Hi Alex
I have a theory about the 12AT7 as a signal tube. Maybe the fact that it does not try to be an X7 or a U7, just a triode, means that it also seems to hit some kind of sweet spot. I have replaced both X7 and 6922 (usually Sovtek or Electro-Harmonix) where possible and easily preferred the results. But of course, not all circuits are suitable or too difficult to adapt. Those I have done it for, not one has been dissatisfied. If Allen had been around today, I would have pushed it to be used in the next generation of the SVP in the Line Stage, but keep 6922 in the Phono Stage.
Allen and I discussed about how the FVP/SVP improved from more humble origins and hence greater progress was made when comparing to the RTP which started from a much more exalted position. Gradually the FVP/SVP became a serious un-balanced alternative to the RTP.
But they do share the same DNA.
I have observed that gradually the T7 has been getting more popular. The late David Manley was a fan going back to hearing this in the 80's.
When T7 is used in the SLCF buffer part, it won't have the same output 'grunt' as a 6922/ECC88/6DJ8. But most people don't even know that. I recommend using about 5mA.
But when you build it, then it does need tweaking, especially if you want to operate it at only 200V. It can be done, but adjustments need to be made once it has been built. In this case 220-230V makes it a lot easier to work.
Also, this is important, you need to have a clean negative supply, anywhere between -6V to -20V will do.
So, do you have an existing FVP Line Stage to update, or are you thinking starting from new? Let me know and I can tailor things from there.
Only a single twin-triode T7 will be needed per channel. Let le know re above and I will reveal.
I have a theory about the 12AT7 as a signal tube. Maybe the fact that it does not try to be an X7 or a U7, just a triode, means that it also seems to hit some kind of sweet spot. I have replaced both X7 and 6922 (usually Sovtek or Electro-Harmonix) where possible and easily preferred the results. But of course, not all circuits are suitable or too difficult to adapt. Those I have done it for, not one has been dissatisfied. If Allen had been around today, I would have pushed it to be used in the next generation of the SVP in the Line Stage, but keep 6922 in the Phono Stage.
Allen and I discussed about how the FVP/SVP improved from more humble origins and hence greater progress was made when comparing to the RTP which started from a much more exalted position. Gradually the FVP/SVP became a serious un-balanced alternative to the RTP.
But they do share the same DNA.
I have observed that gradually the T7 has been getting more popular. The late David Manley was a fan going back to hearing this in the 80's.
When T7 is used in the SLCF buffer part, it won't have the same output 'grunt' as a 6922/ECC88/6DJ8. But most people don't even know that. I recommend using about 5mA.
But when you build it, then it does need tweaking, especially if you want to operate it at only 200V. It can be done, but adjustments need to be made once it has been built. In this case 220-230V makes it a lot easier to work.
Also, this is important, you need to have a clean negative supply, anywhere between -6V to -20V will do.
So, do you have an existing FVP Line Stage to update, or are you thinking starting from new? Let me know and I can tailor things from there.
Only a single twin-triode T7 will be needed per channel. Let le know re above and I will reveal.
Hi Joe, thanks for your reply.
I built an FVP5a from the schematic on the VSE webpage. I have already replaced the 6922 in the voltage amplifier stage of the line stage with a 5687 to reduce the gain slightly. This uses a 250V SuperRegged HT supply as in the schematic.
Any suggestions would be welcome!
Alex
I built an FVP5a from the schematic on the VSE webpage. I have already replaced the 6922 in the voltage amplifier stage of the line stage with a 5687 to reduce the gain slightly. This uses a 250V SuperRegged HT supply as in the schematic.
Any suggestions would be welcome!
Alex
@Joe Rasmussen I love to hear more about FVP-6. If there is a interest then what would be the plan for that?
Give me a little time, but I will post an FVP-6 based on 12AT7 schematic that experienced DIY'ers can try out. If you are doing +200V DC circuits already, then I know you are able to do it.
The big thing is that for this circuit to work, you need an additional low tension negative power supply, something like -15V. Where it will be needed is obvious when you see it, but it is about a current source that can be made up from a simple Fet and two resistors. It will also allow some control over gain, so four will be good for some, but up to ten is also possible. The lower gain will have lower noise.
Requires:
+220 to 230V Regulated. Up to 250V can be accommodated with few changes.
-15V Regulated.
Not today, maybe tomorrow.
The big thing is that for this circuit to work, you need an additional low tension negative power supply, something like -15V. Where it will be needed is obvious when you see it, but it is about a current source that can be made up from a simple Fet and two resistors. It will also allow some control over gain, so four will be good for some, but up to ten is also possible. The lower gain will have lower noise.
Requires:
+220 to 230V Regulated. Up to 250V can be accommodated with few changes.
-15V Regulated.
Not today, maybe tomorrow.
I apologize for intruding on a very interesting discussion. But my question is about Vacuum State FVP5A. And here is a place where I can meet people who directly contributed to the creation of FVP5A. Unfortunately I was not able to get Allen Wright „The TubePreamp CookBook“. I believe there is a lot of answers. Also hope that somehow Mr. Joe Rasmussen can help with it…
I have not grown to RTP3C balanced version yet so I started from FVP5A and built some test version from schematic from VSE web page. My power supply is not Allens Wright Super Regulator but I used RAW DC with EZ81 tube and Salas SSHV2 shunt regulator. Filaments powered with DC and separate for each tube current regulator. Grounds are lifted as suggested by Allen. I did not use 50k potentiometer, instead I use two resistors 20k and 30k to set fixed gain. Here some photos:
FVP5A will be used for MM or high output MC, so I set R2 to 100R and second cathode resistor to 1k6. Transistor is 2SK170BL. First run and with some delay because of soft start of power supply I heard nice music from the speaker 😊. For tests I am using inverse RIAA circuit from https://hifisonix.com/wp-content/uploads/2016/12/Accurate-Inverse-RIAA.pdf
So, first music was played from CD and not from LP. Then I performed some tests with function generator and oscilloscope. I performed tests with square waveform and was not very happy. I did not get correct waveform at the output at 100Hz, 1kHz and 10kHz. Does it mean that RIAA is not correctly reproduced?
But I was surprised when I get almost identical results with LTSpice simulation.
Also noticed that frequency response from 19kHz is starting to lift very steep. So, is there any underwater stones or I just doing something wrong or I do not understand basics?
Any comments are greatly appreciated. Thank you in advance 😉
I have not grown to RTP3C balanced version yet so I started from FVP5A and built some test version from schematic from VSE web page. My power supply is not Allens Wright Super Regulator but I used RAW DC with EZ81 tube and Salas SSHV2 shunt regulator. Filaments powered with DC and separate for each tube current regulator. Grounds are lifted as suggested by Allen. I did not use 50k potentiometer, instead I use two resistors 20k and 30k to set fixed gain. Here some photos:
FVP5A will be used for MM or high output MC, so I set R2 to 100R and second cathode resistor to 1k6. Transistor is 2SK170BL. First run and with some delay because of soft start of power supply I heard nice music from the speaker 😊. For tests I am using inverse RIAA circuit from https://hifisonix.com/wp-content/uploads/2016/12/Accurate-Inverse-RIAA.pdf
So, first music was played from CD and not from LP. Then I performed some tests with function generator and oscilloscope. I performed tests with square waveform and was not very happy. I did not get correct waveform at the output at 100Hz, 1kHz and 10kHz. Does it mean that RIAA is not correctly reproduced?
But I was surprised when I get almost identical results with LTSpice simulation.
Also noticed that frequency response from 19kHz is starting to lift very steep. So, is there any underwater stones or I just doing something wrong or I do not understand basics?
Any comments are greatly appreciated. Thank you in advance 😉
I managed to get to work my FVP5 second stage, but the gain is so low and the sound quality is not acceptable. I tried also FVP5A with 2 tube version second stage. The gain is about 16db but the sound quality is not good but better than FVP5 one tube version with mosfets. It lacks details, HD resolution, mid vocals and higher frequencies. I use Russian Tubes 6N23P E88CC/ECC88/6DJ8 Silver Shields REFLEKTOR 1974. Some better tube recommendation?
How to increase gain in FVP5 with one tube verion with MOSFET-s and sound quality?
How to improve FVP5A sound quality?
FVP5A my gain stage 2 tubes.
How to increase gain in FVP5 with one tube verion with MOSFET-s and sound quality?
How to improve FVP5A sound quality?
FVP5A my gain stage 2 tubes.
Hi Viljo,
If you are dissatisfied with the sound of the SVP I would assume it isn't working properly, since I was immediately struck by the clarity and dynamics of both my VSE preamps. A quick question - have you let it warm up for at least an hour before listening? Both Allen's designs have a ridiculously long warmup time, and I have found the RTP3 in particular sounds a bit muffled and loose for the first hour after switch-on.
Have you checked the DC voltages and currents in your circuit? In particular, check the voltages across each of the 2K2 cathode resistors - If I remember correctly, the first stage should pass about 6 mA (about 13V across the resistor) and the second about 12 mA (about 26V across the resistor).
Also check the filament voltages on the valves are close to the data sheet recommendations. Are you using series or shunt regulation for the filaments?
I'm not familiar with the 6N23P. Are its parameters close enough to be a plug-in replacement for the 6922?
And remind me, what HT regulator are you using? Is it Allen's SuperReg?
Alex
If you are dissatisfied with the sound of the SVP I would assume it isn't working properly, since I was immediately struck by the clarity and dynamics of both my VSE preamps. A quick question - have you let it warm up for at least an hour before listening? Both Allen's designs have a ridiculously long warmup time, and I have found the RTP3 in particular sounds a bit muffled and loose for the first hour after switch-on.
Have you checked the DC voltages and currents in your circuit? In particular, check the voltages across each of the 2K2 cathode resistors - If I remember correctly, the first stage should pass about 6 mA (about 13V across the resistor) and the second about 12 mA (about 26V across the resistor).
Also check the filament voltages on the valves are close to the data sheet recommendations. Are you using series or shunt regulation for the filaments?
I'm not familiar with the 6N23P. Are its parameters close enough to be a plug-in replacement for the 6922?
And remind me, what HT regulator are you using? Is it Allen's SuperReg?
Alex
Allen didn't design his RIAA circuits for "accurate" RIAA reproduction - he describes in section 4 here how he includes the "missing time constant" at 2.1kHz, which boosts high frequencies relative to the standard RIAA equalisation response. Whether or not this this is actually justified has been discussed many times, but if your circuit is built as Allen designed it it will indeed have a rising response at higher frequencies.
Your measurements (and simulations) appear to show some ringing on the leading edge of the square wave, which suggests an ultrasonic resonance somewhere. This can happen if you have a poorly loaded transformer, but that obviously isn't the case here. It's a puzzle to me. Maybe Joe can suggest something?
Alex
Hi Alex!
R6 - 4.88 V
R3 - 23.1 V
U1 - pin 1 - 183.7 V
Filament voltage 6.3 V - series very basic regulator.
Heaters are in parallel connection. And +20 V is from very basic series regulator. For HV I use Allen SuperReg output 250 V.
Tube U1 is much much hotter than U2. I do not know if it is important detail or not.
Could be that my schematic with tubes is mixed up? Because R3 and R6 use same tube U1? My schematic is here
This FVP5A has been playing music today already for 5 hours. Right now I have built one channel only for test and to get idea about sound quality. I replaced Wima MKP10 end cap with Miflex 2.2 uF. In another circuit Wima MKP10 sounded so bad that I could not wait better sound. I noticed also that in other circuit one channel input resistor was low end kind a no name resistor from local shop and other channel VIshay Dale cmf55 series and sound quality difference was already huge for me.
I am very sensitive about bad sounding sound.
I'm not sure these readings make sense. If the anode voltage of U1 is 183V and your regulated HT is 250V, the current through the 25K anode resistor is 2.65mA, which strikes me a little low, but not a deal-breaker. The anode voltage is quite critical for the operation of the output stage, though, since the cathode of the U2B will be about 5V positive of the grid, or 188V. That means that U2B and the bootstrap valve U2A will be a little starved of voltage. If I remember correctly, the anode of U1A should be about 150V (which means the current through U1A and R3 should be about 4mA). So this is definitely not consistent with 23V across R3, which would imply a physically impossible (give the value of the cathode resistor R4) current of 10.4mA through U1A.
A voltage of 25V on the grid of U1B implies a current of 13-14 mA through the output stage, which is not consistent with 4.88V across R6. Maybe you recorded the measurements on R3 and R6 the wrong way round?
Alex
@Alex M R3 and R6 are measured voltage between resistor. What you mean by that maybe I recorded the measurements on R3 and R6 the wrong way round? I checked and they are still same.
R4 - 66.4 V
U1B pin 6 and gnd - 183.1 V
U1A between pin 1 and 3 - 179 V
U1B between pin 6 and 8 - 160 V
U2B - pin 6 - 209 V
U2A between pin 1 and 3 - 41 V
U2B between pin 6 and 8 - 26 V
I also do not have another tubes to test.
6n23p datasheet is available here
R4 - 66.4 V
U1B pin 6 and gnd - 183.1 V
U1A between pin 1 and 3 - 179 V
U1B between pin 6 and 8 - 160 V
U2B - pin 6 - 209 V
U2A between pin 1 and 3 - 41 V
U2B between pin 6 and 8 - 26 V
I also do not have another tubes to test.
6n23p datasheet is available here
Thank you, Alex for comments. Actually my introduction with VSE started from this article
Yes, my first simulations were only with three time constants 3180us, 318us and 75us and just later added 3.18us. And with 3.18us time constant the high frequency hump decreased.
Now curve is very flat but this additional 12n cap probably ruined Allens philosophy.
Simulation with Neumann time constant:
But still I am very curios about ringing on the leading edge of the square wave. Simulation do not see my power supply, so probably problem (or maybe not a problem) somewhere else. In real life we do not have square waves. If Joe Rasmussen can comment this FVP5A behavior, would be very nice.
I measured again and got another numbers. I mixed up R3 and R6 values.
R6 - 23 V
R3 - 5.15 V
R4 - 71.4 V
U1 - pin 1 - 178.9 V
U2 - pin 6 - 209 V
U1 - pin 3 ja 8 - 17,9 V
U2 - pin 3 ja 8 - 24,4 V
U1 - pin 3 - 5.15 V
U2 - pin 3 - 202,5 V
U1 - pin 8 - 23 V
U2 - pin 8 - 178 V
If I increased input sound about to 300 mV then output was about 2 V out and I started to hear some crackling and sound was not good. Does it indicates to something or my receiver started clipping? I have old Yamaha RX-V863 receiver.
Any ideas what changes to make to get it work right and that sound quality is also good?
As I said in my earlier post, U2A and U2B need to have enough voltage across them to be able to work properly. Your measurements confirm that the SLCF output stage is starved of voltage: the cathode follower U2B only has 30V across it, which is almost certainly why you appear to be hearing clipping.
This means you need more current through U1A. Looking back at the article on my website I see Allen recommends 6 mA, which means the anode of U1A should be at 100V. You will need to reduce R3 to increase the current - with 2K2 on the cathode I found this stage only had between 2.5 and 2.65 mA through it. I don't know why Allen published the circuit with such a high value - try 1K0 instead of 2K2 as a starting point.
Alex