• 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.

why do wimpy drivers for 2A3 work as well as they do?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Have a look at the TL431 stability graph. I suspect 10 nF would be OK. I have had them sing their heads off without them. Point taken about easy set up. A simple CCSink if avoiding Early efffect is almost totally linear as a current amplifer. In fairness so is an LM317. The LM317 is approximately a 741 with a pass transistor . TL431 is of the same era. In my heart of hearts a LED and a BD139 is as good as it gets ( or whatever transistor ). Any minor set up problem is as nothing compared with the results. I always use LM317 then design a better version. The LED can be dropped below the chassis if temp co is an issue. LED is better than 1N4148 for that. I do find the valves themselves make remarkable CC devices if used with a boring cathode resistor. However for safety a CCSink is great. CCS is a different story. This simulates using a much higher resistance.
 
Is there a technical reason for this typical lack of highs? It was argued that slew rate limitation at the upper limit of hearing was not necessary with 2A3. I would be interested to learn if anyone successfully maried ecc83 to a SET.

I can't give you a technical reason but I dismissed the ECC83 in the circuit I built last week (schematic in this thread). Just could not get it sound "right". The SL7 driver does a better job in the same topology, check out Soulmerchants' work.
 
I was just reading up on this. It's been suggested (elsewhere) that in order to get 'good sounding highs' you want the driver stage to not slew limit on frequencies up to 5x (or maybe 10x) the highest frequency of interest. According to this rule of thumb, if you want to make sure high freqs up to 20kHz 'sound good' then you want to make sure your driver doesn't slew limit into 100kHz (at full drive voltage).

The by-the-specsheet Miller capacitance of a 2A3 is about 75pF or so, but let's say it's 100pF worst case.

Thanks to gingertube, I have this handy formula:

slewrate = 2 pi F Vpeak / 10^6 Volts per microsecond (V/us)

Signal current = Slewrate (V/us) x C (uF) Amps

So if 100kHz is the highest frequency of interest and the input signal to the 2A3 grid is 50V peak, then the driver will need to sink about 3mA into the 2A3 grid. That gives the headroom suggested from the above 'rule of thumb' for 'good sound.' By that measure, you'd want to use something like a 6SN7 to drive your 2A3, not a 6SL7 or 12AX7.

It's the rule of thumb that's debatable here, and beyond my abilities to justify. All I know is that a 2A3 seems to 'sound better to me' if I use something with a plate current of 8 to 15mA to drive it (6SN7, 5687 or something along those lines). 5mA should be fine. 1mA, no.

--
 
Last edited:
It is also suggested elsewhere that the importance of slew rate limitation is exaggerated. Not by me, but at least I went the mile to actually build and test such an amplifier. I can imagine that the effect depends on topology, you can't just put one rule of thumb on each and every circuit. For evidence: three people build the LW variant with SL7 and are happy with it.
 
An LW circuit with a 6SL7 would sink about 1mA max signal current into the load (2A3 grid), correct?

OK, I might as well try the math and see if it tells me something.

slewrate = 2 pi F Vpeak / 10^6 Volts per microsecond (V/us)

Signal current = Slewrate (V/us) x C (uF) Amps

Let's make F = 40kHz, signal V = 50V.

6.283 * 40kHz * 50V / 1,000,000 = 12.6V/uS

12.6 * 0.0001uF = .00126 A (1.26mA)

Yeah, a 6SL7 can drive the 2A3 grid's (approximately) 100pF of Miller C to 40kHz (double the highest frequency of interest). Maybe that's all you need for 'good sound,' in a properly designed circuit.

PS -- A negative feedback loop around the circuit would change that, though -- correct?
--
 
Last edited:
That's all well and good, but one thing I don't understand is this:

rongon calculates that about 1mA value. Ok, the 6SL7 has a DC current of let's say 2mA, and a plate resistor of 100k. If the 2A3 grid eats up that 1mA, the 6SL7 has only 1mA left. This is a significantly different load line than the one at quiescent.

Driving that kind of current, or even a small fraction of it, is going to make the preamp do some very unlinear and nasty things.

A resistor plate loaded tube simply cannot deliver a very big fraction of it's current to a load, without seriously compromising linearity. 50% of quiescent current: not an option in my opinion.
 
Last edited:
Yes Curwen. No idea why it works so well. It shouldn't... my crude tests so far using Rightmark audio analyzer don't show high THD at all either.

btw - The only reason I didn't use 12ax7 was because the holes in my test chassis were too big. :)

Serious though.. it really shouldn't work but it does. I have heard many superb amplifiers. Built a few of them too. This is by no means the first time I build a single end 2a3... all others used far more grid current.

I think if you used a coupling cap or cathode by-pass it would not work nearly as nicely.
 
Last edited:
soulmerchant:

What exactly did you test?

THD at 1kHz. But I need to re-do the measurments recording actual power ouput. The spectrum looked suprisingly good all things considering - such as old laptop soundcard, etc..

That bad thing about laptop soundcards is that laptops are noisy. I really need to decide on either an external sound card or external digital ossiloscope is the best investment.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.