Tube based VU-Buffer

Help Support GroupDIY:

johnheath

Well-known member
Joined
Jul 31, 2014
Messages
890
Location
Sweden
Hi all…

I have been thinking about a design of a tube VU-meter buffer. I know that there are many well-functioning VU-buffers around OP-amps and all, but I really would like to learn how to do this with tubes if possible at all. There are a quite a few questions to be answered around this but still interesting to do a follow up.

First, as far as I know the need for buffers are reducing THD, proper needle response, avoid diode distortion and more, but would a tube solve these issues to the same extent as an OP-Amp for example? If not, what would the differences be?

I suppose low output impedance is desired, so a cathode follower around a 12AU7 is perhaps a suitable choice… with just slightly below unity gain I could split a triode if two VU-meters are desired and no gain needed. If gain is needed I guess a triode gain stage followed by the cathode follower could work?

A question that I have though is how to couple this tube topology to a balanced output of a tube preamp for instance? I mean the positive of the output transformer to the input of the tube but what about the negative of the output transformer? Does this work at all? Or is this technique just for unbalanced outputs?

I suppose an input transformer in front of the VU-buffer could work but now we are well exceeding the budget for this already somewhat dreamy project solution. I can accept a tube, preferably, in the mentioned “stereo” alternative.

Would other topologies with low output impedance work? Like a Mu-follower or a SRPP? To adjust gain I could perhaps just have a trim pot up front.

And last, how does the output impedance of the tube topology affect the value of the resistor, just before the VU-meter, usually 3k6 or 3k9 (depending on who you ask I suppose)?

I must mention that the meters that I am planning to use (can afford and find) are the Sifam AL series. The smallest one, the AL29 is apparently the one that has the behavior closest to “real” meters.


Best regards

/ John
 

johnheath

Well-known member
Joined
Jul 31, 2014
Messages
890
Location
Sweden
I thought that I could add these theoretical... untested circuits.

Are these completely bogus or could they actually work?

I'd prefer if I could use as few triodes as possible of course :)

Best regards

// John
 

Attachments

  • GROUP DIY Tube VU-Buffer.png
    GROUP DIY Tube VU-Buffer.png
    15.5 KB · Views: 49

Rochey

Well-known member
White Market Member
Joined
Jul 2, 2004
Messages
2,694
Location
Dallas, TX - Formerly UK
Hello John,

I admire your desire to make this circuit, it's important to have a vision and an idea of when the job is complete. you have a complete specification (and therefor goal-line) in mind. Excellent stuff.

But let me ask a dumb question.

Why?

Most people use tubes for the sonic influence they have. In a VU meter, you're unlike to see or visualize such sonic influence.
Your replacing the opamp used - which is small, low power and well known with something that is large, expensive, higher power and offers little to the usage experience.

Aren't there better tube circuits to work on where you can have a sonic influence?

Cheers

R
 

johnheath

Well-known member
Joined
Jul 31, 2014
Messages
890
Location
Sweden
Thank you Sir

Well you are right on target... I also mentioned it as a dreamy project as well :)

The thing is that I usually build my own tube preamps and I've made a lot of bad ones and some really good ones as well. I am just very interested in tube technology and I understand that this buffer wont change or improve the sound of any tube circuit that I have done or will do. I have added a few OP-Amp buffers before but the I end up with creating another PSU for that itself... with the tube and the existing PSU it is a sweet go.

Best regards

/John
 

Newmarket

Well-known member
Joined
Oct 10, 2016
Messages
718
Location
Brighton Sussex UK
Rochey said:
But let me ask a dumb question.

Why?

Most people use tubes for the sonic influence they have. In a VU meter, you're unlike to see or visualize such sonic influence.
Your replacing the opamp used - which is small, low power and well known with something that is large, expensive, higher power and offers little to the usage experience.

Aren't there better tube circuits to work on where you can have a sonic influence?

Cheers

R

What he said  :)
Yes I can see the need to generate voltage rails for op amps but surely that's outweighed by the expense and power consumption (cost / heat)
 

johnheath

Well-known member
Joined
Jul 31, 2014
Messages
890
Location
Sweden
Newmarket said:
What he said  :)
Yes I can see the need to generate voltage rails for op amps but surely that's outweighed by the expense and power consumption (cost / heat)

Yes nothing of this is news to me but in this specific case it started with an idea to build a summing mixer with two tube output amps. The outputs consists o two output transformers and I surely would like to have VU-meter at the output just to be able to roughly see where we are. The summing mixer is intended to primarily be used for live performances.

Anyhow I began thinking about using buffers and why not using tubes even there? I have plenty of space, heat won't be a problem because the tubes are mounted on the outside of the chassis and the other two circuits are not so current thirsty.

Best regards

/ John
 

emrr

Well-known member
Joined
Apr 12, 2006
Messages
7,874
Location
NC, USA
Here's a Daven extended range VU meter

49964736912_1231aeec7d_h.jpg


I thought I'd put that in another thread years ago, you might find that thread with some others. 
 

johnheath

Well-known member
Joined
Jul 31, 2014
Messages
890
Location
Sweden
EmRR said:
Here's a Daven extended range VU meter

49964736912_1231aeec7d_h.jpg


I thought I'd put that in another thread years ago, you might find that thread with some others. 

Interesting. But that seems to be a unbalanced solution right? To me the problem was to find a solution for the balanced output.

Best regards

/ John
 

ruffrecords

Well-known member
Joined
Nov 10, 2006
Messages
13,755
Location
Norfolk - UK
The Daven design is balanced - it has a transformer at the input. Indeed if you want a tube based VU buffer then a transformer input followed by an unbalanced buffer is the obvious way yo go. You do not need anything as sophisticated as the Daven design if you want to drive a true VU meter - you just need something with some gain and enough output current capability to drive the meter.

The simplest solution I can think of is to use a typical double triode output tube like the 6DJ8 arranged as an SRPP. With a quiescent current of about 5mA this will happily drive +30dBV into a 2400 ohm load (which is what a typical true VU meter looks like).

Cheers

Ian
 

PRR

Well-known member
Joined
Jan 30, 2010
Messages
11,144
Location
Maine USA
A VU is an "acceptable" compromise between faults (loading, distortion, scale) and passivity (no amplifier).

To further extend range or reduce loading, put any Line Amplifier in front. But it still gives only a rough idea of the loudest parts of speech/music. It "worked" for the radio-line networks of 1938. Post-war EU/UK broadcasters used the (war-)reduced cost of tubes to make PPM displays, with sensible dynamics and range and loading.


 

johnheath

Well-known member
Joined
Jul 31, 2014
Messages
890
Location
Sweden
ruffrecords said:
The Daven design is balanced - it has a transformer at the input. Indeed if you want a tube based VU buffer then a transformer input followed by an unbalanced buffer is the obvious way yo go. You do not need anything as sophisticated as the Daven design if you want to drive a true VU meter - you just need something with some gain and enough output current capability to drive the meter.

Thank you Ian

Yes I have been thinking about the solution with a transformer up front of an unbalanced VU-buffer....but, would you say there's a way to have the (tube) buffer balanced without the use of transformers>

As described in the post about the summing mixer...I need two meters and thereby two buffers, but I surely would like to skip the need for two input transformers :)

Best regards

/John
 

ruffrecords

Well-known member
Joined
Nov 10, 2006
Messages
13,755
Location
Norfolk - UK
johnheath said:
Thank you Ian

Yes I have been thinking about the solution with a transformer up front of an unbalanced VU-buffer....but, would you say there's a way to have the (tube) buffer balanced without the use of transformers>

/John
Yes, there is. TubeCAD Journal published a circuit to do this some years ago. I have never tried it so I cannot comment on how good it is. Schematic attached.

Cheers

Ian
 

Attachments

  • Vertical Difference Amplifier.png
    Vertical Difference Amplifier.png
    16 KB · Views: 33

emrr

Well-known member
Joined
Apr 12, 2006
Messages
7,874
Location
NC, USA
Simplified version.  Still more complex because it's an extended range meter, -40 to +26 dB.  Note the 6 dB calibration trim between 6SN7 stages.  There are real Weston VU meters in these.  Note also this is a hi-Z bridging input transformer to minimize loading. 

Not suggesting it be copied, but there it is. 

As far as I can tell, the desire to buffer is entirely to reduce distortion in the audio path, outside of any ability to select meter range.  The original bridging resistances reduced distortion to what was considered acceptable at the time. 

Haven't thought through this, but why does it need to be balanced?  Take one leg only for input, compensate. 

the Bell System specification called for a meter system input impedance of 7500 Ω, made up of a 3900 Ω instrument, a 3900 Ω attenuator (if used), and a 3600 ohm series resistor.  Commercially-made volume indicators usually have this 7500 ohm input impedance.

The original 1942 Standard recommended a source impedance of 300 Ω (that is, a circuit impedance of 600 Ω terminated by a load of 600 Ω) but the “dBm”, “vu” and Standard Volume Indicator are not now (and never were) restricted to 600 ohm circuits by ANSI or IEEE.

  - Jay McKnight, Magnetic Reference Laboratory - from the AES archives

Note the 7500 Ω cathode resistor at the meter.   

A typical passive VU meter panel would use a 7100/3900 Ω 'T' attenuator to allow for greater input levels, with a 100Ω series calibration trim pot.

I'm still looking for a VU meter attenuator cookbook chart, I know I've seen one somewhere.  For those times you want to throw together the correct 3 resistor T directly to a meter. 

49968546117_9cb4152134_h.jpg
 

johnheath

Well-known member
Joined
Jul 31, 2014
Messages
890
Location
Sweden
ruffrecords said:
Yes, there is. TubeCAD Journal published a circuit to do this some years ago. I have never tried it so I cannot comment on how good it is. Schematic attached.

Cheers

Ian

Thank you Ian

Yes that seems to be a start with it. I guess it needs a little elaboration and head scratching  first :)

I drew this out of my head... not yet tested at all. two cath followers ... just slightly less that unity gain... do you think it could work?

Best regards

/John
 

Attachments

  • GDIY VU-Buffer.png
    GDIY VU-Buffer.png
    7.5 KB · Views: 32

johnheath

Well-known member
Joined
Jul 31, 2014
Messages
890
Location
Sweden
EmRR said:
Simplified version.  Still more complex because it's an extended range meter, -40 to +26 dB.  Note the 6 dB calibration trim between 6SN7 stages.  There are real Weston VU meters in these.  Note also this is a hi-Z bridging input transformer to minimize loading. 

Not suggesting it be copied, but there it is. 

As far as I can tell, the desire to buffer is entirely to reduce distortion in the audio path, outside of any ability to select meter range.  The original bridging resistances reduced distortion to what was considered acceptable at the time. 

Haven't thought through this, but why does it need to be balanced?  Take one leg only for input, compensate. 


Thank you Doug

Yes, actually I have been thinking about it and I guess there's little practical use for reading the meter level after the output transformer compared to right before... thus I can use an unbalanced version of a buffer.

Compensation? That would require some sort of PAD? Or?

Best regards

// John
 

mjrippe

Well-known member
Joined
Dec 27, 2011
Messages
2,564
Location
Brooklyn, NY
johnheath said:
Yes, actually I have been thinking about it and I guess there's little practical use for reading the meter level after the output transformer compared to right before... thus I can use an unbalanced version of a buffer.

As long as you can adjust for any level loss from the output transformer it should work fine.  However you will be starting at a higher impedance node, so loading may again be an issue.
 

johnheath

Well-known member
Joined
Jul 31, 2014
Messages
890
Location
Sweden
mjrippe said:
As long as you can adjust for any level loss from the output transformer it should work fine.  However you will be starting at a higher impedance node, so loading may again be an issue.

Thank you Sir... good point.

I'll have that in mind.

Best regards

/ John
 

Tubetec

Well-known member
Joined
Nov 18, 2015
Messages
3,111
The Brenell STB tape deck had a relatively simple meter buffer based on the EABC80 tube driving an edgewise meter .


Not sure if it ticks all the boxes for you or not , but it does have simplicity in its favour . I would tend to agree with using tubes where they matter most, in the audio signal path . Still if you really want tube driven meters ,why not .
 

MidnightArrakis

Well-known member
Joined
Sep 25, 2015
Messages
103
[if you really want tube-driven meters, why not] -- Sometimes you just want to do something because you - DON'T - have any reason to do so!!!

/
 
Top