EQ thread from now on

Cheers

 

[ruffrecords]

I was just wondering as well, why is the bottom resistor always bigger then the one above it? For example with 47k, 0 to -40dB the bottom 3 resistors are 436, 275 and 470, but you would think as all the other resistors are lower than the previous one, the last resistor would be lower than 275? It just seems strange and I couldn't see a reason :s

It is simply because the last resistor defines the lowest level, rather than being one of the fixed size steps. In the 40dB example it must be 1/100 of the total resistance i.. 470 ohms

Look at it this way, with 11 steps starting from 0dB attenuation, each step adds 4dB of attenuation, EXCEPT for the last step (the 470) which adds a lot more than 4dB of attenuation (it adds from -40dB to -infinitydB - however much that is!)

Cheers

Ian