Snail Phono Pre

What people are working on at the moment
Post Reply
little eddy
Old Hand
Posts: 625
Joined: Sun Nov 09, 2008 2:06 pm
Location: Manchester

Snail Phono Pre

Post by little eddy » Sat Nov 13, 2010 1:10 am

First the name - relative to you guys I think this describes the pace of my projects quite well. Plus with other things to do and a need to get back to my DAOB speakers, it may be a while before it gets kicked off in earnest.

Background - I was gven a phono/pre-amp in return for some services rendered. I brought this to Eggy last year but no one was impressed and I haven't managed to get it sounding anything like I hoped. I shared the design with Nick who was quite frank and suggested I rebuild the phono and pre-amp stages from scratch.

What's OK. It looks like I have plenty of 12.6V heater capacity but with the currrent amp using 6 ECC83s and the builder saying that the 280V regulated HT would be OK for a further 20mA, I may be limited to 26 to 30mA in total so some of the CCS designs etc are out of the question.

Design. So Nick's pointed me in a direction but as usual I've left my compass at home. If I understand him correctly the phono would consist of a parallel ECC83 input and then 12AU7 stage, and the pre a 12AU7 followed by 5687 CF.

So I've looked on the internet and the two nearest schematic I've found are a Conrad Johnson PV-12 for the phono and an AN M-7 for the pre: http://www.drtube.com/audioamp.htm

I'm going to start on the pre-stage so have taken the M-7 as a starting point. It uses the 6072a in both stages but I'm proposing a 5687 CF as per Nick's advice.

Am I right in thinking the 12AY7 and 6072a and compatable? If so, I think I've worked through the 6072a stage and have Ia of 1.0mA, Vgk of -3V and gain of 28.5.

I've seen both in MJ and on the net a recommendation to apply a cathode bias to the CF so that is where I'm at now, aiming for around 10mA. Am I correct in now thinking that the Vg needs to be quite large to take the voltage swing after the gain of the 1st stage? If I understand correctly, with CD at full output and no volume attenuation, the 5687 CF would clip. A shunt mod on the volume control might half the output but what should my target grid voltage be? Looking at the curves I was thinking of Vg at around -10V and with the proposed potential divider, Rbias would be 1k and with Rk of 20k, Vk would be 210V. Does this sound right/feasible?

Please ignore the phono for now as it is just drawn as a schmatic at the moment with no science or thought given to the component values other than what could be copied from the PV-12 scheme.

Pics to follow when I start the disassembly.
Attachments
10-11-11 Snail ECC83, 6072a, 6072a & 5687 CF  Phono & Line Stage Schematic.jpg
TD-125/RB250/MC25FL & 'Snail' phono, NAS/SBT with CS4398 DAC, 41MP pre & MoFo Power, still messing with OBs.

User avatar
Nick
Site Admin
Posts: 11886
Joined: Sun May 06, 2007 10:20 am
Location: West Yorkshire

Post by Nick » Sat Nov 13, 2010 1:25 am

've seen both in MJ and on the net a recommendation to apply a cathode bias to the CF so that is where I'm at now, aiming for around 10mA. Am I correct in now thinking that the Vg needs to be quite large to take the voltage swing after the gain of the 1st stage? If I understand correctly, with CD at full output and no volume attenuation, the 5687 CF would clip. A shunt mod on the volume control might half the output but what should my target grid voltage be?
Remember the current and so voltage across the 20k load resistor will also change. So for a 10v p/p output swing, the voltage on the CF's grid will only change by ( 10 * 1000 ) / 20000 or 0.5v
Resistance isn't futile it's V / I.

little eddy
Old Hand
Posts: 625
Joined: Sun Nov 09, 2008 2:06 pm
Location: Manchester

Post by little eddy » Sat Nov 13, 2010 1:40 pm

Sorry I probably didn't explain mu question well. If I have a 2V CD input signal to the 6072a and this stage has a voltage gain of 28.5, I can't possibly expect the grid of the 5687 CF to swing this much. But there again, this is probably far mor gain than I would reasonably need.

If I have a Vg of -10V then in theory the max p-p V input is 20V, correct?
TD-125/RB250/MC25FL & 'Snail' phono, NAS/SBT with CS4398 DAC, 41MP pre & MoFo Power, still messing with OBs.

User avatar
Nick
Site Admin
Posts: 11886
Joined: Sun May 06, 2007 10:20 am
Location: West Yorkshire

Post by Nick » Sat Nov 13, 2010 3:33 pm

Yes, normally you would need no more than unity gain for CD input. I would think as it is though it could swing 30v or so.
Resistance isn't futile it's V / I.

little eddy
Old Hand
Posts: 625
Joined: Sun Nov 09, 2008 2:06 pm
Location: Manchester

Post by little eddy » Sat Nov 13, 2010 5:13 pm

Just looking at the output coupling cap, I think this needs to be much bigger. Using MJ's f-3db recommendation of 2Hz, this should be say 3.9 uF.
TD-125/RB250/MC25FL & 'Snail' phono, NAS/SBT with CS4398 DAC, 41MP pre & MoFo Power, still messing with OBs.

User avatar
Mike H
Amstrad Tower of Power
Posts: 18326
Joined: Sat Oct 04, 2008 5:38 pm
Location: The Fens
Contact:

Post by Mike H » Sun Nov 14, 2010 1:34 pm

Hi Mike,

I agree with that, especially if you wanted to be sure it would work into something less than 50k for example. I have studied this schematic and other sugggestions have occurred, back shortly :D



 
 
I never make the same mistake twice. Instead I prefer to fill my day with a variety of different mistakes

User avatar
Mike H
Amstrad Tower of Power
Posts: 18326
Joined: Sat Oct 04, 2008 5:38 pm
Location: The Fens
Contact:

Post by Mike H » Sun Nov 14, 2010 2:53 pm

OK examination complete :D

Observations:
1. input Miller capacitance ~ ECC83 is high mu and with a 100k anode resistor, say, followed by 1 Meg as a following "load", typically has a Voltage gain of X 53. Say the grid - anode capacitance is 2pF (for the sake of argument), Miller capacitance is 106pF. Two fo them in parallel, 212pF. Add cable capacitance, > 300pF or more. If the cartridge is OK with that, then super but is something to bear in mind!

2. Is the grid bias for V5 45V? Else the 1mA for V4 and the 200V 10mA for V5 don't seem to add up. The 125k anode resistor for V4 will drop 125V, leaving 155V at the anode.

If you think the gain of that stage is too much that resistor could be reduced, but will have to re-bias V4 to compensate for DC Voltage change.



 
Attachments
10-11-11 Snail ECC83, 6072a, 6072a & 5687 CF  Phono & Line Stage Schematic[2].gif
 
I never make the same mistake twice. Instead I prefer to fill my day with a variety of different mistakes

little eddy
Old Hand
Posts: 625
Joined: Sun Nov 09, 2008 2:06 pm
Location: Manchester

Post by little eddy » Sun Nov 14, 2010 10:28 pm

MikeH,

I was doing some reading of my second book by Rozenblit today and the deliberate error was all too apparent during the 2 hours I had to spend watching my daughters swim.

Yes you are correct in that I have been mixed up in terms of an ac or dc-biassed CF. MJ jumps straight into the cathode-biased or ac coupled variant whereas Rozenblit starts with the bog standard dc coupled.

So with me just having won a cheap 12BH7 on slease-bay, I'll consider this with the simplest DC bias first. The previous stage I estimate to have a Vout of 145V so if I want 10mA and the grid can swing between 0 and -16V, my bias should be around -8V and therefore Vk should be 153V??? Assuming I am along the right lines, RL should be 15k3 and say 2W. Cout would have to be around 5u5 for 2Hz f-3dB point.

If I now think of the cathode biassed or ac coupled variant supported by MJ, I must add an input capacitor. If again I=10mA and I want a bias of -8V, Rk1 is 800R. MJ doesn't explain why in his ECC88 example he chose Rk2 of 100k but it is obviously too large in this instance and a 20k 3W would give Vout of 208V. Cout would be around 3u9. Cin ???

So am I on the right lines?

So which of the two CF output stages would be recommended?

I have a couple of questions:

The Zin of the cathode biased/ac-coupled is higher but is this worth needing the additional ac-coupling capacitor? Is the linearity worth considering?

If I go for the cathode biased/ac-coupled option, it looks as if the size of Rk2 is arbitary but I read this should be kept reasonably high (Rosenblit suggests at least 2*Rp), and also it determines what power would be dissipated across the valve and resistor, so guess it is less stressful to have a larger resistor hence my choice of 20k???

In terms of the CF after the phono, I want to keep things as simple as possible and therefore do not intend having a phono or rec out. Also Nick suggested that with this then effectively being close coupled, the CF would not be necessary.

I had not looked at the phono in any detail other than copy the Conrad Johnson circuit but will consider your observations after I've read MJ RIAA section 10 times. I did however see something from him about moving the series cap in the RIAA equalisation to where I had it!

EDIT - I asked Nick about using the ECC803? I think following comments in MJ regarding capacitance. Nick asked me what cartridge so I said MC25FL. His response was that capacitance was therefore not an issue due to the input step-up txs and the ECC83 in parallel would be fine.
Attachments
10-11-11 Snail ECC83, 6072a, 6072a & 12BH7 CF  Phono & Line Stage Schematic.jpg
TD-125/RB250/MC25FL & 'Snail' phono, NAS/SBT with CS4398 DAC, 41MP pre & MoFo Power, still messing with OBs.

User avatar
Nick
Site Admin
Posts: 11886
Joined: Sun May 06, 2007 10:20 am
Location: West Yorkshire

Post by Nick » Mon Nov 15, 2010 11:50 am

EDIT - I asked Nick about using the ECC803? I think following comments in MJ regarding capacitance. Nick asked me what cartridge so I said MC25FL. His response was that capacitance was therefore not an issue due to the input step-up txs and the ECC83 in parallel would be fine.
Dont mind me, I just make it up as I go along anyway :-)
Resistance isn't futile it's V / I.

User avatar
Mike H
Amstrad Tower of Power
Posts: 18326
Joined: Sat Oct 04, 2008 5:38 pm
Location: The Fens
Contact:

Post by Mike H » Mon Nov 15, 2010 1:59 pm

lol :D


little eddy wrote:So which of the two CF output stages would be recommended?
Personally I prefer the DC coupled version, if you're asking me :D as it means fewer signal coupling capacitors. Can always tweak the preceding cathode resistors to get the anode DC you want.

I have a couple of questions:

The Zin of the cathode biased/ac-coupled is higher but is this worth needing the additional ac-coupling capacitor? Is the linearity worth considering?
What should happen is V5 will simply follow the anode of V4 ~ its input impedance is even more negligible if DC coupled, provided a negative DC offset for grid bias is maintained. The 20k resistor seems OK, same as what I'm using in my own pre, also DC coupled CF. Previous anode is 110V or something like IIRC.

In terms of the CF after the phono, I want to keep things as simple as possible and therefore do not intend having a phono or rec out. Also Nick suggested that with this then effectively being close coupled, the CF would not be necessary.
I appreciate that, I was thinking in terms of where you usually want as much gain as you can get out of a phono pre, so it would help if a CF was added so that V3 has no restrictions. Also, wiring from that point (to your selector switch) will be lower impedance so less chance of hum pick up and so on. I have been on this path before :D A duplicate of V5 stage would do.


I did however see something from him about moving the series cap in the RIAA equalisation to where I had it!
OK this is the problem ~ as it took me a long time to discover ~ you have the 39.2k and the 0.008 cap doing the bass boost part, problem then is if you then put a grid leak resistor across it you don't get the full gain you should get at 20 Hz. What's interesting here is the adoption of 2 x 2.2M in series for the grid leak, obviously to minimise this effect. That in itself is a bit worrying cos there's usually a limit on how large a value a grid leak resistor can be as well. Simplest answer is move the grid leak to somewhere else where it isn't going to interfere with the bass part of the EQ, and can be a sensible value, 2 problems solved at a stroke. :D
 
I never make the same mistake twice. Instead I prefer to fill my day with a variety of different mistakes

little eddy
Old Hand
Posts: 625
Joined: Sun Nov 09, 2008 2:06 pm
Location: Manchester

Post by little eddy » Tue Nov 16, 2010 11:07 pm

OK so looking on e-bay for 12ay7 or 6702a, older valves that I usually prefer seem to be rare and/or expensive.

When trawling the net for information, I have seen quite a few posts and schematics that use the ECC82 for the gain and CF stages of pre-amps.

Why would the general guideance on this forum be for the 12ay7 or 6702a? What do they offer that the ECC82 doesn't?

What would be the drawbacks of the ECC82 in either of the gain stages, and/or the CF position of the pre?

The reason I ask is that I have quite a selection of ECC82 from my WAD KT88, including a pair of RCA black plate 5814s and Amperex 12au7 Bugle Boys sat doing nothing. With these being good examples of their type and having them sat doing nothing, might your expert advice change?

They don't seem to be recommended for phono stages except in some instances for a CF output stage, and they are used in the Cornet2 output stage design. Generally though even here they seem not to be the best choice.
TD-125/RB250/MC25FL & 'Snail' phono, NAS/SBT with CS4398 DAC, 41MP pre & MoFo Power, still messing with OBs.

User avatar
Nick
Site Admin
Posts: 11886
Joined: Sun May 06, 2007 10:20 am
Location: West Yorkshire

Post by Nick » Wed Nov 17, 2010 8:52 am

Lack of gain is the issue with the 82
Resistance isn't futile it's V / I.

User avatar
Mike H
Amstrad Tower of Power
Posts: 18326
Joined: Sat Oct 04, 2008 5:38 pm
Location: The Fens
Contact:

Post by Mike H » Wed Nov 17, 2010 3:03 pm

Nick wrote:Lack of gain is the issue with the 82

For the phono amplifier stages certainly, although is one of the first things I would think of for the line pre-amp part. After the selector / vol control. Do make useful CF's as well.


There is of course Richard Brice's phono pre from Wireless World 1985, it uses a cascoded ECC82 for the first stage. Only recently did the penny drop as to why this might be a good idea.



 
 
I never make the same mistake twice. Instead I prefer to fill my day with a variety of different mistakes

little eddy
Old Hand
Posts: 625
Joined: Sun Nov 09, 2008 2:06 pm
Location: Manchester

Post by little eddy » Wed Nov 17, 2010 11:01 pm

Having done some more surfing on cathode followers, there seems to be quite a few comentators who suggest that sonically the basic dc-coupled CF is not ideal.

Obviously with my basic design, both pre stages equire an elevated heater supply but why would a split rail power supply help otherwise, perhaps in terms of its audio quality as per the attachment?

If the split rail does offers some advantages, might it be possible to adopt this over the whole phono/pre design?

I think falling asleep reading MJ is on the cards again.
Attachments
split rail 5687 buffer.gif
split rail 5687 buffer.gif (5.52 KiB) Viewed 9959 times
TD-125/RB250/MC25FL & 'Snail' phono, NAS/SBT with CS4398 DAC, 41MP pre & MoFo Power, still messing with OBs.

User avatar
Nick
Site Admin
Posts: 11886
Joined: Sun May 06, 2007 10:20 am
Location: West Yorkshire

Post by Nick » Wed Nov 17, 2010 11:49 pm

Main things is it

1. Avoids the input cap
2. Allows a higher valve load resistor

Its very easy to ignore the load resistor and just pick what fits the voltage, but if you draw the load line you find that its just as important as if its in the anode. using a active load (CCS of whatever sort) sounds vastly better to my ears.
Resistance isn't futile it's V / I.

Post Reply