I'm gathering the parts for this OTL Headphone Amp:
My question is on C1 and C2. The minimum value for C1 is 10uf. How whould I determine the right value to get the best frequency response?
My headphones are 48, 55 and 250 Ohms
An externally hosted image should be here but it was not working when we last tested it.
My question is on C1 and C2. The minimum value for C1 is 10uf. How whould I determine the right value to get the best frequency response?
My headphones are 48, 55 and 250 Ohms
You are probably going to want a 470uF cap which is going to be an expensive part as you'll need a pretty high voltage. For how to figure this out, see my article here: http://www.ecp.cc/cap-notes.html
That article explains a bunch for me.
Just so that I can see I'm using the formula right
F = 1/(2 * pi * C * R)
what is the value obtained if C = 10uf and R = 250 Ohms
Just so that I can see I'm using the formula right
F = 1/(2 * pi * C * R)
what is the value obtained if C = 10uf and R = 250 Ohms
It's about 64Hz (10uF is .00001F).
Really, your caps are probably going to cost more and sound worse than a couple of cheap output transformers. I don't get the OTL thing in general, but I don't think it is particuarly appropriate for headphones, at least not low impedence ones.
One option might be to use the small cap, and add an impedence matching transformer -- something with a 10K primary and a 32 ohm secondary -- after the cap. Or, you could put a solid state buffer after the output cap. These tend to have a very high input impedence which would allow you to get away with a little cap -- something like 0.1uF perhaps?
Really, your caps are probably going to cost more and sound worse than a couple of cheap output transformers. I don't get the OTL thing in general, but I don't think it is particuarly appropriate for headphones, at least not low impedence ones.
One option might be to use the small cap, and add an impedence matching transformer -- something with a 10K primary and a 32 ohm secondary -- after the cap. Or, you could put a solid state buffer after the output cap. These tend to have a very high input impedence which would allow you to get away with a little cap -- something like 0.1uF perhaps?
Now the real question is how low can I actually hear? I should get my ears cleaned and hearing checked anyway!
To calculate the cap size for a given low frequency response..
You need to know two R values...
The (2*pi*R*C) is an ideal equation that assumes you have an ideal source, which is never the case...
You need to know the equivelent impedance looking back into the amplifer from before the cap...and then ahead of the cap which you already know as you headphones.. I would use the smallest headphone impedance for the calculation, since anything higher is gravy...
Keep in mind you typically use a -3dB frequency much lower than 20 Hz, the reason is that in audio you are concerned with phase shifting.... Since the Phase shift at -3dB for a -20db/decade slope is 45 degrees, not really good for audio... Drop the freq then examine the phase shift until it is satisfactory.. I personally design for 6Hz or lower to obtain least phase shifting at 20Hz..
Chris
You need to know two R values...
The (2*pi*R*C) is an ideal equation that assumes you have an ideal source, which is never the case...
You need to know the equivelent impedance looking back into the amplifer from before the cap...and then ahead of the cap which you already know as you headphones.. I would use the smallest headphone impedance for the calculation, since anything higher is gravy...
Keep in mind you typically use a -3dB frequency much lower than 20 Hz, the reason is that in audio you are concerned with phase shifting.... Since the Phase shift at -3dB for a -20db/decade slope is 45 degrees, not really good for audio... Drop the freq then examine the phase shift until it is satisfactory.. I personally design for 6Hz or lower to obtain least phase shifting at 20Hz..
Chris
Did some number crunching...
The output Z of the circuit is roughly 220 ohms.....
Driving a 48 ohm load:
You would need a 100uF to get a -3dB at 6Hz.....roughly...
Driving 250 ohm load:
You would need 57uF to get the same -3dB at 6Hz....
Since they don't sell 57uF, you can parallel caps or use a 68uF that will bring you to -3dB at 5Hz...
Using 6Hz as your -3dB point will roughly bring you in to 12 degrees of phase shift....
Chris
The output Z of the circuit is roughly 220 ohms.....
Driving a 48 ohm load:
You would need a 100uF to get a -3dB at 6Hz.....roughly...
Driving 250 ohm load:
You would need 57uF to get the same -3dB at 6Hz....
Since they don't sell 57uF, you can parallel caps or use a 68uF that will bring you to -3dB at 5Hz...
Using 6Hz as your -3dB point will roughly bring you in to 12 degrees of phase shift....
Chris
That's a great explanation.... How does the first value of resistance figure in?
By using the Ideal formula I get around 33.1Hz from a 48 Ohm load using 100uf
How would I figure out the output Z or another circuit?
By using the Ideal formula I get around 33.1Hz from a 48 Ohm load using 100uf
How would I figure out the output Z or another circuit?
sbelyo said:
OK....
Looks like your using 48 0hms as your R ....This would be fine if you have an ideal voltage source, meaning ZERO source impedance as you do in basic textbook examples.... But in real life your R is actually the source impedance in SERIES with the load resistance for your time constant... The output resistance(source resistance) is roughly 220 ohms from your pre-amp stage.....
Add the 220 + 48 and your R is 268 ohms... Now re-figure..
To figure the output resistance, you first figure the impedance looking into the 12AU7 cathode at that operating point which is 454 Ohms...you have two paralelled sections so that is 227 ohms...then you parallel that with the plate resistors in series with the plate resistance from the 6CG7..which is slightly increased from the unbypassed cathode resistor.... Long story short you have 220 ohms... Remeber it's AC impedance not DC resistance....Then if you had feedback involved you would account for the feedabck ratio...but no feedback....
Keep in mind that your over-all Gain takes a dip as you go lower with the load...
Chris
Darn... try as I might, I just don't seem to be able to figure out how to apply the above explanation.
The plate resistance of the 12au7 at 250V is 7.7K Ohms and it's the same for the 6CG7. I understand parallel and series resistance, but I don't know how to make the equation.
The plate resistance of the 12au7 at 250V is 7.7K Ohms and it's the same for the 6CG7. I understand parallel and series resistance, but I don't know how to make the equation.
sbelyo said:
Keep in mind that the plate resistance and the gm change depending on the operting points of the valves...Refer to the plate curves to really find the correct numbers..but most time with triodes the data sheet is approx and close enough....
In your case the 6CG7 is figured by using the un-bypassed cathode equation...since the plate resistance rises when cathode is un-bypassed... It is Ra= ra+(u+1) * ( Kr) .......
Kr= cathode resistance.... then add is the plate resistors on top..
Then put that in parallel with the 12AU7 .....
Basically the 12AU7 will dominate and no need to really do the 6CG7, but will do just for excercise...
The 12AU7 is a follower...so the impedance looking in is at the cathode....so that is roughly 1/gm ..... In your case your gm= .0022 A/V ....So 1/gm= 454 ohms looking into the cathode of one 12AU7....since the two sides of the 12AU7 are paralleled you would get 227 ohms.... If you put that in parallel with the 6CG7 section it brings it down to 220 ohms.... for your output impedance....then this is in series with your load of 48 ohms to make 268 for your R used for your time constant...
Chris
- Status
- Not open for further replies.