This discussion of grid leak bias recalls design considerations for low noise that I’ve seen in some professional tube audio equipment, and also in my own RIAA/MC head amp work experiments. Basically, coupling caps might increase noise if noise equivalent circuits are not also considered. I’ll present this issue here for what it’s generally worth, and not necessarily as a recommendation to change anything in the design being discussed.
The nice part about a very large grid leak bias resistor (the 10M grid resistor on the ECC88) is that it allows the use of a tiny coupling cap, here a 10nF, which can be of very high quality. The pole frequency is set to a reasonable 1.6Hz, even with only 10nF.
However, there is a risk of additional noise generated by the 10M resistor itself and also by grid current fluctuations acting on that 10M resistor to create additional noise voltages. The latter will be predominantly of the flicker or 1/f^n variety and is similar in principle to noise currents in an opamp. One hopes that the preceding components will provide a low enough source impedance at all audio frequencies to essentially “short out” these noise sources. But we need to test that assumption by looking backwards through the coupling cap. In this case the noise voltage must “see” through the small coupling cap before being attenuated by the preceding RIAA filter components. But the cap’s reactance is no longer small compared to the impedances that precede it in this backwards view, especially in the lower frequencies. So there will be some excess noise from the last stage when using a small coupling cap compared to using a larger coupling cap. I whipped up a quick sim of two cases of coupling cap size, 10nF and 1000nF (a big cap, I realize) using the relevant components in the RIAA filter discussed here. See the image below, which shows the DIFFERENCE between the two cases.
So, interestingly, in some cases you may actually want to use a larger coupling capacitor for noise considerations than you would just for adequate low frequency response.
The original cap choice here may well represent the optimal sonic trade-off because of the more nearly ideal behavior of the smaller value cap. There is also a reasonable charging time constant for setting the ECC88’s grid bias. There are too many variables to know for sure whether a larger coupling cap is worth it without actually trying both a small coupling cap and a larger one. In the most sensitive range of human hearing, there would be only a small difference. But if you notice an undue amount of rumble or woofer movement, it might worth experimenting with larger values.
The nice part about a very large grid leak bias resistor (the 10M grid resistor on the ECC88) is that it allows the use of a tiny coupling cap, here a 10nF, which can be of very high quality. The pole frequency is set to a reasonable 1.6Hz, even with only 10nF.
However, there is a risk of additional noise generated by the 10M resistor itself and also by grid current fluctuations acting on that 10M resistor to create additional noise voltages. The latter will be predominantly of the flicker or 1/f^n variety and is similar in principle to noise currents in an opamp. One hopes that the preceding components will provide a low enough source impedance at all audio frequencies to essentially “short out” these noise sources. But we need to test that assumption by looking backwards through the coupling cap. In this case the noise voltage must “see” through the small coupling cap before being attenuated by the preceding RIAA filter components. But the cap’s reactance is no longer small compared to the impedances that precede it in this backwards view, especially in the lower frequencies. So there will be some excess noise from the last stage when using a small coupling cap compared to using a larger coupling cap. I whipped up a quick sim of two cases of coupling cap size, 10nF and 1000nF (a big cap, I realize) using the relevant components in the RIAA filter discussed here. See the image below, which shows the DIFFERENCE between the two cases.
So, interestingly, in some cases you may actually want to use a larger coupling capacitor for noise considerations than you would just for adequate low frequency response.
The original cap choice here may well represent the optimal sonic trade-off because of the more nearly ideal behavior of the smaller value cap. There is also a reasonable charging time constant for setting the ECC88’s grid bias. There are too many variables to know for sure whether a larger coupling cap is worth it without actually trying both a small coupling cap and a larger one. In the most sensitive range of human hearing, there would be only a small difference. But if you notice an undue amount of rumble or woofer movement, it might worth experimenting with larger values.
Attachments
mus said:
I have. I don't have tools to measure the difference objectively but sound wise I like it more than using resistors for the anode. At first I thought it sounded horrible until I measured that the anode voltage on the first stage was off by 30V because of the added drain on the PS from the CCS. I adjusted the 100K dropping resistor in the first stage to 47K and the sound improved considerably to the point that I preferred it to a resistor load.
I have yet to use MKPs for the decoupling chain (now using only electrolytics, unbypassed) for either stage, will this really make an audible difference?
Hi Garp,
Bypass the electrolytic caps with small value of MKP caps surely makes huge difference.
Give it a try and you will see.
I did it to my clone Marantz 7C line amp. It employs Blackgate caps and bypassed with Jensen's 0.1uF PIO caps. It now has the booming dynamics of Blackgate and smooth voices and highs.
Great, it sounds like a total different amp to me.
Johnny
Bypass the electrolytic caps with small value of MKP caps surely makes huge difference.
Give it a try and you will see.
I did it to my clone Marantz 7C line amp. It employs Blackgate caps and bypassed with Jensen's 0.1uF PIO caps. It now has the booming dynamics of Blackgate and smooth voices and highs.
Great, it sounds like a total different amp to me.
Johnny
fdegrove said:Hi,
I still have some.
They're skirted, noval and for chassis mounting.
Let me know how many you'd need by e-mail if you'd like some.
Cheers,
Just gone through the whole thread, probably will build one soon, after my projects currently on hand.
Just a question: What is the purpose of the skirt on the tube socket? And what is a reasonable price to pay for a piece excluding shipping?
thanks!
plate choke on ecc88?
Hello,
Since KYW is busy in his new venture, maybe others could share their thoughts wrt his Valve El Cheapo design. I intend to build this, and maybe later on try a plate choke (MQ iron etc.) on the ECC88 output (parafeed) just to hear the difference Vs. 12k plate load R. Maybe I could use a 'tuning resistor' before the choke to get the intended operating points?... The question is, will the riaa eq response be off the mark if I use a plate choke on the output? Thanks.
http://www.diyaudio.com/forums/attachment.php?s=&postid=335624&stamp=1077837811
fred
Hello,
Since KYW is busy in his new venture, maybe others could share their thoughts wrt his Valve El Cheapo design. I intend to build this, and maybe later on try a plate choke (MQ iron etc.) on the ECC88 output (parafeed) just to hear the difference Vs. 12k plate load R. Maybe I could use a 'tuning resistor' before the choke to get the intended operating points?... The question is, will the riaa eq response be off the mark if I use a plate choke on the output? Thanks.
http://www.diyaudio.com/forums/attachment.php?s=&postid=335624&stamp=1077837811
fred
I finally have the time to begin building this DC phono stage. I'll connect it to a DL 103 via mc transformer, to JEL type 300B SET amplifier and Pi Theater Four speakers. I'm collecting all the components but I have some difficulties in finding the 180K 7W, 12K 12W and 100K 12W power resistors. Could you suggest me a source in Europe for them?
Another question is if I can replace resistors in PSU with chockes having the same resistive value.In Italy Nuova Elettronica sells cheap double 30H/100Ohm chockes. Are they suitable? Which current value I need?
I'm considering the use of MKP runcaps for PSU and ASC MKP in oil last two caps in riaa cabinet. Is it a good upgrade?
Is it necessary a tube shielding for 6922 and ECC83? I have Siemens E88CC but I need an advice for a good ECC83...
Which transformer is suitable for my Denon? A NOS mic TX is a good option?
Thanks to all in advance!
Vincenzo
Another question is if I can replace resistors in PSU with chockes having the same resistive value.In Italy Nuova Elettronica sells cheap double 30H/100Ohm chockes. Are they suitable? Which current value I need?
I'm considering the use of MKP runcaps for PSU and ASC MKP in oil last two caps in riaa cabinet. Is it a good upgrade?
Is it necessary a tube shielding for 6922 and ECC83? I have Siemens E88CC but I need an advice for a good ECC83...
Which transformer is suitable for my Denon? A NOS mic TX is a good option?
Thanks to all in advance!
Vincenzo
Its worth screening the ECC83 - and running its heaters on well filtered DC.
The ECC88 doesn't need DC or screening.
Be careful with chokes. They may introduce audable resonances, though I would have no way of predicting whether the ones you suggest would.
Farnell is your best source for components in Europe.
Shoog
The ECC88 doesn't need DC or screening.
Be careful with chokes. They may introduce audable resonances, though I would have no way of predicting whether the ones you suggest would.
Farnell is your best source for components in Europe.
Shoog
regulated PSU for phono stages
Hmmm, I also think there is a dangerous potential (pun) problem with the ECL85 regulator cathode-to-filament voltage UNLESS it has its own 6.3v winding and that winding is tied to some potential around about half the output voltage. I really think solid state voltage regulators have far less problems to watch out for, and (if correctly designed) can introduce no sonic differences... and having separate regulated outputs for each stage of each channel is way less impractical than their valve/tube counterparts.
P.S. DC heater supplies for sensitive input stages is a pretty good idea, and that is one supply where solid state regulators are essential.
fdegrove said:Hi,
(responding to: Vkf of ECL85 = 200V. I think I'll have a problem with 250V at the cathode of pentode)
No, not if you build it as drawn.
Cheers,
Hmmm, I also think there is a dangerous potential (pun) problem with the ECL85 regulator cathode-to-filament voltage UNLESS it has its own 6.3v winding and that winding is tied to some potential around about half the output voltage. I really think solid state voltage regulators have far less problems to watch out for, and (if correctly designed) can introduce no sonic differences... and having separate regulated outputs for each stage of each channel is way less impractical than their valve/tube counterparts.
P.S. DC heater supplies for sensitive input stages is a pretty good idea, and that is one supply where solid state regulators are essential.
low output, low quality sound...
This has been a very interesting thread, even for a relative newbie like me. So I don't mean the subject title to be negative. Far from it!
I just have built this phono preamp together with its accompanying linestage (both proposed by Mr T (Kwei YW) and they are in the same box (very tight as it turned out) with a selector switch and 100K volume pot in between the two circuits. The combo is powered from a 4-section outboard power supply and noise is almost inexistent, apart from valve noise.
The line preamp sounds very good, with an authoritative full bodied, sweet character. The phono is however problematic. It has low output and the sound is very dull (closed in – almost like a pocket radio!).
I haven’t used fancy components, just what I had available. The line stage is terminated with a 4.7uF polyprop audio cap, which maybe gives it that sound.
For the phono stage I used standard 1% resistors (except for the high powered anode load resistors on the 2nd stage (ECC88); polystyrene caps in the RIAA network and coupling cap between 1st and 2nd stage (10n)/160V);; polypropylene p/s decoupling cap for the 1st stage (10uF) and electrolytic for the other decoupling cap (22uF);; and standard polyesters 2.2uF for the signal output. This is the link to the phono circuit:
http://diyaudio.com/forums/showthread.php?postid=335624#post335624
I checked the wiring and measured resistances but I can't put my finger on the offending component/s. Can some of the more knowledgeable members (or builders) on this forum indicate the possible areas to look for in order to identify the fault/s.
Thanks in advance.
Joe A
This has been a very interesting thread, even for a relative newbie like me. So I don't mean the subject title to be negative. Far from it!
I just have built this phono preamp together with its accompanying linestage (both proposed by Mr T (Kwei YW) and they are in the same box (very tight as it turned out) with a selector switch and 100K volume pot in between the two circuits. The combo is powered from a 4-section outboard power supply and noise is almost inexistent, apart from valve noise.
The line preamp sounds very good, with an authoritative full bodied, sweet character. The phono is however problematic. It has low output and the sound is very dull (closed in – almost like a pocket radio!).
I haven’t used fancy components, just what I had available. The line stage is terminated with a 4.7uF polyprop audio cap, which maybe gives it that sound.
For the phono stage I used standard 1% resistors (except for the high powered anode load resistors on the 2nd stage (ECC88); polystyrene caps in the RIAA network and coupling cap between 1st and 2nd stage (10n)/160V);; polypropylene p/s decoupling cap for the 1st stage (10uF) and electrolytic for the other decoupling cap (22uF);; and standard polyesters 2.2uF for the signal output. This is the link to the phono circuit:
http://diyaudio.com/forums/showthread.php?postid=335624#post335624
I checked the wiring and measured resistances but I can't put my finger on the offending component/s. Can some of the more knowledgeable members (or builders) on this forum indicate the possible areas to look for in order to identify the fault/s.
Thanks in advance.
Joe A
sonata149 said:
There is a small time window...between ON and the valves drawing current...here your 160V caps can run into their limits. 250V is good, 400V is on the safe site. Please measure.
Not every ECC88 valve runs well in the riaa with 10M in this circuit...but with 3M there is minimal impact to riaa-curve and nearly every ECC88 performs well. Read the former pages, TL says this...AMEN. Else...try another ECC88
FYI:
http://www.jogis-roehrenbude.de/Leserbriefe/Carsten-RIAA-3/Riaa-3.htm
http://www.jogis-roehrenbude.de/Leserbriefe/Martin-Ruppel-Phonamp-2/Phono_II.htm
Sorry folk, german only...please try babblefish for a funny translation!