My implementation of the Cordell Distortion Analyser

Status
Not open for further replies.
Jump to Latest
#206
As a somewhat new member of this list, I've been reading the posts and I've noticed a number of comments about good JFETs for AGC. I'm using a J105 at the moment in Bob's SV oscillator, and it seems to work well -- I'd like to be more specific, but lacking the Cordell DM, I don't have the analyzer resolution at the moment. J105's have Rds(on) of around 5 ohms and pinch-off at 6-10V. They're also fairly cheap.
 
#207
richiem said:
As a somewhat new member of this list, I've been reading the posts and I've noticed a number of comments about good JFETs for AGC. I'm using a J105 at the moment in Bob's SV oscillator, and it seems to work well -- I'd like to be more specific, but lacking the Cordell DM, I don't have the analyzer resolution at the moment. J105's have Rds(on) of around 5 ohms and pinch-off at 6-10V. They're also fairly cheap.
Click to expand...

Krohn Hite uses the 2N5555 in the AVC of the 4400A. I don't know if that's the most recent version. The part is available from the usual suspects.
 
#208
Last night, after posting my note on the J105 above, I realized the full implications of the write-up by Bob on the Distortion Magnifier. My "aha" was that the DM has to be used with a device that has an input and an output. I kept reading Bob's write-up, unfortunately sans schematic, trying to invent for myself how he might be doing auto-tuning, when it all came together. It's just summation, no filtering. The reason that the noise and distortion of the oscillator don't much matter -- only that of the DUT -- is because *all* (or nearly, anyway) of the signal components from the oscillator are cancelled in the DM's summing amp, with only the low-level fundamental re-injected remaining, and having (presumably) noise and distortion much lower than the DUT's. I had hoped for something that would enable more resolution for looking at oscillator outputs.

Seems now that I'll build a simple spot-frequency Twin-T filter with active feedback ala John L. Linsley Hood's 1970's design, implemented with the National 4562 or a couple of LME-whatever-they-are's. With appropriate low-pass filtering and careful tuning, 2nd harmonic attenuation will be less than 0.5dB. I don't see why noise and THD floors below -120dB can't be easily achieved in the audio band. Then, with such effective fundamental notch filtering, adding a spectrum analyzer to the output will reveal all...

It's not as easy as with a Twin-T, but Linsley Hood's active feedback can be used with a Wien Bridge, and I'm in the process of greening an HP 334A, with hopes of relatively easily improving its resolution, and its noise and THD floor, by 10 or 20dB.

Returning to Bob's SV oscillator, I've found that the exact value of the 750k FET feedback resistor has a significant effect on the residual 2nd harmonic. Its value seems to be dependent on the precise output level of the oscillator. The 750k resistor can be easily replaced with a small 1Mohm pot. Then using a spectrum analyzer will let you fine-tune the best value for the resistance. In my version of the oscillator, the value of the 750k resistor has more effect on residual distortion than the rectifier balance pot does.
 
#209
alexcp said:
Thank you, jackinnj and Phil. I did not know that PN4091 is the same thing. Here in Moscow, Russia, both PN4091 and J111 are available from Mouser or DigiKey with shipping cost at $120 per package, or from a local distributor with a 4-week delivery time. It's a choice between paying or waiting 🙂
Click to expand...

Hello Alexcp,
You may try some of old Russian FET's if they are still available in Moscow. КП302Б or КП302B should be good enough. We used КП302Б as a control element for audio limiter with the same kind of 2 harmonic compensation. Worked well. Fet's with higher current at 0 GS voltage should be better. КП302Б should have Rds around 40 Ohm.
 
#210
Eugene Dvoskin said:
Hello Alexcp,
You may try some of old Russian FET's if they are still available in Moscow. КП302Б or КП302B should be good enough. We used КП302Б as a control element for audio limiter with the same kind of 2 harmonic compensation. Worked well. Fet's with higher current at 0 GS voltage should be better. КП302Б should have Rds around 40 Ohm.
Click to expand...

Thank you Eugene, those FETs are still easily available.
 
#211
Euvl: I am interested in building a Cordell Distortion Magnifier. In your schematic for the DM in post #126, should R24a be 39K rather than 3.9K in order to have 10X gain in the output? Unlike the DM shown in Bob Cordell's article, your implementation appears to provide 1x and 10x magnification options rather than 10x and 100x. Why did you make that choice?
 
#212
Question on Bob Cordell's Circuit Design:

In the integrator stages of the state-variable oscillator, as well as the state-variable tracking filters, the '5534 opamps have a 220 ohm resistor in series with the inverting inputs (pin 2). Once upon a time I knew why they are there, but can't recall the reason now. Can somebody give me a hint, or point me to an analysis that explains the function of these resistors? Is this a feature that would be necessary with most opamps, or is there some particular characteristics of the '5534 that makes the resistor necessary for this specific opamp?

(I want to say it has to do with the pin-to-pin capacitance across the inverting and non-inverting inputs. But that doesn't make sense - with the 22 pF compensation caps, the stage should be unity-gain stable without any other measures.)
 
#214
dchisholm said:
Question on Bob Cordell's Circuit Design:

In the integrator stages of the state-variable oscillator, as well as the state-variable tracking filters, the '5534 opamps have a 220 ohm resistor in series with the inverting inputs (pin 2). Once upon a time I knew why they are there, but can't recall the reason now. Can somebody give me a hint, or point me to an analysis that explains the function of these resistors? Is this a feature that would be necessary with most opamps, or is there some particular characteristics of the '5534 that makes the resistor necessary for this specific opamp?

(I want to say it has to do with the pin-to-pin capacitance across the inverting and non-inverting inputs. But that doesn't make sense - with the 22 pF compensation caps, the stage should be unity-gain stable without any other measures.)
Click to expand...

Hi dchisholm,

Sorry I missed this question. Sometimes I had found that the 5534 whould oscillate if the input was not isolated by some resistance, like 100 or 220 ohms, especially if the driving wire ran several inches. I just did this out of paranoia 🙂.

Cheers,
Bob
 
#219
jackinnj said:
It will make you go blind...some of the Harris chips are no longer available...but check the Test and Measurement pages of diytube.com :: Index
Click to expand...

Thanks! I like the notch filter, pretty much what I had in mind here but with FETs instead of relays:
http://www.diyaudio.com/forums/equipment-tools/194243-audio-test-equipment-computer-control.html

And they mention a VB control program at diytube which is what I had in mind also.
Have you tried the program?

They used a clever trick so that the freq setting resistors do not get too large in the notch filter,
but then didn't use it in the oscillator in the 1120. Then they made them both the same in the 1121,
also using the trick in the oscillator.

I've done a good amount of work with state variable filters including voltage controlled but not at this
low distortion level. I studied under Bernie Hutchins during my graduate work and he discusses
state variable filters in Ch 6 of his electronotes, EN-194:
http://electronotes.netfirms.com/EN194.pdf
His entire series on active filter design is excellent:
http://electronotes.netfirms.com/free.htm
We used CA3080 transconductance amps for voltage control but I doubt they are good enough
for these low distortion applications.
 
Last edited:
Status
Not open for further replies.
Top Bottom