How Distortion Free are the Distortion Measurers? - Page 8 - 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 4th August 2012, 03:12 AM   #71
klewis is offline klewis  United States
diyAudio Member
 
klewis's Avatar
 
Join Date: Sep 2007
Blog Entries: 36
here is a similar thd spread sheet except in ".xls" format.
Attached Files
File Type: zip thd calculator.zip (1.9 KB, 24 views)
  Reply With Quote
Old 17th August 2012, 02:51 AM   #72
klewis is offline klewis  United States
diyAudio Member
 
klewis's Avatar
 
Join Date: Sep 2007
Blog Entries: 36
not knowing where else to post this, I figured here is probably as good a place as any. I am making some update modifications to an old Quan Tech 2173C Transistor Noise analyzer. I need to update an amp section (image 1 is the schematic) so that it will work with JFETs. I removed the amp section and powered it up with my lab bench supply to make a baseline measurement (image 2). A pretty clean amp and spectrum, all with the board unshielded out in the open air. I then replaced all the carbon resistors with metal film Dale Rn55's figuring that I could improve the noise floor a bit with quieter resistors. Well no such luck, with the same identical setup, measured early in the morning just like the prior measurement, things got much worse! (see image 3).

I've been thinking about this all day, and believe that it might be the ferrious quality of the Dale resistors that pick up the EMI far more readily than the non-ferrious carbon resistors.

So today I ordered up some non metalic Halco resistors and will install them to see if thing return to the baseline. If that doesn't work, I will put in some NOS Allen Bradley's. Any other thoughts would be welcome.

Ken
Attached Images
File Type: png quantech_amp_1_10khz_signal.png (39.4 KB, 159 views)
File Type: png quantech_amp_1_10khz_signal-2.png (40.7 KB, 158 views)
Attached Files
File Type: pdf quantech schematic.pdf (28.1 KB, 32 views)
  Reply With Quote
Old 17th August 2012, 02:54 AM   #73
klewis is offline klewis  United States
diyAudio Member
 
klewis's Avatar
 
Join Date: Sep 2007
Blog Entries: 36
Just to be clear, the dale resistor leads don't seem to be ferrous, only the body of the resistor, the carbon resistors are not attracted to the magnet. I will also do a good job of cleaning the flux off the board and retest, just in case.

Ken

Last edited by klewis; 17th August 2012 at 02:57 AM.
  Reply With Quote
Old 17th August 2012, 03:26 AM   #74
1audio is offline 1audio  United States
diyAudio Member
 
Join Date: Mar 2004
Location: SF Bay Area
Blog Entries: 3
Quote:
Originally Posted by klewis View Post
Just to be clear, the dale resistor leads don't seem to be ferrous, only the body of the resistor, the carbon resistors are not attracted to the magnet. I will also do a good job of cleaning the flux off the board and retest, just in case.

Ken
All of the resistors in the Quantech are tested for noise. You won't accomplish much trying to second guess Quantech's choices. Good metal film resistors will be more than good enough. Your noise issues are probably environmental, one of the down sides of testing on a bench. You are limited by the device noise. The noise of the preamp will not limit measuring any FET you can buy since it uses gain from the FET so unless the noise of the FET is below .1 nV/rthz it probably won't matter. There are a few locations that are noise sensitive but most metal film resistors will be OK even there.

I'm not sure what you need to change to support FETs. Every one I have played with (3 now) have supported FETs. There is a switch on the front panel that selects fet-bipolar as I remember. I'll dig out the manual tonight.

The parts that age are the electrolytic caps. Current ones are orders of magnitude better than the originals. If I were to invest the effort it would start with new caps on the supplies (not a lot bigger than the originals or some other problem will surface).
__________________
Demian Martin
Product Design Services
  Reply With Quote
Old 17th August 2012, 03:59 AM   #75
klewis is offline klewis  United States
diyAudio Member
 
klewis's Avatar
 
Join Date: Sep 2007
Blog Entries: 36
Quote:
Originally Posted by 1audio View Post
All of the resistors in the Quantech are tested for noise. You won't accomplish much trying to second guess Quantech's choices. Good metal film resistors will be more than good enough. Your noise issues are probably environmental, one of the down sides of testing on a bench. You are limited by the device noise. The noise of the preamp will not limit measuring any FET you can buy since it uses gain from the FET so unless the noise of the FET is below .1 nV/rthz it probably won't matter. There are a few locations that are noise sensitive but most metal film resistors will be OK even there.

I'm not sure what you need to change to support FETs. Every one I have played with (3 now) have supported FETs. There is a switch on the front panel that selects fet-bipolar as I remember. I'll dig out the manual tonight.

The parts that age are the electrolytic caps. Current ones are orders of magnitude better than the originals. If I were to invest the effort it would start with new caps on the supplies (not a lot bigger than the originals or some other problem will surface).
Demian,
My Quan Tech is older than jfets, no switch on the front. I'm updating the device to include the switch. It appears to be a pretty simple modification.

I was just struck by how different the spectrum was with the resistor change, when tested under identical conditions. My theory is that the ferrous quality of the resistors makes them subject to significant emi pick up.

Ken.
  Reply With Quote
Old 17th August 2012, 04:41 AM   #76
richiem is offline richiem  United States
diyAudio Member
 
richiem's Avatar
 
Join Date: Sep 2010
Location: Grapeview, WA
Ken, that's a huge difference in the plots and I can't believe that the change is solely due to changing resistors, especially to metal films. Something else is going on here. I assume you're using good old fashioned 60-40 tin-lead solder? Mixing Lead-free with regular can be a nightmare....
__________________
...................
Dick Moore
  Reply With Quote
Old 17th August 2012, 05:01 AM   #77
klewis is offline klewis  United States
diyAudio Member
 
klewis's Avatar
 
Join Date: Sep 2007
Blog Entries: 36
Quote:
Originally Posted by richiem View Post
Ken, that's a huge difference in the plots and I can't believe that the change is solely due to changing resistors, especially to metal films. Something else is going on here. I assume you're using good old fashioned 60-40 tin-lead solder? Mixing Lead-free with regular can be a nightmare....
Hi Dick,
Yes it seems like too big of a swing. I hope it doesn't indicate over heating of the transistors during the desolder. Using cardias eutectic solder.

Ken
  Reply With Quote
Old 17th August 2012, 05:42 AM   #78
1audio is offline 1audio  United States
diyAudio Member
 
Join Date: Mar 2004
Location: SF Bay Area
Blog Entries: 3
Help me with the plots. The noise floor seems to be around -110 in both. The first plot has a 10 KHz peak in it. Where is that from? The second shows a lot of harmonics from the peak and some 48 kHz and some 65 kHz. The latter two look like switching supply or monitor or CF noise to me.

Which preamp are you tweaking? There are three, depending on the current setting. Or is it the amp (Front cage 4th board from the left)? There is a fair amount of gain in front of the amp so noise there is not too consequential.

The calibrating oscillator runs at 4 KHz.

I would do the following- first put it back together and feed a signal into the noise cal test point at the transistor jig. Look at what comes out with a good transistor in the jig. A 2n3904 running at 100 uA should get down to 1-4 nV /rtHz or so at 1 KHz or less if the base resistance is set to zero. If you add a signal via the noise cal it will be processed and should be visible at the output BNC at around 10X the level it went in at. The input Z is 600 Ohms so measure at the noise cal terminal to get the gain right..

Do you have a manual? I may have a scan here if I can find it. The calibration is pretty straightforward, if you have a noise generator calibrated in mV/rtHz. Otherwise its more difficult to get everything aligned and meaningful numbers.

The manual I have is for serial number 203 and really old. The internals of John Curl's are the same and his is 20 years newer than the manual. The unit has a gate voltage control for setting a fixed gate bias. I pretty much never use that and use the drain/collector current to set the biases. For matching transistors you can connect a more precise voltmeter (DVM) to the gate (base) terminal (D4) of the base supply board and readout the voltage for a target drain current which would be the operating point. Its a very clean way to match JFETs.

edits-
My manual is definitely older than the one you are looking at. It does not have the switch or relay and the connections are not brought out of the card. Your unit predates the relay and switch. I'm not sure that prevents it from measuring JFets but if you are up for the changes have at it. The older drawing will help identify what changed (attached)

Your measurement software indicates that the THD+N went down with the change even though it doesn't look like it. The 10 KHz is also higher in the second plot for what seems to be the same signal level.
Attached Images
File Type: jpg Old Quantech amp.jpg (504.7 KB, 147 views)
__________________
Demian Martin
Product Design Services

Last edited by 1audio; 17th August 2012 at 05:58 AM. Reason: read everything more carefully.
  Reply With Quote
Old 17th August 2012, 04:46 PM   #79
klewis is offline klewis  United States
diyAudio Member
 
klewis's Avatar
 
Join Date: Sep 2007
Blog Entries: 36
Quote:
Originally Posted by 1audio View Post
Help me with the plots. The noise floor seems to be around -110 in both. The first plot has a 10 KHz peak in it. Where is that from? The second shows a lot of harmonics from the peak and some 48 kHz and some 65 kHz. The latter two look like switching supply or monitor or CF noise to me. .
The 10KHz peak is the fundamental frequency, it has gone through a notch filter (Dick's twin -T ) and is down about 50db - the amp output is at 1 Vrms. Using the notch filter drops the floor of the spectrum analyzer to about -130db.

I agree about harmonics being from environmental factors, like the switching supply. The interesting thing is that the board was relatively immune to these factors when first tested.

Quote:
Originally Posted by 1audio View Post
Which preamp are you tweaking? There are three, depending on the current setting. Or is it the amp (Front cage 4th board from the left)? There is a fair amount of gain in front of the amp so noise there is not too consequential.
It's Amp#1 as shown on the right side of your posted schematic. The manual I have is a bit newer than yours, yours exactly matches my board. My manual show things like the JFET switch, which is a part of the AMP #1 schematic. (not shown on your schematic - see lower left corner of AMP #1 schematic.)

Quote:
Originally Posted by 1audio View Post
The calibrating oscillator runs at 4 KHz.

I would do the following- first put it back together and feed a signal into the noise cal test point at the transistor jig. Look at what comes out with a good transistor in the jig. A 2n3904 running at 100 uA should get down to 1-4 nV /rtHz or so at 1 KHz or less if the base resistance is set to zero. If you add a signal via the noise cal it will be processed and should be visible at the output BNC at around 10X the level it went in at. The input Z is 600 Ohms so measure at the noise cal terminal to get the gain right..

Do you have a manual? I may have a scan here if I can find it. The calibration is pretty straightforward, if you have a noise generator calibrated in mV/rtHz. Otherwise its more difficult to get everything aligned and meaningful numbers.

The manual I have is for serial number 203 and really old. The internals of John Curl's are the same and his is 20 years newer than the manual. The unit has a gate voltage control for setting a fixed gate bias. I pretty much never use that and use the drain/collector current to set the biases. For matching transistors you can connect a more precise voltmeter (DVM) to the gate (base) terminal (D4) of the base supply board and readout the voltage for a target drain current which would be the operating point. Its a very clean way to match JFETs.

edits-
My manual is definitely older than the one you are looking at. It does not have the switch or relay and the connections are not brought out of the card. Your unit predates the relay and switch. I'm not sure that prevents it from measuring JFets but if you are up for the changes have at it. The older drawing will help identify what changed (attached)

Your measurement software indicates that the THD+N went down with the change even though it doesn't look like it. The 10 KHz is also higher in the second plot for what seems to be the same signal level.
I would be interested in a scan of your manual if you can find it.

Thanks for the tip on transistor matching!

Regarding modification for the JFET switch, the board came with the D(4) circuit and the connector is empty at that location, so, the modification will be quite easy. Regarding the THD+N, it's read off my HP8903A, I don't know if any filters were turned on, nor to I particularly trust it. I normally hand calculate the THD from the Harmonics - note that I am able to extract and display them from the spectrum data.

Ken
  Reply With Quote
Old 17th August 2012, 05:52 PM   #80
klewis is offline klewis  United States
diyAudio Member
 
klewis's Avatar
 
Join Date: Sep 2007
Blog Entries: 36
So I made a quick check with a spice simulation using 2N5401 and 2N5551 Cordell models. The FFT plot is quite similar to the second spectrum plot. I don't know what to think... maybe the first plot was a total fluke.

I thought that maybe I had run the first plot at a lower input voltage, but, the voltage measurement shown on the plot is taken from the HP8903A and it reads 1vrms on both plots.

Given the spice results, the quan tech will probably run fine as modified, but, I am still going to try to reproduce the first spectrum result. I will build a replacement board and try populating it with carbon resistor just like the original. Looks like a Fairchild SS8050 or BC63916 are both reasonable subsistution for the 2n3644 device.

Ken
Attached Images
File Type: png quan_tech_amp_1_fft.png (51.1 KB, 130 views)
  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
FREE Distortion Analyzer imix500 Swap Meet 3 28th July 2010 12:15 AM
Free to good home..HP 302a distortion analyzer-needs work boywonder Swap Meet 8 12th March 2010 02:58 PM
Distortion. bigwill Tubes / Valves 8 14th July 2007 11:47 PM
Non Linear Distortion testing - Harmonic Distortion JMB Multi-Way 0 20th July 2004 03:54 PM
Cable Distortion: Materials Science Free For All Steve Eddy Everything Else 21 24th December 2003 03:01 AM


New To Site? Need Help?

All times are GMT. The time now is 11:19 PM.


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