Tube simulated inductor EQ question

So I've been playing around with the circuit attached. The problem with it can be seen in the frequency graph. The frequency range above the bell is boosted or cut a couple dB's along with the bell. Below the bell is fine. And its additive as you boost more frequencies.

I've replaced the tube with a regular inductor and it works as expected with the frequencies above the bell being flat.
I thought it might have to do with miller capacitance and tried it with a pentode but no different than the triode.

I've tried all different resistor values and capacitor values. tried shunting grid to ground, to plate, to cathode with small capacitance but it only seemed to make things worse.

Any tube guys out there understand how this works and have any ideas on how to bring that high end down?
Thanks!

Could it be that the mid and hi bands are interacting with each other?
What happends when you only boost/cut the hi band..?

Thanks 5v333. No thats not the problem. As you can see in the graph with just one band active the high end above it still is boosted a couple db like there is an added high shelf. It has something to do with the simulated inductor "leaking" high end to ground. the surrounding tube circuit is basically just an opamp. So the boost/cut pot's two ends are connected to the + and - of the op amp.

It works just like the circuit below but with tubes. The simulated inductors just arn't working properly...

bluebird said:

So I've been playing around with the circuit attached. The problem with it can be seen in the frequency graph. The frequency range above the bell is boosted or cut a couple dB's along with the bell. Below the bell is fine. And its additive as you boost more frequencies.

I've replaced the tube with a regular inductor and it works as expected with the frequencies above the bell being flat.
I thought it might have to do with miller capacitance and tried it with a pentode but no different than the triode.

I've tried all different resistor values and capacitor values. tried shunting grid to ground, to plate, to cathode with small capacitance but it only seemed to make things worse.

Any tube guys out there understand how this works and have any ideas on how to bring that high end down?
Thanks!

Click to expand...

You don't specifically need a "tube guy" to understand the issue. It's a network analysis issue.
This problem is quite common in simulated inductors. At HF, the circuit becomes an RC series circuit, constituted of the grid resistor and the input cap.
The common way of solving the problem is to increase the grid resistor. There is no reason for making it so low. You can increase it tenfold without any issue. Increasing it more is possible but you may run in grid current issue that would change the bias point.
When changing the grid resistor, you must change the 2nd cap (the one that goes to the grid) accordingly in order to maintain the tuning frequency.
I would think the proposed schematic comes from a cookbook and untested.

Thanks Abbey, yes I've changed the value of the grid resistor and related cap. It is true that the high end does go down but so does the amount of boost the bell can achieve. Seems the extra high end shelf and bell are linked in a way.

bluebird said:

Thanks Abbey, yes I've changed the value of the grid resistor and related cap. It is true that the high end does go down but so does the amount of boost the bell can achieve. Seems the extra high end shelf and bell are linked in a way.

Click to expand...

That's because the tube's output impedance interferes (it would not with an opamp). The amount of boost/cut is primarily governed by the 10k resistor. You may want to try a smaller value and adjust the other cap for re-adjusting frequency.
I would suggest you simulate the circuit with LTspice for quicker assessment.

abbey road d enfer said:

That's because the tube's output impedance interferes (it would not with an opamp). .

Click to expand...

I tried to buffer the grid with an op amp actually and still couldn't get it to flatten out! I've never tried to use LT spice. Perhaps its time...

bluebird said:

I tried to buffer the grid with an op amp actually and still couldn't get it to flatten out! I've never tried to use LT spice. Perhaps its time...

Click to expand...

I would think it's the output that needs to be buffered.

bluebird said:

I've never tried to use LT spice. Perhaps its time...

Click to expand...

Did a quick sim. It seems changing to the attached circuit would improve interaction significantly.

Wow thanks Abbey! I will try on the breadboard today and report back...

Does using the 12AT7 follower make any difference...... ?

scott2000 said:

Does using the 12AT7 follower make any difference...... ?

Click to expand...

Simulation shows imperceptible.

> Does using the 12AT7 follower make any difference...... ?

All the usual suspects, at the same current, will have very-very similar cathode impedance.

12AX7 may be better at 1mA. 12AU7 is better at 10mA which a 12AX7 won't do. But there's rarely a need to run 10mA in any small-signal audio. And you don't want a filter-bank to eat more power than your power amp.

As seen in the graph the new circuit didn't change the outcome. I had a feeling because I tried changing around values for hours with no success.

The only reason I wanted to try this is that the tube simulated inductor will add distortion that an actual inductor won't. Theres probably another way to do it.

scott2000 said:

Does using the 12AT7 follower make any difference...... ?

Click to expand...

I'm actually using a regular NE5534 op amp instead of the tube circuit for the amplifier...

So a takeaway is that this was an overlooked issue in the original design? 

i don't know since I never did it with the tube op amp circuit. I was just interested in the simulated inductor part. But like Abbey said, it was probably never really used as a complete working circuit but more of a breadboard idea. it would probably sound cool as is if you don't mind a lot of high end, or lack there of ;D

:-\

Fascinating ...  I've always been an 'AngelFire'  [tube and radio guy]  fan myself

I always wanted to try this particular circuit, which he spoke of as being 'scaled' from an opamp  eq  application  ...


..  but I stayed with my first 'angelfire'  build, which was his  tube based 'bax type eq' circuit, also done with '3 bands'.

As always is the case, my own version varied somewhat from the published tutorial, and toward the 'downside' ; but it did really get me going on the circuit topology [which was very motivating and so on].

I hope you get the performance going ...  my own questions would be related to more detail on the implementation and how you are measuring.  I only say because I'm interested, and that measurement techniques can have a bearing on the situation.

This particular  circuit in tubes has always been on my 'to do list' 

..

And just on a related topic, anyone ever done that 'other tube eq' ?  ...  the 'Phutney-Creech'  [sic]  tube graphic eq with bundles of ax7 s   

..

And the other question I would have, is 'cant we get some inductors in there somehow ? '  [shielded of course]

Good luck with your endevours ..

Yes I've seen that article. So the circuit is found in two different schematics but it really doesn't work without the high end being added. I've given up. I mean it actually sounds pretty cool with the added high shelf but it a bummer you HAVE to have it along with the bell.

I'm measuring it with an Audio Precision system one. I'm sure I got all that right. its just the nature of the beast.

alexc said:

And the other question I would have, is 'cant we get some inductors in there somehow ? '  [shielded of course]

Click to expand...

Yes actual inductors work perfectly. So if you just want to use the surrounding tube amplifier with an inductor instead of the tube simulated one it would be great. But the whole point for me was to get the sound of the tube simulated inductor in there and all its tubey eccentricities. But not the high end boost. Oh well. I will leave it to someone else to carry on, and on, and on...

A Mosfet follower would have higher Gm and lower output Z.  Granted you would want to source it with more than your 1mA, but it would be closer to the ideal Z than the usual valve suspects. 

Edit:  or am I a day late and another dollar short having misunderstood the issue?