First phono needs QC

[bobhayes]

Hello everyone,

This is my first attempt at designing a phono stage from scratch, so it might have a few teething problems.
I was rather hoping that some of the more experienced members might cast an eye over the schematic for me and point out any problems, and/or improvements.
I have many of the parts required to build this circuit, hence the tube line up, so I probably won't change them!


Thanks to all respondants in advance.

Bob.

[SY]

12AX7 and 12AU7 are not the best choices for linearity or noise, especially with a low plate load. To boot, the input tube will have a stunningly high input capacitance. Gain is also likely to be marginal here- figure the first stage is something like A = 40, the second stage is about A = 12, for a total A(DC) of 480. Now knock off 20dB, and A(1kHz) = 48. For a 5mV nominal output cartridge, that's only 250mV output from the preamp; ideally, you'd like 5 times more output. A 12AT7 as the second tube will get you better linearity and about 3 times more gain.

I haven't rechecked the EQ components- you did annotate the source resistance of the input tube and the Miller capacitance of the second tube, but I'd like to know where the 21k came from. I get a significantly higher number because of the cathode degeneration.

[bobhayes]

Thanks for the expert input Sy.

I forgot to draw the bypass caps onto the legible redraw! There should be 100uF bypasses across both 910 R's. That should add gain to the input, and lower the output resistance somewhat. Sorry for the misinformation.
I calculate 21k by:
Plate resistance of 60k deduced from tangent drawn on curve (from data sheet), at theoretical operating point. Maybe not perfectly accurate, but close enough for first sketch up.
RL in parallel with Rp's (68k//60k//60k).

I know other tubes would be more suited to the application, but I'd like to try this out at little cost, hence the compromises. I can try more suitable tubes later, if this one at least works! It's just a first step on a long road.

In my current phono, I use 12ax7 input, and I don't find the noise too intrusive. The crackle 'n' pop disguise it. (record cleaner is on the cards)
The purpose of the parallel input was to reduce noise.
I suspect my Grado cartridge will benefit from the increased input capacitance.

Do you think I should increase the load (instead of 68K) to increase gain?
I can do the same on both common cathode stages, and re calculate the other values to spec.
Alternatively I could make the final stage common cathode? I don't need a lot of drive current.
One final thought was to use the 12AU7 for input, and the 12AX7 for gain and output. Would this benefit linearity?

If you see no reason to waste your brain power, and my time on this, I will have to bite the bullet and blow the cobwebs off my wallet, in which case I might as well use 6922's, or such.

Thanks Sy.

[bobhayes]

Another bypass cap (220u) should also be on the second stage. Should find the gain is around 14.5
It's late here! zzzz

[kevinkr]

I designed and built similar phono stages early in my audio career. It will work, but you definitely need all of the gain you can get out of that first stage - I'd increase the plate load resistance to at least 88.7K and recalculate the eq components.

I would either substitute a higher mu tube for the 12AU7A or consider a mu-follower for the second stage instead. Gain will be close the theoretical mu, and output impedance should be under 1K with a 12AU7A - note that you need to run current through this tube to get decent linearity, say something >5mA. Some 12AU7A have significantly better linearity than the ones Morgan Jones wrote about, but most are not as good as other types - at these signal levels I doubt it is going to be a big issue. (I found Mazda, Mullard and Dutch or German Amperex to be exceptionally good sounding compared to anything else - they also measure better too. JJ's do sound ok, most vintage US made ones seem to have poor linearity, and sound mediocre to awful.)

I find that the minimum useful gain for a phono stage is typically about 40dB @ 1kHz so this implies at least 60dB (1000) of gain prior to the application of eq.

I used up to 4 12AX7A/5751 sections in parallel in early designs for noise reduction, and independent cathode bias for each one is absolutely the right thing to do. My cartridges at the time were all moving coil types and not too sensitive to capacitive loading - a possible issue with some induced iron and mm types. These days I use really high transconductance tubes like a triode connected D3A for this task.

[SY]

What Kevin said.

You only get a 3dB decrease in noise by paralleling. Check and see what the Grado wants for capacitive loading- the two 12AX7 could represent 300pF before you've even added the cable.

[zarniwoop]

Originally posted by SY
What Kevin said.

You only get a 3dB decrease in noise by paralleling. Check and see what the Grado wants for capacitive loading- the two 12AX7 could represent 300pF before you've even added the cable.

I thought that carts. usually even LIKE some load capacitance. I've seen people adding a cap. at the input even.

[SY]

They do, but the right amount, not too much. You might investigate what your cartridge needs, then see if that extra input tube is a help or detriment.

[Giaime]

I suggest (other than the great help you have here) adding the fourth time constant Very good results.

Just parallel that 191k resistor (or you could also use the 56k resistor, series resistors in the RIAA correction) with a 50pF-100pF capacitor. The right value could be calculated with much hassle: SPICE gives great help here

http://palimpsest.stanford.edu/byfor.../msg00424.html

[bobhayes]

Thank you all for the great information.

I can't find what the specified input capacitance should be for my grado prestige black cartridge. All I can find is a passage mentioning that the cartridge inductance is low, and it is relatively immune to cable and input capacitance (compared presumably to other similar cartridges).
So the general concensus is that the input capacitance is too high for most cartridges, the gain is too low, and the 12AU7 should be popped in the rubbbish where it belongs?
I will investigate the fourth time constant, but I'll need to do some more reading first.
It looks like it may well be back to the drawing board for this one then, with different tubes. I know this might turn the thread into a battle ground, but does anyone have recommendations for reasonably priced (common) tubes for phono applications? I would at least like to use the 9 pin bases. How about 6922/6DJ8 etc? They aren't too dear.

Thanks again for all the input.