D'lator NFB LTSPice simulations: true or artifact??

[markusw]

Hi all,

did some LTSpice simulations trying to understand the NFB loop at the recovery stage.

As far as I understand the attenuation by the NFB loop is defined by the ratio of the output impedance of the previous stage and the 20M of the NFB loop.
With no fx in the loop the output impedance should be the output impedance of the cathode follower plus the series resistance of the send pot (in total roughly 250k with a 250A pot set to half, or about 25k with the send pot at 100%).

With an fx unit in the D'lator the output impedance should be much lower than the 250k, basically the output impedance of the fx unit. For the simulations for simplicity I only added an opamp buffer into the D'lator's send/return using an input impedance of 1M and an output impedance of 100R (basically the values of the Suhr Minimix II kindly provided by bluesfendermanblues in the schematics section). Thus with an fx unit in the loop, the setting of the send pot shouldn't influence the function of the NFB loop.

The images show the frequency response of the D'lator at the recovery out pot after the recovery stage.

The top image shows the frequency response with the send pot at half.
The lowest curve (blue n016) is the D'lator with no fx in the loop. The curve above (green n049) is the D'lator with no fx in the loop again but lacking the NFB . The NFB clearly causes an attenuation of the signal.
The two curves on top (red n083 and light blue n119) show the D'lator with the opamp buffer of the Suhr Minimix II in the loop +/- NFB loop. The low output impedance of the Suhr Minimix II prevents the attenuation by NFB loop.

With the send pot at 100% (lower image) the NFB doesn't significantly attenuate the signal even in the absence of the opamp buffer.

Do I miss something here?
What do you guys think?

Peace,

Markus

[markusw]

Here's a cleaned up schemo of the simulation (D'lator with one preceding triode stage including opamp buffer in the loop).
So the simulations include one triode stage in front of the D'lator.

Hopefully, this makes it easier to follow my previous posting.

Peace,

Markus

[markusw]

No one?

Is it that dumb/ridiculous ?

Peace,

Markus

[bluesfendermanblues]

Markus, this is great work.

I dont know The SPICE software, so I cannot chime in.
How about putting pink noise through a dumbleator + minimix and check The output signal ón a spectrum analyzer (SW) ?

[markusw]

Claus, thanks a lot for your answer!
I already thought about testing it in reality with pink/white noise. At the moment, I don't have my D'lator at home. It's in our rehearsal room.
Next time I'm there I will try to take it home for a couple of days provided our guitar players don't need it.

What do think about the theoretical stuff? Am I right that the signal attenuation by the NFB loop depends on the output impedance of the previous stage?

Regards,

Markus

[bluesfendermanblues]

Regarding The theory, my impression of the dlator is that it Rolls off a little bit of low end at the input (50n / 1027k). And that the second half, with the lnfb also rolls off bottom end - plus it adds a bit of compression if you have the recovery pot up high (mine is always maxed).

I haven't though about the effect of the output impedance of the previous stage.

[gilgalad101]

What do think about the theoretical stuff? Am I right that the signal attenuation by the NFB loop depends on the output impedance of the previous stage?

Yes, you are right. I believe you could model this triode stage as an inverting op amp with feedback. In this case, the voltage gain would be the feedback impedance divided by the output impedance of the previous stage. I know this won't be the exact voltage gain, but it at least shows that the voltage gain is dependent on the output impedance of the previous stage. The Aiken Amps site has a pretty good article on single-stage inverting feedback amps: http://aikenamps.com/FeedbackAmp.htm

[markusw]

Yes, you are right. I believe you could model this triode stage as an inverting op amp with feedback. In this case, the voltage gain would be the feedback impedance divided by the output impedance of the previous stage. I know this won't be the exact voltage gain, but it at least shows that the voltage gain is dependent on the output impedance of the previous stage. The Aiken Amps site has a pretty good article on single-stage inverting feedback amps: http://aikenamps.com/FeedbackAmp.htm

Thanks a lot! Interesting read!

[glasman]

Regarding The theory, my impression of the dlator is that it Rolls off a little bit of low end at the input (50n / 1027k). And that the second half, with the lnfb also rolls off bottom end - plus it adds a bit of compression if you have the recovery pot up high (mine is always maxed).

I haven't though about the effect of the output impedance of the previous stage.

The results I have measured with the cathode follower are flat response from about 10hz up. Remember that the CF is a bootstrapped section and the input impedance is multipled many times over the 1Meg / 27K combination. Upwards of 15Meg or so. So the CF input break point is very low.

You are also correct, the recovery amps break at about 1000 hz and fall off from there. Substituting a 12AT7 will give a MUCH broader frequency response, which may or may not be wanted. But it is an option.

Also dumping the 220K/500pf and FB loop can give a virtually transparent loop. Changing the output section of the CF to drop the output impedance also helps to widen and flatten the overall response. This is nice when you are incorporating digital effects internally in the box.



Gary

[markusw]

Hey Gary,

thanks a lot for sharing your observations!

Did you by chance ever check the attenuation by the NFB loop with some fx unit connected to the D'lator?

Peace,

Markus