Medium Power SET, EL156 based

[Joshua_G]

As I wrote in earlier post:
Recently I purchased an amazing low cost SET integrated amplifier: MP-401 EL156 Integrated Amplifier [MP-401] - $599.99 : Musical Paradise - Audio Art . It's the first SET I owned or listened to and I'm highly impressed by it's sound quality, which, as I understand, is to be expected from good SET amps. This amp comes with Chinese EL156 and it can accept various tubes. I tried it with current production 6550 and KT88, but I like best the sound of the EL156.

Alas, this amp with it's declared 15Watt doesn't have enough power to drive my speakers to their full ability. Also, in high listening volume it distorts.

Being retired I cannot afford buying a commercial good SET with enough power (>20W). Such amps are way beyond my means.

Consulting a guy I know with over 20 years of experience building tube amps, he told me that the problems I encountered with this amp are probably due to poor design and low quality OPT. He said that EL156 SET, deigned properly and with good OPT, should give me enough power to drive my speakers which have sensitivity of 91dB and only 2 elements and it should sound great.

So, I decided to build an amp based on EL156 and pay the cost of good quality OPT's and chokes. That much I can afford.

As for the topology, he suggested 2 stages only. Also he said that the best approach is anode choke loaded driver, since anode CCS adds high order harmonic distortion.

The schematic of my initial idea is posted below.
From the triode mode anode curves, what I got is over 30W output.
PSU schematic will follow.
All comments and suggestions are welcomed.

[Arnulf]

You've got 50W quiescent dissipation. 30W output sounds very optimistic figure for triode mode. Can you post the loadline plot (which I assume you have handy since you calculated the relevant values for your schematic) ?

Preamplifier cathode resistor, something isn't quite right there (U = R * I, but not in your case of 1.4V, 50R and 22 mA).

By the way, don't worry about calculating things out to two digits past decimal point, your real life tube is likely to deviate from your calculation by 10% or so anyway

[Joshua_G]

In the driver, Ia is 22mA and Ig2 is 6mA, so Ik is 28mA which drops 1.4V over 50R.
A copy of anode curves for triode mode with 3K dynamic load is attached.
On the 3K load line I plotted the delta voltage and delta current between Vg=0V and VG=-48V, multiplied the 2 which gave me peak power and divided that by 2 which gave me approximate RMS power.

[Arnulf]

Originally Posted by Joshua_G

On the 3K load line I plotted the delta voltage and delta current between Vg=0V and VG=-48V, multiplied the 2 which gave me peak power and divided that by 2 which gave me approximate RMS power.

You've got 450V peak to peak and 140 mA peak to peak. In order to do the "divide by 2 to get RMS value" thingy you need to use peak values (wrt quiescent point) only, not peak to peak.

450 p-t-p = 225V peak = 160V RMS

140 mA p-t-p = 70 mA peak = 50 mA RMS

Pout(RMS) = V(RMS) * I(RMS) = 8W in your case

The "divide by 2" shortcut works because there is a 2^(1/2) factor in each term (U and I) when calculating RMS value from peak, and 2^(1/2) * 2^(1/2) = 2. When using p-t-p values you should be dividing by 8 to get RMS power value.

[revintage]

Sorry Joshua, but you must have some kind of error in your calculations. You will not get over 10W out. Even in pentode mode you will not be near 30W.

You get a p-t-p swing of ca 460V. 460/2=230V peak, rms is 163V. 163*163/3000=9W

To lighten the burden of the tube, don´t go over 40-45W Pd.

If you want as much power as possible with low distortion a Schade D3a/EL156 could be an option.

EDIT: Ooops Arnulf, our posts crossed

[Joshua_G]

On the 3K load line, at Vg=0 Va is about 162.5V and Ia is about 170mA while at Vg=-48 Va is about 620V and Ia is about 20mA. So, dV is 457.5V and dI is about 150mA. Multiplying dV by dI gives 68.62Watt. These numbers are for p2pV and p2pI, thus it is peak instantaneous power. Thus, dividing peak power by 2 gives approximate RMS power for sine wave, which in this case is 34.31Watt. Look here: How to Design an SE Power Amplifier .

I don't know right now what is the idle anode power consumption since I don't know yet the DC resistance of the OPT which is to be ordered from Sowter. Once I'll know the DC resistance of the OPT, the idle working point will be finalized and I intend not to go over 40Watt at idle. While listening to actual music, peak output power is very seldom reached, so safety power margin is usually given at idle.

[Joshua_G]

Originally Posted by revintage

If you want as much power as possible with low distortion a Schade D3a/EL156 could be an option.

What do you mean by Schade? Do you mean omitting the coupling capacitor? I don't think it's possible with driver's anode choke loaded. Right now I don't want to use anode CCS.

[Allen Wright]

Originally Posted by Arnulf

You've got 50W quiescent dissipation. 30W output sounds very optimistic figure for triode mode.

Optimistic isn't the word, to be blunt, dreaming is closer.

There is NO WAY you will have over 50% efficiency from a single SE triode in Class A.

I don't care what your curves may say, I predict from experience that with 50 watts anode + screen dissapation, you's be lucky to actually get 25% efficiency, after OPT losses etc, which equals 12 1/2 watts, unless you are happy with 10% distortion.

If you force the EL156 into Class A2 you might get more power, but a pentode drive can't do that without the addition of something like a CF driver.

So the 15 watts claimed for your Chinese amp is also doubtful, but as you have it there, run a power test on it. Find the max output before clipping, then back down until the distortion is acceptable. I expect you will be in for a disapointment.

The other point is if you do have 15 watts now and it's not loud enough, then 20 watts will not make any difference at all. To double your volume you need 10X the power - TEN TIMES - so a big beefy P-P amp with 4 KT88s per ch (or 4 x EL156) is more like it.

Sorry to be so down on your plans, but as money is short, you can't afford to take too many blind alleys.

Regards, Allen

[Joshua_G]

Thanks Allen.
I don't lack loudness with the present amp, it only distorts and seem to be choked on loud listening levels. I have a power amp with PP 6550 and I don't like its sound at all. It seems to me that no PP amp can give the details and micro-dynamics that a SET amp gives.

[Allen Wright]

Originally Posted by Joshua_G

It seems to me that no PP amp can give the details and micro-dynamics that a SET amp gives.

Then you have been listening to inferior P-P amps.

If you send me (aw@vacuumstate.com) your email addess, I'll send you the schematic of a KT88/EL156 P-P amp that will match and most likely better the sound of any SE amp you are likelyto hear. Maybe you can rebuild your P-P amp to this new circuit.

Regards, Allen