Low-distortion Audio-range Oscillator

Jump to Latest
#5,762
1audio said:
Thanks for your work. I thought the AG16 had an analog multiplier. I can't find my photo's right now. The AGC is quite involved and an almost discrete sample and hold type circuit is very involved. I'll need some time to digest it. The model of the simple oscillator above doesn't quite work accurately since it stabilizes in less that 2 seconds, not the 15 seconds the simulator needs, which is typical of this stuff.
Click to expand...

How stable is the phase of the harmonics at these ultra low levels?
 
#5,763
All seems stable... doesnt change with time is minor. The gen used has THD of <-130dB re 1v/100K.

But, the A-P FFT firmware may be different than the QA401 uses. Maybe differences in FFT algor? It is closer reading to the others when its monitor output is also used with QA401.




THx-RNMarsh
 
Last edited:
#5,764
Richard -- I'm very much a learner on the Panasonic 7722A. I'm looping the oscillator of the 7722A back to input with a 600 ohm termination. I'm between houses and workshop so not a great set-up. The BNC connectors on the front panel need cleaned or replaced, and I've jury-rigged the load but getting 0,0005% THD, 1 v RMS @ 1 kHz, no filters. Both inputs read about same. Relative about -124 dB, plus minus X. Is this in-line with your units? Hope this provides another data point here.

Thanks
Phil
 
#5,768
rsavas said:
https://www.pasternack.com

Biggest difference is temperature of operation, thus power rating because of the higher temp. As you see they are not all created equal, even for the same RG #. Usually the PTFE types feature silver plating. I recall that when I was at Motorola they used RG-400/U in the high end gear. We used a Wiltron return loss bridge setup to test our cable builds for return loss when we had critical measurements.

Generic Name RG174
Flex Type Flexible
Impedance 50 Ohm
Dielectric Type PE (LD)
Velocity of Propagation 66 %
Jacket Diameter 0.1 in
Jacket Material PVC
No. of Shields 1
Attenuation at 1 Ghz. 32 dB
Power, Max at 1 Ghz. 16 Watts
Frequency, Max 1,000 MHz
Max Operating Temperature 80 deg C
Center Conductor Type Stranded
Inner Conductor, Number of Strands 7
Coax Type Coax

Generic Name RG174 Type
Flex Type Flexible
Impedance 50 Ohm
Dielectric Type PE
Velocity of Propagation 66 %
Jacket Diameter 0.109 in
Jacket Material LSZH
No. of Shields 2
Attenuation at 1 Ghz. 25 dB
RF Shielding 90 dB
Frequency, Max 5 GHz
Max Operating Temperature 85 deg C
Center Conductor Type Solid
Inner Conductor, Number of Strands 1
Minimum Bend Radius, One Time 0.25 in
Minimum Bend Radius, Repeated 1 in
Coax Type Coax

Generic Name RG316
Flex Type Flexible
Impedance 50 Ohm
Dielectric Type PTFE
Velocity of Propagation 70 %
Jacket Diameter 0.114 in
Jacket Material FEP
No. of Shields 2
Attenuation at 1 Ghz. 26.1 dB
Power, Max at 1 Ghz. 160 Watts
Frequency, Max 3 GHz
Max Operating Temperature 200 deg C
Center Conductor Type Stranded
Inner Conductor, Number of Strands 7
Minimum Bend Radius, One Time 0

Generic Name RG316
Flex Type Flexible
Impedance 50 Ohm
Dielectric Type PTFE
Jacket Diameter 0.098 in
Jacket Material FEP
No. of Shields 1
Attenuation at 1 Ghz. 38 dB
Power, Max at 1 Ghz. 160 Watts
Frequency, Max 3 GHz
Center Conductor Type Stranded
Inner Conductor, Number of Strands 7
Coax Type Coax
Click to expand...

rsavas,

Thank you for this detailed information. Looks like main differences that would matter are dielectric (probably for the higher temp of RG316) and whether or not center conductor is silver plated. Jacket material probably doesn't matter much (316 different for higher temp). I would only be interested in the stranded types. Either type looks like it could have one or two shields. As you say, much seems to depend on the manufacturer.

I have some high-end RG174-like coax that we used at work up to 10 GHz, but it does not say on it whether it is 174 or 316 or something else. It is marked Belden 83265 MIL-C-17. The center conductor appears to be silver plated with 7 strands. Looks like a single silver-plated braided shield. The cable is white and fairly flexible.

I looked up the specs. Belden spec sheet says conductor is 30 AWG stranded (7 X 38 AWG) silver-plated copper-covered steel conductor (SPCCS). Dielectric is Teflon (TFE). Shield is Silver-plated copper braid (SPC), 95% coverage. Outer jacket is FEP. Operating temperature range to 200C. RG type 178B/U. 69.5% velocity, 25 pF/ft, 72 nH/ft.

Cheers,
Bob
 
#5,769
RNMarsh said:
Something i had explored was the accuracy of "sound card" type/based test systems. I found that below -100db the harmonics shown had no relationship to any reality.
recently, some have done a little better ---- -105-110dB is Ok.

Now I checked the relative numbers of three very expensive analyzers to see to what degree they agree. The three are Audio-Precision 2722, Panasonic 2277 (Matsushita Comm Ind) and ShibaSoku 725D. On each of these, i connected the QA401 FFT to their respective monitor output.

I used a generator which had 2H and 3H dominant. The only difference is found in the level of the 2H. of course, the 2H is closest to the notch circuits influence etc.

The levels are < -120dB r 1v rms

With the 7722, the 2H was +6db above 3H;
With the 725D, the 2H was +4dB above 3H;
With the AP, the 2H and 3H were same level.

However, the FFT displayed on the computer interface/screen... not at the monitor output via QA401.... showed the 2H to be much lower than the 3H. By -2 -4db
(2H seemed to go down with time -- heat related? drift? )

So, even with expensive analyzers, the numbers are not exactly the same, either. There is up to a 6dB or 10dB difference. But not wildly out of whack like cheap units. No ghost or added harmonics either. Of course, this is based on tests well below -100dB. Their limit in accuracy appears to start to become nuts below -130dB though fairly useful/consistent to -140 to -150dB.

The 2H seems to be most affected. higher fundamental freqs a little more so... null spreading/Q change etc?

I tend to think that the analyzer which gives the higher 2H is a more accurate unit..... tentatively.


THx-RNMarsh
Click to expand...

Were thse measurements done at 1 kHz fundamental?

Cheers,
Bob
 
#5,772
Richard and Demian -- thanks for the tips. And yes I dropped a zero. At 1 kHz fundamental, the VP-7722A reads about -123 dB THD re 1V/100K. With either 30 or 80 KHz filter in, this improves by about 2 dB.

Does anyone know of a handbook or paper with the nut and bolts of distortion measurement - good and bad practices, tricks of the trade, info on the actual taking of the measurement? The fundamentals are covered well elsewhere and I have the manuals but looking for something with some hands-on advice to cut the learning curve. Thanks!
 
#5,773
PH104 said:
Does anyone know of a handbook or paper with the nut and bolts of distortion measurement - good and bad practices, tricks of the trade, info on the actual taking of the measurement? The fundamentals are covered well elsewhere and I have the manuals but looking for something with some hands-on advice to cut the learning curve.
Click to expand...

Here is a great "starter" for me circa 1975 by Tektronix. The techniques are unchanged although the tools available today are better, including your VP-7722A.

https://www.google.com/url?sa=t&rct...gcoW8e0kw&sig2=G2iyZq8D092w7Wff-lF7UQ&cad=rja

It will bring back lots of memories and a few yuks too. Have fun and enjoy.
 
#5,774
PH104 said:
Does anyone know of a handbook or paper with the nut and bolts of distortion measurement - good and bad practices, tricks of the trade, info on the actual taking of the measurement? The fundamentals are covered well elsewhere and I have the manuals but looking for something with some hands-on advice to cut the learning curve. Thanks!
Click to expand...

Audio Precision have "The Audio Measurement Handbook" available for free download as a PDF on their site, you will need to set up an AP account. It's an excellent overview for audio measurement in general, I finally managed to find a hard copy recently.
 
#5,778
Richard: For non-harmonic accuracy at low levels you can build a mixer/attenuator and add a spur at -120dB or -130dB (not trivial to get it right). I did this for the 725 but I needed the spur to be right on frequency. Fortunately the Boonton has 5 digits of resolution and crystal stability so it was possible. I did that down to -120 dB. It seems its time to figure out how to get to -140 dB.
 
#5,779
here is what might be termed a tracking caused error on the A-P 2722 analyzer --

The freq is 1KHz. The 2H and 3H is affected inversely.... but most affected is the 2H; The 2H level decreased at each freq change of 1% higher. At +5% change the 2Harmonics went back up.


freq tracking error.JPG



THx-RNMarsh
 
Last edited:
#5,780
Well, it has been awhile since i had to trouble shoot to the component level of a complex instrument.....

I got both A and B input circuits working again and they match within a fraction of a dB.

AP THD match.JPG

This picture is using the AP's internal oscillator/generator. The picture above [line # 2579] was using the ShibaSoku AG16B generator.



THx-RNMarsh
 
Last edited:
You must log in or register to reply here.
Top Bottom