Low-distortion Audio-range Oscillator - Page 104 - diyAudio
Go Back   Home > Forums > Design & Build > Equipment & Tools

Equipment & Tools From test equipment to hand tools

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 5th December 2012, 07:33 PM   #1031
davada is offline davada  Canada
diyAudio Member
 
Join Date: Jul 2004
Location: Fort St John, BC Canada
Quote:
Originally Posted by richiem View Post
In my experience active filters are just noisy; don't know if it has to be that way... but for me the crucial filters are 400Hz and 80kHz (for LP, I'd rather have 10kHz and 100kHz, and may make those changes). In the 339, the 80k filter is the worst for added noise -- the 400Hz HP actually doesn't seem to add much to the baseline noise, and it really is needed to control the line-related spurs.

The guy with the recently acquired 7L5 won't be looking inside his 339 until after the holidays and on into the new year a ways, maybe February. Sorry folks, just how it is. This isn't such a deal for David, who wants his 339 to be a front end for spectrum analysis, but for those who just want to see a believable meter reading, it's a different matter. Still, David could benefit from the 400Hz HP.
Hi Dick,

I am referring to the noise when the filters are switched out, bypassed.

The quad op amp for the meter's filters etc. was changed out about a year and a half ago with no improvement at all. There is more to this.


A brick wall filter such as what's in the 3581C would do well.

I think we might have some parasitic stuff going on.

Cheers,
__________________
David.
  Reply With Quote
Old 5th December 2012, 07:37 PM   #1032
diyAudio Member
 
Join Date: Jun 2005
Location: Zürich
Quote:
In the 339, the 80k filter is the worst for added noise.
With the 80 kHz filter, the residual shows more noise because the measurement bandwidth is greater, not because the low-pass filter adds any significant noise! In a well-designed analyzer, there is, depending on the range, 20-60 dB gain ahead of the bandwidth-limiting filter. With this, it isn't hard to make its noise contribution entirely negligible.

I don't know the exact situation in the older HP analyzers under discussion, but I'd be surprised if they had fallen into such a basic trap.

Samuel
  Reply With Quote
Old 5th December 2012, 08:16 PM   #1033
richiem is offline richiem  United States
diyAudio Member
 
richiem's Avatar
 
Join Date: Sep 2010
Location: Grapeview, WA
My point is that, looking at the monitor output with a spectrum analyzer, switching in the stock 80kHz filter raises the noise floor significantly over leaving it out, which should be the higher noise case. The effect is less pronounced with the 30kHz filter, thanks primarily to the much more restricted bandwidth, just as Sam points out.

Yes, getting rid of the noise that's running around post-notch is needed; but a filter that increases noise instead of reducing it, isn't really that helpful if you want some filtering before the ADC. You're right David, we have to reduce the broadband noise somehow, regardless of whether we want to read a meter or drive other gear.

Maybe Rick's filter amp swap-out will tell us something, as I think he's chosen the right amp for the filters... Can't remember David, which quad you used.
__________________
...................
Dick Moore
  Reply With Quote
Old 6th December 2012, 12:18 AM   #1034
RNMarsh is offline RNMarsh  United States
diyAudio Member
 
RNMarsh's Avatar
 
Join Date: May 2012
Location: 2457 Cascade Trail; Cool, CA. 95614
Default Passive changes only to Analyzer -

Without working on the source - davada has that covered.....
I added a socket for the quad filter and dropped in a TL074 as a place holder. Its still quieter than the original Quad IC. Then --> I changed the large value caps out with 105 degree C NP (non-polar) types and retuned for null. The reson for doing this is that lower esr means lower noise and deeper null. Also, the NP types have lower thd than the tant types in the first place. Now I can see some 5th at -145dB (!). Top of screen is -80dB. and 10dB/div. All filters IN.
Now, I am going to work on the sine wave source and start on those issues davada has done. Noted also that filter switch contacts must be very clean or contacts will raise thd a small but trouble-some amount. Thx-RNMarsh
Attached Images
File Type: jpg cap changes2.jpg (856.5 KB, 232 views)
File Type: jpg cap changes.jpg (562.8 KB, 199 views)

Last edited by RNMarsh; 6th December 2012 at 12:45 AM.
  Reply With Quote
Old 6th December 2012, 12:54 AM   #1035
davada is offline davada  Canada
diyAudio Member
 
Join Date: Jul 2004
Location: Fort St John, BC Canada
Default 339a grounding

With just the QA400 connected to the 339a notch amp output.
The spectra looks like this.
And with just the ground to my scope connected tot he ground post of the 339a monitor output the spectra looks like this.

I would suspect something wrong with the scope but I get RF parasitic oscillation just connecting the ground of my HP3581C to the ground post of the monitor output.
Attached Images
File Type: png No gnd.PNG (440.4 KB, 188 views)
File Type: png scope gnd.PNG (455.3 KB, 172 views)
__________________
David.
  Reply With Quote
Old 6th December 2012, 01:02 AM   #1036
davada is offline davada  Canada
diyAudio Member
 
Join Date: Jul 2004
Location: Fort St John, BC Canada
Well maybe it is the scope gnd.

This is with the scope ground connected to the QA400.
Could try floating the scope gnd.

Cheers,
Attached Images
File Type: png Scope.PNG (452.4 KB, 163 views)
__________________
David.
  Reply With Quote
Old 6th December 2012, 01:18 AM   #1037
davada is offline davada  Canada
diyAudio Member
 
Join Date: Jul 2004
Location: Fort St John, BC Canada
Quote:
Originally Posted by richiem View Post
My point is that, looking at the monitor output with a spectrum analyzer, switching in the stock 80kHz filter raises the noise floor significantly over leaving it out, which should be the higher noise case. The effect is less pronounced with the 30kHz filter, thanks primarily to the much more restricted bandwidth, just as Sam points out.

Yes, getting rid of the noise that's running around post-notch is needed; but a filter that increases noise instead of reducing it, isn't really that helpful if you want some filtering before the ADC. You're right David, we have to reduce the broadband noise somehow, regardless of whether we want to read a meter or drive other gear.

Maybe Rick's filter amp swap-out will tell us something, as I think he's chosen the right amp for the filters... Can't remember David, which quad you used.
That op amp is deep. Wouldn't you know it, can't find a working flashlight around the house.
I think it was a quad low noise jfet type.

I think what we need to do is isolate each stage and look at them individually. With such a long chain it's hard to tell just exactly where the noise originates. Speculating is good but it not getting us anywhere. We might find everything is where it should be and the problem is fundamental to the design. I do know that there is large amount of high frequency noise viewed on my scope off the monitor output which significantly drops when either of the LP filters are switched in.
__________________
David.
  Reply With Quote
Old 6th December 2012, 02:09 AM   #1038
davada is offline davada  Canada
diyAudio Member
 
Join Date: Jul 2004
Location: Fort St John, BC Canada
If we look at the schematic, we can see the post amplifier after the filters is the one that has a gain 100. This means that the filters noise is being amplifier by that amount. One could reason that there would be better SNR if the filters came after the x100 amp. There is a trade off here. With higher level we would see higher distortion from the filters and that would be added to the measurement. Instead HP went for the higher noise. It does satisfy there base line with the filter in.

We can test this by rearranging the filters and post amplifier. Make the x100 a preamplifier to the filters and see what happens.

But the bulk of the noise isn't coming from the filters it's from the analyzer it self. There is a huge difference in noise level between level mode and analyzer mode but both have the same scale.

I just notice something. I have the 339a with just the oscillator connected to the analyzer.
The meter is reading near 0.004% with the LP filter off. The range is -80dB with 1Vrms at the notch filter input, what I've been using with all my measurements. With nothing changed from the above, when I connect the QA400 to the monitor output the meter pins at over full scale. Even with just the QA400 ground connected, the meter pins. With the LP filters switch in the meter goes to a reading of 0.0015%.

So what does everyone make of this. I say there is something very wrong with the grounding in this 339a

I am curious of why I don't see the HP function when the LP filters are switched out. THe HP filter has been in all this time.

Cheers,
Attached Images
File Type: png LP out Disto.PNG (445.8 KB, 47 views)
File Type: png LP out level.PNG (444.4 KB, 36 views)
File Type: png LP IN disto.PNG (445.5 KB, 36 views)
File Type: png LP IN level.PNG (444.2 KB, 36 views)
__________________
David.

Last edited by davada; 6th December 2012 at 02:15 AM.
  Reply With Quote
Old 6th December 2012, 02:40 AM   #1039
davada is offline davada  Canada
diyAudio Member
 
Join Date: Jul 2004
Location: Fort St John, BC Canada
I guess the schematics would help.

These are third order filters.
Attached Images
File Type: png SCH1.PNG (105.8 KB, 94 views)
File Type: png SCH2.PNG (117.5 KB, 100 views)
__________________
David.
  Reply With Quote
Old 6th December 2012, 03:10 AM   #1040
davada is offline davada  Canada
diyAudio Member
 
Join Date: Jul 2004
Location: Fort St John, BC Canada
Default Grounding problem

It seems I have misinterpreted the problem.

As it turns out this is a low frequency problem. The ground selector switch is not doing
a very good job of grounding the in chassis to the outer earth grounded chassis.
This can also trigger RF parasitic oscillation.

Look at the difference when I ground the inner chassis to the outer chassis with a small piece of wire.

The meter reading dropped 1.5 orders of magnitude with the LP filter out.
Seems more normal now at 0.004% reading LP filters out.

With the 80KHz filter in it reads 0.0018% and the 30KHZ the usual 0.0014%.

I can't get to the ground switch because all the set screws for all the knobs are frozen.

Nothing quite like a earth ground loop.

Cheers
Attached Images
File Type: png Without gnd wire.PNG (450.7 KB, 89 views)
File Type: png With gnd wire.PNG (450.4 KB, 46 views)
__________________
David.

Last edited by davada; 6th December 2012 at 03:21 AM.
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
Radford Low Distortion Oscillator Series 2 audiomik Equipment & Tools 21 19th February 2014 11:46 AM
ultra-low distortion audio oscillator geekysuavo Analog Line Level 16 26th March 2013 04:04 PM
Low distortion oscillator? rjm Equipment & Tools 30 4th May 2011 11:45 PM
Can we improve this low distortion sine oscillator ? gaetan8888 Solid State 22 29th March 2009 01:30 PM
Simple, low distortion 1kHz oscillator jackinnj Solid State 4 6th October 2003 04:58 AM


New To Site? Need Help?

All times are GMT. The time now is 03:15 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright ©1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2