iv output cap - value

[lordearl]

Looking at taking out this 47uF electrolytic cap and replacing with a film or oil cap - I assume it is just there to block DC. What value is OK (would 10uF work)?

[Mooly]

The value depends entirely on what load you hang on the end of it and what you wish the cut off frequency to be. And yes, it is to block DC either from the stage itself or from whatever it connects too.

The cap and the load form a 1st order filter,
High-pass filter - Wikipedia, the free encyclopedia

10 uf should be OK for feeding any "normal" preamp.

[lordearl]

Interesting article! I can't see a calculation method there which factors in the input impedance of the amplifier. I am running this dac straight into a 6V6 power amplifier (with a pot installed) that has a 100K input impedance. Although I assume the input impedance changes depending on what volume the pot is set at.

I do have a 10uF film cap, but is the cut off frequency adversely affected by dropping the value from 47uF to 10uF?

[Ken Newton]

BTW, that White source-follower JFET output buffer has a design flaw. That is, assuming you were seeking symmetrical push-pull drive from the stage. The current-source formed by Q12 should be removed and R11 connected directly to the positive supply rail. The correct value for R11 would be equal to 1/[Transconductance of Q10 + effective resistance of VR2]. This will cause the anti-phased currents of Q10 and Q11 to balance, giving the greatest reduction in distortion. Details on this are explained by John Broskie on his TubeCAD website, under 'White Cathode-Followers'.

[lordearl]

Hi Ken & Mooly,

Thanks for the help. This IV stage comes pre-attached to an AD1865 board, so I have little choice in how I can modify it. The sound is already quite good (preferable to my CEC DX71 dac) but I'd like to try to change a few small things if possible.

The highest uF film cap I can find is 10uF, will this be OK if I am putting it into a power amp with input impedance 100K? If not, I also have a preamp with input impedance of 2K.

In relation to the design flaw on the white source follower, do you mean I simply remove Q12? Then try to determine what value resistor to replace R11 with, and connect the resistor in place (ie. over) of Q12?

[Ken Newton]

10uF will be more than enough. In fact, if the load is 100k ohms, you could go to a much smaller cap. Say, just a few microfarads. I've read very good user reports regarding the Dayton branded film & foil capacitors (not to be confused with their metalized film cousins) sold by PartsExpress, the largest of which is 0.47uF. You would want to parallel about 5 of them per channel. Effectively, creating a 2.35uF composite capacitor. They cost $1.40ea. if you buy 10 or more - a bargain by premium film cap. standards.

[lordearl]

Excellent, if 3.3uF should be enough, I'm sure I can find something which will make a good improvement on the existing electrolytic caps.

Not sure about the Q12/R11 issue, anyone else out there got this i/v stage + ad1865 board?

[commstech]

my amp has 100k input impedance. i experimented with film output caps to replace some 100uf electrolytics in my discrete iv stage too. 3uf~3.3uf should be the bare minimum and 4uf ~ 5uf will be excellent, bass is as low i like it to be. anything lower than 3uf, be prepared to compromise with bass rolling off, loss of music ambience etc.

[lordearl]

thanks mate - what brand/model film caps did you try?

[Ken Newton]

There is an error in my above formula for computing the optimum value for R11. It should be: R11 = (1/Transconductance of Q10) + the effective resistance of VR2.

Sorry, for any inconvenience.