On Tue, 13 May 2008 03:58:36 +0100, "Ian Iveson"
<IanIveson.home@[EMAIL PROTECTED]
> wrote:
>flipper wrote
>
>> Discussion about the headphone interface on the 6AW8 PC
>> Speaker amp
>> got me to thinking so I whipped up a little OTL circuit,
>> just for
>> headphones, using the 6CY7 (essentially a 'half power'
>> 6EM7)..
>>
>> Why? Well, when researching the series resistor thing I
>> noticed a lot
>> of the so called 'tube' headphone amps have solid state
>> outputs and a
>> lot of the all tube ones couldn't drive 32 ohm headphones
>> very well..
>> The 6CY7 should take care of both those issues, and with
>> the IEC
>> specified 120 ohm source impedance
>>
>> Schematic and pics of the breadboard (I only did one
>> channel) are up
>> on my site http://flipperhome.dyndns.org
>>
>> It's not really fleshed out more than needed to breadboard
>> and see the
>> circuit work, but that it did.
>>
>> There's plenty of room for tweaking it. Spice indicated
>> excellent open
>> loop performance, in case that's what floats your boat,
>> but I wrapped
>> 13dB around it anyway for the exercise. And there's room
>> for a bit
>> more gain since, for the sake of simplicity, I didn't
>> bypass the gain
>> triode Rk.
>>
>> I figured some might get a few laughs at my sloppy
>> breadboard so I
>> included a pic of it for your amusement.
>
>Exactly the same as my breadboard, except you use screws
>where I use solder.
You can tell from the pic I end up with a mix, especially with a
circuit like the headphone amp, and that 'two tube' board, where it's
not 'just' a pair of PP tubes (like 6BQ5s) into an OPT.
>I started a thread a while ago about valve headphone amps,
>and I don't think any different now, except thanks for the
>detail of the IEC standard (regulation, guideline...what are
>these things and what are they for?).
Well, an interconnect standard, for example, is so things
interconnect.
A dramatic example is Apollo 13. One 'simple' solution to the oxygen
problem was to swap control valves with the working unit in the Lander
but they were built by two different manufacturers and the threads
were different.
If there had been a 'standard' defined then it would have worked.
In the case of headphones, it's so if you buy a 'standard' headphone
and have a 'standard' amplifier then they should work together. At
least within the parameters defined in the standard.
>For what it's worth, my current design aim is for an
>open-loop output impedance of 5 ohms, with feedback to
>reduce that to a couple of ohms at most.
>
>At the time of that thread I was looking into the consequent
>desire to limit power output for safety reasons, and
>wondered how.
As I mentioned in the PC Speaker thread, the 120 ohm source impedance
helps that issue, over differing impedance headphones anyway. By that
I mean, if you set the volume for 1V out, driving 300 and 32 ohm
headphones directly, with 'low impedance' results in an almost 10dB Po
difference but (increasing Vo to compensate) it's less than 1dB with
120 ohm source impedance.
> I will experiment once I've built the amp,
>which will be some time after I've found some money to pay
>for the transformers.
How are you planning to solve the wide range of impedances presented
by differing headphones?
>I think you will find, if you look at decent high fidelity
>headphone amp designs, and transformers intended for decent
>valved examples, they achieve an impedance ratio of 1:10 or
>lower.
I did look and found just about 'everything' imaginable, as well as
units that meet IEC specifications. Of course, those have 120 ohm
source impedance.
> One problem of traditional valved designs is that
>they are intended for phones of 70 ohms or more.
I don't know. Before I finally found one or two with one I had thought
the 'norm' was *no* headphone output on valve amps.
> 32 ohms is
>increasingly the norm, it seems.
FWIW, my circa 1970 Sansui headphones are 8 ohm. My new ones are 32
ohms.
> I can't see how driving a
>non-linear voltage-controlled device of 32 ohms impedance
>with a 120+ ohm source is ever going to be a good idea.
Well, you have imbedded assumptions in there but a low source
impedance won't magically turn a non linear device into a linear one.
>Perhaps if a high output impedance is enforced, high
>impedance phones may become common,
I don't know how much you consider 'high impedance'. Seems to be a lot
of 300 ohm phones out there. Is that 'high'?
> but surely that would
>entrench a compromise to the detriment of fidelity?
Again you're operating from the imbedded assumption of some source to
driven impedance ratio but you don't explain the reasoning for it.
I get the impression it's an extrapolation from driving speakers but
headphones are not speakers. Mass is, by comparison, virtually
non-existent, they move almost no air, inductance is miniscule, and so
on. I don't have math to evaluate all that but they're so dramatically
different that simple extrapolations are highly suspect, at best.
Btw, the OTL doesn't 'require' the 120 ohm resistor and it works just
fine as a 'low impedance' output.
You could just remove the 120 ohm, and then adjust the FB resistor for
the changed Vo, but there's also a nice trick can be played. Instead
of the FB R and blocking cap. as shown, put a 12ohm under the existing
Rk on V2A and take the headphone return to that junction. That
eliminates the DC blocking cap.
With low impedance headphones there's a teensy improvement in
distortion because, without the 120 ohm, output swing is less but
it's negligible with 300 ohm headphones.
Like I said, plenty of room for tweaking ;) But the basic topology
looks good and 6CY7s are rather cheap.
You felt the 6AW8s looked 'gloomy'. Well, lemme tell ya, the 6CY7s are
anything but and those suckers glow like Christmas tree lights.
Oh, something I noticed after posting everything. When I compromised
with 200V on the phase splitter I forgot to readjust the R3 current
soak resistor. It's there just to suck up the idle current V1A puts
onto the output rail, and at 200V it's less than at 300V, so R3 should
be more like 250k. Doesn't really matter all that much, though.
I kept idle current low for the wallwart but gain goes up, and
distortion goes down, a bit if you increase idle and if you use a
conventional AC main supply you can go up to 50mA, which increases
idle from 4 Watts to a little over 10 Watts (per channel). Max Po also
goes up to 200mW. That's into 32 ohm headphones but it's more with 300
ohm phones and it'll do over a half watt with those (with 300V on the
phase splitter and, perhaps, a bit of optimization).
>cheers, Ian
>
|
