• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Effect of lower G2, EL84 PP

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Thanks for that PDF Ian, I haven't seen it before. Straight to my collection ;)

It seems there are some benefits from the shared resistor, but I think I will reduce its value and see if I can get a little more power from the output stage. I'll take it easy, though, as I don't want the G2 getting wrecked!

Interesting that this amp has no grid stoppers at all on the output stage, although the schematic as provided by the seller shows them. I haven't had any stability problems with these amps, is oscillation the only reason for the stoppers? AFAIK it is, but still learning plenty about valve topologies.

cheers, Jacob
 
That Retrovox site has some other interesting Application Reports noteably for 807.

A bit OFF Thread (OT) but for your possible interest - OZ tube history.
STC was one of only 3 Australian Tube Factories (AWV in Sydney and Philips in Adelaide were the other 2). STC Engineers were trying to keep up with technical articles coming out of the AWV factory where Fritz Langford-Smith was Chief Engineer and was busy editing the Radiotron Designers Handbook. STC factory was also a Western Electric manufacturer and were making STC4300B (300B) etc.

Cheers,
Ian
 
Grid stoppers are used in hi-fi amps to prevent oscillation. They are used in some guitar amps to provide crude tone controls, and compensate for poor layout.

Bear in mind when looking at rise in current with signal that this is caused by second-order distortion in the valves, and for this the DC increase equals the 2nd harmonic. 10% DC rise means 10% 2nd, other things being equal. You don't hear this 2nd as it cancels in the balanced stage. It does mean that the valve is being quite non-linear, so re-entrant distortion is possible. To support this, look at the graphs of power output vs input voltage. They start as a parabola (power rising as square of voltage) but soon turn to almost a straight line. This shows compression i.e. odd order distortion (mainly 3rd). Some of this may be re-entrant distortion, as re-entrant tends to be compressive.
 
Look at this Application Report for the 6BQ5 which is an EL84:
http://www.retrovox.com.au/STC6BQ5.pdf
See the last couple of graphs for Class AB1 push pull. The total screen current (two tubes) does not change much until you get to very high power output levels. That is because as the screen current on one tube of the push pull conducts more on one half cycle the other conducts less and then they swap for the other half cycle - so the TOTAL screen current from both tubes is fairly constant. The implication of that is that you can use a common screen resistor for both tubes and get almost screen regulator performance.

What's the point of doing a higher power amp if you cannot use it because it doesn't work and sound right at medium and high levels? You' are better off with a simpler triode or triode-connected PP! If you simplify so much the power supply a pentode connected amp cannot compete. Also that data sheet refers to an amp without big screen resistor. Distortion gets worse with a big screen resistor, especially intemodulation. You can use as much feedback as you want to reduce distortion a posteriori but the sound will not improve that much. The triode connected amp will still sound a lot better.
I agree with DF96, the sound is quite compressed and it seems that the amp hasn't got the power you measure on the bench.
 
Last edited:
I am running my current class A amp with a 8K5 shared screen resistor on the output el86's. It runs in Class A and has plate to grid feedback. No hint of compression in the presentation. Sounds considerably better than an amp which is almost the same but uses triodes as outputs. The two amps are not exactly comparable since the pentode one has more feedback and lower output impedance.
However I would like to see some actual measurements (which I cannot perform) to settle the matter somewhat.

i might try scoping up the screens to see how stable they are under signal load.

Shoog
 
Last edited:
Sounds considerably better than an amp which is almost the same but uses triodes as outputs. The two amps are not exactly comparable since the pentode one has more feedback and lower output impedance.
Shoog
If fact they are not comparable in general. It becomes rather arbitrary....
You should compare triode and pentode configurations of the same amp (although I cannot see the point of using plate to grid feedback for triode connected pentode) or even better to regulate the G2 voltage and compare the two G2 configurations.
I have only seen that sort of high G2 shared resistor in guitar amps which are completely different animals.
 
Last edited:
As you can see from the Mullard data-sheet, probably the most complete, the shared resistor should be 1K8. In that case the average result will be Pout = 15W with 3.4% THD. The equivalent 300V amplifier with self-bias, same 8K load, same anode current and without G2 resistors gets to 17W at 4% THD and at 15W THD is less than 2%.

http://frank.pocnet.net/sheets/129/e/EL84.pdf
 
Last edited:
I hate to be picky - but that datasheet describes operation deep into Class AB.

The two amps I am comparing use exactly the same circuit configuration apart from that they use different output tubes. The point of using plate to grid feedback in a triode amp is that it still lowers output impedance and linearizes the output - it just does it to a lesser extent than in a pentode amp since there is less mu to convert to gm.

Shoog
 
I hate to be picky - but that datasheet describes operation deep into Class AB.


Shoog

I am afraid it is not deep class AB. It works in ultrapure class A up to 6W. From my point of view it is class A up 8W as the total current draw is nearly constant. The quiescent current for each valve is rather high at 35-36 mA for 300V plate voltage which is close to max plate dissipation.
I can also tell you that the ultra-linear version, using the right speakers, doesn' t need any feedback because THD and IMD can be surprisingly low and crossover distortion negligible if the amp is well sorted-out. If they do not recommend pure class A for the push-pull there is a reason: there is little to gain from it. You just get less power with higher distortion for the same Pout, in particular when you are close to Pmax for the class A amp.
Class A is not immune from the regulation problem as both G2 and anode currents increase quite a bit even in this case. The behaviour of G2 currents as function of input signal in the two cases is not such that in one case is a problem and in the other an advantage....
 
Last edited:
Thanks all & thanks Shoog for following up.

Amp1 of the pair is undergoing a major re-engineering into a new chassis that will allow much easier experimentation, as the original is quite compact.

So amp2 is in use, being the totally unmodified (except for safety) it's an interesting comparison. Looking forward to being able to really get into Amp1 mods once I have it all assembled, the new chassis will have plenty of space for playing with G2 supplies, cathode bias schemes, etc :)

cheers, Jacob
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.