Auto-tune modifications?

[KingJL]

I think I have discovered the one modification that greatly improves the operation of the HH. An active filter, 2 pole Bessel with 3.5hZ cutoff, followed by an inverter (to get the signal back to the correct polarity for the autotune circuit), placed after the output of the the differential integrator (IC8a). The output of the inverter feeds the existing autotune circuit. Solves all my instability problems and give smooth operation. Sweeeet!

Used a TL072 to implement the two stages (active filter and inverter). The values for the active filter used were determined by FILTER PRO (see attachment).

[KingJL]

A clearer attachment!

[Tinkerer]

The TI Filter Pro is a great tool and the low pass filter after the integrator reduces the noise considerably.

Thanks for the tip

Tinkerer

[Eugene51]

Auto-tune modifications?

KingJL , I have not seen you on Getech Forum in a few weeks , Hope you are doing O.K !!! I was wondering if you posted the "Final Modifications" on the Hammerhead improovements you made . Also , I remember you corrected something with the "Corner Frequency" . And I can not remember if you posted final images of your prototype Hammerhead machine ??? Thank-You , for any information you might have.........Take care.........Eugene51

[KingJL]

KingJL , I have not seen you on Getech Forum in a few weeks , Hope you are doing O.K !!!

I am here and doing well!

I was wondering if you posted the "Final Modifications" on the Hammerhead improovements you made . Also , I remember you corrected something with the "Corner Frequency" .

Nothing since adding low pass just before autotune circuit. Did raise the corner up to 9Hz.

And I can not remember if you posted final images of your prototype Hammerhead machine ???

Haven't taken any more pics of the beast. With the number of conections between boards, I try not to take in and out of the 'can' too much.

Right now, I am studying different approaches to MD implementation.

Regards,
J. L. King

[Tinkerer]

Low pass filter

I am here and doing well!

Nothing since adding low pass just before autotune circuit. Did raise the corner up to 9Hz.

Haven't taken any more pics of the beast. With the number of conections between boards, I try not to take in and out of the 'can' too much.

Right now, I am studying different approaches to MD implementation.

Regards,
J. L. King

Glad to see you on the forum again.
I made a similar experience with the low pass filter. Upped the corner frequency to 7Hz
Got some oscillations when using the Sallen & Key configuration, so changed to MBF single ended and it works fine now.

Got some problems with using the AD797 for pre-amp, so for now use OP37, sampling at 3us.

Tinkerer

[KingJL]

Got some oscillations when using the Sallen & Key configuration, so changed to MBF single ended and it works fine now.

I tend to stay with MFB Butterworth or Bessel models. The Bessel has the smoothest response of the two.

Got some problems with using the AD797 for pre-amp, so for now use OP37, sampling at 3us.

I feel that the AD797 is only viable in a 2 stage pre-amp configuration. And then the gain of the AD797 must be kept below 10 (about 8x is ideal). If the gain is kept below 10 you can end up with a better noise performance than with a 5534 (about 1/6th the noise of a 5534). and still have the bandwidth that will not destroy the decay slope (at a gain of 10x the BW is 1.1MHz). Anything above 10x really limits the upper side of the BW.

I have been experimenting with a OP847. It exhibits about the same noise charactaristics as the AD797 but has much higher GBW. The noise charactaristic at very low frequencies (<1000Hz) is not so great, but can be canceled out with a HPF in the input biasing. I have run it with gains up to and including 1000x.

[Tinkerer]

I tend to stay with MFB Butterworth or Bessel models. The Bessel has the smoothest response of the two.
I feel that the AD797 is only viable in a 2 stage pre-amp configuration. And then the gain of the AD797 must be kept below 10 (about 8x is ideal). If the gain is kept below 10 you can end up with a better noise performance than with a 5534 (about 1/6th the noise of a 5534). and still have the bandwidth that will not destroy the decay slope (at a gain of 10x the BW is 1.1MHz). Anything above 10x really limits the upper side of the BW.

I have been experimenting with a OP847. It exhibits about the same noise charactaristics as the AD797 but has much higher GBW. The noise charactaristic at very low frequencies (<1000Hz) is not so great, but can be canceled out with a HPF in the input biasing. I have run it with gains up to and including 1000x.

Thanks for the input.
At present I am using a single preamp with a gain of 1000 non-inverting. In this configuration the OP37 works fine, the LT 1037 also but with some tweaking.

Combining the LP filter with a HP filter to zero the output, after the S&H, it looks as if a Chevychev LP would give the best bandpass.

What are the disadvantages?

Tinkerer

[Carl-NC]

What's the OP847?

[KingJL]

What's the OP847?

OOPs, my mistake!! Correction, OPA847. Data Sheet attached. I got mine from TI (free sample).
Regards,
J. L. King

[Tinkerer]

OPA847

OOPs, my mistake!! Correction, OPA847. Data Sheet attached. I got mine from TI (free sample).
Regards,
J. L. King

Thanks for the information. Looks very promising.
Tinkerer

[Carl-NC]

Nice looking opamp, except for the supply current.

[KingJL]

Nice looking opamp, except for the supply current.

Yes, that seems to be the downside of all ultra low noise solutions that I have looked at so far. Another solution is 2 or 3 MAT03's in parallel (see attached, figure 3a, page 7) feeding an op-amp, but here again the downside is supply current (2ma for each transistor plus the op-amp) as well as total footprint real estate. The tradeoff seems to be high current for ultra-low noise front end and use ultra low current devices (e.g. TLC07x, TLC08x) elsewhere to make up the difference.

Regards,
J. L. King

[cko47]

Could you show it (HPF ) on the circuit,
In version Hammer Rev D1 is this HPF(R22 & C16). What value should be for R22 &C16?
Excuse me. I have those problems with my HH You wrote ..... TP 5 indicates plenty of dynamic range remains (that follows target strength) between 0 and +5 left even after the signal at tp9 becomes erratic.

[cko47]

1 K and 150 pf for R22 and C16? And how is purpose of Threshold?