Kofi Annan said:
I still say KISS. I see no benefit to constant current filament supplies here.
Sheldon
Between us, there is a different sound with them that he will like or not. But I told him too to keep it simple initially. After everything works OK, and he becomes familiar enough, he can try secondary things like that, to look for their merit. 😉
Kofi,
I've built this phono and I've also been puzzled with filament configuration,so I'll tell how I solve it.
First thing for me to understand is that 12AY7 has two filaments inside (one for eatch triode) so can be configured to work with 12,6V (two filaments of the same tube in series) connecting V+ to pin 4 and V- to pin 5,pin 9 floating.They can also be configured to work with 6,3 V straping pin 4 and pin 5 together feeding V+ to both and pin 9 with V-.
6n23/6dj8/ecc88 can only be configured as 6.3v pin 4 with V+ pin 5 with V- pin 9 either floating or to ground.
What I did was to configure each 12AY7 to work with 6,3 v and then connected both tubes in series(one tube's 4+5 with V+ ,same tube's pin 9 to other tube's pin 4+5 and same tube's pin 9 with V-) feeding them from a 12.6 v voltage regulator.
6n23/6dj8/ecc88 also connected in series (one tube's pin 4 with V+ same tube's pin 5 with other tube's pin 4 ) and feed them from a different 12,6 V voltage regulator.
So you need to rectify your AC once then make two 12,6 V voltage regulators (i.e LM 317) and that's it.
I've built this phono and I've also been puzzled with filament configuration,so I'll tell how I solve it.
First thing for me to understand is that 12AY7 has two filaments inside (one for eatch triode) so can be configured to work with 12,6V (two filaments of the same tube in series) connecting V+ to pin 4 and V- to pin 5,pin 9 floating.They can also be configured to work with 6,3 V straping pin 4 and pin 5 together feeding V+ to both and pin 9 with V-.
6n23/6dj8/ecc88 can only be configured as 6.3v pin 4 with V+ pin 5 with V- pin 9 either floating or to ground.
What I did was to configure each 12AY7 to work with 6,3 v and then connected both tubes in series(one tube's 4+5 with V+ ,same tube's pin 9 to other tube's pin 4+5 and same tube's pin 9 with V-) feeding them from a 12.6 v voltage regulator.
6n23/6dj8/ecc88 also connected in series (one tube's pin 4 with V+ same tube's pin 5 with other tube's pin 4 ) and feed them from a different 12,6 V voltage regulator.
So you need to rectify your AC once then make two 12,6 V voltage regulators (i.e LM 317) and that's it.
You guys are making this waaay too complicated.
Steve's schematic shows you exactly how to wire it for 6.3V. Just follow the schematic.
Sheldon
Steve's schematic shows you exactly how to wire it for 6.3V. Just follow the schematic.Sheldon
RIAA AND HEATERS
Maybe I missed something... but I would like to propose to whoever is building it (or any other RIAA phono) to try my RPA. The RPA circuit can be found on my site: www.tubeaudio.8m.com
The circuit is basically MC, but can be easily "adapted" to MM use, by slight changes and adjustments to fit other tubes. I.e. mine uses an E180CC as first tube, while the second is ECC81 and the line stage is 5670.
Further with the 6.3 in series vs 12.6 in line: I have done that myself and can share the experience. In practice, seldom two halves of the same double triode draw the same heater current, let alone two tubes (like 5670 x 2 in series). In the end, I connected all the center pins of the 12.6V heaters with the mid point between two 6.3V heated tubes in series to balance the heater current draw (and therefore the voltage, and the sound as well, since the sound of the same tube heated at 6.0 and 6.6V is not the same...).
While there is an important current advantage in heating tubes at 12.6V (2x less current, easier voltage regulation), on the other hand there can be several practical disadvantages.
Inquireies and comments welcome.
Regards,
Alex
Maybe I missed something... but I would like to propose to whoever is building it (or any other RIAA phono) to try my RPA. The RPA circuit can be found on my site: www.tubeaudio.8m.com
The circuit is basically MC, but can be easily "adapted" to MM use, by slight changes and adjustments to fit other tubes. I.e. mine uses an E180CC as first tube, while the second is ECC81 and the line stage is 5670.
Further with the 6.3 in series vs 12.6 in line: I have done that myself and can share the experience. In practice, seldom two halves of the same double triode draw the same heater current, let alone two tubes (like 5670 x 2 in series). In the end, I connected all the center pins of the 12.6V heaters with the mid point between two 6.3V heated tubes in series to balance the heater current draw (and therefore the voltage, and the sound as well, since the sound of the same tube heated at 6.0 and 6.6V is not the same...).
While there is an important current advantage in heating tubes at 12.6V (2x less current, easier voltage regulation), on the other hand there can be several practical disadvantages.
Inquireies and comments welcome.
Regards,
Alex
salas said:
Salas,
There is a difference between current regulated, and voltage regulated filaments for DHT's. The signal sees the impedance of the filament and supply. No such thing occurs with unipotential tubes. If there is no signal passing through the filament, no voltage potential change occurs. The filament presents a constant load to the supply. Given that, there is no inherent difference between voltage regulation and current regulation here. If you hear a difference, it's much more likely due to noise issues than supply type.
Sheldon
edit: I agree with Alex. More to be gained by adjusting voltage than regulator type.
I have experienced it with both my old cascade noval preamp and with my 6V6 single stage preamp. Both indirectly heated, both no loop feedback. They got a mellower tone with more differentiation on bass notes with ICCS heaters. I know what you mean technically about DHTs and unipotential, but I would not be honest if not to inform my personal experience. Its a marginal gain. Other people may not hear it, maybe debated as an illusion, but that is what I know. They could be coupling less noise to the cathode somehow or not. But as I say to Kofi too, there isn't much about it so to must plan it.
salas said:
Sure, one supply might couple less noise than another. But that's just a function of the supply quality. One can build a poor current supply or a poor voltage supply. My main point is that one can rely on basic physics first. I see too much emphasis on chasing subtle effects before the darn amp is even built - especially effects that don't have a clear cause. "Better" is often the enemy of "good". Keep it simple, build it, then worry about (as you say) "marginal" effects.
Sheldon
Its the parts collecting through post idle period, and the bad tubes that Kofi got and has to get new, that give him time to think about how to make it best possible in one shot I guess. If he had em all parts on his bench, he would have slapped it together as normal and fast as possible...😉
OK-- I hope I haven't caused any problems here. I understand the CCS issue now, I believe, but I will start by implementing a simple regulated 6.3V supply for all heaters. I haven't ordered parts yet, but I will do so today.
Part ordering delays were courtesy of Mrs. Annan. She wants a new house. I say, what's a house without a phono stage? She has lots to say in response that would result in ******s on the forum.
Anyway, I'll advise once I order parts, then when they arrive. In the meantime, I'll be working on layout chassis modeling.
Thanks for all the input.
Stay tuned.
Kofi
Part ordering delays were courtesy of Mrs. Annan. She wants a new house. I say, what's a house without a phono stage? She has lots to say in response that would result in ******s on the forum.
Anyway, I'll advise once I order parts, then when they arrive. In the meantime, I'll be working on layout chassis modeling.
Thanks for all the input.
Stay tuned.
Kofi
No problems at all, constructive dialogue! Will be happy to see you starting it with pictures and all! In the meantime I will do some head to head measurements and listening again with a very good build of my NJfet Riaa with double mono shunts that Mikvous did very recently, and if I end up in any tweaks I will let you know. My best to Mrs. Annan!
Kofi Annan said:
If that's the only way you can wire the filament tranny, then you are better off going with 12.6V (I'd do 12, or a little less, actually). Wire the two 6922 in series and the two AY7 (wired for 12.6), in parallel with the series wired 6922. Lift the whole shebang to about 50V. I'd start with a simple voltage regulated supply. Go ahead and leave enough room to play with other options down the road, if you like. That won't be hard to do, since your supply should be a separate unit anyway. Just leave some extra filament wiring in your umbilical cable.
Get it built, play with your tubes, cartridge loading, get the set up singing and quiet, and get the RIAA curve correct and balanced between channels. Then play with refinements.
Sheldon
Hi all,
just want to join the party ;-)
Did i understand it right: you are discussing two versions of
steve benches riaa amp for mm systems and
Salas uses the original circuit with just C1 changed and
sheldon uses slightly modded verison with added buffer?
What outputimpedance do both versions have? Or better
what input impedance is needed from the following linepre?
Greetings Ulf
just want to join the party ;-)
Did i understand it right: you are discussing two versions of
steve benches riaa amp for mm systems and
Salas uses the original circuit with just C1 changed and
sheldon uses slightly modded verison with added buffer?
What outputimpedance do both versions have? Or better
what input impedance is needed from the following linepre?
Greetings Ulf
le´flu said:
Yes, that's about right. I moved the position of the coupling cap between the cascode to the following stage, and changed the resistor and cap values to match parts that I had on hand. Like Salas, I also recalculated the second RIAA cap and added the extra HF compensation.
The original version is calculated for a load of 100k or higher. My load is about 50k, so I added the buffer. The grounded cathode amp and buffer actually comprise a version of John Broskie's Aikido (note that the tube with unbypassed cathode has an ri that equals the plate resistor). The TubeCad site has a table of output impedances, for various tubes and cathode resistors.
Sheldon
Ahhh, thanks for information.
One question concering the original circuit....
is it possible or recommendet to leave out cap
c 12? I havn´t seen inputcaps on tube riaas
until now. What problems could hapen?
Sheldon, your version with the input transformer
is for mc isn´t it?
Ulf
One question concering the original circuit....
is it possible or recommendet to leave out cap
c 12? I havn´t seen inputcaps on tube riaas
until now. What problems could hapen?
Sheldon, your version with the input transformer
is for mc isn´t it?
Ulf
I'd guess that the input cap and grid resistor serve as an extra pole to roll off LF response - rumble filter, if you will. If your amp won't blow your speakers with some very low frequency content, then you can leave it out. I'd not worry at all about tube amps. Most SS amps can go all the way to DC and often have high enough power to damage the LF driver.
Yes, mine is configured with an MC step up. The step up will not pass very low frequencies.
Sheldon
Yes, mine is configured with an MC step up. The step up will not pass very low frequencies.
Sheldon
The one I have shown on the picture (post 68 page 3) has no input cap and a naked Dynavector DV6/10x silver step up inside. No problems. For MM I don't know, it never played with one. It never played with SS amps either.
I have investigated deeper what goes on with the RIAA curve and input cap for this cascode phono. My new card has also a high Z input that I measured for 1.18 MegOhm. So the loading does not affect the non buffered Bench. I think that Sheldon uses a buffer stage?
The answer for why I kept on tweaking its Riaa emerged. Whatever we do, its gain drops below 100hZ gradually. Either MM or with MC step up, the loss is about 6dB @ 50Hz. With input cap and one Meg grid to ground as per original, it only gains 1.5dB VS 47-68k Rload - grid to ground and no cap. Off course the cap there impairs the resolution and still nowhere rectifies the situation. The posted picture shows it versus red ref Riaa with 8n2 C1/C9. The absolutely right HF is with 6n8, but it deviates in upper mid. Alas, without proper bass level, it sounds and measures tipped up with any cartridge. By drooping the highs, enough subjective balance is restored from 8n2 to 11n4 range DEPENDING on the speaker - room interface.
The answer for why I kept on tweaking its Riaa emerged. Whatever we do, its gain drops below 100hZ gradually. Either MM or with MC step up, the loss is about 6dB @ 50Hz. With input cap and one Meg grid to ground as per original, it only gains 1.5dB VS 47-68k Rload - grid to ground and no cap. Off course the cap there impairs the resolution and still nowhere rectifies the situation. The posted picture shows it versus red ref Riaa with 8n2 C1/C9. The absolutely right HF is with 6n8, but it deviates in upper mid. Alas, without proper bass level, it sounds and measures tipped up with any cartridge. By drooping the highs, enough subjective balance is restored from 8n2 to 11n4 range DEPENDING on the speaker - room interface.
- Status
- Not open for further replies.