Wessel Dirksen wrote:
Patrick Turner schreef:
Wessel Dirksen wrote:
Hi RATs,
I'm considering a grounded grid stage as the first input stage in
my
continuating first power amp project. Considering that my
soundcards
(Lynx 2 and Terratec EWX-24/96) both have low output impedances
(100
and 200 ohm resp), it seems that this would be possible from a load
perspective. I am considering a rather high gain config with a
12AU7
that according to my calculations should yield an input load of
about
1.4K. So a few questions:
How did you calculate the Rin to a 12AU7 with grounded grid was 1.4k?
As it turns out calculated incorrectly. I have rp+Ra/mu+1 as the input
impedance formula and I'm only getting about 300 ohms or so now.
Disappointed though. I was hoping I was on to something.
If you can't remember all formulas, think in basic terms.
Say your grounded grid input stage has 1/2 a 12AU7 with 47k anode load and
2.7k Rk, and you wish to connect to the cathode.
The tube open loop gain is µ x RL / ( RL +Ra )
For 12 AU7, A = 14 approx.
So for 47 volts of change at the anode, you'd get 1 mA of of I change.
Since A = 14, you'd need ( 47 / 14 ) v at the cathode, = 3.35, and since
there *must* be 1 mA change there too the Rin = ( 3,35 / 1.0 ) k ohms =
3.35k
In other words, Ri = LL at a / A.
BUT, we also have 2,7 k in parallel with Rkin, so Ri = 3.35k parallel to
2.7k
= 1.498k, or say 1.5k.
If you were to use a CCS from oV to the cathode, then Ri at k
is simply 3.35k.
1)Seems like a reasonable impedance matching for soundcards but
will
this also be compatible for other pre-amps ect?
No.
For future reference, what is a typical output impedance for a preamp
or CD player?
Ro = low = 600 ohms typically,
Ri = high = over 10 kohms typically.
Output Z should be low so the voltage output is unaffected
by the load of something connected to it.
All devices should follow the rule of low Ro and high Rin.
2)I believe that the input would see the low DC bias voltage of a
few
volts. Can I direct couple this without the soundcard getting
unhappy,
or would a soundcard need to be shielded from this small voltage?
Use a coupling cap to isolate the tube DC from the sound card output.
What value C will you need to get a LF pole at 7 Hz?
Ok, also for future reference then. I guess you are suggesting that it
is a good idea to cap couple the input. Primary 1st order pole at 7 hz
would seem like a good filter point. Good LF bandwidth but not too
rediculous high value of C.
-3 dB point at LF is where the recatance of the cap in ohms
= Rout + Rin.
If you have 600 ohms feeding 1.5k, you have a total R = 2.1k.
To get a reduction of signal current by 3 dB you need C =
1,000,000 / ( 2,100 x 6.28 x 7 ) uF = 10.8 uF.
So you'd ask for a 100 volt rated 12 uF cap at the counter of the parts
store.
3)I want to run feedback to this stage. It seems to me that you
could
maybe put a small resistance between the grid and ground and put
the
feedback between them. Is this possible?
Yes, this is possible, and phase of the FB has to be correct.
The Rin to the cathode will rise, and so will the required
input voltage.
Humm, this would raise Zin then as well but not enough to help Zin
without killing all the gain.
Oh well . . .
Your power amp should be configured to make full power at a volt
input including NFB.
The soundcard can make up to a volt output, no?
Thanks. I'm coming along in this tube stuff. Lots of fun.
Before you award yourself a medal for your achievements,
be sure you know what you are doing :-).
Considering my nonsense calculation of 1.4k for Zin, I can understand
this caveat.
Yet the addage " . . . know just enough to get yourself into trouble"
is something I thrive on during the learning process.
Thanks for your input Patrick.
Wessel
In time, you will become use to sorting out all the factors affecting a
simple
circuit.
Learning the basics about triodes and some RLC network theory
are the most important things when building analog amp stages.
I tell Mr Yeager that all the time, and finally I hope to not have to tell
him anything;
he'll always get it right.
Patrick Turner.
Patrick Turner.
Wessel
|