Low-distortion Audio-range Oscillator

Do you have the B1 oscillator in the UPL? I wonder what you get from that with 2VRMS / 1KHz? My UPVs vary between -112 to -115dbV (they average -112 to -114, but one of them will hit -115 if the wind is blowing in the right direction). The SYS-2722 can get -115dBV when feeding the UPV.

The HP339A performance is stellar for such an ancient unit.
 
  • Like
Reactions: anatech
Hi thermionic,

the THD and THD+N measurements I took yesterday were from the HP339A generator. Unfortunately my UPL does not have a B1 low distortion oscillator option. This was too expensive for me as a hobby device. The UPL with B1 option still costs over 6,000 euros today. Not to mention the UPV with B1, which costs > 10,000 euros. The DAC oscillator from my UPL delivers about -112dB THD and -105dB THD+N at 1kHz 1-10V RMS. But I can also measure up to -118dB THD+N with my UPL analyzer (with an external low distortion oscillator connected, if I go up to just over 3V RMS input voltage, i.e. manually set the input sensitivity to 3V). For example, I measured the oscillators I developed myself. In fact, they deliver up to 4 dB better THD+N, i.e. around -122dB (at 1KHz-3V 22kHz BW) see also my post on page 506 in this thread, where I had an oscillator of mine measured with the APX555. If you measure the UPV B1 generator with an APX555, it could well be that it also achieves a THD+N of perhaps -116 to-118dB.

Best regards
Helmut Sell
 
  • Like
Reactions: thermionic
My experience with the UPV-B1 is that the oscillator averages 1.5-2dB worse THD+N combined than the SYS-2722. My worse UPV-B1 struggles to get -112dB THD+N, but the best can achieve -115dB (the other two get around -113dB - which is what I've always expected from a UPV-B1). The SYS-2722 gets a solid -115dB THD+N, but can nudge -116dB on the right day. I get the impression that the input section on the UPV is capable of better THD+N than the B1 oscillator. I also get the impression that THD+N specs for the UPL and UPV are pretty much the same. I suspect that they focused on features when the UPV came out and didn't feel the need to get into a race for THD+N. If R+S ever released a new AA I would have G.A.S. I don't get G.A.S over the 555 TBH. Too many add-on expenses. It seems like a wallet-gouging activity to me, great though it is.

NB - there is definitely a little LF noise inherent to the UPV... I have always wondered if putting the PSUs in an external enclosure would do anything... (there are 2: a large linear board, and a SMPS)
 
The reason is more the time it was designed and the components available. That's "age" in this case. We've learned more ad can design better oscillators today. However, what HP did is truly top of class given it was mass produced as well. All this 50 years ago approx.

I have a couple 339A, and 33xA THD meters plus 65xA oscillator models also. Most other brands don't come close.
 
I have just fired up an HP 339A that does not appear to have ever been opened. Calibration seals are still intact. Pots + switches are noisy, and need a little nursing to give best results. Note that this UPV has the worst digital oscillator of any of our UPVs - they usually get better than this (it seems to give a better THD+N than THD alone, which is a little odd). The B1 oscillator is the optional low-THD option and UPV works nicely in this mode. All measurements are taken via UPV inputs (I haven't bothered testing the inputs on the 339A TBH).

@ 1KHz

THD @ 2VRMS UPV standard oscillator = 105dB

THD @ 2VRMS HP339A = -112.5-113 dB

THD @ 2VRMS UPV-B1 = -118dB


THD+N@ 2VRMS UPV standard oscillator = 106.5dB

THD+N @ 2VRMS HP339A = -111.4 dB

THD @ 2VRMS UPV-B1 = -114dB

@ 20KHz / 2VMRS the 339A gives -103dB THD (80K BW) and -100 THD+N (80K BW)

(in truth is was -99.9 THD+N, but I rounded it down out of respect for a 1977 veteran)

How good will the 339A be with a good service and the modifications described by Helmut? For a 1977 machine to have better THD+N than the UPV's digital oscillator is quite something. I tested it against another UPV and saw that it still beat the digital oscillator by 1-2dB THD+N, so not as much as in the above test - but impressive nevertheless.

I wonder how the 339A's oscillator compares to later HP AAs?

It would seem that my measurements of the 339A mirror Helmut's nicely. I should get some work done and stop geeking around with test gear...
 
  • Like
Reactions: anatech