RC filter cutoff frequency

Discussion in 'General Electronics Chat' started by coinmaster, Feb 7, 2016.

  1. coinmaster

    Thread Starter Member

    Dec 24, 2015
    350
    6
    Hello, I am trying to make a 600v RC filter power supply for a tube amp but high voltage capacitors are quite expensive.
    I was wondering, for an audio amplifier does it matter whether the cutoff frequency is 9hz or 1hz? Is there any reason to care about filtering below 10hz?
     
  2. Mike33

    AAC Fanatic!

    Feb 4, 2005
    349
    25
    No, for all practical purposes in tube amp construction, 9Hz is fine!

    But - don't you need to get cap(s) rated to your PS, anyway? Where's the cap - power supply, or coupling? It's just a bigger cap to get you to the magical 1Hz, isn't it? That's the target to shoot for, altho it makes little difference in a practical circuit. But we do try for the ideal, right? ;)

    Post a schematic of what you've got going on for more specific info...if it were my build, I wouldn't worry about it, many classic designs sure didn't... ;)
     
  3. coinmaster

    Thread Starter Member

    Dec 24, 2015
    350
    6
    Yeah but bigger caps are high voltages cost $$$.

    Although, I just realized, RC filters aren't brick walls so maybe a 9hz cutoff will leak into higher frequencies?
    Maybe I should use capacitors in series, it seems way cheaper to get high voltage caps this way.
    I've read that if you put caps in series that you need to use leak resistors to balance them in case one capacitor has more leakage than the other which would cause one capacitor to exceed it's voltage rating due to imbalance.
    Here is the schematic series capacitors-1.png I am unsure how to calculate the correct leak resistor values and I wonder if the leak resistors affect the filtering somehow.
     
    Last edited: Feb 7, 2016
    Tech7 likes this.
  4. Mike33

    AAC Fanatic!

    Feb 4, 2005
    349
    25
    Dude, they are not going to cost THAT much more...where do you live?

    By the time you put equalizing resistors and all, I believe you'll have complicated the design enough to maybe make trouble for you down the line. I'd just buy the correct caps. Think about 10-20 yrs. from now, when someone (you) has to change the caps...

    Some other info to think about: what are your current demands - how is the amp designed (how many preamp stages, what're you doing for a power amp - how much do you need to feed it, as a load)?

    Anyway, if you do want to try to 'series' them, do some reading: http://www.valvewizard.co.uk/smoothing.html

    I don't think they do much re. filtering, and 220k to 470k are typical values (1/2W). 10x leakage current of the caps.

    You ARE going to have an RC network for each triode down the line, right? That will take care of anything that sneaks thru :)
     
    Tech7 likes this.
  5. coinmaster

    Thread Starter Member

    Dec 24, 2015
    350
    6
    I plan on making quite a few of these power supplies, it's worth the trouble to series them. It costs like $40 in capacitors to make that filter with series capacitors and multiple hundreds of dollars with single capacitors.

    I need to output of the filter to to have some overhead for the shunt regulator so the input to the filter will be at around 700v. 500uf 700v capacitors are not cheap.

    Maximum current is going to be about 100ma. This power supply is for the input/amplification stage. The power stage will be on a completely different power supply at much lower voltages so that's not a problem at all.

    Yeah, each triode is going to have it's own RC filter. The input stage needs to be super quiet though so I have to make sure I do it right. I need to figure out how to make a super stable and quiet shunt regulator that will function on 600+v as well but I think I will make a separate thread for that.
     
    Last edited: Feb 7, 2016
    Tech7 likes this.
  6. Mike33

    AAC Fanatic!

    Feb 4, 2005
    349
    25
    Why so high a B+, by the way? Just asking...if this was more conventional, 400 to 600V caps are all over, and certainly not hundreds of dollars!!!

    But, ok, you did ask. To do your caps in series, of course it's the series formula (C1*C2/C1+C2) for total capacitance. I.E., 2 1000uF caps = 500uF. For just 2 400V rated caps in series, I would use a 470k 1/2W resistor across each one. This will also make a nice bleeder network, as well. The comparative effects on ripple will be negligible. Simulate this with LT Spice or what have you.

    Wow, 100mA preamp? That is SOME preamp!! Your network for each B+ should make this more than quiet.
     
    Tech7 likes this.
  7. coinmaster

    Thread Starter Member

    Dec 24, 2015
    350
    6
    Well the tubes will be 300B tubes and possibly some high voltage sweep tubes as well. The 300bs like to run at around 350v. It seems to be a common rule to drop 1/2 the B+ supply when powering the plate for headroom or whatnot. The Shunt regulator will need over 600v at its input so it has room to function.
    Adding capacitors in series seems to be a swell solution, it downgrades the cost of the power supply by multitudes. Sadly the diminishing returns hit hard after 2 in series.

    You have no idea :cool:
     
    Tech7 and Mike33 like this.
  8. Mike33

    AAC Fanatic!

    Feb 4, 2005
    349
    25
    Perhaps the series caps are the way to go for sure then, ha ha ha...this sounds a little larger than what I'm used to working on, just guitar amps and a little radio gear.
    What is the whole project, if you care to talk about what you're working on? Hifi system or something? Radar? :)
     
    Tech7 likes this.
  9. coinmaster

    Thread Starter Member

    Dec 24, 2015
    350
    6
    It's for a hi-fi system. It is a small piece to a greater concept. I would love to talk about what I am working on but the idea is so good I think it would be foolish to do so. I'm working on something that I can put on the market.
     
    Tech7 likes this.
  10. AnalogKid

    Distinguished Member

    Aug 1, 2013
    4,546
    1,252
    Yes.

    For a single-pole R-C filter, the attenuation is only 3 dB at the cutoff frequency. As you move the corner frequency lower and lower, the attenuation at the frequency of interest increases. Starting at 120 Hz, each octave lower corner frequency adds 6 dB of attenuation, or each decade lower adds 20 dB.

    120 Hz -3 dB
    60 Hz -6 dB
    30 Hz -12 dB
    15 Hz -18 dB
    7.5 Hz -24 dB

    120 Hz -3 dB
    12 Hz -20 dB
    1.2 Hz -40 dB

    From 9 Hz to 1 Hz is greater than three octaves (-18 dB) but less than one decade (-20 dB) worth of extra attenuation at some higher frequency.

    ak
     
    Tech7 likes this.
  11. coinmaster

    Thread Starter Member

    Dec 24, 2015
    350
    6
    I wonder what the practical affect is on a preamp stage though.
     
    Last edited: Feb 7, 2016
  12. AnalogKid

    Distinguished Member

    Aug 1, 2013
    4,546
    1,252
    All preamps have hum and noise on the output. The question is, how much is too much? The less ripple there is on the power supply output, the less the preamp is dependent on its power supply rejection (a form of common mode rejection) to keep that hum out of the output.

    ak
     
Loading...