Poor Man's Tube Gain Make Up Stage

[JW]

Hey guys,
I thought I'd tag this question onto this thread, as I'm wanting to build up a makeup amp for my Pultec filter boards.
Did the solid state pultecs have an interstage transformer? In researching so far, I haven't seen a S.S. pultec schematic.

I have a couple 2520 opamps that I'd like to use in an API style makeup amp, and am mostly wondering about what the output of the filter going into the makeup amp would look like. To add interstage transformer or not? Any recommendations?

 

[MatthisD]

I think you can run the Pultec filter section straight into the 325 lineamp without any fuss. If you use the Poor Man's tube stage then you can use the bypass pad as it is and workout a smaller pot value to replace the 470k gain trimmer.

Some EQP-1A3 info here:

http://danalexanderaudio.com/OutboardPic/PultecEqp1A3.jpg
http://www.groupdiy.com/index.php?topic=29742.0

 

[ej_whyte]

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.

The mu follower is remarkably tolerant of dc operating point and the unloaded distortion is not very dependent on the plate current - the output drive capability is, of course, directly proportional to dc current.

Cheers

Ian

Did anyone ever try this out?

Would a stepped pot with a high value resistor in parrallel still pop as it switched?

Cheers

 

[ruffrecords]

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.

The mu follower is remarkably tolerant of dc operating point and the unloaded distortion is not very dependent on the plate current - the output drive capability is, of course, directly proportional to dc current.

Cheers

Ian

Did anyone ever try this out?

Would a stepped pot with a high value resistor in parrallel still pop as it switched?

Cheers

I had forgotten all about that.No, I never got round to trying it.Don't know if anyone else did. Why not have a go yourself?

Cheers

Ian

 

[ej_whyte]

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.

The mu follower is remarkably tolerant of dc operating point and the unloaded distortion is not very dependent on the plate current - the output drive capability is, of course, directly proportional to dc current.

Cheers

Ian

Did anyone ever try this out?

Would a stepped pot with a high value resistor in parrallel still pop as it switched?

Cheers

I had forgotten all about that.No, I never got round to trying it.Don't know if anyone else did. Why not have a go yourself?

Cheers

Ian

Wouldn't you have to jump the 100n cap and 220k res as well?

Elliott

 

[ruffrecords]

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.

The mu follower is remarkably tolerant of dc operating point and the unloaded distortion is not very dependent on the plate current - the output drive capability is, of course, directly proportional to dc current.

Cheers

Ian

Did anyone ever try this out?

Would a stepped pot with a high value resistor in parrallel still pop as it switched?

Cheers

I had forgotten all about that.No, I never got round to trying it.Don't know if anyone else did. Why not have a go yourself?

Cheers

Ian

Wouldn't you have to jump the 100n cap and 220k res as well?

Elliott

I wondered about this too, but I don't think so. When the pot is at zero the 100n is directly in parallel with the 220K. This is just like a regular SRPP with a large grid stopper bypassed by 100n so it ought to act just like an SRPP. The low frequency response might suffer a little as the 220K is no longer bootstrapped. With the pot at 10K it is exactly my regular mu follower. Quite what it is in between is anyone's guess but as the pot value is reduced it become a poorer mu follower so the plate load for the lower tube gets less and the distortion gets higher until with the pot at zero you get the SRPP sound.  Thank's for bringing this up again as it might make a useful 'tube tone' effect so I think I mihght just build one.

Cheers

Ian

 

[ej_whyte]

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.

The mu follower is remarkably tolerant of dc operating point and the unloaded distortion is not very dependent on the plate current - the output drive capability is, of course, directly proportional to dc current.

Cheers

Ian

Did anyone ever try this out?

Would a stepped pot with a high value resistor in parrallel still pop as it switched?

Cheers

I had forgotten all about that.No, I never got round to trying it.Don't know if anyone else did. Why not have a go yourself?

Cheers

Ian

Wouldn't you have to jump the 100n cap and 220k res as well?

Elliott

I wondered about this too, but I don't think so. When the pot is at zero the 100n is directly in parallel with the 220K. This is just like a regular SRPP with a large grid stopper bypassed by 100n so it ought to act just like an SRPP. The low frequency response might suffer a little as the 220K is no longer bootstrapped. With the pot at 10K it is exactly my regular mu follower. Quite what it is in between is anyone's guess but as the pot value is reduced it become a poorer mu follower so the plate load for the lower tube gets less and the distortion gets higher until with the pot at zero you get the SRPP sound.  Thank's for bringing this up again as it might make a useful 'tube tone' effect so I think I mihght just build one.

Cheers

Ian

More thanks to you for suggesting it in the first place!

That is exactly the reason I am looking in to this. For a unit I am working on where I need a 20-26dB line stage the mu follower seems to meet all the specs perfectly, but after reading around a lot (no experience with building tube gear yet) I am slightly worried that the mu follower is too clean for what I want. I'm stuck deciding between the mu follower or SRPP. In the end I will probably just work the PCB so that it can be configured for either, but I'm still quite interested in being able to vary the tone.

If you do get round to building it before me, please let us know how it goes, and whether the pot is scratchy.

Thanks
Elliott

 

[ruffrecords]

snip

Wouldn't you have to jump the 100n cap and 220k res as well?

Elliott

I wondered about this too, but I don't think so. When the pot is at zero the 100n is directly in parallel with the 220K. This is just like a regular SRPP with a large grid stopper bypassed by 100n so it ought to act just like an SRPP. The low frequency response might suffer a little as the 220K is no longer bootstrapped. With the pot at 10K it is exactly my regular mu follower. Quite what it is in between is anyone's guess but as the pot value is reduced it become a poorer mu follower so the plate load for the lower tube gets less and the distortion gets higher until with the pot at zero you get the SRPP sound.  Thank's for bringing this up again as it might make a useful 'tube tone' effect so I think I mihght just build one.

Cheers

Ian

More thanks to you for suggesting it in the first place!

That is exactly the reason I am looking in to this. For a unit I am working on where I need a 20-26dB line stage the mu follower seems to meet all the specs perfectly, but after reading around a lot (no experience with building tube gear yet) I am slightly worried that the mu follower is too clean for what I want. I'm stuck deciding between the mu follower or SRPP. In the end I will probably just work the PCB so that it can be configured for either, but I'm still quite interested in being able to vary the tone.

If you do get round to building it before me, please let us know how it goes, and whether the pot is scratchy.

Thanks
Elliott

A mu follower is certainly clean which is why I use them. Using a 6CG7 or 6SN7 into a 10K load you will get about 0.04% distortion at 2V rms out (+6dBV) rising to about 0.4% at 20V rms (+26dBV) and the distortion is almost entirely second harmonic (3rd is usually 20dB lower).

At the other end of the scale, a 6922 (6DJ8) SRPP will give you about 1.8% into 10K at 20V rms. I suggest 6922 because it has a higher max plate voltage then the regular 6DJ8. If you want more distortion you should try a tube with greater intrinsic distortion like the good old 12AU7.

Cheers

Ian

 

[MrZpliff]

Hi there ! 

I like to check a couple of things out before starting to build.

A PSU question:
Since I have most of the parts for a G9 psu I thought I'd use it for the PM Tube Gain Makeup.

Any comments ?

For the heaters in the G9 there are 4 ECC82 of 150 mA each (12,6V - series) = 600 mA.
If I connect the 2 heaters in series for the 6CG7 (6,3V, 600 mA) I would end up with the same current and 6,3V for each tube.

Should work, or am I missing something ?

Edit:

Well I guess that question's already answered  :-[

That would probably do. The 6CG7 requires 0.6amp at 6.3V and you can run the heaters of a pair in series without problems. For HT you only need 240V ac at 15 to 20mA and the mu follower stage is very tolerant of actual HT voltage so something like the G9 back to backs should work.

One more question, though...
I thought I'd use the PM as makeup in a passive summing mixer - balanced, also.

For best result, can I use the same type of 1:1 transformer for both input (debalancing) and output.

I thought I'd use a Lundahl LL1540:
Static resistance of each primary: 610W
Static resistance of each secondary: 800W
Distortion (source impedance 600W ): + 20 dBU < 0.1% @ 50 Hz
+30 dBU < 1 % @ 50 Hz
Self resonance point : > 60 kHz
Recommended load for best square-wave response: 22 kW in series with 1nF
Frequency response (source 600W, load 15 k W ) 5 Hz -- 50 kHz +/- 0.2 dB
Loss across transformer (at 1 kHz with above termination): 0.5 dB

Best regards
/Magnus

 

[ruffrecords]

I thought I'd use the PM as makeup in a passive summing mixer - balanced, also.

For best result, can I use the same type of 1:1 transformer for both input (debalancing) and output.

I thought I'd use a Lundahl LL1540:
Static resistance of each primary: 610W
Static resistance of each secondary: 800W
Distortion (source impedance 600W ): + 20 dBU < 0.1% @ 50 Hz
+30 dBU < 1 % @ 50 Hz
Self resonance point : > 60 kHz
Recommended load for best square-wave response: 22 kW in series with 1nF
Frequency response (source 600W, load 15 k W ) 5 Hz -- 50 kHz +/- 0.2 dB
Loss across transformer (at 1 kHz with above termination): 0.5 dB

Best regards
/Magnus

This transformer will certainly work at the input to interface to your balanced passive mix network to the PMTGMU as long as your bus impedance is 600 ohms or less, It will probably work at the output as well but transformers like this that are primarily intended for inputs tend to have smaller core sizes and hence are more prone to distortion at low frequencies. However, provided you don't need more than +20dBu at 50Hz it should not be a problem.

Cheers

Ian

 

[MrZpliff]

Thank's Ian !

I thought I'd use Ytstetsef's passive mixer network

If I've done my math correctly it's output impedance should be something like 200 ohms.
I'll probably use a little beefier transformer for the output, like the Lundahl LL1517.
I'll check out the Sowter's also, but Lundahl's convenient for me since i live in sweden...

 

[ruffrecords]

Thank's Ian !

I thought I'd use Ytstetsef's passive mixer network

If I've done my math correctly it's output impedance should be something like 200 ohms.

You probably do not need the 237 ohm resistors; they just lower then output impedance to 200 ohms so it is suitable for plugging into a mic pre. Since you plan to connect this to the PMTGMU via a 1:1 10K bridging transformer you do not need so low an impedance. Without the 237 ohm resistors I calculate the bus impedance is 980 ohms which should be fine for a 10K bridging transformer. The bus loss will be close to 20dB which is easily made up by the 26dB gain of the PMTGMU.

You should be aware that passive mixing only works properly if unused inputs are shorted i.e.hot is shorted to cold - which probably explains why a lot of passive mixers have TRS inputs rather than XLR.

Cheers

Ian

 

[MrZpliff]

You probably do not need the 237 ohm resistors; they just lower then output impedance to 200 ohms so it is suitable for plugging into a mic pre. Since you plan to connect this to the PMTGMU via a 1:1 10K bridging transformer you do not need so low an impedance. Without the 237 ohm resistors I calculate the bus impedance is 980 ohms which should be fine for a 10K bridging transformer. The bus loss will be close to 20dB which is easily made up by the 26dB gain of the PMTGMU.

You should be aware that passive mixing only works properly if unused inputs are shorted i.e.hot is shorted to cold - which probably explains why a lot of passive mixers have TRS inputs rather than XLR.

Cheers

Ian

I can try it without the  237 ohm resistors or maybe double the value of them ...

I thought I'd wire it up this way (which seems to be how others have done it):

Soundcard Out - Patchbay - Passive sum network -Makeupgain - Soundcard.
The patchbay will be used for busscomps and Eq's.

Where is the best place for the mainbus comp (a turbo/crc GSSL) ? Before or after makeup gain. I dunno the input impedance of the GSSL for the moment.

I'll have all the channels connected, but will probaly install a muteswitch which shorts hot and cold.

One idea I have is to use 16 EZ Tube Mixer channels (and let's say 4 of them with the mic trafo, the rest just for line level) for summing, but that's for later.

Best

 

[ruffrecords]

I can try it without the  237 ohm resistors or maybe double the value of them ...

For overall unity gain you want 26dB of loss. Without the 237 ohm resistors the bus impedance is 980 ohms which means the bus loss is  close to 21dB. For 26dB loss the bus impedance needs to be reduced to 526 ohms which you can do by using a 1K2 resistor instead of the 237 ohms.

I thought I'd wire it up this way (which seems to be how others have done it):

Soundcard Out - Patchbay - Passive sum network -Makeupgain - Soundcard.
The patchbay will be used for busscomps and Eq's.

Where is the best place for the mainbus comp (a turbo/crc GSSL) ? Before or after makeup gain. I dunno the input impedance of the GSSL for the moment.

I'll have all the channels connected, but will probaly install a muteswitch which shorts hot and cold.

That sounds OK. The bus comp should come after the gain make up.

One idea I have is to use 16 EZ Tube Mixer channels (and let's say 4 of them with the mic trafo, the rest just for line level) for summing, but that's for later.

Best

Now that sound like  an excellent project!

Cheers

Ian

 

[MrZpliff]

Hi again...  :

Just want to check. What's the necessary voltage of the 100nF cap ?
I'm not really shure of how to interpret your handwriting on the schematic .... 

 

[ruffrecords]

Hi again...  :

Just want to check. What's the necessary voltage of the 100nF cap ?
I'm not really shure of how to interpret your handwriting on the schematic .... 

I use 400V rated parts for this component. This is what I normally fit:

http://www.rapidonline.com/Electronic-Components/100n-400v-368-Series-Polyester-Capacitor-10-4134

Cheers

Ian

 

[MrZpliff]

Thank's Ian.

I found a suitable power transformer (for us living in 230v countries) for your PSU:
http://www.tube-town.net/ttstore/product_info.php/info/p5391_Torodial-47VA.html

a decent price I think..so I've changed my mind and will use it instead of the G9-PSU.

 

[ruffrecords]

Thank's Ian.

I found a suitable power transformer (for us living in 230v countries) for your PSU:
http://www.tube-town.net/ttstore/product_info.php/info/p5391_Torodial-47VA.html

a decent price I think..so I've changed my mind and will use it instead of the G9-PSU.

That is an interesting transformer especially as it has 12.6 and 6.3V windings. It does not say what the current capability is for the 250V secondary, but it is said to be a 47VA transformer and the two heater windings account for only 22VA so there's a good chance the 250V winding will do close to 100mA which is more than adequate for the tube gain make up stage. As you say, the price is good and I need to build a general purpose power supply for the EZ tube mixer boards so I think I might just get one of these for myself.

Thanks for the tip.

Cheers

Ian

 

[MrZpliff]

Hi !

Got a couple of questions again.

When using this for gain make up for a passive mixer network (as in the posts above) - could one go straight to the tube at the 1M resistor after the transformer ? I obviously don't need the pad as I don't have a bypasswitch. I suppose I need the 470K trimmer if I want to adjust the volume between the different channels, but can I just omit it otherwise ?

I happen to have some 10uF Wima PP caps as leftovers from another project. Is there a reason not to use them as output caps ?
Overkill, sure, but can the value get too large for some reason ?

 

[ruffrecords]

Hi !

Got a couple of questions again.

When using this for gain make up for a passive mixer network (as in the posts above) - could one go straight to the tube at the 1M resistor after the transformer ? I obviously don't need the pad as I don't have a bypasswitch. I suppose I need the 470K trimmer if I want to adjust the volume between the different channels, but can I just omit it otherwise ?

For gain make up for a passive mix bus you can certainly go straight in across the 1M from grid to 0V. I often put the master level control here too. So your mix bus is connected to the top of the master level pot and its wiper goes the grid/1M as before.

I happen to have some 10uF Wima PP caps as leftovers from another project. Is there a reason not to use them as output caps ?
Overkill, sure, but can the value get too large for some reason ?

They should be fine as long as their voltage rating is sufficient. Remember these could see the full supply voltage worst case so I normally use 400V rated parts.

Cheers

Ian