SRPP with buffer - 1MOhm impedance?

[revintage]

The refs should be decoupled to ground with caps, typical 22u/0,1u(plastic). Connect the LL1690 with its windings in series. Connect + and - directly to the outer ends of the primary, no coupling caps. The termination should probably be between 5 and 10kohm, depends on secondary load.
Check the Twisted Pear Opus schematic.

 

[Bertel]

Yes, I have the REFs decoupled to GND as you say, with 22uF and 0.1uF, but the latter being multilayer ceramic caps X7R 50V, not plastic, would you say this is ok?

I still don't understand it: So you would not use the REFs from the DAC? Don't they kind of set the mid-rail "zero" reference and therefore are important as, well, the reference onto which + and - relate (not to GND directly)...? Why is this approach with the LL1690 better...?

Sorry for my ignorance, and thanks for your help... ;-)

 

[revintage]

Yes, I have the REFs decoupled to GND as you say, with 22uF and 0.1uF, but the latter being multilayer ceramic caps X7R 50V, not plastic, would you say this is ok?

OK

I still don't understand it: So you would not use the REFs from the DAC? Don't they kind of set the mid-rail "zero" reference and therefore are important as, well, the reference onto which + and - relate (not to GND directly)...?

The DAC has two signal outputs, + and - , use them. Don´t know if you even has to AC ground the REF,s when using a transformer. DACs and sand isn´t my speciality so maybe you should consult the datasheet. There are also quite a few treads here at diyaudio about it.

Why is this approach with the LL1690 better...?

About LL1690, its a recommendation and a very straightforward way of going balanced to unbalanced, you must decide yourself. You could also go 2:1 by parallelling the output windings. Gives you four times lower Zout and still 1Vrms.

Maybe we have lost focus here. Lets go back to the tubedriven phone-amp.

 

[Bertel]

Lars,

many thanks, I've got it now - Wolfson themselves in their data sheet recommend to decouple the REFs to GND, and I'll float the following amp stage for galvanic isolation. (They btw only mention "differential analog voltage outputs" but nowhere talk of "balanced", it isn't...)

Well, Benny now has recommended the LL1676 as his favorite over the LL1690... ;-)

Maybe we have lost focus here. Lets go back to the tubedriven phone-amp.

You're right Now I think the setup is clear from DAC chip all the way to transformer secondaries. The following amp stage, well, probably will just be the straightforward one-triode amp as you have suggested, plus the appropriate output transformer right at the anode. Benny has recommended the LL1630 for this, I have also seen the 1660...

 

[revintage]

Yes, LL1676 is Bennys favourite as well as mine. But then you should go for 2:1 to get it right. Not a problem as I see it, as 1Vrms normally is enough. Note the transformer should be connected 2+2:1+1 (2:1) or 2+2:1//1 (4:1).

You shouldn´t parallell the primaries 2//2:1+1 to get 1:1 according to Per Lundahl. But you can get away with it in a lowZ application like this. Also 1+1:2//2 is then possible.

LL1630 would probably be the appropriate due to its lower secondary RDC. If using the 20mA you should use one parallelled E182CC/channel due to the low inductance.

 

[Bertel]

Amp stage

Ok, it's been decided then - LL1676 connected as 2:1 !

Ah, I didn't see your edit - I initially intended simply to go like this (see schematic).

Values are:
R1: 100k
R2: 50-100k
C1: 220uF
C2: 10uF
C3: 1uF
C4: 220nF
C5: 0.7nF

But to your edit: Is one parallelled E182CC per channel anyway better than "one half" per channel = one E182CC for stereo, or is this just for the LL1630 and we should try to find an alternative for it?

 

[ErikdeBest]

Bertel

I know it is off-topic, but want to ask anyway. Are you using the UE-10 PRO for hi-fi at home? Have you compared it to 'traditional' headphones (stax, sennheiser, etc)? I want a real good headphone and read comments that the UE-10 PRO beats these 'traditional headphones'... but I am a little worried about comfort if used for long listening sessions in a home setup.

Thanks for your comments! Erik

 

[Bertel]

UE-10pro

Hi Erik,

thanks for asking! Yes, I had many of them - Sennheisers (HD650 etc.), Beyerdynamics (DT880, DT990), AKG (K701, K240), Grado (SRs and RSs), though no Stax. I was shocked at the beginning how "bad" the UEs sounded - flat, dull, compressed, uninvolving - until I found out one step after the other that this was due to (i) my listening experience from before, (ii) flaws in my audio chain that went unrecognized before, and (iii) limitations in the audio material which really is revealed in a clear and often brutal manner. But this of course is also true vice versa for excellent material which truly comes through fantastic.

I love my UEs for what I consider as their transparent, accurate and extremly neutral sonic reproduction. There are certainly many headphones out there that play more "musical" and involving, but when it comes to neutrality and accuracy, hearing every bit and detail exactly as it is on the recording etc., IMHO there is no better choice.

Having those "plugs in the ears" was a little strange at first, but I quickly got used to this (they're custom molded so the fit is perfect), and i now listen for several hours without even noticing them.

Please feel free to email me directly if you wish to discuss this further, I'll be pleased to help with my experience and heavily biased perception of how good they are

Best,
Robert

 

[revintage]

This is an idea of how it could be done. Have already recommended a substitute: triode-strapped 6E5P. But if you have the 182´s, why not use them?

 

[Bertel]

Many thanks Lars - that looks good!!

Yes sure, I love going with the E182CCs, just thought you maybe had suggested the parallelled E182CCs just because we had focussed on the LL1630, while there might be a better tubes for this recommended solution if we only chose a different output transformer. But that does not seem to be the case, right?

A few questions to the design please:

- You omitted the snubber to the grid at R2, I had thought it would be good for resonance elimination?

- Only 120 ohm for R1? Is that enough...? One usually finds values in the 50k-10k range there...

- What would you recommend for C2? 10uF? C3 and C4 are not good?

- And finally: Do I have to put a resistor across the headphone output...? I am afraid it does harm to the sound in some ways...

Thanks again
Robert

 

[Bertel]

New thread on amp

Hi Lars,

since we have long moved away from the initial 'SRPP with buffer' topic, I have moved/created a new thread here: http://www.diyaudio.com/forums/tube...2cc-lundahl-transformers-8ohm-headphones.html

I hope you don't mind - and I'd appreciate if you could find some time to look into my questions above and help me understand, that would really be very helpful for me.

Many thanks!
Robert