Unholy Alliance Phono Amp

[Sheldon]

Combining glass, sand and iron (I'll be using an MC step up), and various ideas picked up here. The basic schematic is from Dmitry (don't know if he posts here, but has some good stuff on his site): www.dmitrynizh.com/3a5-phono.htm.

The HV supplies are based on Salas' shunt reg., and the transistors in the filament supply are those recommended by Rod Coleman. Gyrator from discussions among Wavebourn, Michael Koster, Kenpeter, Revintage, Smoking amp, etc..

I still have to build the case work, so I haven't given it a listen. As shown, the response is within 0.3dB from 20Hz to 20kHz. I can tweak the network resistor values a bit, to get it even closer Channel to channel it's within 0.1dB over the entire range, if I match the tubes. (BTW, the Hagerman calculator seems off to me. I actually measured the Rp of the tubes in the circuit (which were very close to the values measured from the curves), so my calculations there are accurate, and I was a little suprised that my initial measurements were off at the 20kHz end by 0.75dB. When I calculated the time constants from the Hagtec calculator values, they were lower than the expected 75 and 3180us.)

The circuit values are as tested.

Sheldon

[Sheldon]

Pics. Note that I have installed only one of the 10mF filament filtering caps. I'll see if they are necessary. The raw filament supply has about 6mV of ripple. The CCS should give me at least 80dB PSRR. We'll see.

[Sheldon]

Bigger picture of the amp bottom. This method is compact, but a time consuming to solder up.

[Sheldon]

Oops! Raw filament supply should look like this:

[Sheldon]

Thought occurs that I maybe should have started this tread in the analog source section. At any rate, I adjusted the RIAA resistor values. I now get within 0.2dB, from 12Hz to 100kHz. That takes away the 50kHz compensation, but I'm not so sure it's needed. I can't hear the difference with or without it, in my other phono amp (but, given my hearing, that can't be assumed to be the universal experience).

The series resistor for the first stage was increased to 46.5k and the shunt resistor was changed from 2k to 1k. The second stage series resistor was increased to 39k and the shunt resistor from 4.7k to 5k.

[Sheldon]

Final schematic, with actual common connections shown. Almost a star, except that it was a little easier to connect the RIAA compensation shunts and the 2nd and 3rd stage grid leak resistors to the filament buss. It's a 16gauge copper wire, and the final segment is short, so it shouldn't cause problems.

Lashed it up and plugged it into the system to check for subjective noise performance. Haven't scoped it as I tripped over my scope probe and broke it a few days ago. Noise is all white noise and similar in level to my other pre (modification of Steve Bench design). At any rate, well below record surface noise. The simple CCS for the filaments works well. I compared it with battery power and could detect no difference in noise level. Confirmed (no shock here) that mechanical isolation will be necessary to manage microphonics. For now, a thick strip of felt for it to sit on tames it.

Haven't yet listened to music. Will do that when I get the case for the preamp section done, and can mount the step up trannies. Will also try a head amp, but need a few parts.

I haven't prompted any response, so must not be anything controversial here.

Sheldon

[SY]

OK, I'll be controversial. Why only 35V on the first stage plate? What's the second harmonic distortion look like?

[Sheldon]

Originally Posted by SY

OK, I'll be controversial. Why only 35V on the first stage plate? What's the second harmonic distortion look like?

Nice to see you here, thanks for the comments.

I just followed Dmitri's recipe for operating points.

The voltage on the first stage is a consequence of the filament biasing used. Haven't measured the harmonic distortion (fried the input on my external sound card a while ago - might be a trend starting on broken test gear), but looking at the plate curves suggests that I'm on up on a reasonably linear part of the 0V grid curve, and I'm not swinging much. It's a little odd interpreting this, as one side of the filament is at 1.3V (slightly starved) and common side at 0V. But I assume the grid curves are derived as referenced to common.

I can certainly play with it, as it's a simple matter to adjust the B+ with a trim pot. That will change the third stage plate voltage as well, as the gyrator reference is the first stage B+. But, higher plate voltage there should be no problem. Or, I can put a pot before the bias connection and trim it down.

Sheldon

[Sheldon]

Everything seems to be checking out fine, with one exception - so far. I get a significant turn on bump. I checked DC at the output and it seems to spike almost to the full B+, swing down, then rise gradually as the bias cap charges.

The B+ itself, rises quickly - within a second or so - but without overshoot.

Looking at the schematic, I'm stumped. The only thing I can think of is the safety diode from the gate of the SK170 to the B+ for the gyrator. Does it have enough capacitance to send the gate of the 170 momentarily to B+? Or does it have something to do with the gate capacitance of the 10m45?

Sheldon

[SemperFi]

Interesting application of DHTs. I have a bunch of 3A5s since I had a plan to use them in guitar effects pedals. Turns out they are too microphonic in that application. Rubber mounting the tubes and having silicone rings around the tubes helpes of course, but when a guitar is at full that still wasn't enough.

I look forward to your final conclusion regarding a phono pre with these tubes. I wouldn't mind putting mine to good use.

If the tubes are cold on turn-on the CCS would surely drive the output to b+. My guess.