Low-distortion Audio-range Oscillator

[1audio]

I pulled my other Shibasoku oscillator, 590AR, from storage and took a photo of its insides. This is the top view. The bottom has the frequency and level networks. Both this the AG15 and AG16 are remote controllable. This older one uses a parallel interface, the newer ones GPIB, otherwise they are identical.

There are several opamps, all TL07X types. There is also an LM710CH (The can) and an MC1495. I suspect the AGC uses the small metal can device 2SK15 (??). The tree trimmers are +1 dB, 0 dB and -1 dB for some function. My guess would be a steering for the AGC so the active part doesn't need much range to stabilize it. The supplies are +/-30V.

[davada]

Very nice.
What's the op amp in the can?

[1audio]

And here is why you won't want to duplicate this thing.

There are a jillion PCB's and each have a bazillion relays. When Richard gets the manual calibration should be simple since the key nodes and adjustments are clearly marked in good English.

The detector board is actually the input board. The device in the middle is a dual Jfet I think. The relays handle the input attenuation.

Design was an enormous undertaking but its still in the lineup 30 years later.

[davada]

[QUOTE=1audio;3393204]And here is why you won't want to duplicate this thing.

There are a jillion PCB's and each have a bazillion relays. When Richard gets the manual calibration should be simple since the key nodes and adjustments are clearly marked in good English.

The detector board is actually the input board. The device in the middle is a dual Jfet I think. The relays handle the input attenuation.

Design was an enormous undertaking but its still in the lineup 30 years later.[/QUOTE

Considering all the board artwork was done by hand too.

[RNMarsh]

Originally Posted by 1audio

I pulled my other Shibasoku oscillator, 590AR, from storage and took a photo of its insides. This is the top view. The bottom has the frequency and level networks. Both this the AG15 and AG16 are remote controllable. This older one uses a parallel interface, the newer ones GPIB, otherwise they are identical.

There are several opamps, all TL07X types. There is also an LM710CH (The can) and an MC1495. I suspect the AGC uses the small metal can device 2SK15 (??). The tree trimmers are +1 dB, 0 dB and -1 dB for some function. My guess would be a steering for the AGC so the active part doesn't need much range to stabilize it. The supplies are +/-30V.

It appears the oscillator(s) are all descrete transistors design.

[1audio]

I just posted the first rev on the QA400 interface here: QuantAsylum QA400 We should probably move discussions about the QA400 there.

[RNMarsh]

Attached are some harmonic pictures of 4 ultra-low distortion oscillators. The noise from the oscillators is between 140-150dB (no averaging used -- eg FFT). The ShibaSoku 725 notch shows itself at -155-160dB. All Harmonics are below -130, worst case harmonics. Typically -140 or more. The AG16 looks like it has higher harmonics but its just the noise floor is 10dB lowe than the others... and fairly flat (not rising at lower freq end).
Top left is Viktor's 1Khz, Top right KH 4402B, Bottom left is HP 339A, and bottom right is AG16. All excellent. However the HP339A cannot keep its lowest thd at other freqs as well as the others without individual fine tuning.... Thx-RNMarsh

[davada]

Hi Guys,

I've done some mods to the EMU 0204 to improve dynamic range and lower the measurement noise floor. I started a new thread for this because others have commented to me that it's too difficult to plow through hundreds of post to find things. As it is for the work on the 339A.

Have a look.

EMU0204 mods for FFT measurements

Cheers.

[RNMarsh]

Great, Davada --- More spin-off's. Demians work can also apply to EMu or others. I am interested in both. I think the HP339A is done. It is what it is. Can be useful to many. But we would need a notch filter and FFT. All together, with a low cost ultra-pure oscillator source, a DIY'er can measure without breaking the bank. Thus, these spin-off forums. Some already before us.

We might summarize where to find the How-To changes for the 339A... which line #'s. Then, it wont be so arduous for others as it was for us to do it first for them.

And, I have reached my goal of at least -140 THD+N or better but thru other hardware and more $$. I wouldnt have got to that point without first exploring the 339A which lead to my own spin-off directions.

Thx all -- Richard Marsh

[Telstar]

I discovered this project last night and I think it can be of extreme interest here:
An Audio T.H.D. Analyser

I don't think the author follows this forum.