Valve Itch phono

diyAudio Chief Moderator
Joined 2002
Paid Member
After my NJFET phono, the shunt Mosfet LV and HV psus, I felt too deeply buried in sand. Got an itch. The valve itch. :clown: Had to mess with glass again.
Having the Simplistic NJFET and the Steve Bench 12AY7 & 6DJ8 cascode already, both on Mosfet shunts of mine, I knew it would be a tough call. Comparisons would come inevitably. So I thought to compare along testing and tweaking phases so to make something that has some sonic merit to stand there. To have something compelling to tell in its own tone way.
The design had to use preferably cheap and available but very silent tubes. Had 6N2PEV Russian tubes that I wanted to test. I liked their characteristics. Also 6N8 NOS Soviet black anode 6SN7 equivalents. I want the most gain there is available on first stage, I don't want extra resistors there, want good PSRR, and lower gain at second stage. Also low impedance Riaa I prefer, so some good drive is needed. Two mu follower stages with passive Riaa between, seemed appealing for my concepts. I have seen the Phono Dude doing that too, have never had one, but it was good to know it already works as phono with some acclaim. Put it together on the simulator, tweaked this and that, seemed nice. I decided not to look for how low THD, as long as it would be predominantly second. Wanted some triode like tone this time since the NJFET and Bench do the monitor style already.
Good DIYer mate Mikvous liked the idea and decided to make one in some spare boxes that used to house a Thorsten phono, so we can test and tweak. It took us about 2 weeks so to decide its what we wanted.
The noise floor is amazingly low for 61dB gain of step up and valve phono. Pitch black background. The THD is 0.2% 2nd, and the Riaa curve is identical to the NJFET Riaa. No metal shield caps on the input valves for such a low noise floor, and we can happily knock em on their glass while the volume is set loud and hear nothing. The sound is great. Lush but has it all. Best valve phono we ever had bar none. And we had tested about 7 DIY ones.:cool:
 

Attachments

  • Itching.JPG
    Itching.JPG
    136.5 KB · Views: 4,930
  • DSCF1713.JPG
    DSCF1713.JPG
    87 KB · Views: 4,797
diyAudio Chief Moderator
Joined 2002
Paid Member
The PSU circuit.

This one is regulated with a Mosfet Maida reg. (link has gone, see more)

A Mosfet HV shunt can be used, but we did the quicker, cheaper Maida for testing, since mu follower has nice PSRR too. Normally there is more to listen with the HV shunt. Delete the 0.1uF parallel to the B+ connector if with the SSHV shunt reg.
 

Attachments

  • valveitchpsusm.jpg
    valveitchpsusm.jpg
    107.4 KB · Views: 5,180
diyAudio Chief Moderator
Joined 2002
Paid Member
The Riaa components values have been tested both by measurement and subjective tone. Silver Mica bypassed PP is used for the 100nF, K72 33nF Teflon + 2200pF Siver Mica for the 35.2nF. Resistors are a mix of Kiwame and cheap carbon 2W. Buss bar grounding has been used, silver content solder, and Mogami thin shielded coax for signal. Lundahl 9206 step up is included in Riaa box, set at 10X. If someone wants to use it for high MC or MM, the gain without the step up is adequate, but has to block the bare input with some 2.2uF capacitor since there always be some mV offset due to grid leak that may fry some cartridge.
 

Attachments

  • valveitchriaacurve.gif
    valveitchriaacurve.gif
    19.8 KB · Views: 8,040
diyAudio Chief Moderator
Joined 2002
Paid Member
Some details.

To check for channel balance, the voltage drop must be measured on one of the series resistors of the first mu stage. Either on the 560R or the 47k. If it is the same, then the channels have the same gain. This is important so to check for sections gain of 6N2PEV used. They seem like following well though. For example, by randomly putting one couple, I was able to see if the channel RIAA curves met on the FFT screen for level, when the generator was set for pink noise and 1/12 octave scaling. By just swapping those valves between sockets, the two curves matched, as well as the voltage drops on the respective resistors. I did not need to go through a batch. Maybe a bit lucky, but non the less its a hint to good manufacturing tolerances. Pin 9 is screen between sections on 6N2PEV and must be grounded near on the ground copper buss bar. The gain of first section is about 85X and this gives 200pF input capacitance. No weird interactions with the Amorphous Lundahl 9206's and it looks it can still do OK with many MM, and Grado does not care. Someone must measure the input for mV DC before attempting to use the Valve Itch without a transformer as MM or High MC stage (44dB gain) because there can be grid current leak, although the 6N2PEV is made to produce low grid current. A coupling capacitor is safety in any case. Of course its best use is with MC due to the really low noise floor and beguiling sound quality. The step up won't pass DC. With the standard 61dB measured gain (Lundahl set at 10X) even a Denon 103 can be used if there is normal gain in the rest of the system. Still there is 5X and 20X option in the Lundahl 9206. I measured 30% less than claimed step up at 1X10. I would fearlessly try even the most low output MC cartridge with this phono, since it is the quietest valve phono I have ever had. Don't forget, the cart load is the phono input resistor, divided by step up ratio squared. This phono needs about a week of enough everyday play so to break in, as it proved in practice.

6N2P-EV
Grid current alternative solution to AC coupled input
 
diyAudio Chief Moderator
Joined 2002
Paid Member
Did that side by side with the other phonos. The interesting thing is that it has the same balance with the NJFET. You don't listen to some fictional thermionic rendering of the recording. Also has the noise floor of well executed solid state. So there is a fair platform to compare. I will run a session of exchanging the Itch, the Simplistic, and the Bench. I am waiting for it to fully break in. Mikvous got perplexed when initially exchanging with the Simplistic. He told me, ''maybe now I hear what the different devices do, not the designs''. I don't think it has the depth resolution of the Simplistic, but has different breath. Its much less linear (0.2% THD mainly 2nd, ouch!), but benign. On the other hand, it can swing much more dynamically of course. Careful designs of different nature must not sound strongly wrong or correct in my view, but in the long run must provide a different aftertaste for suitability choice to different TTs, speakers etc.
The Riaa values are different to the Simplistic due to different source impedance. Their curves just overlap on FFT. I started it with the trendy 50kHz fix included, but the classic Riaa just sounded better, as it happened some other times before. Will tell more about the ''Valve Itch'' after a while. For now, I am sure its a worthy vacuum state device.
 
diyAudio Chief Moderator
Joined 2002
Paid Member
Hi Allen,

Those 1.4 & 4mA on the schematic are actually measured as far as I recall but could have made a typo as well.

I would just measure its safe if I had an MM to hook up and exchange some tubes if leaky, so to avoid the input cap. But you know how it goes with safety recommendations when publishing ccts. Tubes age etc.

Best
 
I like that low impedance RIAA. I was never comfortable with very high impedance circuits.
Why do you not use a cathode resistor at the input ? Noise ? Some lokal degeneration could improve TDH. Some second harmonic can sound very attractive though. I learned that when i build the PSG buffer. Another option whould be to put the transformer in the cathode.
 
diyAudio Chief Moderator
Joined 2002
Paid Member
Yes, noise. Interestingly after a year I made it, I read Vogel's silence and he is pointing out the SRPP & mu topologies and no grid resistor as best for noise. He he, nice. I would think about just an IR LED the most on cathode. The THD probably comes from driving the 22K from the high rp follower upper section though... Paralleling 2 sections up & down the mu would do better for noise and drive, those 6N2PEV are not costly. A tube like E810F trioded or D3A would be best noise wise. I just had the Russians and tested. I got surprised by the non existent microphony from those.

6N2P
 
Those 1.4 & 4mA on the schematic are actually measured as far as I recall but could have made a typo as well.

Doesn't make sense to me.

You have the first tube without any cathode R, yet get 1.4mA. The second tube has substantial cathode bias yet you get three times as much current at 4mA. And in both cases the bias R on the upper tube is the same.

???????

Regards, Allen
 
diyAudio Chief Moderator
Joined 2002
Paid Member
Don't remember, really. I have to get it back or visit the friend I gave it to and see again the drop on 1st stage RC filter's 3k3 inline with PSU. Maybe I mixed something back then and noted those numbers. Second tube is different, its 6H8/6SN7.
 
diyAudio Chief Moderator
Joined 2002
Paid Member
Got intrigued and since I haven't got it in hand to measure the first stage current, I made a simulation on LT Spice. I used a 12AX7 model which is not far from 6N2 just to be in the ballpark. It roughly follows that 1.4 mA first stage schematic note.
 

Attachments

  • uBias.gif
    uBias.gif
    10.3 KB · Views: 2,952