I finally measured this thing a bit. Using Rightmark, I get -3db at 30hz and -1db at 20khz. Both roll off evenly with very flat output between those values. I don't know how to really do distortion with this program, as it seems to average the two channels. Assuming that, I seem to get IM and THD about 4 times that for the card itself. The card with just a loop connection gives about 0.006% each. The amp in one channel and the card in the other give about 0.015 THD and IMD.
Sheldon
Sheldon
I tried again today. I wouldn't put too much stock in the results, as I'm not sure how the RightMark software computes distortion. If I just test the soundcard, I get THD and IMD of around 0.005%. If I have the amp on one channel, it's about 0.015%. If I only record the amp, I get about 0.3% each, but I don't see any way to compute for one channel only. A way to record only the amp, is to unplug the other input after the levels are set and you click ok on the level dialog box. Otherwise, as Sy said in his Jiggy thread (unfortunately the program he used is not available now), it won't let you continue. I'm not sure what it then does for the disconnected channel as there is a baseline level recorded for that.
Anyway, those levels seem reasonable, if in fact they're based on anything real. I do trust the frequency response graph, as I can read that.
The measurements were made at a reasonable listening level (87 db speakers), so I'd guess around 2 watts.
Sheldon
Anyway, those levels seem reasonable, if in fact they're based on anything real. I do trust the frequency response graph, as I can read that.
The measurements were made at a reasonable listening level (87 db speakers), so I'd guess around 2 watts.
Sheldon
measurement results
Here's the data table for several conditions:
card only card/amp amp/amp low amp/amp loud
response .03,-.01 .2,-2 1.6,-1 1.6,-1
Dy. Range -93db 87 69 82
THD% .0013 .0023 .023 .14
IM% .007 .013 .3 .14
crosstalk -86 83 69 75
Card only is just loop out to in for both sides of the card.
Card/amp is card on one channel, one amp side on the other channel. Amp output is through 8ohm resistor, no speaker.
Amp/amp low, is through amp on both channels at moderate to low listening volume. Amp is driving speakers on this and next.
Amp/amp high, is same but at higher amp output - fairly loud listening level. A couple of watts on 87 db speakers?
Card/amp outputs attached:
Here's the data table for several conditions:
card only card/amp amp/amp low amp/amp loud
response .03,-.01 .2,-2 1.6,-1 1.6,-1
Dy. Range -93db 87 69 82
THD% .0013 .0023 .023 .14
IM% .007 .013 .3 .14
crosstalk -86 83 69 75
Card only is just loop out to in for both sides of the card.
Card/amp is card on one channel, one amp side on the other channel. Amp output is through 8ohm resistor, no speaker.
Amp/amp low, is through amp on both channels at moderate to low listening volume. Amp is driving speakers on this and next.
Amp/amp high, is same but at higher amp output - fairly loud listening level. A couple of watts on 87 db speakers?
Card/amp outputs attached:
Hmmm. I've just read through this thread. Comments on the test measurements, though. Looking at your THD, you have a crap power supply. That's why you have sidebands around the distortion spectra. Aside from that, the spectrum is very nice indeed. In fact, it's suspiciously nice. What level were you using? What does the spectrum of the measurement system look like (do you have confidence in your measurements)?
EC8010 said:
The power supply schematic is shown above. I don't know how to exactly model it in psud, but an approximation shows about 7V of ripple for the B+ of the output tubes. The driver tubes are much lower. The output obviously has good psrr, as I get less than 1mv ac on the output, but I don't know how the ripple will modulate distortion.
With this simple setup, I don't know how to get the output level, as it's not a steady tone and it's difficult to measure with a digital mm. But I was estimating a watt or two, based on the listening level, as I had the speakers hooked up during some of the measurements (as indicated in the text). I don't know what the spectrum of the measurement system looks like, other than to say that in the card/amp examples, you can see the output of the system without the amp. That is, the line out loops directly back into line in for the "card" channel. I do hear the output though. One of the tones is identical to the MLS for Speaker Workshop. there are at least two additional tones, including one at lower frequency. If you've got better ideas on how to measure, I'd be happy to try. If I keep at this long term, I'll get proper measuring gear.
Sheldon
What I'm getting at is that you shouldn't be seeing that cluster of sidebands spaced 120Hz(?) away from the 1kHz fundamental and its harmonics. In my experience, that indicates a poor power supply, and often that the amplifier is going into Class B. That's why I wondered about what power level you had.
But you've lost me by saying that the output isn't a steady tone. Am I misinterpreting your images? I had assumed that your THD spectrum was the result of applying a steady 1kHz tone and then doing an FFT on the output of the amplifier?
But you've lost me by saying that the output isn't a steady tone. Am I misinterpreting your images? I had assumed that your THD spectrum was the result of applying a steady 1kHz tone and then doing an FFT on the output of the amplifier?
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.