Low-distortion Audio-range Oscillator

Jump to Latest
#1,741
richiem said:
The K-H 4400 has a very low-Q filter, with not much notch depth. Not a good candidate i my opinion. Most oscillators use relatively low-Q filters for tuning. The 339 and 239 probably have the highest Q filters, but even so their notch depth would not be much over 30dB. None of the commercial oscillators I've seen use a Twin-T for tuning -- most use a Wien or bridged-T -- inherently low-Q.
Click to expand...

I was thinking of pressing my parts spare 339A into notch service this way. Maybe a slight mod would boost the dB's of the notch... but 30 might be Ok for somethings. Any ideas on how to do this?

Notching out a dominant 2H or 3H from a source would be another use to make the source even better... as I have done using passive notch filter.

The advantage is the multiple notch freq that are selectable with all the parts and chassis, pwr supply etc already done.

Thx-RNMarsh
 
Last edited:
#1,742
RNMarsh said:
I was thinking of pressing my parts spare 339A into notch service this way. Maybe a slight mod would boost the dB's of the notch... but 30 might be Ok for somethings. Any ideas on how to do this?

Notching out a dominant 2H or 3H from a source would be another use to make the source even better... as I have done using passive notch filter.

The advantage is the multiple notch freq that are selectable with all the parts and chassis, pwr supply etc already done.

Thx-RNMarsh
Click to expand...

Rick the 339a is already a notch filter. Pull out all the auto stuff and manually tune it.
This will probably get you a 20dB reduction in distortion and maybe a reduction of noise as well.
 
#1,743
1audio said:
The problems with the state variable filter used that way is the intrinsic distortion of the amplifiers and the tuning. That is the core of the analyzer in several analyzers, the Boonton and the Amber are two examples. The passive notch, like the B&K, are better. You could modify the B&K with active circuitry to make a tuneable high q notch with it. We just need to reverse engineer its circuitry, not a small task.
Click to expand...

I could do that....
are you back from Hawaii? I'll come see you when you are back and also bring Auraliti for upgrade.

RNM
 
#1,744
davada said:
Rick the 339a is already a notch filter. Pull out all the auto stuff and manually tune it.
This will probably get you a 20dB reduction in distortion and maybe a reduction of noise as well.
Click to expand...

I know... and that is why I was thinking of using it's switchable notch filter in my spare 339A.

Right now I have manually tuned it at one freq - 1KHz and it is super low thd. Its actually as low as Viktors on the ShabaSoku.

I need to get the QA400 working better for FFT and the notch comes in for that. Because noise is now in the way in obtaining the true thd readings.

-RNM
 
Last edited:
#1,745
RNMarsh said:
I know... and that is why I was thinking of using it's switchable notch filter in my spare 339A.

Right now I have manually tuned it at one freq - 1KHz and it is super low thd. Its actually as low as Viktors on the ShabaSoku.

I need to get the QA400 working better for FFT and the notch comes in for that. Because noise is now in the way in obtaining the true thd readings.

-RNM
Click to expand...

Hi Rick,

I listed all the mods you need to do to use the QA400 off the 339a notch filter many pages ago. It's not that hard. Just go back and read. You'll be up and running under an hour.
 
#1,746
Right. I am multi-tasking right now. I know it can be done off the notch -- I tried it with the HP3580A spectrum analyzer. When using the ADC, I found its limitations which has caused a sharp left turn.

When measuring an audio headphone amplifier, I found limitations of the AQ400 and it needs a notch filter to do better... that is the notch I am referring to using a second 339a for.

Also, the active twin-T circuit/pcb I am loading up can be put inside my variable, passive twin-T to make it active and higerh Q. Fun.

Thx, Richard
 
Last edited:
#1,747
I'll look back into what I need to do to use the spare 339a's internal Notch Filter (pcb A3) for this ADC app.
OR complete the active twin-T circuit and interface it with the B&K variable passive twin-t.

Either way, I get a flexible notch filter for use with an ADC with a minimum of time. Thx-RNMarsh
 
#1,749
On the Radford schematic there is a reference to a Morica MCD-521 control cell. It looks to be an incandescent to CDS cell device but that is a guess. Does anyone know? The reference to -120 dB distortion would be pretty impressive if that is all it is.
 
#1,758
I have found that only my stabilized oscillator (Boonton) comes close to a digital source on its own. I have modified my KH4400 for injection locking and that does get a really clean spectral line when locked to an external crystal derived source.

How would you propose to measure the phase noise? My mixers don't go that low. The notch filters are too broad to see much really close in.
 
You must log in or register to reply here.
Top Bottom