• Welcome to The Audio Annex! If you have any trouble logging in or signing up, please contact 'admin - at - theaudioannex.com'. Enjoy!
  • HTTPS (secure web browser connection) has been enabled - just add "https://" to the start of the URL in your address bar, e.g. "https://theaudioannex.com/forum/"
  • Congratulations! If you're seeing this notice, it means you're connected to the new server. Go ahead and post as usual, enjoy!
  • I've just upgraded the forum software to Xenforo 2.0. Please let me know if you have any problems with it. I'm still working on installing styles... coming soon.

How to build amp measurement gear?

DIYer said:
soundhound said:
I think the high noise floor of your interface is obscuring most of the view of the distortion products. The Dynaco MkIV is a bit surprising because the sidebands on the sides of each distortion product is intermodulation distortion (IM).
Does it have something to do with bias voltage? I have it set at 1.1VDC instead of standard 1.56VDC. I've read it somewhere that the original design was made when typical residential wall outlet was supplying around 115VAC 40 years or so ago.


The only thing that having the bias lower than usual would do is to make the amp clip at a lower wattage than otherwise. The bias voltage reading is independent of the AC line voltage (which was 117 VAC in the 1950s-60s rather than 120 VAC today). I think there's something else going on because those distortion sidebands are not normal operation.
 
DIYer said:
As for Tubelab SE, here is measurement at 1W 8 Ohm (vs. 2W above).

Yeah, those distortion components above the 7th shouldn't be there.

Here is the 1 watt distortion spectra of my SET amp; the distortion components are basically gone after the 3rd harmonic, which is extremely clean.
 

Attachments

  • SET Amp 1 Watt.jpg
    SET Amp 1 Watt.jpg
    195.8 KB · Views: 532
This thread is getting to be a one-on-one dialogue.....where are the posts from other people? Aren't they interested at all in this stuff??? :scared-eek:
 
Sorry SH, but I don't know anything about this subject. I'm not an electrical engineer. I can't even come up with a dumb question to ask.
 
soundhound said:
Aren't they interested at all in this stuff??? :scared-eek:
It could be the combination of St. Patrick's Day, NCAA basketball and warming weather thing... :confusion-shrug:


I'll redo the setup with my sound card and see if noise level changes.
 
I am actively reading every post. Please continue speaking so I can advance my knowledge.
 
Reading and absorbing... Spending too much time with the damn camera to do much with audio. :oops:
 
soundhound said:
I don't know. I have a USB interface in my shop (M-Audio Audiophile USB), and its noise is almost as low as my ASUS interface. Could be noise coming from your computer over the USB power?
SH, I forgot to ask one thing, did you happened to remember the voltage going into the sound card after the dummy load resistor and volume pot? I put multimeter at the IC cable leads to monitor the voltage and it ranges from 0.5 to 0.85VAC depending on the wattage the amp is putting out. Would that be too high?
 
DIYer said:
soundhound said:
I don't know. I have a USB interface in my shop (M-Audio Audiophile USB), and its noise is almost as low as my ASUS interface. Could be noise coming from your computer over the USB power?
SH, I forgot to ask one thing, did you happened to remember the voltage going into the sound card after the dummy load resistor and volume pot? I put multimeter at the IC cable leads to monitor the voltage and it ranges from 0.5 to 0.85VAC depending on the wattage the amp is putting out. Would that be too high?

That depends on the sound card / interface. Some interfaces can't take more than .2 volts and some can take considerably more; usually the maximum is about 1 volt. You should go by the input meters on Multi-Instrument Pro and avoid clipping. Also, make sure you are not clipping the output of your soundcard or interface.
 
soundhound said:
The only thing that having the bias lower than usual would do is to make the amp clip at a lower wattage than otherwise. The bias voltage reading is independent of the AC line voltage (which was 117 VAC in the 1950s-60s rather than 120 VAC today). I think there's something else going on because those distortion sidebands are not normal operation.
I found out what happened. I made a goof. :oops: Couple months ago, I replaced output tubes (EL34) of my Dynaco Mark IV. I lowered the bias all the way down before putting the new ones in, then I was supposed to raise the bias voltage slowly after output tubes warm up. For some reason (don't recall) I never went through the raising part and left it at the lowest setting. :angry-banghead:

I brought up the bias to 1VDC and here is the result at 2W 8 Ohm.
 

Attachments

  • 2W 8 Ohm Hann 8192.jpg
    2W 8 Ohm Hann 8192.jpg
    223.9 KB · Views: 494
DIYer said:
I found out what happened. I made a goof. :oops: Couple months ago, I replaced output tubes (EL34) of my Dynaco Mark IV. I lowered the bias all the way down before putting the new ones in, then I was supposed to raise the bias voltage slowly after output tubes warm up. For some reason (don't recall) I never went through the raising part and left it at the lowest setting. :angry-banghead:

I brought up the bias to 1VDC and here is the result at 2W 8 Ohm.

Kind of surprising the distortion goes way out like that at 2 watts. I only have measurements done with TrueRTA of those amps; I might do them again with Multi-Insturment. At any rate, you can see the even order harmonic distortion component "supression" going on, which is happening in the output stage. Also, the Ultra-Linear transformer probably has something to do with that, and the distortion components way up in the harmonic series.

Pentode and tetrode tubes will always display more (and higher in the harmonic series) distortion components than triode tubes. In the 1950s the big debate was between those who believed that triodes were best because they were inherently more linear and those who believed that tetrtodes and pentodes were best, using negative feedback to supress their higher distortion. I firmly fall on the triode side. They didn't have sophisticated FFT equipment in the 1950s, but now that we do, I think its pretty obvious that triodes display the more preferable distortion characteristics. The Dynaco amps use tetrodes, with about 20dB of negative feedback, and SET amps use triodes with no feedback because of the linearity of the tubes making feedback unnecessary.
 
soundhound said:
Kind of surprising the distortion goes way out like that at 2 watts.
It's possible that my computer is having some issues. Lately, the noise floor of my computer speaker has gone up. I tried different USB port, different wire arrangement at the back and etc to no avail. I heard about Dell computer power supply caps going bad or something like that in the news couple years ago... :think:
 
DIYer said:
soundhound said:
Kind of surprising the distortion goes way out like that at 2 watts.
It's possible that my computer is having some issues. Lately, the noise floor of my computer speaker has gone up. I tried different USB port, different wire arrangement at the back and etc to no avail. I heard about Dell computer power supply caps going bad or something like that in the news couple years ago... :think:

Dirty power from the USB jack would not cause harmonic distortion components exactly related to the 1kHz fundamental test frequency as seen in the graph; they would be unrelated in frequency. Various tubes from different manufacturers and build lots can have very different distortion characteristics.

As for Dell computers; I use a newly purchased T5500 "Precision Workstation" for engineering use at work; it is junk, and clients in the field also report unfavorably on their experiences with Dell computers. I have a Gateway gaming computer at home that is loaded to the gills with added PCIe cards and demanding software; it has always run perfectly.
 
:text-bump:
soundhound said:
Here is the 1 watt distortion spectra of my SET amp; the distortion components are basically gone after the 3rd harmonic, which is extremely clean.
Here is my other SET amp measurement. The previous SET is my center channel mono and this is my stereo SET tube amp.


Here's a comparison view of soundhound's and my SET tube amp. I'm not sure how critical it is to measure with same computer and same version software. :confusion-scratchheadblue: :confusion-shrug:
 

Attachments

  • 1W 8 Ohm Hann 8192 - full screen.jpg
    1W 8 Ohm Hann 8192 - full screen.jpg
    245.1 KB · Views: 466
  • 1W 8 Ohm - sh vs mine.jpg
    1W 8 Ohm - sh vs mine.jpg
    382.7 KB · Views: 466
Back
Top