| cowanrg |
well, ive finally resumed progress on my big aleph2's.
my transformers will arrive tomorrow (1.7kva per mono). they have 45v secondaries. i am having a hard time finding caps that are large and rated above 50v. the ones i want to use are only 50v.
my question, after rectifiers, im sure i will be over 50v... with current draw, i should be right at that, or a bit under. ive heard of using a pi filter to lower power a bit? can i do something to make sure i stay under 50v?
or better yet, can somoene tell me where to get some reasonably priced caps that are 63 or 75v rated? im going to do about 250,000uf per channel, so they need to be BIG.
thanks! im ordering the metal here in a few days and will be re-designing the case. the new case will be 0.25" top, bottom and back, and 0.5" sides. its all going to be CNC milled to fit together like a glove. i didnt like the last one, looked too much like sheet metal. |
|
|
| AuroraB |
You will need at least 70V cap's.
45V secondaries will give you 60V++ after rectifiers, so a bleeder will have to take out at least 15V to be on the safe side!
If your trafo is a toroid, at least here in Europe, toroids are spec'ed at full load, which means that your idle voltage will be even higher. You should also have enough safety margin to handle slight overvoltage,- also within the allowed +10% from your power company.
I think I would settle for 80-100 V myself................ |
|
|
| byteboy |
If you have the space available for it, maybe using cheaper 50 Volt types in series (double number of capacitors, because capacity would be halved) would be a cheaper solution than buying (much-) more expensive single 80V-100V types?..........
What types did you plan to use and what is the price / source?
(I' am looking myself too for a good/cheaper source for capacitors) |
|
|
| jackinnj |
your power company is going to love you. what do you pay for a KWH out there in Colorado (it's about $0.12 here in NJ).
63.6 volts rectified and filtered -- plus a 15% margin for line voltage variation -- so you might as well breathe in deeply and order 80V (Panasonic TSHA) or 100V caps.
If you are going to load up a quarter of a farad, you should put a slow-start circuit on the power supply line -- I use a 100 ohm 50 watt resistor which is shorted out after a couple seconds -- this reduces the surge current -- remember that 1/4 of a farad looks like a dead short for some milliseconds. 50 watts seems to be OK since the resistor is only heating for a couple seconds. There are resistors in this flavor available in TO-220 packages == a little more expensive than the surplus store, but a lot smaller and they can be heatsinked.
Lastly, excercise extreme car when "tooling" around this much capacitance. I've stated before that you want the power supply voltage to decline by 90% in the time it takes you to unscrew the cabinet -- this may not be conservative enough in the case of so much capacitance since the reservoir will still be capable of supplying amps of current. It's time for "belt and suspenders".
When the music stops you will want some form of "bleeding" to kick in -- you can work the math -- if it were my project I would use the reservoir caps to drive a fan on an active load. |
|
|
| cowanrg |
thanks everyone for your input!
to start, my power only costs $0.071 per KWH. its a bit cheaper. lost month's power bill, for the WHOLE house with taxes (3 bedroom house with AC, and all electric appliances), was under $45. we never leave on lights that we dont need, etc... so power isnt an issue :)
as for bleeding the cap bank before working around it, im used to it. my aleph3's have 208,000uf per channel. this is almost 1/4 farad. as SOON as i opened up a case to work on it, or just turned it off to check a problem, i reached for this motor i have. its a pretty big one, but will operate on as low as a volt, but not burn up with 50v... so i just touched it to the cap bank, and it runs for a half a minute or so, and when it stops, i KNOW the bank is dead. it will still barely turn at 1v. i could bleed each rail down to about 0.7v or so. i did this every time i would play around.
i have no problem loading my aleph3's up. they are operated by a 12v trigger from my pre-amp, and they both kick on at the same time. i have a thermistor on them and it seems to work well. the lights dont really dim all that much when BOTH kick on at once.
i think the best solutions so far would be to use a TON of the little ones MikeW suggested (i would want like 30 or so per channel), OR do what byteboy suggested and just double up. this would effectively get me 100v from a 50v cap, which would be playing it safest.
my xformer is a toroid. loaded it will most likely be a bit lower, but i would still be running close to 60v if i wasnt bleeding off a few volts. so, i would need 63-70 or higher. so, let me see if i have room inside the case for 50 small caps, or a few bigger ones.
byteboy,
this is what i was planning on using before i found that 50v wouldnt work. they are 50v, but are 68,000uf, so you dont need many:
http://www.apexjr.com/capacitorsR.html
scroll down to computer grade. at that price, i can get several. |
|
|
| GRollins |
Burning off 15V or so per rail at ca. 3A is an awful lot of waste heat. If you're serious about building Aleph 2s using those transformers, I see three choices:
--Regulated rails.
--CRCRCRCRC...
--Variac
If anything else occurs to me, I'll check back in, but I'd recommend biting the bullet and building 1.2s.
Grey |
|
|
| cowanrg |
hum, ok.
none of those are options i really want to do.
i think my best bet for now is to keep the rails at their voltage, and get twice as many 50v caps as i had thought, wire them in series, and then have 100v caps. that would allow me higher rail voltages, AND keep my capacitance.
i was thinking of doing 8 of the 68,000 uf caps though. so now i need 16 per channel. thats a whole lot. |
|
|
| byteboy |
Maybe instead of CRC you could better use RCRC....etc. (or LCLC.... )?
It could lower the voltage at the first cap(s) enough, if correctly chosen, to a voltage you could get away with using 63 Volts caps? |
|
|
| cowanrg |
true, but i havent really seen any 63v caps out there either. PLUS, i do like the idea of actually having extra power on the rails. i plan to run these with a little more bias than normal, and always at 4 ohm, so the extra voltage cant hurt...
im seeing if i can fit twice as many caps in my case. looks like i can. |
|
|
| dshortt9 |
By the way, those are 43-0-43 V toroids so your theoretical after the bridge is 59V. I get 55 after a 120uf - 2mh -120uf pi filter for my Aleph 2. You could use a 2 ohm 25 watt resistor before the bridge to drop another 6
Dave |
|
|
| cowanrg |
hum, so IN THEORY, i could get less than 50v (barely). maybe ill wait until they show up today and do some testing...
im actually home sick with a cold, so ive got the time. im playing around in solidworks right now, and in a few hours i will have something to show for it! |
|
|
| Oil |
| Just use several smaller capacitors in parallel to make up your desired smoothing capacitance. It also has the added bonus of lowering the impedance of the supply, the ripple current rating will be higher AND it will cost less. There is a point of diminishing returns however... |
|
|
| janneman |
| quote: | Originally posted by dshortt9
By the way, those are 43-0-43 V toroids so your theoretical after the bridge is 59V. I get 55 after a 120uf - 2mh -120uf pi filter for my Aleph 2. You could use a 2 ohm 25 watt resistor before the bridge to drop another 6
Dave |
You could also use an impedance at the primary side, say a small transformer used as a choke. Less heat in your amp.
If you play it smart, like using the primary of a transformer in series with the primary of your main transformer, you can accurately set the final voltage by putting a resistor across the secondary of the small transformer, although that DOES genarate some heat again, but less than with the brute-force dropping resistor. That's VERY unelegant!
Jan Didden |
|
|
| Geoff |
| quote: | Originally posted by cowanrg
i think my best bet for now is to keep the rails at their voltage, and get twice as many 50v caps as i had thought, wire them in series, and then have 100v caps. that would allow me higher rail voltages, AND keep my capacitance. |
NO! You will need FOUR times as many capacitors to double the voltage (two in series) whilst retaining the same overall capacitance.
Also, you will not double the voltage rating with two in series since parameter variations will not give an identical voltage over each capacitor in a pair (unless you also fit a resitive divider). |
|
|
| cowanrg |
wow!
this seems like i need to get the xformers and find out just how much voltage im getting out of them. sounds like the best way would be to just lower the voltage somehow.
dshortt9 sold them to me, so he knows them well. if he is right about getting 55v rails, i could deal with that. all i need is another 6v or so drop to be safe. i do hate running caps AT their limits though. hum... |
|
|
| Oil |
| i think most people would agree here that trying to reduce the output voltage somehow is not a very good solution, but almost an inefficient 'bodge'. If you have not already bought the capacitors, it would be easiest, cheapest, most reliable and performance wise the best solution to use a bank of smaller sized caps, maybe 15,000uF per cap for example. |
|
|
| cowanrg |
ok.
i do seem to be getting that general feeling (but no one except you has said explicitally NOT to do it).
i guess ill have to look around for some high value, high voltage caps. i was hoping to do at least 250,000uf. that will take a LOT of 15,000uf caps. |
|
|
| Geoff |
| quote: | Originally posted by cowanrg
this seems like i need to get the xformers and find out just how much voltage im getting out of them. sounds like the best way would be to just lower the voltage somehow. |
I'm surprised that no-one else has mentioned this but, as you seem to be happy with a lower rail voltage, rather than burn-off the excess you could always remove some turns from the transformer secondary windings (assuming the toroids have been constructed in the normal manner) thus reducing the output voltage to a level that is within your proposed capacitor's voltage limits. |
|
|
| woody |
| Or what about using a bucking transformer to run the transformer at say 100v instead of 115v? |
|
|
| GRollins |
Does this transformer happen to have a multi-tap primary? I just pulled a stunt raising the voltage on the secondary by judiciously playing with the taps on the primary. Note that this is only good for about 10% or so at most, and that most transformers today have dual primaries, but not multi-tapped ones. Whether you'll be able to raise or lower the voltage on the secondary will depend on the line voltage in your area. Come to think of it, it's likely that you'll only be able to raise the voltage.
Geoff's solution is a good one if you don't mind doing surgery on the transformers.
Grey |
|
|
| cowanrg |
hum, i would rather not unwind the xformer. if its the last option, i can do it though.
here is my thought... i talked with some people at the electronics surplus shop i goto. they seem to think that taking two caps and wiring them in series will give you twice the voltage, and half the capacitance. so, a 1,000uf cap (50v) wired in series with the same cap will give you two 5,000uf 100v caps. is that right? someone on here said you woud need 4, they disagreed with that, and from what ive learned, i disagree too.
if that's the case, i could just go with double the caps in series, i would be fine with that.
i would rather keep the rails high, as i will have pretty good heatsinking, and the case will be nice and thick for sinking as well. i actually just got some renderings done and threw them up on my site. here is one:
http://www.cowanrg.mesanetworks.net...leph2/case3.jpg
this is just the basics, i dont have screw holes, etc... goto my site for more info if you wish. but the side plates that the heatsinks connect to are 1" thick, plus the huge heatsinks.
anyways, thats what im thinking. if i can do that, im willing to just buy double the caps. im (believe it or not) a good amount UNDER budget for this project, and i have little left to buy (just little odds and ends at this point).
thanks again guys. |
|
|
| GRollins |
Two 1000uF 50V caps in series equals one 500uF 100V cap so, yes, it will take four times as many caps to equal the original capacitance.
And as Geoff pointed out earlier, you should use resistors in parallel with the caps in order to ensure that the voltage stays put. Otherwise, the voltage midpoint between the caps will tend to drift, with potentially disasterous consequences. No need to lose a lot of current to these resistors--anywhere from 10k to 100k will do just fine. One resistor to each cap. Tube people do it all the time. In fact, I'm running 575V rails on two 350V caps in series (with, as I recall, 100k 2W resistors--you won't need 2W resistors, 1/2 to 1W will do, depending on what value you choose).
Grey |
|
|
| cowanrg |
sigh.
well, thre goes that idea. i actually found some really nice 120,000uf caps that were 25v, for $2 each. wish i could just do those.
maybe ill think about unwinding the xformer. ill have to see what voltage i have after rectification. |
|
|
| Magura |
You could get 120.000uF 75V from those caps, for 18 USD a bank.
All you gotta do is to put 3 in series, and 3 times that in parallel.
Magura:) |
|
|
| cowanrg |
| hum. so if i had room, and wanted to get a bit more capacitance, i could put a few banks of those... cheap, but a LOT of caps. |
|
|
| Magura |
| quote: | Originally posted by cowanrg
but a LOT of caps. |
LOL....sure, as in 36 caps per PSU :D
Magura:) |
|
|
| cowanrg |
| how much is that? thats 36 caps for 270,000uf? |
|
|
| Magura |
I expected you to run a CRC or CLC, in that case its 240.000uF on each rail, 480.000uF per monoblock.
Magura:) |
|
|
| cowanrg |
| hum... 36 big can caps in one case. thats a whole lot. ill have to measure them and see if it can be done. i have a lot of extra room. most of the circuit boards will be embedded into the case, so the xformer and rectifiers will be inside the case in addition to the caps, and thats about it. |
|
|
| Magura |
It sure would be a cheap solution, and 480.000uF per monoblock ought to make do ;)
Magura:) |
|
|
| cowanrg |
yeah, but if i remember, these are pretty big cans. 36 might be a LOT of them.
they are screw in, thats 72 screws for each channel! its gonna weigh a ton just for the hardware! |
|
|
| Magura |
| quote: | Originally posted by cowanrg
they are screw in, thats 72 screws for each channel! its gonna weigh a ton just for the hardware! |
Any decent amp is at least 100kg (~200 lb)..... per monoblock (integrated transportation system anyone??) , so you got nothing to fear ;)
Magura:) |
|
|
| cowanrg |
i dont see how they can be that heavy. ive estimated mine, and its going to be less than 100lbs per mono. and thats with no part of the case less than 0.25". the face plate is 1.5", and the side panels are 1".
i just need to see if it will fit. |
|
|
| paulb |
How about a "CQC"? Relatively small capacitor up front, with whatever voltage rating you need, then a capacitance multiplier (transistor) with a zener diode to drop the voltage, then a big capacitor at lower voltage on the output.
This is only practical if you don't need to drop that much voltage. You may want to consider getting or building a dummy load to do some experiments.
Stay away from those people in the capacitor store; that's pretty bad advice they're giving.
Or how about a light bulb in series with the primary? |
|
|
| cowanrg |
hum. you guys give me too much credit.
yeah, those guys were wierd... they kept suggesting a switching supply. he said he didnt know of any good amps that used linear supplies, they all used switching. i didnt even argue. i didnt even come in for advice, just for caps.
i think if i can fit enough caps in there, i will just go that route. at $2 each, its a good price. i just hope they have 72 of them :( |
|
|
|
