1st stage tube considerations
The phono stage is intended for both MM and MC, probably with SUT (step-up transformer).
It looks to me that the most important characteristic for that tube should be very low noise and microphonic.
On top of that, reasonably low miller capacitance is necessary (for MM cartridges).
Tough signal levels at this stage are low, good linearity is preferred, especially low 3rd and higher odd harmonics distortion levels.
Long life, reasonably low plate resistance, availability and reasonable cost are preferred.
I hope I didn't miss anything here.
All the above relate to anode resistor loading, capacitor coupling to the next stage (including the RIAA equalization network).
What are your recommendations?
Replies based on experience, especially listening evaluations between different tubes, will be appreciated.
The phono stage is intended for both MM and MC, probably with SUT (step-up transformer).
It looks to me that the most important characteristic for that tube should be very low noise and microphonic.
On top of that, reasonably low miller capacitance is necessary (for MM cartridges).
Tough signal levels at this stage are low, good linearity is preferred, especially low 3rd and higher odd harmonics distortion levels.
Long life, reasonably low plate resistance, availability and reasonable cost are preferred.
I hope I didn't miss anything here.
All the above relate to anode resistor loading, capacitor coupling to the next stage (including the RIAA equalization network).
What are your recommendations?
Replies based on experience, especially listening evaluations between different tubes, will be appreciated.
High gain is essential to deal with the 20 dB. roll off in the EQ network.
Low CMiller, while highly desirable, is not absolutely essential. All sorts of 12AX7/ECC83 based circuits are available and the majority of MM carts. and SUTs work well with them.
Low noise definitely is important. Because of the partition factor, it's difficult to find a pentode that's quiet enough. The 6AC7 could work out. Definitely use silicon rubber "O" rings on the metal envelope to tame microphonics.
Because of its μ of 33, things can get "dicey" with the 6922, as gain without CCS loading is definitely marginal. All in all for a project where you will experiment to "nail" RIAA EQ down, the 6GK5 seems hard to beat, with its high μ, low noise, high gm, and low RP.
Low CMiller, while highly desirable, is not absolutely essential. All sorts of 12AX7/ECC83 based circuits are available and the majority of MM carts. and SUTs work well with them.
Low noise definitely is important. Because of the partition factor, it's difficult to find a pentode that's quiet enough. The 6AC7 could work out. Definitely use silicon rubber "O" rings on the metal envelope to tame microphonics.
Because of its μ of 33, things can get "dicey" with the 6922, as gain without CCS loading is definitely marginal. All in all for a project where you will experiment to "nail" RIAA EQ down, the 6GK5 seems hard to beat, with its high μ, low noise, high gm, and low RP.
Hi,
Thank you.
Indeed.
Tubes with about 1mA anode current are outside my list.
Low noise with high gain is probably the most important parameter for phono 1st stage.
I've seen phono stages with D3a, C3g and E810F pentodes on the 1st stage. I wonder how they compare.
That goes without saying.
Yet, the tubes, especially the one in the 1st stage, should be of very low noise and microphonic.
Indeed, it looks like it is so.
The datasheet statess noise figure for RF, not for audio frequencies, so I wonder what is the tube's noise in audio frequencies.
Also, I'd love to hear from people who have experience with this tube about how its' microphonic it is.
Thank you.
Indeed.
Tubes with about 1mA anode current are outside my list.
Low noise with high gain is probably the most important parameter for phono 1st stage.
I've seen phono stages with D3a, C3g and E810F pentodes on the 1st stage. I wonder how they compare.
That goes without saying.
Yet, the tubes, especially the one in the 1st stage, should be of very low noise and microphonic.
Indeed, it looks like it is so.
The datasheet statess noise figure for RF, not for audio frequencies, so I wonder what is the tube's noise in audio frequencies.
Also, I'd love to hear from people who have experience with this tube about how its' microphonic it is.
If I may chime in; I was wondering if Eli would reply, and he did. He made me aware of the 6GK5 a few years ago, and it is a quite good choice (also not that expensive - but don't go buying up all of them now!).
It may be somewhat more microphonic than the ECC83, but that is not the most important factor. It has a frame grid, which makes it comparatively stable to drift. You seem to know that noise is invertedly proportional to gm, thus working at a high plate current has benefits. I doubt that one will find a more suitable input tube for a phono preamp. I have also used an E88CC if one needs to be economical of tube count, but found the 6GK5 superior noise-wise.
It may be somewhat more microphonic than the ECC83, but that is not the most important factor. It has a frame grid, which makes it comparatively stable to drift. You seem to know that noise is invertedly proportional to gm, thus working at a high plate current has benefits. I doubt that one will find a more suitable input tube for a phono preamp. I have also used an E88CC if one needs to be economical of tube count, but found the 6GK5 superior noise-wise.
The gain does not have to be that high. Assuming passive EQ with only half of it being performed after the first stage, you only have 20dB of loss to contend with. A gain of a few dBs more than 20dB will therefore be enough to overcome the noise contribution of the second stage (particularly since flicker noise dominates), and most triodes can manage this easily. A mu of about 30 is probably as low as you would want to go, so the ECC88 does hover at that end of the spectrum.
CCS loading is a universally poor choice, as the noise current generated by any CCS will be similar to or even greater than that of the triode itself, and will only degrade the SNR. You really can't beat passive loading for best SNR.
CCS loading is a universally poor choice, as the noise current generated by any CCS will be similar to or even greater than that of the triode itself, and will only degrade the SNR. You really can't beat passive loading for best SNR.
Last edited:
Yes, the 'X7 triode is a wimp. That's easily dealt with.
Enter the ZVN0545A. A fair number of people have built the tweaked RCA setup that I've uploaded. Comments have been favorable and continue to be so. Are better designs available? Certainly. However, the tweaked version of RCA's seminal passive EQ design gets the job done quite well, indeed.Attachments
Hi,
Thank you.
Indeed.
On top of that, CCS loading increases odd and high order harmonics.
Thank you.
Indeed.
On top of that, CCS loading increases odd and high order harmonics.
Hi Eli,
Thank you.
I'm not going to use SS devices, save for rectifying the heaters voltage, nor am I going to use questionable tubes.
I'm looking for the best possible design. Getting the job done well isn't what I'm looking for.
So, please, let's go back to square one. What tube(s) is/are the best for a phono 1st stage? Are there any contenders to the 6GK5?
Thank you.
Yes, the 'X7 triode is a wimp. That's easily dealt with.
I'm not going to use SS devices, save for rectifying the heaters voltage, nor am I going to use questionable tubes.
I'm looking for the best possible design. Getting the job done well isn't what I'm looking for.
So, please, let's go back to square one. What tube(s) is/are the best for a phono 1st stage? Are there any contenders to the 6GK5?
Where in the world did you get that idea?
If the rp of the tube is low, the in of the CCS is not a significant contributor, as long as it's reasonable low. Case in point: my MC preamp, which uses CCS loading and is ultra-quiet for a tube design.
If I may be blunt, Joshua, you're coming at this with a lot of preconceptions. That is not likely to get you a top-performing result.
Morgan Jones book 'Valve Amplifiers Third Edition'. Shall I look for the page number?
The above is supplement by listening evaluations of some people whose taste I trust.
I give serious considerations to engineering/technical aspects, however I wouldn't follow an engineering consideration that isn't supported by listening evaluations. For example, tubes datasheets don't specify microphonics. Therefore, I like to have as full picture as possible – both the engineering considerations and the listening evaluations.
Of course, I wouldn't and I cannot design any amp based on listening only. That goes without saying. However not all audio designers, or engineers, consider listening evaluations. Which is why I will not copy blindly a design based on engineering considerations only, without supplemented listening evaluations. Nor would I copy blindly a design which has apparent engineering flaws.
I've been using a modified Dynaco PAS MM Phono stage(12AX7) also using a regulated power supply and DC heaters. So far I have been very pleased with the sound! It's very quiet. I have some Sheffield Lab vinyl I've been playing, these recordings are very unforgiving of any phono stage, they will thoroughly bring out the best and the worst in any phono stage.
One of my previous phono stages was based on 12AX7 tubes. It sounded enjoyable, however I want do something much better.
Please. Morgan and I enjoy catching one another out, and it's rare that I can find a whopper like that.
Page 160.
Actually, I stand corrected. What Morgan Jones says there is that because of active load supresses the 2nd harmonic and barely change higher harmonics, the effect of the higher harmonics become more significant, with the result that some triodes don't have distortion that is proportional to level.
This is enough for me to rule out CCS, or active load, anywhere in a phono stage.
Also, cathode followers are ruled out for me.
For me it isn't business, it's only personal.
So the CCS is ruled out because it lowers distortion?
And the cathode follower is ruled out just because?
I think you'd do best to rule out voltage regulators, too, cause they work really well.
P.S.,
The fact that with active load the higher harmonics become more prominent may well explain the listening evaluations that didn't like CCS in tubes amps.
I have no wishes to be 'modern', I'm only after the best possible sound quality.
To the best of my knowledge, good sound quality calls for certain proportions between the second harmonic and the higher ones.
Possibly the second harmonics masks, at least to some degree, the higher harmonics. I don’t care much why and how. All I care about is the sound quality.
This is why engineering considerations, which aren't supported by listening evaluations, aren't sufficient for me to copy blindly a design made under such considerations.
Should Morgan Jones not mention that paragraph, the effect of CCS may escape the eyes of many engineers, or audio designers, who don't support their considerations with listening evaluations.
The fact that with active load the higher harmonics become more prominent may well explain the listening evaluations that didn't like CCS in tubes amps.
I have no wishes to be 'modern', I'm only after the best possible sound quality.
To the best of my knowledge, good sound quality calls for certain proportions between the second harmonic and the higher ones.
Possibly the second harmonics masks, at least to some degree, the higher harmonics. I don’t care much why and how. All I care about is the sound quality.
This is why engineering considerations, which aren't supported by listening evaluations, aren't sufficient for me to copy blindly a design made under such considerations.
Should Morgan Jones not mention that paragraph, the effect of CCS may escape the eyes of many engineers, or audio designers, who don't support their considerations with listening evaluations.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.