First some backstory: picked up Tektronix AA501A Audio Distortion analyzer (with the IM option); a dual-bay TN502 cabinet and a few other plug-in’s for the 500 series. The 501 seems to work fine. Problem is I don’t have a matching low distortion oscillator to “test” its 0.00X HD measuring capabilities. Best I have is a Tenma with a (measured) HD of 0.03%.
Only other audio plug in I have for the 500 series is the Tek AF501: a multipurpose amplifier/bandpass filter /oscillator covering 3hz to35Khz. The (sine wave) oscillator distortion is spec’d at 0.5% HD. It measures around 0.3%. A copy of the unit’s schematic is attached
The oscillator function uses a wien bridge (components in and around U150) with its amplitude stabilized by op amp U160 and FET Q168. The output is buffered by U200.
I’ve read up on wien bridges here and on other websites. They mention various methods of amplitude stabilization and supposedly the FET approach can result in distortions well below this current unit. Don’t want to contravene Tektronix here, but can anyone suggest modifications to this circuit to get the distortion down by >10X or so? Don’t care what happens to the useful frequency range nor if it impacts the other plug-in functions
THANKS very much
Only other audio plug in I have for the 500 series is the Tek AF501: a multipurpose amplifier/bandpass filter /oscillator covering 3hz to35Khz. The (sine wave) oscillator distortion is spec’d at 0.5% HD. It measures around 0.3%. A copy of the unit’s schematic is attached
The oscillator function uses a wien bridge (components in and around U150) with its amplitude stabilized by op amp U160 and FET Q168. The output is buffered by U200.
I’ve read up on wien bridges here and on other websites. They mention various methods of amplitude stabilization and supposedly the FET approach can result in distortions well below this current unit. Don’t want to contravene Tektronix here, but can anyone suggest modifications to this circuit to get the distortion down by >10X or so? Don’t care what happens to the useful frequency range nor if it impacts the other plug-in functions
THANKS very much
For low distortion requeriment in Wien Bridges is the matching between pots. Small differences between them in the carbon tracks creates issues in the transfer function.
A friend told me that actually there are good and cheap DDS devices capables of very fine tunning and very low distortion, however I didn't try myself. Perhaps you can.
A friend told me that actually there are good and cheap DDS devices capables of very fine tunning and very low distortion, however I didn't try myself. Perhaps you can.
You will get a lot further quicker getting one of Victor's single frequency oscillators. (https://www.diyaudio.com/community/members/vicnic.19397/) They are quite reasonably priced as well. The Tek analyzer is very good and the Tek oscillator needs too much modification to even get close. Meanwhile the Tek oscillator is fine for frequency response testing. You might be able to mix Victors oscillator with the Tenma for IM, but a 1 KHz may not work for the HF with the Tek AA501. DDS oscillators ar OK but not in the .0005% or better that the AA501 can do.
Thanks,
Got some ideas from a Linear Technology App note (AN-43) that I'm going to try.
The AA501 IS quite the instrument.
Got some ideas from a Linear Technology App note (AN-43) that I'm going to try.
The AA501 IS quite the instrument.
I'd recommend organising a soundcard input with software like REW, as the main bench tool to guide your measurements of harmonic levels, and benchmarking HD variation with circuit changes/tweaks.
I'd also start with measurement and trimming of existing C and R parts that are obvious 'select on test' paired parts on the schematic, and measure HD reduction as they are balanced with better precision.
Distortion from the ganged pots may also be an issue, as better vintage low distortion oscillators used a large precision WW pot (unobtanium today).
I'd also start with measurement and trimming of existing C and R parts that are obvious 'select on test' paired parts on the schematic, and measure HD reduction as they are balanced with better precision.
Distortion from the ganged pots may also be an issue, as better vintage low distortion oscillators used a large precision WW pot (unobtanium today).
Success:
With help from AN-43 mentioned above, I "bypassed" The stock FET and IC stabilization circuit (U160/Q168) and substituted an incandescent lamp. Spent a few hours playing with lamps and feedback resistors. Ended up with an HD around 0.01% - over a 60hz to 20Khz range. Was then able to reduce that HD to nearly the residual of the AA501 (~0.007%) by swapping the output buffer IC.
Best distortion below 60hz, down to 20Hz is ~0.01%.
Oscillator takes a while to settle upon turn on or range change, but I've got the time! Nice to have a "one-box HD solution".
With help from AN-43 mentioned above, I "bypassed" The stock FET and IC stabilization circuit (U160/Q168) and substituted an incandescent lamp. Spent a few hours playing with lamps and feedback resistors. Ended up with an HD around 0.01% - over a 60hz to 20Khz range. Was then able to reduce that HD to nearly the residual of the AA501 (~0.007%) by swapping the output buffer IC.
Best distortion below 60hz, down to 20Hz is ~0.01%.
Oscillator takes a while to settle upon turn on or range change, but I've got the time! Nice to have a "one-box HD solution".