I finished this project over Christmas and am now ready to post some details.
Here's a description; cct will be posted very soon
MC stepup by Lundahl LL9206 transformer.
First stage is a Kimmel mu stage of E180CC triode, loaded by D3A pentodes.
Single stage RIAA, with series resistor of 13.6Kohm, C1 of .160uF, R2 of 1985 ohms; then C2 of .056uF and (optional) R3 of 56 ohms.
Second stage is again a Kimmel mu stage, this time of 6SL7 loaded by E180F pentodes. Then direct to output via .33uF caps.
Details, Philosophy, Issues:
I went round in circles on the overall design but then a lot of things came together.
Use of good sounding valves! Fundamental. Why start any other way?
Use of a highish Gm valve in the first stage, to get some 'energy' into the system.
Running this valve at highish current for low noise.
The E180CC seemed to fit the bill; enough gain and good Gm, said to sound good at over 6mA (I run it at 10mA).
I *really* wanted to ensure there was enough drive, guts, whatever left in the signal to be able to drive a 6SL7 for stage 2. This suggested that the RIAA compensation should be done at a lower impedance than usual; i.e. with lower series resistance, and higher capacitance, than normal.
This needed a low output impedance from stage 1 so as to drive the RIAA, and so that impedance changes would be negligible compared to the RIAA series resistor. Hence the choice of active loading.
This gave a big issue ... would a pentode load be too noisy for a phono stage?
I chose to do the RIAA at 13.6kohm series resistance. This is a lot lower than just about any design I've seen out there, and I've looked at many.
Doing this seemed to have many advantages. Lower resistance may reduce noise; high capacitance values will swamp out any stray capacitance; the Miller capacitance of the 6SL7 is small compared to the RIAA capacitors. But the main reason was to ensure that the 6SL7 could be driven really well; no compromises here if possible!
I wanted dynamics, drive, music, power, drama; but also sweetness, delicacy, tonality. I like Herb's phrase, 'Flesh and Blood'. I'd say you need really good *drive* between stages to get this, and avoid a wimpy or half hearted sound.
Another issue: I'd not seen RIAA done at this low a resistance; if I was right, it could be a major breakthrough! Most likely I expected to learn why not ....
The values were chosen to match Russian teflon caps which I bought in bulk and then selected.
The second stage is again a mu stage, as I wanted to avoid a cathode follower stage.
The Noise Issue:
Noise in the first stage was minimised in several ways. Using the E180CC of high Gm; running it at high current (10mA) and low bias (-1V); and using the D3A pentode load which is very low noise for a pentode.
Note that the bias point was very carefully considered so as to give just enough headroom (I hope).
Results:
This phono stage is quiet!
That is a big, big result. It's hard to say exactly but it's very similar to my Sonic frontiers SFP-1 in noise level; i.e. only significant when turned up appreciably above normal listening levels. I have a feeling it may end up a touch quieter when all is done and run in.
Sound quality is excellent (with very minor issues which I expect to be fixed by running in).
My only comparator is my SFP-1 (upgraded with FREDs, film PSU cap, teflon signal caps, Auricap output cap); not a bad unit.
Isabelle has better: dynamics; transient attack; tonality; resolution (low level detail); clarity.
Very musical, lively, engaging.
The margin of improvement is significant.
The issues are:
A slight loss of fundamentals/undertones on voice. I had this, only much more so, with 'Nina' until she had run in, and that took over a month; so I'm hoping for a result here.
A touch of harshness, very slight ring, almost metallic sound, sometimes on female vocals. I guess that this is microphonics; if so, tube dampers and better feet on the unit should be an easy fix.
Image width is fine, but depth isn't great; it may improve with running in so I'll just wait and see.
The issues are *small*; sound quality is very good, more than meeting my objective of bettering the SFP-1.
Oh, there is another issue; I broke a lug off a PSU choke so I'm using a resistor of the same dcr as a temp fix; so there is just a bit of noise on one channel.
The way she plays Mozart piano concertos is just amazing!
Here's a description; cct will be posted very soon
MC stepup by Lundahl LL9206 transformer.
First stage is a Kimmel mu stage of E180CC triode, loaded by D3A pentodes.
Single stage RIAA, with series resistor of 13.6Kohm, C1 of .160uF, R2 of 1985 ohms; then C2 of .056uF and (optional) R3 of 56 ohms.
Second stage is again a Kimmel mu stage, this time of 6SL7 loaded by E180F pentodes. Then direct to output via .33uF caps.
Details, Philosophy, Issues:
I went round in circles on the overall design but then a lot of things came together.
Use of good sounding valves! Fundamental. Why start any other way?
Use of a highish Gm valve in the first stage, to get some 'energy' into the system.
Running this valve at highish current for low noise.
The E180CC seemed to fit the bill; enough gain and good Gm, said to sound good at over 6mA (I run it at 10mA).
I *really* wanted to ensure there was enough drive, guts, whatever left in the signal to be able to drive a 6SL7 for stage 2. This suggested that the RIAA compensation should be done at a lower impedance than usual; i.e. with lower series resistance, and higher capacitance, than normal.
This needed a low output impedance from stage 1 so as to drive the RIAA, and so that impedance changes would be negligible compared to the RIAA series resistor. Hence the choice of active loading.
This gave a big issue ... would a pentode load be too noisy for a phono stage?
I chose to do the RIAA at 13.6kohm series resistance. This is a lot lower than just about any design I've seen out there, and I've looked at many.
Doing this seemed to have many advantages. Lower resistance may reduce noise; high capacitance values will swamp out any stray capacitance; the Miller capacitance of the 6SL7 is small compared to the RIAA capacitors. But the main reason was to ensure that the 6SL7 could be driven really well; no compromises here if possible!
I wanted dynamics, drive, music, power, drama; but also sweetness, delicacy, tonality. I like Herb's phrase, 'Flesh and Blood'. I'd say you need really good *drive* between stages to get this, and avoid a wimpy or half hearted sound.
Another issue: I'd not seen RIAA done at this low a resistance; if I was right, it could be a major breakthrough! Most likely I expected to learn why not ....
The values were chosen to match Russian teflon caps which I bought in bulk and then selected.
The second stage is again a mu stage, as I wanted to avoid a cathode follower stage.
The Noise Issue:
Noise in the first stage was minimised in several ways. Using the E180CC of high Gm; running it at high current (10mA) and low bias (-1V); and using the D3A pentode load which is very low noise for a pentode.
Note that the bias point was very carefully considered so as to give just enough headroom (I hope).
Results:
This phono stage is quiet!
That is a big, big result. It's hard to say exactly but it's very similar to my Sonic frontiers SFP-1 in noise level; i.e. only significant when turned up appreciably above normal listening levels. I have a feeling it may end up a touch quieter when all is done and run in.
Sound quality is excellent (with very minor issues which I expect to be fixed by running in).
My only comparator is my SFP-1 (upgraded with FREDs, film PSU cap, teflon signal caps, Auricap output cap); not a bad unit.
Isabelle has better: dynamics; transient attack; tonality; resolution (low level detail); clarity.
Very musical, lively, engaging.
The margin of improvement is significant.
The issues are:
A slight loss of fundamentals/undertones on voice. I had this, only much more so, with 'Nina' until she had run in, and that took over a month; so I'm hoping for a result here.
A touch of harshness, very slight ring, almost metallic sound, sometimes on female vocals. I guess that this is microphonics; if so, tube dampers and better feet on the unit should be an easy fix.
Image width is fine, but depth isn't great; it may improve with running in so I'll just wait and see.
The issues are *small*; sound quality is very good, more than meeting my objective of bettering the SFP-1.
Oh, there is another issue; I broke a lug off a PSU choke so I'm using a resistor of the same dcr as a temp fix; so there is just a bit of noise on one channel.
The way she plays Mozart piano concertos is just amazing!
PSU and Earthing Details
Easier to describe it than draw it ...
Toroid mains TX, 330-0-330 100mA and 0-6.3 4A for rectifier heater.
6CJ3 damper diode rectifiers.
2 pi PSU.
1st stage 1.5uF, 20H 170 ohm, 20uF
This then splits into 4 legs for left and right, stage 1 and 2.
Each leg 30H 560ohm, 20uF.
I analyzed the PSU in two ways; as a second stage on its own, and as all four second legs parallelled up. Results were:
Alone: pole frequency 9.4 Hz, Q 3.1
Parallel: pole frequency 14 Hz, Q 4.7
These results look OK to me, with a low enough pole and with moderate Q ... low enough for stability and high enough not to give a 'slow' sound. PSUD simulation also looked promising.
I planned to use about 3uF for the first cap but needed to reduce this to 1.5uF to get the right voltage. I think this gives a 'transitional' PSU with some of the qualities of a choke input supply?
Anyway, the first choke seems up to the job as it's well over rated.
Components:
cap 1, 1.5uF, general film
choke 1: Sowter
all 20uF caps: GE 97F series poly/oil motor run caps
30H chokes: Bartolucci S6
Other:
I use resistor bridges to float the heater PSUs. One across each 2nd stage psu leg. One floats the triode heaters to 30V, the other floats the pentode heaters to about 270V (i.e. 30V above the cathode). The bridges also act as bleeds. And yes, I did take great care over power ratings of the resistors and allowing plenty of space around them.
Earthing:
Left and right are isolated until they meet at cap 2.
Earth node in stage 1 cct; input earth and PSU cap earth connect to this node.
Node is linked to one end of ...
Bus bar of thick wire (14GA?) across the ends of the RIAA caps. Other end of bus links to ...
Earth node in stage 2 cct; output socket earth and PSU cap connect to this node.
Middle of bus bar connects to earthy side of cap 2.
Cap 2 and cap 1 connect to physical earth.
This gives tight coupling within each stage.
Also, total length of earth wiring is kept short so as to minimise pickup of stray signals, RFI etc.
And noisy PSU sub-circuits are kept away from signal earths (I think!).
Easier to describe it than draw it ...
Toroid mains TX, 330-0-330 100mA and 0-6.3 4A for rectifier heater.
6CJ3 damper diode rectifiers.
2 pi PSU.
1st stage 1.5uF, 20H 170 ohm, 20uF
This then splits into 4 legs for left and right, stage 1 and 2.
Each leg 30H 560ohm, 20uF.
I analyzed the PSU in two ways; as a second stage on its own, and as all four second legs parallelled up. Results were:
Alone: pole frequency 9.4 Hz, Q 3.1
Parallel: pole frequency 14 Hz, Q 4.7
These results look OK to me, with a low enough pole and with moderate Q ... low enough for stability and high enough not to give a 'slow' sound. PSUD simulation also looked promising.
I planned to use about 3uF for the first cap but needed to reduce this to 1.5uF to get the right voltage. I think this gives a 'transitional' PSU with some of the qualities of a choke input supply?
Anyway, the first choke seems up to the job as it's well over rated.
Components:
cap 1, 1.5uF, general film
choke 1: Sowter
all 20uF caps: GE 97F series poly/oil motor run caps
30H chokes: Bartolucci S6
Other:
I use resistor bridges to float the heater PSUs. One across each 2nd stage psu leg. One floats the triode heaters to 30V, the other floats the pentode heaters to about 270V (i.e. 30V above the cathode). The bridges also act as bleeds. And yes, I did take great care over power ratings of the resistors and allowing plenty of space around them.
Earthing:
Left and right are isolated until they meet at cap 2.
Earth node in stage 1 cct; input earth and PSU cap earth connect to this node.
Node is linked to one end of ...
Bus bar of thick wire (14GA?) across the ends of the RIAA caps. Other end of bus links to ...
Earth node in stage 2 cct; output socket earth and PSU cap connect to this node.
Middle of bus bar connects to earthy side of cap 2.
Cap 2 and cap 1 connect to physical earth.
This gives tight coupling within each stage.
Also, total length of earth wiring is kept short so as to minimise pickup of stray signals, RFI etc.
And noisy PSU sub-circuits are kept away from signal earths (I think!).
More Info - E180CC; the name; microphonics
The E180CC's came from Billington (as did all the other valves). Nice Mullards.
The name Isabelle:
Those of you who remember where the name 'Nina' came from (for the line stage) should know which Isabelle I mean ... mmmm!
An alternative spelling 'Isobel' was promising, as a pun could have been made on it in the form of 'Iso' and 'Bel' to mean 'of constant loudness' ...
But it's no good, her name is spelt 'Isabelle'. Anyway, it looks OK in script, on a light oak front panel natching 'Nina', but the unit is just an ugly box really. I'll have some photos soon.
Microphonics:
Last night I repeated the tap test and found that one D3A was more microphonic than the other; replaced with a better one.
I also made up some tube dampers; I had some silicone O rings that were a bit too big so I wound ptfe tape around them (fiddly!) until they fitted. Put 2 on each stage 1 valve and one on each E180F ... didn't have one for the 6SL7.
Results are promising; the slight edge seems to have gone, vocals are generally better all round, still with excellent resolution but with most (all?) of the fundamental now back, and sounding a bit 'liquid'. Results were very emotionally persuasive!
To be honest I'll reserve judgement until I've listened a few times; I find that memory effects and possible placebo effects can cloud my judgement on a first listen, usually OK after 3.
But I'm very optimistic; time now to relax, stop trying to evaluate and just listen to music as, with luck, it runs in some more. It now has about 30 hours over 2 weeks and I expect improvements up to about 100 hours.
The E180CC's came from Billington (as did all the other valves). Nice Mullards.
The name Isabelle:
Those of you who remember where the name 'Nina' came from (for the line stage) should know which Isabelle I mean ... mmmm!
An alternative spelling 'Isobel' was promising, as a pun could have been made on it in the form of 'Iso' and 'Bel' to mean 'of constant loudness' ...
But it's no good, her name is spelt 'Isabelle'. Anyway, it looks OK in script, on a light oak front panel natching 'Nina', but the unit is just an ugly box really. I'll have some photos soon.
Microphonics:
Last night I repeated the tap test and found that one D3A was more microphonic than the other; replaced with a better one.
I also made up some tube dampers; I had some silicone O rings that were a bit too big so I wound ptfe tape around them (fiddly!) until they fitted. Put 2 on each stage 1 valve and one on each E180F ... didn't have one for the 6SL7.
Results are promising; the slight edge seems to have gone, vocals are generally better all round, still with excellent resolution but with most (all?) of the fundamental now back, and sounding a bit 'liquid'. Results were very emotionally persuasive!
To be honest I'll reserve judgement until I've listened a few times; I find that memory effects and possible placebo effects can cloud my judgement on a first listen, usually OK after 3.
But I'm very optimistic; time now to relax, stop trying to evaluate and just listen to music as, with luck, it runs in some more. It now has about 30 hours over 2 weeks and I expect improvements up to about 100 hours.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.