I can point at two models for this.
The Boonton oscillator uses a digital loop PLL feeding onto the analog multipliers to tune it and it has 4+ digit tuning accuracy and crystal stability. it can be done but the analog multiplier is the challenge. Davada's multiplying DAC concept may be the answer for this.
The Shibasoku uses 3 cascaded bridged T notch filters and are externally tuned with a jillion (technical term) relays. No active tuning elements. The distortion floor is very low and that is with their input amp/impedance converter and active feedback on the notch filters. The notches are wide enough and stable enough to get 130dB + attenuation for the fundamental. Its very difficult to know if the distortion floor of the Shibasoku is from the input amp (all discrete) or the opamps in the filters. I do know that upgrading the opamp in the feedback for the first filter to a now unobtainable LME49990 made a significant difference. none of the other opamps accomplished this.
When looking at these really low distortion figure you need to be sure you aren't dealing with harmonic cancellation effects, which can alter distortion levels a lot. Ideally you want a 10 dB margin between the DUT and the measurement system but in this case how can you know?
I have an LTspice models of the Shibasoku if anyone wants to play with it. I was trying to reverse engineer calibration process etc.
My point was that its no longer in production. The OPA1612 is not a direct replacement although similar despite TI's claims. The datasheets are not quite the same conditions and the LME49990 has a GBW of 110 MHz and similar conditions for the OPA1612 are 80 MHz. And input characteristics are not fully spec'ed.
That is true, unfortunately.
I tried a number of op-amps in my Audio Analyzer, including the OPA1611. The performance was not as good as with the LME49990. With one exception. The AD797 performance was very similar to the LME49990 (once I got the AD797 stable!).
I tried a number of op-amps in my Audio Analyzer, including the OPA1611. The performance was not as good as with the LME49990. With one exception. The AD797 performance was very similar to the LME49990 (once I got the AD797 stable!).
The AD797 performance is just as good but a two to three times the cost CAD.
The LME49990 is essentially the LME4562 on steroids.
The LME49990 is essentially the LME4562 on steroids.
Last edited:
Hi,
I'm guessing CAD might mean something other than Computer Aided Design in this context. I'm probably being dense here--- would you educate me? Thanks.
The AD797 sure does get expensive in a hurry if you want to sprinkle them around in distortion-critical spots.
The story I heard goes like: "Well, if no one comes up with anything else, I'm gonna put all the letters of the alphabet into this hat, and the first three letters I pull out will be the name of this new country."
"C, eh?"
"N, eh?"
"D, eh?
Measured the OPA1656 in this schematic-
https://content23-foto.inbox.lv/albums/v/viccc/PhonoSC/OPA1656-5Vrms.jpg
The winner.
🙂 Now I am surprised.
More measurements with opamps as followers before the passive notch. 1kHz 2,5V RMS at the follower input. The source impedance 600 ohm. The scale is properly adjusted at 3kHz. Needs to add +4dB at 2kHz.
Reference measurement without the follower before the notch-
https://content23-foto.inbox.lv/albums/v/viccc/Followers/Ref.jpg
With followers:
TL072-
https://content23-foto.inbox.lv/albums/v/viccc/Followers/TL072.jpg
OPA2134-
https://content23-foto.inbox.lv/albums/v/viccc/Followers/OPA2134.jpg
LM4562-
https://content23-foto.inbox.lv/albums/v/viccc/Followers/LM4562.jpg
NE5532-
https://content23-foto.inbox.lv/albums/v/viccc/Followers/NE5532.jpg
... and the winner OPA1656-
https://content23-foto.inbox.lv/albums/v/viccc/Followers/OPA1656a.jpg
Last edited:
New friends: NE5532 & OPA1656! I will order a few OPA at Mouser...
Does the THD improve when you lower the 600 ohms to 300 or 50?
Last edited:
Redundant info makes for confusion ;-)
It is also 2 or 3 times the cost in KZT or RMB 😎
Jan
The OPA has to be completely measured. With different supply voltages, input levels, loads, source impedances, gains, non-inv, inv. 🙂
impressive opamp, that 1656, thanks for caring out the measurements. Did you compare those pampas when driven with a higher source impedance like 60k ? That way, the common mode distortion becomes more apparent, will the different results stay in the same league ?
Do I surmise correctly that your oscillator now uses OPA1656 in the reference trace you posted in #8033? Reference measurement without the follower before the notch-
https://content23-foto.inbox.lv/albu...lowers/Ref.jpg
Your distortion residuals are astonishing! Since it's absent in the oscillator output, the very small distortion shown in the last voltage follower trace probably arises from common-mode distortion errors? They are not present in the oscillator because the opamps are inverting?
Outstanding work!
for those that missed it, here is a thread on the 1656, lead by one of the developers.
OPA1656: High-Performance CMOS Audio Op Amp
Cheers
Alan
OPA1656: High-Performance CMOS Audio Op Amp
Cheers
Alan