It's nothing wrong with Arcotronics, possibly the 1500 uF/35 V seems to be not so good. 1000 uF/35 V has worked like clockwork. The customer support though.... from one person in particular could be better.... 
Hi Peranders,
have you told them about this thread and pointed out the bad publicity they are creating for themselves?
It would be very easy for them to give you the support you need and then you could come back and tell how good they really are.
have you told them about this thread and pointed out the bad publicity they are creating for themselves?
It would be very easy for them to give you the support you need and then you could come back and tell how good they really are.
LOL, that's some pretty extreme stuff, I vote for the railgun. Or are you making an electric car ? The voltages seem prety low for that, though.
Anyway, with probability of failure p, having N caps gives a probability of failure for the assembly of 1-(1-p)^N, if p is small you can approcimate this to N*p...
So if you want to have 1 in 1000 of your boards fail with 3000 caps, you need a failure p = 3.335e-7, that is one in 3 million caps should fail !
But the result of your experiments is even more interesting. I bought Panasonic Low-ESR before because of the performance, so I'll keep buying that also because it doesn't blow up !!!
Anyway, with probability of failure p, having N caps gives a probability of failure for the assembly of 1-(1-p)^N, if p is small you can approcimate this to N*p...
So if you want to have 1 in 1000 of your boards fail with 3000 caps, you need a failure p = 3.335e-7, that is one in 3 million caps should fail !
But the result of your experiments is even more interesting. I bought Panasonic Low-ESR before because of the performance, so I'll keep buying that also because it doesn't blow up !!!
Yeah, I had guessed that was a power supply, but the bet is on... what is it supplying ?
An aluminum electrolysis plant ?
A super secret fusion experiment ?
A million LEDs ?
A RADAR ?
An aluminum electrolysis plant ?
A super secret fusion experiment ?
A million LEDs ?
A RADAR ?
an EMP generator?...........THUMP! ......"hey, what gives, my pc quit all of a sudden????? heck, power went out too......".......
i have used banks of large PS caps, a coil of 10ga wire and a motor relay to magnetize screwdrivers (and it shoots screwdriver bits too).......
i have used banks of large PS caps, a coil of 10ga wire and a motor relay to magnetize screwdrivers (and it shoots screwdriver bits too).......
2304 mosfets, 220 A type a piece are working together.... class D, weight around 2000 kg, water cooled.... a DIY'ers wet dream. 
Hi Peranders,
we have recently started using Relex software to determine MTBF information. The biggest issue we have had is capacitors (especialy electros) working near their working voltage. A few caps working near working voltage, mainly in the power supply section was giving us figures of 20-30 hours before failure, doubling the working voltage started giving more acceptable figures. We use Panasonics for quality, but have to use generic figures for the reliability figures. We work to the ultimate level of reliability prediction, but even so it is a worrying trend, that capacitors are regarded as so unreliable.
400Vac!! is it a starter supply for a jet engine?
we have recently started using Relex software to determine MTBF information. The biggest issue we have had is capacitors (especialy electros) working near their working voltage. A few caps working near working voltage, mainly in the power supply section was giving us figures of 20-30 hours before failure, doubling the working voltage started giving more acceptable figures. We use Panasonics for quality, but have to use generic figures for the reliability figures. We work to the ultimate level of reliability prediction, but even so it is a worrying trend, that capacitors are regarded as so unreliable.
400Vac!! is it a starter supply for a jet engine?
Interesting problem. Being in the domestic appliance repair game at present I can only agree the quality of electrolytic caps has fallen.
I've never considered multiple caps in parallel to have issues. But in this situation there is a huge energy supply. If one cap in the bank developed low leakage there's a huge current source right beside it. In a typical PSU the current available is limited by the inductance of the transformer and the rise time of the mains or SMPS transformer and tuning caps. That leaky cap may heal under those conditions, but go BANG in the bank. Just my thoughts.
But then, why just caps with that voltage rating?
Last week I found a few 35v 1000u caps in the bin that were the same physical size as 25v 1000u of the same brand.
Are you lighting up Atlantis with that supply?
I've never considered multiple caps in parallel to have issues. But in this situation there is a huge energy supply. If one cap in the bank developed low leakage there's a huge current source right beside it. In a typical PSU the current available is limited by the inductance of the transformer and the rise time of the mains or SMPS transformer and tuning caps. That leaky cap may heal under those conditions, but go BANG in the bank. Just my thoughts.
But then, why just caps with that voltage rating?
Last week I found a few 35v 1000u caps in the bin that were the same physical size as 25v 1000u of the same brand.
Are you lighting up Atlantis with that supply?
My vote: You are building power supplies for the Large Hadron Collider Proton Beam Magnets...
For space electronics there is limitation to the amount of self-healing capacitors one can place in parallel. We never use electrolytics, but the theory is the same.
The limitation of capacitors to be placed in a single bank is expressed in Joules contained in the capacitors. There is also a minimum specified to guarantee proper self-healing, because the process actually needs a substantial amount of energy (about 2 - 3 Joules).
My guess is, if the Panasonics are of better quality, you might just be in the situation that you are lucky no single cap has induced failure yet.
Try to induce a failure in one of the panasonic caps and see if the self-healing process becomes destructive. We did similar tests to see if a capacitor bank fails when it is connected to a li-ion battery and results are interesting enough (no destructive failure)
Put a large diode in series between large capacitor bank and "device to be destroyed". Put a diode and a switch between "device to be destroyed" and power supply. Put a substantial overvoltage on the power supply to induce failure in the test-cap, and let the magic happen....
For space electronics there is limitation to the amount of self-healing capacitors one can place in parallel. We never use electrolytics, but the theory is the same.
The limitation of capacitors to be placed in a single bank is expressed in Joules contained in the capacitors. There is also a minimum specified to guarantee proper self-healing, because the process actually needs a substantial amount of energy (about 2 - 3 Joules).
My guess is, if the Panasonics are of better quality, you might just be in the situation that you are lucky no single cap has induced failure yet.
Try to induce a failure in one of the panasonic caps and see if the self-healing process becomes destructive. We did similar tests to see if a capacitor bank fails when it is connected to a li-ion battery and results are interesting enough (no destructive failure)
Put a large diode in series between large capacitor bank and "device to be destroyed". Put a diode and a switch between "device to be destroyed" and power supply. Put a substantial overvoltage on the power supply to induce failure in the test-cap, and let the magic happen....
i once repaired a Fluke voltage standard that had a large bank of 4700uf/50V caps in it's power supply. half of them had burned open, and some of the bad ones continued to arc internally, causing a large amount of random spiking in the output voltage. a lot of strange things tend to happen with capacitor banks.
Sounds like a pretty neat supply for a HIFI amp
Maybe you should consider using one per channel to avoid
crosstalk
Maybe you should consider using one per channel to avoid
crosstalk
- Status
- Not open for further replies.