AIYIMA T20 Tube Pre-amplifier Review

[KL....]

This is the simulated THD+N distortion vs. output voltage. Please compare with @amirm measurement in post #1. I think we are very close.

View attachment 484117

Thank you @pma, do you mean this one, IMD vs output voltage, where your graph/simulation indicates Benchmark performance level to about 1Vout, doesn't it :=)

• If Yes, then this tends to indicate that the T20 does not utilise SRPP, doesn't it, but if the T20 does then the implementation could be better (as shown)?
• Your graph/simulation still looks decent at 5Vout, doesn't it?
• With SRPP, what would you expect the Crosstalk to be?

 

[pma]

Thank you @pma, do you mean this one, IMD vs output voltage, where your graph/simulation indicates Benchmark performance level to about 1Vout, doesn't it :=)

Good catch, I have overlooked that the plot shows IMD vs. level, not THD+N.

I really do not know the circuit design of the T20, I just showed one possibility how to put the opamp into the tube feedback loop, and the two dual triodes indicated to SRPP. The actual parameters can be widely modificated, depending on the overall gain and tube used. My simulation uses 12AU7 (ecc82) and overall gain 20dB. AIYIMA uses 12AX7 (ecc83) and overall gain is unknown to me. With lower overall gain, distortion is reduced.

Re crosstalk, it depends on the PCB design and plate supply design. The channels are independent, one tube per channel.

 

[morillon]

Uhhhh
For a change, on the subjective side... :
Has anyone with a bit of experience, on a well-known system, been able to listen to it with demanding music, acoustic type, etc.?

(no answer for gain??? and output impedance ? someone from aiyama seems to be following this thread..... ;-) )


;-)

 

[pma]

With SRPP, what would you expect the Crosstalk to be?

Re crosstalk, it depends on the PCB design and plate supply design. The channels are independent, one tube per channel.

Crosstalk, measured with the tube SRPP board (no op-amps), is good enough:

BTW, dynamic range is 117dB A-weighted. This is really "not bad".

 

[mhardy6647]

With the opamp buffer output impedance should be low. Gain should be easy to measure.

 

[SMen]

Uhhhh
For a change, on the subjective side... :
Has anyone with a bit of experience, on a well-known system, been able to listen to it with demanding music, acoustic type, etc.?

(no answer for gain??? and output impedance ? someone from aiyama seems to be following this thread..... ;-) )


;-)

you might need to reference Aiyima in your post ... so type @ before aiyima.

I have - 3e A7 driving Triangle loudspeakers. Being driven by SMSL Raw DAC. Compared with high quality balanced stepped attenuator...
Please forgive all clichés.
Stock Form: Yes, quite 'warm', and not as detailed as with attenuator.
Tung Sol tubes: tighter base, cooler, more detailed than stock.
Teradak linear power supply - I switched back to the stock power supply to check and now I am back to this TeraDak - better dynamics and clarity for me - I found myself chasing detail by increasing the volume with the stock power supply.

Basic sound: good and very solid left / right / center images. Greater sense of body on voices and instruments. I like my stepped attenuator but this is nice to have as an alternative. Wish it had a pass through / solid state option for comparison and better understanding as to what is due to balanced driving from a pre-amplifier vs. passive, and what is due to the valve topology in place here.


Jazz, latin, classical guitar - pop - speach on poor quality broadcasts all fare well. Yes - on many recordings natural sense of depth.

Apologies for subjective ... but @morillon did ask ... and if Morillon is any reference to the alpine village I know well, it is a lovely place!

 

[Ropeburn]

View attachment 483847
Looks like one of them magically disappeared in your attached image, there.
Perhaps the Thermionic Rapture? One was taken, the other left behind.

It rolled into the sunset...

 

[KL....]

Good catch, I have overlooked that the plot shows IMD vs. level, not THD+N.

I really do not know the circuit design of the T20, I just showed one possibility how to put the opamp into the tube feedback loop, and the two dual triodes indicated to SRPP. The actual parameters can be widely modificated, depending on the overall gain and tube used. My simulation uses 12AU7 (ecc82) and overall gain 20dB. AIYIMA uses 12AX7 (ecc83) and overall gain is unknown to me. With lower overall gain, distortion is reduced.

Re crosstalk, it depends on the PCB design and plate supply design. The channels are independent, one tube per channel.

Although all of the BA9 base twin triodes are "compatible" in the sense that the pinout's identical, there are differences in heater current, gain, and other important operating characteristics that render them not necessarily swappable! Usually any 12A*7 will work in place of another, in the sense that it won't simply blow something up, but that in no way means that it will work properly. A foolhardy swap may endanger the well-being of the tube and/or the component into which it installed.

B9A Pinout:
View attachment 483332

I'd (strongly) encourage anyone who wants to fiddle around blithely swapping* 9 pin twin triodes to at the very least consult a good tube manual for the specifications, limits, optimal parameters, and curves for candidate substitutes.

Plenty of receiving tube manuals to be found at:

BOOKSHELF RCA: Radi Corporation of America tech guides

BOOKSHELF RCA: Radi Corporation of America tech guides

www.worldradiohistory.com

Here's a good one for most folks interested in typical audio tubes.

https://www.worldradiohistory.com/BOOKSHELF-ARH/Technology/RCA-Books/RCA-Receiving-Tube-Manual-1963-RC-22.pdf

Alternatively, one can look 'em up, e.g., at https://tdsl.duncanamps.com/index.php

Here's some extremely basic information I pulled together some years back for some reason or another...

View attachment 484264View attachment 483331

Note that the 12AU7 has a much lower amplification factor than the 12AX7, e.g.
Even the 5751 twin triode (not shown above), which is almost a 12AX7 and often seen in hifi circuits, has a somewhat lower mu than does a 12AX7/ECC83.

____________________
* N.B. I am NOT referring to @Barrelhouse Solly's example, which appears to be very specific. I am speaking more generally of just F'ing around and finding out with the B9A twin triodes.

Thank you, @pma, 12AU7/ecc82/20db mu/Gain v 12AX7/ecc83/100db mu/Gain is not apple v apple , is it, and not a reasonable comparison to the T20, is it? Are you able to redo your graph/simulation useing the 12AX7 Tube?

Note that the 12AT7 is quite popular, 12AU7/ecc83/60db mu/Gain but the 12AX7/ecc83/100db mu/Gain has more mu/Gain is less cost and usually more noise? Perhaps the Psvane Horizon 12AX7-AT is an exception :=)

 

[Ropeburn]

Thank you, @pma, 12AU7/ecc82/20db-mu Gain v 12AX7/ecc83/100db-mu Gain is not apple v apple , is it, and not a reasonable comparison to the T20, is it? Are you able to redo your graph/simulation useing the 12AX7 Tube?

Why wouldn't it be a relevant comparison? Line preamps are a typical application where not much gain is needed, because functionally they are "attenuators with a bit of gain on the side". 20dB is factor 10 and easily enough here.

 

[pma]

Are you able to redo your graph/simulation useing the 12AX7 Tube?

Of course, easily, but why should I do it? For someone’s fun or mocking comment?

 

[KL....]

someone from aiyama seems to be following this thread..... ;-) )

Thank you, who are they/somone?

 

[KL....]

Of course, easily, but why should I do it? For someone’s fun or mocking comment?

Thank you, @pma, the scientific process.... hope that makes sense (note, they mock me all the time/often, don't they? Thank you that you educate rather than mock, very much appreciated :=)).

I do not know that I have ever mocked you and if I have, let me know (Unintentional/Apologys), thank you :=)

 

[KL....]

Why wouldn't it be a relevant comparison? Line preamps are a typical application where not much gain is needed, because functionally they are "attenuators with a bit of gain on the side". 20dB is factor 10 and easily enough here.

Thank you, Yes, relevant when declared before/at posting, otherwise not apple v apple, reasonable, is it?....

 

[SSS]

Good catch, I have overlooked that the plot shows IMD vs. level, not THD+N.

I really do not know the circuit design of the T20, I just showed one possibility how to put the opamp into the tube feedback loop, and the two dual triodes indicated to SRPP. The actual parameters can be widely modificated, depending on the overall gain and tube used. My simulation uses 12AU7 (ecc82) and overall gain 20dB. AIYIMA uses 12AX7 (ecc83) and overall gain is unknown to me. With lower overall gain, distortion is reduced.

Re crosstalk, it depends on the PCB design and plate supply design. The channels are independent, one tube per channel.

You created a good circuit. But this is not the only way to use the double triode. One is DC coupled common cathode amplification with cathode follower. Another is the triodes paralled for lower output impedance for feeding the op-amp circuit.

 

[mhardy6647]

It rolled into the sunset...

"Git along little tubie, git along..."

The song of the American tubeboy...

 

[pma]

I’ll leave the party for you guys. You know better.

 

[KL....]

I’ll leave the party for you guys. You know better.

Hi @pma, Thank you that you (seek to) educate rather than mock, very much appreciated :=)

Apologys, if the Delivery (apple v apple) of my post (above) was not (sufficiently) Respectful/Helpful, especially, know that was not the Intent/Desire :=)

 

[morillon]

Uhhhh
For a change, on the subjective side... :
Has anyone with a bit of experience, on a well-known system, been able to listen to it with demanding music, acoustic type, etc.?

(no answer for gain??? and output impedance ? someone from aiyama seems to be following this thread..... ;-) )


;-)

you might need to reference Aiyima in your post ... so type @ before aiyima.

I have - 3e A7 driving Triangle loudspeakers. Being driven by SMSL Raw DAC. Compared with high quality balanced stepped attenuator...
Please forgive all clichés.
Stock Form: Yes, quite 'warm', and not as detailed as with attenuator.
Tung Sol tubes: tighter base, cooler, more detailed than stock.
Teradak linear power supply - I switched back to the stock power supply to check and now I am back to this TeraDak - better dynamics and clarity for me - I found myself chasing detail by increasing the volume with the stock power supply.

Basic sound: good and very solid left / right / center images. Greater sense of body on voices and instruments. I like my stepped attenuator but this is nice to have as an alternative. Wish it had a pass through / solid state option for comparison and better understanding as to what is due to balanced driving from a pre-amplifier vs. passive, and what is due to the valve topology in place here.


Jazz, latin, classical guitar - pop - speach on poor quality broadcasts all fare well. Yes - on many recordings natural sense of depth.

Apologies for subjective ... but @morillon did ask ... and if Morillon is any reference to the alpine village I know well, it is a lovely place!

Thanks... But I was wondering about the little preamp in this thread. ;-)

(Morillon could also be...a duck. ;-) Aythya fuligula)

 

[KL....]

You created a good circuit. But this is not the only way to use the double triode. One is DC coupled common cathode amplification with cathode follower. Another is the triodes paralled for lower output impedance for feeding the op-amp circuit.

Thank you, @SSS, perhaps you could show us the circuit diagram that you have in mind that utilises the 12AX7 in the FB Loop with the opamp (as the T20 overview shows) and measurements?

This is an excellent circuit (SRPP), as provided by @pma (thank you :=)), isn't it, which utilises the 12AU7 in the FB Loop with the opamp (as shown in the (above) circuit diagram and as the T20 overview shows), and graph/simulation of measurements show excellent measurements, don't they?

• Utilises both sides of 1 (dual Triode) Tube/Valve. With SRPP, matching is not critical but nice/preferable.
• Very low output impedance. Feeding Loads down to 600 ohm/Input Impedance, is no problem.
  • This implys that it will/should impedance match with any Device/Amp that it is Feeding, doesn't it?
• H2 but not too much (for those who enjoy, a little extra, H2 but perhaps inaudable), H3 below -120dB and no Higher H

The graph/simulation indicates Benchmark performance level to about 1Vout, decent 1 to about 3/3.5Vout, and still looks decent even 3.5 to 5Vout, doesn't it?

Crosstalk

 

[SMen]

Thanks... But I was wondering about the little preamp in this thread. ;-)

(Morillon could also be...a duck. ;-) Aythya fuligula)

Surely this should be in a different thread? This is supposed to be about the T20...