Capacitor for DC servo

Status
Not open for further replies.
Jump to Latest
B

borges

Member
Joined 2003
#1
Hi,

I'm putting a 1st-order lowpass filter into my circuit to block DC output. Question is: what type of capacitor?

I figure out that I need a bipolar cap that can handle signals of +-4V. It needs to be >100µF. Low currents, up to +-8mA.

Any tips? Os-cons?


Børge
 
#2
Is this an active servo, or just the common RC filter in the feedback path? IMO, the main reason for using an active servo is to avoid the use of large caps. You can use a quality film cap. In the normal RC filter, you're stuck with large values. Films aren't practical there. IMO again, bipolar caps are pretty crummy. I'd use an ordinary electrolytic, as there shouldn't be any significant voltage across it. I'd also use a larger value than theory says is necessary, to put any adverse effects far out of the audio band. -3dB at a few Hz is reasonable. Oscons have much better HF performance and lower DF than other electrolytics, but higher DA by my measurements. Since you'll likely have a series resistor, the DF of the cap doesn't mean much. Not sure about the DA. I'd lean towards a high quality standard electrolytic, but why not listen to both? Just don't be too distressed if they sound the same. Or if they don't, be sure to post back with your results.
 
#3
Conrad,

my initial thought is to replace R by (R+1/sC) in the input branch of an inverting opamp circuit. That is equivalent to a passive lowpass filter when it comes to component values.

The feedback path around this inverting opamp is a simple (R||1/sC). I'd like to avoid microphonics in the capacitor. So I guess it's back to the desk to find a nice active solution.


Børge
 
#5
Lost your post for a while- sorry! Understand that I've never met a bipolar cap I liked, so I've never paid much attention to the more recent ones. Any that I've measured in the past were pretty awful. I won't even use 'em in crossovers. Thus, the three little words that are so hard to say- "I don't know." Your application is one that might bring out flaws you wouldn't hear in a high impedance coupling application, which actually makes it interesting as a test case. One design cheat is to run two polarized caps back to back, but put a high value resistor to the junction and bias them. This only works if you have low impedances on the other sides, and if the caps have very low DC leakage. I found no great advantage to this in higher impedance coupling circuits, but your case is different. Caps are cheap, so I'd get a few and try 'em!
 
#6
What size caps do you need?


I have referenced some available large-valued film caps that are fairly-reasonably priced, here:

*** LARGE-VALUE FILM CAPS:

http://www.diyaudio.com/forums/showthread.php?s=&threadid=103336


And I recently posted a schematic that is a pretty good starting point for a DC Servo design, here:

*** DC SERVO SCHEMATIC:

http://www.diyaudio.com/forums/showthread.php?s=&threadid=103308


If you'd like to do LTspice modeling and include the caps' ESR that varies with frequency, one method is posted here:

*** SPICE MODEL FOR CAPS' ESR VARYING WITH FREQUENCY:

http://www.diyaudio.com/forums/showthread.php?s=&threadid=102208&perpage=10&pagenumber=3


Regards,

- Tom Gootee

http://www.fullnet.com/~tomg/index.html
 
Status
Not open for further replies.
Top Bottom