|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
184.70.169.107
In Reply to: RE: 2A3 Operating Points posted by maxhifi on November 25, 2014 at 11:03:16
Last night I decided to have some fun on the bench.. I connected my 2A3 amplifier to an 8 ohm dummy load, and fed it with a signal from my audio oscillator. I used 60Hz, just so that my DMM would be capable of measuring output voltage. Interestingly, the amplifier has output to a very high frequency, over 80 kHz, the Sowter output transformer is truly amazing.
The current operating point is with 290V on the plate, 5000 ohm output transformer primary, and 44V of bias voltage (53mA of idle current), and 246V on plate to filament, I get the following.
1.77W RMS just before the grid of the 2A3 starts drawing current, and clips the top of the waveform. I measured distortion with a Heathkit THD meter, and at this power level it is about 5%. I can get a little more power if I allow the THD to get up to 10%, but the waveform looks too ugly to count this. On the scope it looks like mostly second harmonic.
I am thinking Jeff's operating point makes sense, with my output transformer and plate voltage. I could gain a little headroom by reducing the idle current to his value of 42mA, possibly at the expense of very low level distortion.
On the whole this is the worst amplifier I have ever measured, by a long shot! Right now I am deliberating if it is worth it or not to change the power transformer and raise the B+ up to a more useful level. If I do that, then I can at least get 3W out as per Paul Joppa's suggested operating point from his Robin Hood amplifier. The problem with that, is my whole amplifier is constructed with 350V electrolytic capacitors, so they would all have to go.
The electrical engineer in me says that this SET thing is a little crazy... all this effort and a pitiful result, with loads of distortion at all frequencies... however listening to music after my measuring session was finished and the amp back at its new home in my listening room tells me otherwise.
Follow Ups:
"and clips the top of the waveform."
The top of the output voltage waveform is the bottom of the load line to the right of the idle point, where the plate current is the lowest and the plate voltage is at it's highest.
Simply lowering the idle current will cause the "top of the waveform" to clip sooner.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
I must have a phase reversal happening in the output transformer, or it is also possible that I didn't return the "phase" knob on my scope to "normal" between tests. The clipping behaviour is definitely related to the onset of grid current in the 2A3, because the signal at the grid of the 2A3 looks identical to that at the output.
At first I thought it's a problem with the biasing of the driver, but it is not, it's just grid current.
I see the problem.With 246vdc across the tube and your load impedance of 5k you should be running 36ma.
246vdc, at 53ma would call for a 2700 ohm load.
53ma would be fine with a 5k load impedance if you had a 2a3 that would stand up with 365 volts across it. But there's no such thing.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
Edits: 11/26/14
I think the JJ 2A3s I bought are good for 400V
Check out the attachment I added - THD is clearly lower for the higher current (well, as it should be since it's in a more linear part of the curve), however biasing it hotter does lose a bit of power output.
I think that 42mA bias is looking not so bad
What is the speaker load going to be? What is it's sensitivity, you know, dB per one Watt input?If you have a high sensitivity speaker, the amp will be used at 50 to 500 milli Watts, MOST of the time - 99% of the time, and hardly ever over 3/4 s of a Watt.
Why design for maximum power at clipping, when what you really need in real life is "the best performing one Watt amp you ever heard".
I can send you email addresses of people who reduced 2A3 current from 250 VDC P-K and 60 mA. to 250 VDC at 42 mA. and they will tell you its nicer sounding. This is not tube curve theory, but the real deal, takes into account the most important ( and overlooked by you people, except for Dennis Fraker ) consideration of all, that at 42 mA., (a) there is LESS standing current in the tube, that the grid has to "buck" when reproducing the first Watt !!!! (b) reduced thermal stress which positively effects sonics and longevity and finally, (c) reduced unbalanced DC on the SE output's core, which has a profound effect on the SE output.
Today, there are new FILM capacitor technologies that you should investigate, called DC LINK, made by many companies, WIMA works well, this we KNOW. I am trying VISHAYs also. The voltage ratings are great, 600 to 1200 VDC, prices are very FAIR, and the measured and listening performance is wonderful, very wideband !! Check it out for yourself.
SE two-stage gets to the SOUL of the music. Push-Pull is more of a processor. DC two-stage is even better.
Have fun.
Jeff
Edits: 11/26/14 11/26/14 11/27/14
Hi Jeff
As per my above post, I think your operating point is probably not only best empirically, but it does seem to look the best theoretically too, given my other constraints. Looks like the best compromise between output, and distortion. Easy to make the adjustment too, just have to add some ohms to my cathode resistors.
My speakers unfortunately are not very sensitive, on the order of 90dB/W. This is a situation I intend to change, but change takes time, and now that I have had a taste of SET sound on my 90dB speakers you would probably be correct if you predicted I will have speakers which are 6 - 9 dB more efficient within the next few months :)
Best Regards
Max
Just to sort out a possible mis-impression, Jeff is NOT using a 5000 ohm transformer. This is important because you cannot separate the load from the voltage and current when looking at operating points. I think Jeff is using a 2500 ohm transformer - not sure though.
I agree with Jeff that running lower tube temperature gives longer life. I believe you are into the realm of diminishing returns when the dissipation falls below 50-80% of the maximum - depending on how conservative the original rating may have been.
I also believe that Jeff's experience of improved sound at lower current (relative to the conditions in RCA's original data sheet) is specific to the amps he has listened to, and cannot be generalized to other designs, nor ascribed to the output tube in isolation.
"operating point" is voltage and current, operating condition is voltage, current and load impedance.
At 42ma 250vdc the load would need to be 3.9k.
While a 5k load will work at 42ma 250vdc, it's not optimum.
Thankfully tubes are pretty forgiving.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
I orginally bought these transformers for a single ended EL84 project. I put this amplifier together using things I already had on hand, to get a taste of what real triodes are capable of. If doing it from scratch, there's no way I would pick a 5k impedance transformer but at this point it's no longer a variable. I am just trying to optimize what I have rather than spend money on things like new transformers, as my longer term project is a 300B amplifier and I would prefer to put funds in that direction.
I hate to say this, but convert it to a 45 amp? Just throwing that out there.
I love your driver stage. It's what I would want.
Jamie
Big speakers and little amps blew my mind!
I actually did have that in mind when I first built it in 2006, the driver stage was inspired by a Japanese 45 amp. I went with the 2A3 becasuse they are available new. With the JJs on the way I am more or less committed to the 2A3 for the time being. To be honest this little amplifier really sounds good as is, I am just in the process of optimizing it. The operating point may well not be ideal, but it has as much power as a 45 would make and it is extremely enjoyable to listen to. That said it may have a future application driving a tweeter or mid horn in which case the 45 option will be revisited.
"With the JJs on the way I am more or less committed to the 2A3 for the time being."
No your'e not. If you can design and build your own amps you can do anything you want. $120 mistakes are cheap in the high-end world.
Jamie
Big speakers and little amps blew my mind!
Is the secondary tapped?
A 8 ohm speaker on a 16 ohm tap gives you a 2500 ohm primary.
Just thinking out loud.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
There is the extra primary inductance and better HF but the disadvantage are the copper losses of a 5k primary and the 16 ohm secondary tap which would be much higher than a 2k5:8 transformer . Sowter's 5K SE TX already has high copper losses , especially on the primary .
Al
Yes, I certainly could give that a try. The only reason I haven't gone down that path, is the transformer will be more lossy that way. Us engineers can tend to hit a mental road block when entering the "use outside of design intent" territory, I am happy for this suggestion, especially since it should have been an obvious one.
This extra transformer loss may be more than compensated for though by the tube being loaded properly and making more power. In either case this is sure easy to test.. will try it tonight and see what happens. I suspect I will get better power output and a slightly worse damping factor.
What am I missing?
Dennis/ Jeff have been criticised for suggesting a 2k5 load was suitable for the ~ 42mA 250V operating point... and that was into a pair of speakers that likely did not need a low impedance drive. Now 2k5 is considered a more suitable load than 5k?
Well, it will net you higher power... at the cost of distortion and lower damping factor, which I imagine could be an issue with your speakers (it is with most speakers). OTOH, I guess it is "proven" is Dennis' amps.
I think I need a break...
Cheers.
"In the beginner's mind there are many possibilities, in the expert's mind there are few." Shunryo Suzuki
When I said "A 8 ohm speaker on a 16 ohm tap gives you a 2500 ohm primary."
I was just saying he wasn't stuck with 5k. He could go for 250vdc 60ma. 2500 ohms for instance.
But at 250 42ma, 5k is too high (but might be fine assuming the power output is enough for his situation) and 2.5k is too low leading to high HD at all power levels.
I really wasn't trying to confuse anyone.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
I am just a little (lot!) sleep deprived and therefore have pulled my cranky pants on.
Sorry, I did not mean to hassle.
Cheers.
"In the beginner's mind there are many possibilities, in the expert's mind there are few." Shunryo Suzuki
H Tre, can you please clarify/ confirm:
> > > While a 5k load will work at 42ma 250vdc, it's not optimum. < < <
I thought the increasing the load would lower distortion, considerably. I am guessing that by "optimum" you mean that more power can be achieved at clipping or similar?
I need about a Watt for my speakers in my room at my loud listening levels; would a 5k load be more ideal than ~4k?
Thanks Tre.
"In the beginner's mind there are many possibilities, in the expert's mind there are few." Shunryo Suzuki
Optimum (in terms of loading a power tube) is not anything that is cast in stone.
Normally, a designer would want to pick an operating condition that would give a reasonable amount of power at reasonably low HD.
A operating condition (idle current, voltage and load impedance) that gives the straightest transfer curve (lowest HD) wouldn't be the one that gives the most power before clipping.
Only requiring 1 watt should make it easier to achieve low HD.
Tre'
Have Fun and Enjoy the Music
"Still Working the Problem"
.
"In the beginner's mind there are many possibilities, in the expert's mind there are few." Shunryo Suzuki
Thanks for writing.
I would consider two speaker designs as being the most conservative, lowest financial risk choices.
GPA 604 in a MLTL enclosure, or a ALTEC A7-800 two way as a lower cost and easier to source alternative.
You have private mail.
Jeff Medwin
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: