![]() ![]() |
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
Click below for the schematic of my latest choices of parts for my phono stage. Power supply is not included, I get it from my line stage. I can't think of a better way to make this phono stage from the point of maximum transparency, except maybe to change the output tube to a 417A or other low Rp tube with high gain just to drive the autoformer step attenuator better.The tubes are some high frequency low noise high gain lowish Rp single triodes with exceptional detail and almost no sonic signature in this level of signal. The 6ER5 is a fine frame grid triode of that type of tube, and so is the 6GK5, a cousin to it. They have the same pinouts but slightly offsetting sonic signature. If I use two of the same kind I get a washed out sound as similar even order harmonics cancel out.
Most of the resistors are very transparent Caddocks, and not quite as sterile clean as Vishays. Tantalums would be nice I bet. The caps are as much teflon as I could put them in, with one 0.0033uF being too small to get in teflon, and so it's polystyrene. The equalization takes into account the little 3.18 usec time constant. The simulated error in RIAA taking this high time constant into account and neglecting the RC rolloff of the load of the line stage, can be seen at
http://www.sonic.net/~ktstrain/phonoeq.jpgThe power supply splits into the "C4S" or equivalent active constant current source loads with filtering to each CCS. This is necessary to stop small residual ringings that somehow get through to the power supply. Some opportunity to improve on the CCS is there.
If there was a way to avoid the big 249K series resistor between tubes this could sound better, but I have no practical means yet to do that. Something in series has to be there I believe unless you use a lot of transformers and inductors or whatever.
The EAR MC-3 is a very good MC step-up transformer and if you're going to go all out, why not? This adds about $1000 to the bill. The phono preamp after box and ground plane proto board and wiring and parts, etc. costs about $700 to make, plus power supply.
I think this phono preamp can beat most others on the market, but I haven't heard them all. This does beat the following ones I have heard:
Audible Illusions MOD3 and MOD3A with and without gold phono board.
Herron's first preamp offering (forgot that model)
Air Tight ATE-2
EAR 834PThe sound is characterized as very neutral in tone, great dynamics, great microdynamic shadings, and details, details, details, and very low noise. It doesn't impart much on the music at all. For character I use my line stage and driver stage of my amp to get into the finer tones.
Compared to the mentioned brands of phono stages, this one is not as dry as the Herron with better body and tone and seems about equal in detail and probably more so, not as dark as the Air Tight and more detailed, not as bright and sterile as the EAR, and definitely all around better than the Audible Illusions.
So there's some other data points to consider in a DIY phono stage.
Kurt
Follow Ups:
This phono stage design has gone through many iterations and I am very thankful for the top expert help I have received. Good designs are simple, but the details of the design and its implementation took a lot of good people giving inputs. Thanks to Doc B and the folks at Bottlehead for publishing "Groove Thang" which started it all, with elements that topped my more conventional approach. Thanks to DrM for taking it a few steps forward and introducing me to the higher quality sound of unbypassed cathode resistors. Thanks to Dave Slagle for advice on the autotransformer step attenutator that this phono stage has to drive. Of course a big thanks to Thorsten for the tips that really sent this thing cookin': the idea for the transformer attenuator, grid leak bias, 3.18 usec time constant, biased up RIAA capacitors, and whatever else I may add to it later. Wow! I got my money's worth on this board! I am an extremely happy camper.Kurt
Kurt,Wouldn't those 10M and 22M resistors be extremely noisy? Grid leak bias works, but because the leak current is unreliable, the operating point varies with tube sample more so than cathode bias.
You might consider battery bias or just plain bypassed cathode as viable alternatives.
-Richard
Hi,> Wouldn't those 10M and 22M resistors be extremely noisy?
Of course they are terribly noise. But the Impedance in the Grid Circuit is dominated by the RIAA Network, excepting VERY low frequencies, so these resistors are not what makes the circuit tick.
> Grid leak bias works, but because the leak current is unreliable,
> the operating point varies with tube sample more so than cathode
> bias.I have heard similar statements before and found them to be rather untrue in reality. In fact when servicing gear that used gridleak bias I was surprised how well the bias was held over a range of different valve brands.
> You might consider battery bias or just plain bypassed cathode
> as viable alternativesGrid battery bias is an alternative, but it requires rather low resistance grid leak resistors and as a result causes more losses in the RIAA network. The bypassed Cathode Resistor forces you to use a substandard quality electrolytic bypass capacitor in a rather critical position. There are applications where gridleak bias does not work that well, but they tend to happen at fairly high levels. In the second stage of a phonostage it is a really smart idea.
You should try it on your 6DJ8 Phonostage one day, you maybe surprised.... And while you are at it, add the 3.18uS 4th Timeconstant.
Ciao T
Are you using actual C4Ss as current sources or something "equivalent", and if the latter is the case, can you tell us what you ARE using? Thanks.
The only difference between my CCS and the C4S is I chose a 5.6V zener diode in parallel with a 10uF tantalum cap where the top LED of the C4S is. The only difference is that this gives a little more current feedback because the emitter's current setting resistor needs to be a larger value because the zener is 5.6V instead of about 2V for an LED. You need the tantalum cap to filter off the zener noise.Other options are Gary Pimm's pentode CCS if it's quiet enough, or my PMOS cascoded CCS, or whatever else there is that's the latest and greatest. But my modified "C4S" works great in these low level circuits from what I can tell, but I have not heard any other. It's all soldered together on my single proto board and not readily changeable. I will leave it this way probably for a long time.
Kurt
Kurt,Use either Grid or full Grid leak bias on the second valve. The 750 Ohm resistor alone is reflected up to the anode as 750 * Mu or theoretically 60k.... Bridge out (for fun) the cathode resistor on the output and use a 20M Gridleak resistor, plus a 2M7 before the coupling Cap (or just re-adjust the RIAA).
Other than that and the 417A it looks very interresting.... I almost made something similar myseld, but decided to save my shots for a fully LR/LRC equalised WE 396 & we417A phonostage, so I'm just building a little "drilled up" version of the EAR 834P.
BTW, if you drop all that cash on Teflon Cap's, do try a pair of S&B 103's for around $ 200 the pair. In my setup you have to wipe the EAR MC-3's off the wall in small chunks after a "shootout", but that just me and WTFDIK....
Ciao T
Ciao T
I forgot that I took out the bypass caps! That changes the model completely. Now I have to go back to the drawing board. Thanks for the heads up!
I had to make some further changes after being pointed out my mistake about missing the cathode resistors in the model. The links shown in the original posting are now correct.I changed the grid resistor on V2 to 22M and its now completely grid leak biased. Damn, what a difference. I thought it was just awesome and now I have awesome some more. The bass delves deep down to the bottom. It also appears from new simulations that Dave's autoformer step attenuator has about 125H of inductance if it were to work as well as it did with 60K of output impedance driving it! Now I have it back to 6K output impedance and shakes the foundations. In fact it still goes down more than I like, with me now hearing LF noise from the turntable again. The grid leak is a peculiar beast. It's the same plate voltage whether I use 16M or 22M as the grid resistor, about 70-90V depending on tube. I tried upping the CCS to 5mA and it made the sound go dead! Go figure. There's some shunting diode-like element in the grid model or something like that.
Should I go to 417A or 6S45P for the output stage? It doesn't sound like I need more drive, but who knows.
Okay, Thorsten, now what do you think? I'm going to have to sit before thinking about a new step-up transformer or tubes. I'm out of spending money for now.
Kurt
Hi,> Okay, Thorsten, now what do you think?
I have a surprisingly similar drawing (and associated models) in my P-Spice, only difference is the 417A in the output.
> I'm going to have to sit before thinking about a new step-up
> transformer or tubes. I'm out of spending money for now.I know the feeling.... Very well. I just bought another 50pcs 2nd hand records. So only enough cash left to make a EAR 834P copy for now to keep me playing those LP's I bought. No fancy phonostages for me for a while.
Ciao T
This post is made possible by the generous support of people like you and our sponsors: