|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
24.140.98.34
Here is a spectrum graph of the SE DC 2A3 beta follower circuit I had been listening to for several weeks.
I used a 10:1 passive test probe so the soundcard was not overloaded.
8 ohm/50W power resistor was the test load on the OPT.
Spectrum graph @ 1W RMS.
The 2A3 finals are biased at 43maDC.
This conservative current biasing "concept" seems to reduce headroom and increase distortion measured at the output.
My question : even if you are using high EF horn speakers, will dynamic transients clip the output and cause more unpleasant odd order distortion products?
I read somewhere that a scope should be placed across the OPT secondaries and monitored while music is played at typical listening levels.
The idea is to look for clipping behavior which would indicate lack of sufficient amplifier power to drive the speakers cleanly.
I have not tried the scope idea yet.
Any comments?
Edits: 03/28/14Follow Ups:
Using Paul Joppa's output tube operating point rule of thumb, 43ma at 325 volts into a 5.6k plate load would give normal power and distortion specs.
If you're talking about 43ma at 250 volts into 2.5k or so....then I'm not surprised.
The operating point is not right for the tube.
When you lower the current, you have to increase the load impedance to rotate the load line more horizontal to keep the tube from operating in the non-linear cutoff region as seen in the lower right half of the plate curve graph where the grid lines all bunch up.
Looking at the plate curves, if the idle current point is simply dropped from 60ma to 43ma (without adjusting the load impedance to rotate the load line towards the horizontal) the load line just drops smack into the non-linear cutoff region and distortion goes up.
That is to say, the plate voltage still decreases by the same amount for the positive half of the input grid signal but (because of the non-linearity of the tube in the cutoff region that the load line is now trying to operate in) the plate voltage no longer can increase by the same amount for the negative half of the input grid signal.
Any plate voltage movement, that is not symmetrical and linear, + vs. -, is what we call harmonic distortion.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
2A3 plate = 480VDC , 2A3 cathode = 235VDC
245 x 43maDC = 10.5W plate
2.5K primary Z of OPT.
These are close to the recommended 2A3 finals biasing points for a "modern" SE DC 2A3.
I guess if the speakers never use close to 1W of power, the distortion would be a lower and perhaps not as much of an issue.
"I guess if the speakers never use close to 1W of power, the distortion would be a lower and perhaps not as much of an issue."
Perhaps, but it is my belief that the harmonic distortion percentage will be higher at all power levels when the operating point for the output tube is wrong vs. when the operating point for the output tube is correct.
Higher idle current and a higher load impedance will place the tube in a more linear operating condition for all power levels not just at the higher power levels.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Hi Dt,
I use a 5k opt on my dc45 but use 8 ohm taps for 16ohm speakers, this presents a 10k load to the finals.
Could you try something similar to increase the opt load or change the opt to a 5k.
If I understand correctly, this gives lower distortion with only a few mw reduction in power.
Headroom and dynamics are unbelievable on my horns.
Smart
I tried a 5K plate load for the 2A3. Didn't change anything else.
Here are THD measurements :
4.3% @ 1W , 3.1% @ 3/4W, 2.1% @ 1/2W and 1.15% @ 1/4W.
Spectrum at 1W RMS:
Fundamental : 1KHz sine wave @ 0dB
2nd harmonic : -28.41dB
3rd harmonic : -41.50dB
4th harmonic : -67.84dB
5th harmonic : -65.60dB
The performance is a little better, but still not optimal.
I will try again the 2.5K plate load and higher bias current.
-28db is 4%
Something isn't right.
A 2a3 at 1 watt should not have 4% 2nd harmonic distortion.
Is your driver stage clean?
Is you audio signal generator clean?
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
I compared both external AF generator vs. internal signal generator.
@ just over 1W output with 1.327VAC input sine wave:
external AF generator measured : 3.75% THD
internal signal generator measured : 3.41% THD
The internal generator is a bit cleaner.
Still using the 5K plate load and biased at 43mADC on finals.
Why the THD measured lower today than yesterday is a mystery to me.
I might swap in a biplate 2A3 and see anything changes.
I meant you should measure the distortion of just the generator.
Then measure the distortion of the amp at the grid of the output tube so you know how much of the total is attributable to something other than the output tube.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
I have two 2A3 spice models one based on the RCA and a Sofia.
At 43ma (1w) and 2.5K I get 2.2% for the RCA and 3.8% for the sofia.
I set the bias to 43ma and the grid to -50V for each and let the B+ fall where the model put it. (I did not make either of the models but I can trace their curves if anyone wants)
sofia 2A3
RCA 2A3
Run 60mA and repeat your measurements.
it would be interesting to look at the spectrum of 1/4 watt and 1/16th watt to get an idea of how "quickly" this distortion shows up.
Also can you give an honest assessment of your sound for the few weeks you listened to the amp prior to running the FFT?
dave
Dave,
I liked the fact that you mentioned 1/4 Watt and 1/16th of a Watt as power levels. Running GPA 604s in MLTLs, we measure the power used to be under 3/4s of a Watt almost all of the time!!!
If thats the case, power and distortion at 1 Watt, and clipping at maximum power, would be not relevant if the amp does'nt need to operate there, but, say, for 1% of the time.
Regards and cheers,
Jeff Medwin
It's proportional with level if 2nd harmonic distortion is prominent . He measured over 5% at 1W . That's chocolate sauce on your steak as Allen Wright used to say . With a resistive load on the driver there may be some cancellation but that 2k5 load does seem awfully lowAl
Edits: 03/28/14
I had the amp on some Optimus ProLX5's , not full-range or high EF.
They were just workbench monitors for my tube amps played nearfield at lower volumes, hardly demanding by any means.
In that application, the circuit sounded alright.
But I was more interested in the objective performance of the circuit, which I have discovered is poor in terms of THD.
No doubt, the biasing is a problem, as Tre noted.
last test I ran I came up with 5% distortion @ 1 watt with 6ej7 driver tube.
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: