It's time I built a D3a -> 5687 phonostage

[Thermionic Idler]

I have a hankering to try a good valve phonostage again, and having read through several threads here, and on Hi-Fi Haven, we seem to quite regularly come back to the D3a for stage 1, and more often than not a 5687 for the second stage, to get really good results. Yesterday I impulse bought a couple of D3a's and two 5687's to start my 'stash'.

It'll likely be some time before this progresses beyond 'theory' as I've committed to take (and to be honest, I need) a few months off once the Quad-inspired linestage is done (and that should happen this weekend), and when I restart, I ought to tackle the Valve DAC project as that was first on the list. I'm also not giving up on the Millett stage just yet, I'm sure I'll get to the bottom of those response issues, but having already invested in the Lundahl step-ups, it would be churlish not to try something with valves

So at this stage it's a matter of identifying a good circuit to base the design on. The one that's attracted me the most at this stage is Stephen's (izzy-wizzy) current circuit. Clearly a LOT of research and development has gone into this, and the differential output stage and OPT will let me easily balance the output to the rest of my balanced system. We also have Mark's (islandpink), and I remember Thomas's knocking my Whest into a cocked hat at an Owston a few years ago too.

So... yeah. Question to the floor - what do we like when using a D3a / 5687 to amplify a rock scraping down a plastic ditch?

[IslandPink]

Sounds good ... bit of a phono season coming at atm !
Have you already worked out the gain you have, or need, is that why you're suggesting 5687 for the second stage ?
Mine has lots of gain, but the output is just the anode Zout of a D3a, hence about 2 to 2.5k , so i have to be careful to use short interconnects and 100k + 200k inputs on the two amps it feeds.
If I did a re-build it would most likely have a second gain valve and a CF ( maybe 5687 ) on the output.
Has Stephen given the gain of his latest circuit ?

[izzy wizzy]

The hifihaven cct started as an earlier version of mine. I've conversed a lot with Salectric over the last decade or more helping with his deviation. He's still going so his latest is different to the build threads but I think he may have done a build thread for his latest.

Many use the d3a as a triode. Those that do usually include the so called 50k pole. I use it as a pentode and don't use that pole. My feeling is as a triode, it's a bit dark and maybe the 50k pole lightens it. Or it could be just a taste thing. I've gone back and forth and always prefer the pentode. As a triode it has a lot of input capacitance which might be more reactive with a transformer stepup.

Nick pointed out that the diff 2nd stage does a lot of heavy lifting. Many would put a ccs in the tail. I put a pimm design in once and could not really hear the difference. Maybe coz the signal is quite small. Salectric has a SET out stage now. I have the iron to do that but haven't tried it.

Mine has modelled gain of 48dB.

[Thermionic Idler]

I've had to go back to the Whest stage for the moment whilst I try and get the LR stage fixed.

Whilst I can hear its full bandwidth, I immediately missed what the LR stage was doing in the midrange (as heard at Owston) despite its response anomalies. The Whest just isn't as good. I don't know how much of this has to do with the active first stage versus really good MC stepups, or the LR capacitor-less eq - I might try changing the loading on the Whest to see what happens as I think it's 470R at the moment versus the 150R of the Millett.

All of that in the context of this thread... maybe I should go for broke and build a version of the Stephie Bench LR phono... now that would be a project and a half...

[Thermionic Idler]

Right then. I've decided to be ambitious.

So Sowter make the 1280 RIAA 600 Ohm EQ Unit - an off-the-shelf RIAA EQ similar to the classic Tango units.

I'm keen to design a circuit around that. Ha! as if I know anything about phono stage design. But I really like what the Pete Millett LR does, so I want to stick with induction based RIAA, but try to implement it with valves.

Seems to me that one obvious way to get around the low impedance input to the RIAA is to use a step-down interstage in front of it. Though if I were to go for a hybrid solution, an op-amp may be much cheaper. MC step-ups are selectable 1:16 and 1:32, the latter to eventually be used with Don Harper's DH3. Probably don't need anything more than 40 dB overall gain (finger in the air figure after several G&T's).

The Sowter unit is reasonably affordable and means I don't have to think about the RIAA itself, only what sits either side of it, making sure the impedances are matched.

Time to get my thinking cap on.

[Nick]

[Ant]

I thought you had cinemag lcr units in the one you built, could you not take those out instead of shelling out for the sowters?

The s&b ones in my cole are 600r ones iirc, are there different impedance ones available? I always wondered why 600r which seems a little bit of a strange value, but i know nothing

[Thermionic Idler]

I thought you had cinemag lcr units in the one you built, could you not take those out instead of shelling out for the sowters?

The s&b ones in my cole are 600r ones iirc, are there different impedance ones available? I always wondered why 600r which seems a little bit of a strange value, but i know nothing

They are standalone inductors rather than complete LCR 'black box' units - plus getting them off the PCB would be tricky!

600R seems to have been settled on as a standard value, I think a lot of pro gear terminates at that impedance as well (?). It occurred to me that a valve really isn't the best choice of device to drive a load like that, so I'm currently buried in Spice modelling an op-amp instrumentation amplifier for the first stage into the LCR, based on the linestage circuit I just completed. The idea is to put the valves after the LCR to make a kind of hybrid affair.

[Ant]

I looked at the same hybrid idea when the cole broke for the second time and i looked at the inductor spec. 600r does seem abit high for the valves in the cole to be driving.
I came to the conclusion that i was well out of my depth so left the idea
Fwiw i was looking at texas instruments lme49720

[IslandPink]

Seems to me that one obvious way to get around the low impedance input to the RIAA is to use a step-down interstage in front of it. Though if I were to go for a hybrid solution, an op-amp may be much cheaper.

Have you considered Nick's method and rejected it ?
I'm just a bit concerned about transparency for either of the methods you're suggesting.

[Thermionic Idler]

Seems to me that one obvious way to get around the low impedance input to the RIAA is to use a step-down interstage in front of it. Though if I were to go for a hybrid solution, an op-amp may be much cheaper.

Have you considered Nick's method and rejected it ?
I'm just a bit concerned about transparency for either of the methods you're suggesting.

Ahh... is that cathode follower with feedback? (as mentioned in the C3g thread). I haven't really rejected anything at this stage, just exercising various ideas.

In terms of transparency, I basically can't hear the instrumentation amplifier first stage of my linestage, so it seemed sensible to consider something similar for the first stage of the phono (configured to provide the necessary gain).

Of course whether that quality ports over to a phono stage where the voltages are a lot lower and the signal balance is pre-RIAA is still under question, but these things have vanishingly low distortion and huge bandwidth. I'll be using MC stepups so the signal will be in the millivolts rather than microvolts. But, yep, CF with feedback may still be the superior solution.

[IslandPink]

Nick will be able to clarify, i think it's a 5687 CF working over a mosfet as the load, maybe the feedback is working with the previous Pentode mu-stage, to help stabilise it and reduce ageing drift. Nick, is it the top of the mosfet that feeds back to G2 on the pentode ?
Step-down transformers are adding another tricky ( expensive ) part into the signal chain, and it's already tough to get MC step-ups that are really transparent, although they are to be fair working at really low signal levels.
Nick's phono uses either a 1.5k or 2k LCR , so maybe something even lower Zout than a 5687 might be required for 600... ?

[Nick]

Yes. The circuit is somewhere around. And the LCR is 1k5. The 5687 will probably manage 600R.

[Thermionic Idler]

This was what I had in mind for stage 1:



To the right of the RIAA, probably an interstage?

Will hunt around for Nick’s arrangement too.

[Thermionic Idler]

Found it: Link

Wow... that's interesting...

Yeah, that looks DIY-feasible. How hard can it be?

Time to knock up a Spice sim...