I am putting together a Hiwattish amp using a 700v voltage doubling PS design based on MusicMan iron. Hiwatt power supplies have a lot of filtering in the first stage - typically around 100uf. In the DR201 which has a similar high voltage PS design 4X100uf 450v caps are used in parallel/series to get a 100uf 900v rating.
I want to do something similar, but being a cheap ass I would ideally like to use parts on hand and I only have 2 100uf caps but bunches of 47uf caps. So the question is: Is it kosher to use 1-100uf/450v cap and 2-47uf/450v caps in parallel to get 194uf/450v? I would use two parallel sets in series to get 194uf/900v. So far the sources I have read lead me to believe this would work, but is there some downside in using caps of different capacitance but the same voltage in parallel in a tube amp power supply?
Check on the series balancing resistors. I was really more concerned with any possible issues from connecting different value caps in parallel. I do just barely have room, and yes, when I mocked it up in TW stacked cap fashion it does look butt ugly even to non-anal me.
I may chicken out and buy new caps but it's the principle. I have all these cheap 47uf caps from Apex Jr. just begging to be used.
Shouldn't be a problem really no different than mixing resistors of different value long as your voltages will work together and not stress on or the the other.
It doesn't matter "a rats" if you use 100uF , 2 x 47uF in parallel or 5 x 22uF in parallel.
When you start to take them (either a single cap or parallel combination of caps) and series connect them, then there are some things to make sure off.
1st - series connected caps will share voltage according to the inverse of the capacitor sizes. If you want equal voltage across the caps or cap banks then you MUST have equal capacitance. (The charge current is the same for the series connected caps, the lower value capacitor will charge up to a higer voltage than the higher value capacitor).
2nd - you must use voltage share resistors. The "usual" values you see for voltage share resistors are TOO high.
Guidance:
2 x 1000uF in series use 47K max.
2 x 470uF in series use 100K max.
2 x 220uF in series use 220K max.
2 x 47uF in series use 390K max.
2 x 22uF in sereis use 470K max.
What you want is 3 to 5 times the capactor leakage current flowing in the voltage share resistors. The Cap data sheet will generally give you a formula to calculate the leakage current.
If the datasheet does not give the leakage current formula then use this as default:
Ileak = 0.006 CV
where C is the capacitance value (in Farads) and V is the applied voltage (not the caps rated voltage)
In a Tube Amplifier Power Supply, the High Voltage Rail will be the peak value of the secondary voltage untill the tubes start to conduct. The series connected caps should have a voltage rating sufficient to handle this peak voltage.
Also watch the power rating of the voltage share resistors. In general 2W metal film will do for teh 220K or larger but for the 100K and 47K you will need to use higher rated resistors or series parallel connect resistors too.
gingertube wrote:Also watch the power rating of the voltage share resistors. In general 2W metal film will do for teh 220K or larger but for the 100K and 47K you will need to use higher rated resistors or series parallel connect resistors too.
Cheers,
Ian
Something to be said for voltage rating of these resistors too. Many R's top out at 200-250V.
Mouser has a selection of 3 Watt IRC carbon R's that are 1000V rated. I stock 100K and 330K for voltage-share service. Sorry no link & double sorry to make you dig thru Mouser's search BUT these are worth finding.
Leo_Gnardo wrote:
Something to be said for voltage rating of these resistors too. Many R's top out at 200-250V.
Mouser has a selection of 3 Watt IRC carbon R's that are 1000V rated. I stock 100K and 330K for voltage-share service. Sorry no link & double sorry to make you dig thru Mouser's search BUT these are worth finding.
I was planning on using Xicon 220k 3W metal oxide balancing resistors. Now you've given me one more thing to obsess over and one more thing to order. I suspect you are in league with the USPS
One more thing to note, although I don't think you'll be doing it here, is that if you use different values of capacitance in series then the balancing resistors will not be equal in value. I usually find a calculator online to spit out everything (required resistances and wattages) in one shot.
Leo_Gnardo wrote:
Something to be said for voltage rating of these resistors too. Many R's top out at 200-250V.
Mouser has a selection of 3 Watt IRC carbon R's that are 1000V rated. I stock 100K and 330K for voltage-share service. Sorry no link & double sorry to make you dig thru Mouser's search BUT these are worth finding.
I was planning on using Xicon 220k 3W metal oxide balancing resistors. Now you've given me one more thing to obsess over and one more thing to order. I suspect you are in league with the USPS
Those Xicons have a working voltage of 350, so in a 700 volt supply with two equal series caps and two equal balancing resistors, you're right on the edge. The 5W Xicons are rated at 500 volts, so would work more reliably.
Leo_Gnardo wrote:
Something to be said for voltage rating of these resistors too. Many R's top out at 200-250V.
Mouser has a selection of 3 Watt IRC carbon R's that are 1000V rated. I stock 100K and 330K for voltage-share service. Sorry no link & double sorry to make you dig thru Mouser's search BUT these are worth finding.
I was planning on using Xicon 220k 3W metal oxide balancing resistors. Now you've given me one more thing to obsess over and one more thing to order. I suspect you are in league with the USPS
So as randomness allows, it turns out that I have some 220K 3 watt IRC resistors of just the kind that Leo suggested. Not sure how they got in my parts pile - either through my frequent mis-ordering from Mouser or from grabbing plausible looking bargain parts at the local electronic surplus store, but either way I'll go with it. Kind of surprising - I haven't been doing this that long and my spare parts pile isn't that big, but hey..
You'd be surprised at what I've found in my (admittedly way too large) stash of parts. All sorts of parts and projects that I've forgotten about and rediscovered multiple times. Heck, sometimes what I find is even useful!!
