Simple Inductor EQ, with Tubes

[liquids]

Here we have a simple, op-amp driven way to get midrange peaks and notches (thanks R.G for the article):

http://www.geofex.com/article_folders/eqs/parmet1.gif


My basic question/difficulty is like this...
While  I am not an always 'all tube' kind of guy, for the sake of a friend who may be, and/or for implementation in what may be an otherwise all-tube pre-amp/amplifier (where adding low voltage solid state components becomes a complication), how can one use tubes to otherwise duplicate this?

thoughts:
- initial op amp is just normalizing the signal into a low output impedance to 'drive' the circuitry....a 12AU7 stage (etc) can do this, though not as well.  
- A mosfet like the  IRF820 (per RG) is something that might do this well also, depending on how 'undesirable' that might be (can often work off B+, but it isn't a 'tube').
- the final stage is really the issue - or configuring the circuit to work properly.  

Any links would be a great help!

[slacker]

For the last stage you might be able to build an opamp out of tubes, examples here http://headwize.com/projects/opamp_prj.htm, don't know how well they work in that configuration though. Or have a look how Mesa Boogie did it, think they used trannys.

[liquids]

For the last stage you might be able to build an opamp out of tubes, examples here http://headwize.com/projects/opamp_prj.htm, don't know how well they work in that configuration though. Or have a look how Mesa Boogie did it, think they used trannys.

Thanks - and which mesa used transformers for driving an eq?

[slacker]

Sorry, I meant transistors, like the Mark V, which re-reading your original post you probably dont want.

[PRR]

> adding low voltage solid state components becomes a complication

Nevertheless, nearly all tube-amps with non-trivial EQ turn to opamps for the EQ stage.

The plan you cite: tubes will not drive the 2.7K impedances at all well.

You can scale-up to 27K or 270K, but then your inductors must be obscenely large, and your caps become distressingly small (prone to pick-up signals and radio).

That plan shows coils. Even at 2.7K, this is expensive, R.G. goes on to show "coils" faked with R C and buffer. Same thing: while tube buffers can have nice high input impedance, their output impedance is high enough to upset good coil fakery.

You only need about 10mA of 20V-36V to power the whole EQ (two core amps and 2-4 fake-coil buffers). 300V supply, 27K 5W resistor, 30V Zener. Yes it is wasteful, but in context of a 25W-100W tube amp, another 3W is no big deal.

Or: many large tube amps have a 50V tap on the PT to supply output tube bias. Now you only have to drop from 70V, not 300V. Much less waste.

> mesa used transformers

No; tranSIStors as simple op-amps.
http://www.el34world.com/charts/Schematics/files/Misc/M-B_I_II_III_400.pdf

[KMG]

Ampeg implementation of "tube" EQ
http://milas.spb.ru/~kmg/files/schematics/ampeg/SVT3PRO/534XXHB.PDF

[liquids]

> adding low voltage solid state components becomes a complication

Nevertheless, nearly all tube-amps with non-trivial EQ turn to opamps for the EQ stage.

The plan you cite: tubes will not drive the 2.7K impedances at all well.

You can scale-up to 27K or 270K, but then your inductors must be obscenely large, and your caps become distressingly small (prone to pick-up signals and radio).

That plan shows coils. Even at 2.7K, this is expensive, R.G. goes on to show "coils" faked with R C and buffer. Same thing: while tube buffers can have nice high input impedance, their output impedance is high enough to upset good coil fakery.

You only need about 10mA of 20V-36V to power the whole EQ (two core amps and 2-4 fake-coil buffers). 300V supply, 27K 5W resistor, 30V Zener. Yes it is wasteful, but in context of a 25W-100W tube amp, another 3W is no big deal.

Or: many large tube amps have a 50V tap on the PT to supply output tube bias. Now you only have to drop from 70V, not 300V. Much less waste.

> mesa used transformers

No; tranSIStors as simple op-amps.
http://www.el34world.com/charts/Schematics/files/Misc/M-B_I_II_III_400.pdf

Thanks...it's good to know when a futile path is futile.         Question - would aforementioned HV mosfets offer any easy way to make this happen?

As for the ampeg circuit - I'd probably stick to the original SVT/V-series circuit, for which this is close (and I have a torroidal for a V-series chassis I got for spare parts), though it might be nice to get my hands on a current-production torroidal.  In the end, the 3-tube with an inductor and feedback may be the best way outside of going into a op-amp EQ preamp--> into tube PI/PA.  I've been leaning toward building a particular tube power amp as it's own thing, and then building various preamps (all op-amp, all tube, and then maybe also a hybrid and other various preamps) to drive the  power amp, so I could mix and match.  More of a 'long term fun' kind of thing. All the options are there if I can get my mind into building and the like. 

[PRR]

> Ampeg implementation of "tube" EQ

Interesting; thanks.

The plot you show seems to be the "MID" control, V2a+V3 and L1+caps. This goes back to very early SVT-era Ampegs. It is swinging-outputs, opposed to the swinging-inputs topology seen in the top of this thread and most siliconized graphic EQs.

Ampeg only used that plan for a single knob (actually switchable as shown in your plot). It "could" be extended to multiple bands, but with more loading on the not-so-strong tubes, and possibly more internal loss and thus higher hiss level.

This 1999 implementation sure has grown a lot of frills! On-topic: they did add a Graphic EQ, but totally TL074 chips (sheet 3). Even when tubes are "the thing", chips do the job. The cost of the low-volt supplies is spread over a dozen other silicon thingies added to old-SVT design, for buffering and switching. (And the fact that "SVT" means premium price.)