Tube parametric eq-add cut and boost?

[valvusmusicus]

Found this the other day-any idea what the control in the middle, between the 2 valve stages does? And how would you add gain (cut and boost) to the circuit please?

[Chuck H]

That is a silly circuit with way too many phase complications. NEVER include this circuit within a feedback loop. There are plenty of good circuits that don't require sixteen dollars worth of capacitors to execute. I think the unmarked control in the middle is probably a phase error circuit. Intended to remain inside the chassis and not user adjustable.

[Steve Conner]

The two stages look like low and high pass filters. Distant cousins of the tube tremolo oscillator. I guess extra circuitry is needed to make them into an equaliser, and you would have a choice of a shelving response, or a bandpass with the upper and lower limits set by the two sections. In both cases the Q controls add bumps around the cutoff frequencies.

It's not really the same as a modern parametric EQ, which is based on bandpass building blocks rather than lowpass or highpass.

The biggest problem I can see is that it seems to need dual gang pots with one gang 250k and the other 500k.

[valvusmusicus]

A less "silly" circuit

That is a silly circuit with way too many phase complications. NEVER include this circuit within a feedback loop. There are plenty of good circuits that don't require sixteen dollars worth of capacitors to execute. I think the unmarked control in the middle is probably a phase error circuit. Intended to remain inside the chassis and not user adjustable.

Thanks for clearing that up a little more
Perhaps you could suggest a less silly circuit-I am only a newbie to this, and could do with some food for thought I think

[Steve Conner]

If Chuck thinks your circuit is silly, this one will give him a heart attack!
It's my own design, and it works pretty well. Only one band is shown. The only limitation is, the input voltage can't be more than about 20V p-p, or the op-amps will clip.

It has a midband gain of 10 and the output was designed for driving a master volume and a Fender-style PI.

The op-amps run off a +30V supply with a 15V "virtual ground", all derived from the PT bias tap.

I made the following mods: Pot in series with R20 for adjustable Q. Q1, R32, R34, C13 all removed.

[valvusmusicus]

If Chuck thinks your circuit is silly, this one will give him a heart attack!
It's my own design, and it works pretty well. Only one band is shown. The only limitation is, the input voltage can't be more than about 20V p-p, or the op-amps will clip.

It has a midband gain of 10 and the output was designed for driving a master volume and a Fender-style PI.

I made the following mods: Pot in series with R20 for adjustable Q. Q1, R32, R34, C13 all removed.

With my limited knowledge-I think it looks very elegant-using the opamps as gyrators-pretty cool

[valvusmusicus]

With my limited knowledge-I think it looks very elegant-using the opamps as gyrators-pretty cool

Sorry-not gyrators-Duuuh!!

[Steve Conner]

It is pretty much the same idea as gyrators. It's just emulating the complete LC circuit, rather than the inductor only.

[J M Fahey]

Not about this specific circuit but in general lines, I would never sweat and suffer trying to make and stabilize complex tube versions of what can be made 1000 times easier and more perfect with Op Amps.
Or even with discrete Transistors.

As Steve said, this is a phase shift RC oscillator, which as most of its kind, if set with *slightly* less gain than needed to oscillate, and injected some audio signal at an approppriate point, will show a peak (or bump) at the frequency at which it *would* have oscillated if allowed to.
Since we are using it at the edge of the cliff, parts tolerance becomes *very* important.
That´s why chuck h spoke about expensive 1% precision capacitors.
I might add it needs *perfectly* tracking pot sections.
In a typical tremolo oscillator this is not so critical, because we are using it fully as an oscillator, but for EQ use, as I said, it has to stay "on the edge".

A *much* better equalizer can be built around a $2 TL074.

In fact, when Tube technology had to be used for complex circuits, they *had* to make Op Amps too!!
This is just *one* Op Amp.
To make a real parametric equalizer you´ll need 3 or 4 of these, plus necessary feedback components.


The first page of its datasheet:

[valvusmusicus]

Not about this specific circuit but in general lines, I would never sweat and suffer trying to make and stabilize complex tube versions of what can be made 1000 times easier and more perfect with Op Amps.
Or even with discrete Transistors.

As Steve said, this is a phase shift RC oscillator, which as most of its kind, if set with *slightly* less gain than needed to oscillate, and injected some audio signal at an approppriate point, will show a peak (or bump) at the frequency at which it *would* have oscillated if allowed to.
Since we are using it at the edge of the cliff, parts tolerance becomes *very* important.
That´s why chuck h spoke about expensive 1% precision capacitors.
I might add it needs *perfectly* tracking pot sections.
In a typical tremolo oscillator this is not so critical, because we are using it fully as an oscillator, but for EQ use, as I said, it has to stay "on the edge".

A *much* better equalizer can be built around a $2 TL074.

In fact, when Tube technology had to be used for complex circuits, they *had* to make Op Amps too!!
This is just *one* Op Amp.
To make a real parametric equalizer you´ll need 3 or 4 of these, plus necessary feedback components.


The first page of its datasheet:

Can the Neons be replaced with Zener diodes of the appropriate wattage?? If so, is about 90v each OK?
Looks like a great idea

[J M Fahey]

Yes, probably so.
I guess "turn on" voltage is 90V, but "sustaining" voltage is 60/65V.
Very dusty memories, but maybe you can google some NE2 datasheet.

[valvusmusicus]

Yes, probably so.
I guess "turn on" voltage is 90V, but "sustaining" voltage is 60/65V.
Very dusty memories, but maybe you can google some NE2 datasheet.

Having a look again at the tube opamp schematic-why is the cathode bypass cap (C2) 500pf? I presume uuF means pf?

[J M Fahey]

Yes, uuFis oldspeak for pF.
They compensated it for flat response over a wide range.
What´s amazing is that being a tube amp it had flat response down to DC !!!!!!!!!!!!!!
If they needed 500pF somewhere, so be it.
Do not even *ask* how much one of them cost.
Just consider the earlier versions were used as "directors" for radar controlled Anti Aircraft guns and then in Nike Missiles control units. !!!!
Maybe you'll enjoy reading this page:
Life before Transistors

[Steve Conner]

Zeners?! If you're building one of those Philbrick things, you have to use the authentic neon lamps. Any neon lamp should do.

Believe it or not, I considered tube op-amps for an equaliser, but I rejected them for the following reasons.

My circuit would need 4 more tubes per band. The way it is, it offers two bands from one dual triode tube.

We use tubes because of the pleasant sounding distortion. But in a tube op-amp circuit the distortion is suppressed by feedback. In my circuit, the tube has no feedback. OK, the second triode has local feedback at the cathode, but when you boost the EQ, this gets bypassed.

I still think my design is a good solution for a full parametric EQ, with non-interacting frequency and Q controls. You could make a semi-parametric more simply: start with a tube graphic EQ (see the Fender 400PS etc) and replace the inductors with gyrators based on high voltage transistors. The frequencies can be adjusted by replacing one resistor in the gyrator with a pot.

[valvusmusicus]

Zeners?! If you're building one of those Philbrick things, you have to use the authentic neon lamps. Any neon lamp should do.

Believe it or not, I considered tube op-amps for an equaliser, but I rejected them for the following reasons.

My circuit would need 4 more tubes per band. The way it is, it offers two bands from one dual triode tube.

We use tubes because of the pleasant sounding distortion. But in a tube op-amp circuit the distortion is suppressed by feedback. In my circuit, the tube has no feedback. OK, the second triode has local feedback at the cathode, but when you boost the EQ, this gets bypassed.

I still think my design is a good solution for a full parametric EQ, with non-interacting frequency and Q controls. You could make a semi-parametric more simply: start with a tube graphic EQ (see the Fender 400PS etc) and replace the inductors with gyrators based on high voltage transistors. The frequencies can be adjusted by replacing one resistor in the gyrator with a pot.

On reflection, and after doing a little more reading-I agree