• 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.

Tube driver for a SS power amplifier

Status
Not open for further replies.
Hello to everybody!

As I already wrote in a different thread, I need a tube driver for a SS power amplifier

I read almost all the threads about tubes driver, and I received a lot of helps.

I would like to explain my needs for the pilot and my experiments:

Needs:

- I need 40 volt rms at minimum THD+N (0,5%)
- Very low output impedence (max. 100 Ohm)
- A gain around 20
- The voltage for the tube stage is not a problem (300 - 500 volts)
- No feedback
- The simplicity and the "sound" of a triode stage (SRPP - Mu follower or "Gomez")
- No cathode follower ont the stage (100% feedback, I ear the differences)


Premises:

- In my hybrid amplifier, I try a lot of experiments, with the second stage always the same, only changing the first - tubes - stage.
- I can ear in my loudspeaker when the THD is around 1% so I need less;
- I can ear the differences from different tube stages and I like SRPP - Gomez - mu follower
- I can ear the differences (expecially in low frequency) beetween a stage with Z=2K
- I can ear the differences with feedback or not. And I don't want feedback.
- I like the sound also of hybrid SRPP or mu follower
- I don't want any electrolytic capacitor on the signal path!
- I DON'T WANT A High MICROPHONIC TUBE (like 6c45 or 5842 or pentodes, or similar)!!!

Experiment:

1) I tried a Gomez with ECC99, 440 volts, 18 mA each part. The result is a very good output Z (aroud 200 Ohm) a good sound, but still too distorsion 0,9% THD+N at 20 volt rms output. I increased the power to 500 volts, but only a little improvement on THD. I tried different polarization of the different section, I tried bypassing the two cathodes, I tried linking the two cathodes... but the result above was the best.

2) a tried a Parallel SRPP with ECC99, as above all different experiment, but worse behaviour in almost all parameters

3) I tried a mu follower with ECC99, but the THD spectrum is the worst!!

4) I tried the same three types, with E88CC at 300 volts, but very good sound, but still 1& THD+N at 20 volts, and I can ear.

5) I tried an Paralled SRPP of 6072A (12ay7), super linear tube, direct linked to a cathode follower of a parallel 6SN7. Tried all differend configuration. Powering at 440 volt, I have execellent results (the ones I need): 20 volts rms out at 0,2% THD+N; Z out around 100 Ohm - an execellent bass control (slew rate also very very good in consequence) BUT BUT, NOT the same lively, great, sweet sound of SRPP or triode although their weakness in the mesurement.


The question is: May I have a sound like the one I mentioned, better in all categories, without the problems of the different configuration and with the request I mentioned above?

The idea is to use a very good sounding and low Z triode, like (for me) 6922, ECC99, 6H30 (also if... too small gain), 6SN7, alone or paralled...

...Or using of a simple Hybrid stage (with mosfet or BJT as upper dispositive)

I really need help, with if possible (please!! 🙂-))) the schematic

Thank you very much!

G.M.

p.s. Sorry for my really bad English...
 
Thank you!

Sorry, my message come from a non-expertice. So any help is of course welcome!!

According to that, my aim is of course not to teach anything to anybody...

What I wanted to say in one point of my message was simply I don't like the sound coming from a cathode follower when a high swing of volt is needed as in drivers for final stages (SS or VT)

I also said I like the sound of SRPP


Reading my thread, have you any schematic helping me?

Thank you in advance.

G.M.
 
GM,

Maybe you need some more "sand". Gary Pimm's CCS loads completely isolate the gain triode from the load. Since the triode "sees" a GOhm range load, you get both the full mu and VERY low distortion. GP's circuits have a LOW O/P impedance, especially when the tube's plate current is relatively high.
 
Needs:

- I need 40 volt rms at minimum THD+N (0,5%)
- Very low output impedence (max. 100 Ohm)
- A gain around 20
- The voltage for the tube stage is not a problem (300 - 500 volts)
- No feedback
- The simplicity and the "sound" of a triode stage (SRPP - Mu follower or "Gomez")
- No cathode follower ont the stage (100% feedback, I ear the differences)

40 volts rms, 100 Ohms >> 16 watts ??
 
analog_sa said:


I am getting more and more worried that Ohm's law has finally been leaked to the public. Please do not abuse it.

Hi !
I recon I was somewhat "provocative", just because it is not the first time I read such things about "internal impedance" that I can't understand.
It seems that (for exemple) a SRPP built with a 12AX7 exhibits such numbers, that is: low internal impedance around 100 Ohms and hi level output near 40v rms, but is unable to provide both at the same time.

What load is it really able to "cleanly" drive ?

Simply bothered !

Yves.
 
Hi Yves

The output impedance of a stage is just an impedance connected in series with a voltage source. If unloaded there is no power dissipated. In your SRPP example the output impedance is not really dependent on the voltage output. That is if there is no load. As soon as you load it things change dramatically.
 
analog_sa said:
Hi Yves

The output impedance of a stage is just an impedance connected in series with a voltage source. If unloaded there is no power dissipated. In your SRPP example the output impedance is not really dependent on the voltage output. That is if there is no load. As soon as you load it things change dramatically.

Indeed, but obviously, this exemple does not reduce to a 40v voltage generator in serie with a 100 Ohm res.
If so, it could be able to deliver 20v into a 100 Ohm load, that is 200mA current swing, well above the capacity of a 12AX7 😀

This is why I'm bothered with this 100 Ohms claim.

Yves.
 
Hi GianMaria

what is the input resistance (impedance) of the
SS stage? 30k ? 10k ? I need it to make some simulation.

Have you considered a mu-follower
followed by a White- follower.
I think this combination has the lowest distortion and
the lowest output resistance.
It requires two capacitor, however.

Federico
 
Yvesm said:


Indeed, but obviously, this exemple does not reduce to a 40v voltage generator in serie with a 100 Ohm res.
If so, it could be able to deliver 20v into a 100 Ohm load, that is 200mA current swing, well above the capacity of a 12AX7 😀

This is why I'm bothered with this 100 Ohms claim.

Yves.

Actually it reduces to a 40v generator in series with 100ohms//output stage impedence. When the load is very small (very high input impedence to the next stage) you can ignore the load. It then reduces as analog_sa said. I guess the convention is that you can ignore the contribution of the next stage if it's input impedence is as least 10 times higher (YMMV) than the output impedence of the driver stage. Ten times 100ohms in this case.

Analog did not say that your generator had unlimited current capacity or power capacity. You still have to live within those. It's just that IFyou are within the limits of the driver, AND your load is greater than at least 1000ohms, then the load should not cause much more distortion that what is inherent to the driver.
 
Sheldon said:


Actually it reduces to a 40v generator in series with 100ohms//output stage impedence. When the load is very small (very high input impedence to the next stage) you can ignore the load. It then reduces as analog_sa said. I guess the convention is that you can ignore the contribution of the next stage if it's input impedence is as least 10 times higher (YMMV) than the output impedence of the driver stage. Ten times 100ohms in this case.

Analog did not say that your generator had unlimited current capacity or power capacity. You still have to live within those. It's just that IFyou are within the limits of the driver, AND your load is greater than at least 1000ohms, then the load should not cause much more distortion that what is inherent to the driver.

Thanks for trying to enlight.
Now, it remains to me a way to measure internal source impedance by injecting a small signal "backward" to the output of "unit under test".
It looks like you tell: the internal impedance is true only for low levels.
Anyway, the true drive capacity of such a stage is still unknow (to me only ! )

Yves.
 
Yvesm said:

Anyway, the true drive capacity of such a stage is still unknow (to me only ! )
Yves.

I'm by no means an expert, but those here will steer us straight if we get too far off the track. Soo,

I would think that you satisfy drive requirements if, and only if, you can supply the maximum current required by the output stage, the voltage swing needed, and you have an appropriate impedence ratio. So no, just having the proper impedence ratio is not enough. And you probably want some extra in each factor to provide a good operating margin.

For me the trickiest part is the current part, because it depends on the input capacitance of the output stage, and the frequency extension you want. Kuey has a good discussion on that topic here: http://www.fortunecity.com/rivendell/xentar/1179/theory/seamptheory/SEAmplifiertheory.html

Maybe it helps to think of power amps and speakers as a partial analogy (partial because we don't have the capacitance issue). You'd like to have an impedence ratio of about 10:1. Let's say your amp has an impedence of .5 ohms and can deliver 1 watt. Your speakers have an impedence of 6 ohms and have an efficiency of 80db but can handle 250 watts. The impedence ratio is fine, but you still can't drive the speakers to full power. You need more juice.
 
Thank you to everybody!

To fscarpa:

The input impedence of the SS state, is around 10K, but I think is not completely resistive... also because the two stages are capacitor (2,2uF by now) coupling

That's why I need a very low output impedence from the tube stage. A low output impedence it seems to me to sound better. I'm sure it sounds better in this contest. high slew rate is a very fast and clear soud.

But this impedence with a low distortion must be obtained without in-out feedback. So white follower...

40 volts rms are needed from the tube stage because the SS stage delivers at maximum 200 watts on 8 Ohm.

At the moment I'm listening to a very good and simple circuit, also if we have a cathode follower (as I wrote, It seems to me the c.f. "compressed" a little bit the sound).

It is made by:

- SRPP of 6072A non cathode bypassed (gain around 20) (Rk=Rak= 850 OHM) followed and direct coupled with a 6SN7 in cathode follower (rk=15Kohm 15 watts not inductive)

- al tubes are paralled

- power supply 440 volts

- THD + n at 40 volts on 10 K resistive is around 0,8%

- very low output impedence

What do you think?

p.s. And if we put another 6sn7 instead of the 15k resistor with grid at 0 volt? what are... and if... are there advantages?
 
Small signal performance vs large signal performance.

It is not strange that a circuit have a output impedance of e.g 100 ohm but not being able to deliver max output voltage in a similar load impedance, the difference is small signal performance vs large signal performance.

for instance every OP amp is specified this way with an output impedance measured at very low signal level and a max swing measured at a current that doesn't allow the load to be anywhere near the small signal output impedance.

In this case gianmaria is asking for a circuit that can deliver 40V at a load impedance of 10kohm and an unspecified capacitive load. If we knew the capacitance it could be possible to calculate what the drive requirements are and come up with a suitable output impedance but I seriously doubt that as low as 100 ohm is needed.

If OTOH 100 ohm is needed due to the load requirements the circuit has to be able to deliver 40V into that load which will give required current drive ability of the driver.

BTW, the easiest way to measure ouput impedance is by connecting 2 different loads with different impedance and measure the voltage at constant drive, then the output impedance can be calculated easily WTHO Ohm's law.

BTW2, a white follower or any other follower include feedback, it is practically impossible to make a circuit using other than power tubes with an output impedance of 100ohm without feedback.

Regards Hans
 
Hi GianMaria, All

I simulated various topologies using your data and
I found that a Gomez with different tube performs
very well.
It uses a 12AX7 in the input/gain stage
and a Ecc88 in then output.
a 300V supply is sufficient to provide
good results: a gain of about 45
and an output of 40 Vrms with 0.4% Thd , but I suspect
that my 12ax7 model id too optimistic.
Output res is good at mid-high freqs. while the
bass performance is not very good due to the coupling
capacitor.

I've attached the circuit with a couple of graphs.

SRPP of 6072A

Unfortunately I have not a model for that tube
and the process of making one takes me a lot of time.

bye
Federico
 

Attachments

  • gomez.gif
    gomez.gif
    14.1 KB · Views: 460
Gomez stage

I found this circuit very attractive. It appears in different sources, called in different ways: in l'Audiophile no 22 used for a 211 PP stage, in Glass Audio called Hedge voltage amp, in several works of Aloia, in Italy called Supertotem.

For my present understanding, the circuit works better than a SRPP+White cathode follower if the four tubes are equal. In this case the coupling tends to compensated the distortion of each tube.

Is it true? How much imbalance of each tube this stage can tolerate with a modest distortion increase?

Attached there is the calculus of Gomes, that assume 4 12AX7, same mu.

A quick way to find out the Koren's tube parameter, in case one want to derive the spice model of 12AY7:
http://www.geocities.com/dmitrynizh/tubeparams.htm
 

Attachments

  • x_gomes.jpg
    x_gomes.jpg
    87.9 KB · Views: 479
question

Hi fscarpa, all.

I've seen your very interesting simulation on the file GOMEZ.JPG.

UNFORTUNATELY The output impedence will be high exactly where it is needed low: at lower frequencies!!!

So, could we - in that schematic - lead back the output impedance to the same value of the high frequencies (around 100 Ohm) just increasing the value of the capacitor?

What will be the result then at 20 Hz with the new value capacitor?

Could you, or another kindly person, show me the new simulation?

Thanks, suggestions are welcome!

Whishes, G.M.
 
Status
Not open for further replies.