mains dropping capacitor for 6A

[Simon]

OK don't shout at me I only need to drop 16V although my calculation puts that at 1200uF. What voltage  capacitor do I need ?

[MF-jockey]

Nobody can give you a serious answer to such a poor question.
More information about your application is needed!

[Simon]

I need to power 8 13V fans off the mains, this means a 600W power supply, or put them in series, put a diode in series and I get 120V pulsing at 50Hz, I need to be at 104V so need to drop a further 16V but this is with 5-6A going through

[Simon]

for me it works out around 1000uF I actually only need 5A so 16/5 = 3.2R

The capacitance for this is: 1/(2 x pi x 50 x3.2) = just under 1000uF so if I had a 50V 1000uF it would act as a lossless 3.2R resistor to take off my excess 16V. Or course I need a zener across it to make sure it does not go over voltage when the fans start up (it should see on average 32V).

[bktemp]

Why do you only use one half cycle? The fans will probably not like that. The next problem is, a capacitor won't work, because by adding a diode you have no longer ac but dc .
Give us more information, there is probably a much easier and less dangerous solution. Maybe a simple transformer will be enough.

[Simon]

Oh yes a simple chuncky 600W transformer will be fine, but they want compact. A toroidal may well do it if it's small enough. I have a 200VA transformer at home that looks beefy enough without considering something 3x the power.

If I use one diode I will have pulsating DC, i suppose I need something to discharge it again. The other option is literally 80W of resistor, it will be force cooled anyway.

I can't see DC fans having a problem with pulsating DC.

Yes the one big drawback is that every fan is live. I'm yet to find out much detail about the installation so I'm just brainstorming at the moment.

[Ian.M]

Psssssst-BANG is the most likely result. 

Non-polarised caps with a 6A ripple current rating are likely to be bulky and expensive.

Connecting multiple DC fans in series without a bulk decoupling cap + clamping Zener across each one is *NOT* a good idea.   During commutation their instantaneous load current is likely to vary considerably, and it will be very easy to exceed the max Vdd spec of the BLDC controller chips inside the other fan motors.  You can usually get away with it for a while with only two fans, but more is asking for trouble.

[Simon]

these are brushed fans, if we went to the expense of BLDC fans they would be straight 240V AC which I am also looking into.

[bktemp]

Oh yes a simple chuncky 600W transformer will be fine, but they want compact. A toroidal may well do it if it's small enough. I have a 200VA transformer at home that looks beefy enough without considering something 3x the power.

Did you search for a AC rated 1200uF cap? It won't be much smaller than a 600W SMPS...
You do not need a 600W transformer: It must only provide 16V at 6A = 100VA. Wire it in series with the mains voltage to remove 16V.

I can't see DC fans having a problem with pulsating DC.

They will vibrate at 50Hz. It will make acoustic noises and the vibrations might damage the bearings.
Why don't you use a bridge rectifier?

[Simon]

the idea of the single diode was to remove half the voltage and get down to 120V easily.

I've not heard of using a transformer before, will it still take 16V out if it is unloaded ? that sort of implies that transformers use their rated current wether they are used or not or are you suggesting that it will just act as an inductor in which case that is what i need. If I use a transformer and bridge rectify i now have 136V and 680W to remove, not so small anymore.

The next best thing is bridge rectify and PWM at say 1KHz at 43%

[bktemp]

the idea of the single diode was to remove half the voltage and get down to 120V easily.

A diode does not half the voltage! It halfs the power. I assume you have 240V. A diode reduces the voltage to 240V*0.707=170V.

Because of that, a transformer will be much bigger: ~420VA.

The next best thing is bridge rectify and PWM at say 1KHz at 43%

This is probably the best solution.

[Zero999]

I wouldn't worry about the extra voltage.

The half wave AC waveform could be more of a problem though.

How about using an autotransformer to get 120V (this will be half the size of a 600VA transformer) and run the fans off a bridge rectifier.

[Ian.M]

For brushed fans, its not so bad, but I would expect a high risk of insulation breakdown on the armature as the brushes wear.  The whole idea is a nasty hack and unless the fans are installed so it is IMPOSSIBLE to access them with power applied, IMHO is unsafe.   The interlocks and additional enclosures required for safety are likely to cost more than a proper SMPSU or the difference in price for mains voltage AC fans.

[Simon]

there is a 24V version of the same fan so that makes life much easier. A zenner and small filtering capacitor could be put across each fan.

I'm not too concerned about 15V instead of 12V or 30 instead of 24V but we do have a history of problems with the supplier trying to wriggle out of warranty claims because they only rate an automotive fan to 13 or 26V which is mad - bloody iti's

[grumpydoc]

A diode does not half the voltage! It halfs the power. I assume you have 240V. A diode reduces the voltage to 240V*0.707=170V

Surely it will just half-wave rectify and you will still have the peak voltage present - 240*1.414=340V, not 170.
If you stick a capacitor after that what you get depends on the load and value.

I'm afraid it all sounds a bit dodgy not least from the safety angle, is there scope for an SMPS or mains fans.

If you absolutely must go with this idea it should work to put in a cap to drop the voltage, followed by a full wave bridge, smoothiing cap and then your fans in series. Don't forget the "top" one is going to have 100V WRT ground so safety is still going to be a consideration - especially as someone looking at a fan with a 12V label isn't going to be expecting that sort of voltage.

[T3sl4co1l]

I doubt low voltage fans are double insulated approved, either.  (Though they might pass as such.)

What's wrong with a bigass turbine and ducts?

Tim

[Simon]

yes I will probably get more power than I need with one diode, hadn't considered it like that. I have just been finding out more information on the project and spoken to a supplier so I think a direct 240VAC fan will be used.

[amyk]

A 240V mains fan would be the best option here - a single one could probably replace several lower-voltage fans and not need as much current.

Non-polarised caps with a 6A ripple current rating are likely to be bulky and expensive.

They actually do make them that large... and you're right about "bulky and expensive": http://www.mouser.com/ProductDetail/WIMA/DCP6I07103EG00KS0F/?qs=sGAEpiMZZMv1cc3ydrPrFxHYOjNyEwr5YvxDbo6biTg%3d

[Simon]

sounds like it would need to be a smaller value anyway if I'd need to take as much voltage as people suggest. 24V fans do make it needless anyway and I'm hoping for a mains option to come up.

[Andy Watson]

yes I will probably get more power than I need with one diode,

It's not going to work with one diode - the capacitor will simply charge-up.
This idea has disaster written all over it - do yourself a favour and step away from the bodge.

[Dave]

A diode does not half the voltage! It halfs the power. I assume you have 240V. A diode reduces the voltage to 240V*0.707=170V

Surely it will just half-wave rectify and you will still have the peak voltage present - 240*1.414=340V, not 170.
If you stick a capacitor after that what you get depends on the load and value.

Eh, no.
He was speaking of RMS voltage. The RMS voltage of half-wave rectified mains (which is 230V in UK) is 163V. Do the math, it checks out.

AC dropper caps only work when you only have AC currents in your circuit. This is not a trait of a half-wave rectifier.

The capacitance for this is: 1/(2 x pi x 50 x3.2) = just under 1000uF so if I had a 50V 1000uF it would act as a lossless 3.2R resistor to take off my excess 16V. Or course I need a zener across it to make sure it does not go over voltage when the fans start up (it should see on average 32V).

Repeat after me: A capacitor is not a resistor.
There are phase angles involved, something which you clearly haven't taken into account.

Wiring fans in series is a bad idea. A fan is not a predictable load.
Let's say one of the fans gets a bit of crap into the bearings and gets slightly more load. The speed of the fan will drop, reducing EMF, the overall voltage on that fan will drop, meaning every other fan will get a voltage boost. Do you see the problem?

Simon, hate to break it to you, but you clearly don't have the experience to work on a circuit like this. Just get a pre-made SMPS (something like a 12V Meanwell with a trimpot) and do it safely, or switch the fans with ones rated for mains voltage.

[grumpydoc]

A diode does not half the voltage! It halfs the power. I assume you have 240V. A diode reduces the voltage to 240V*0.707=170V

Surely it will just half-wave rectify and you will still have the peak voltage present - 240*1.414=340V, not 170.
If you stick a capacitor after that what you get depends on the load and value.

Eh, no.
He was speaking of RMS voltage. The RMS voltage of half-wave rectified mains (which is 230V in UK) is 163V. Do the math, it checks out.

Ok, true enough but the peak is still 170  325-357V.

AC dropper caps only work when you only have AC currents in your circuit. This is not a trait of a half-wave rectifier.

Agree, but you put the cap before the rect - this seems normal in consumer electronics with transformerless supplies, obviously at less current than 6A though.

I don't think any one has suggested this was a good idea and it sounds like Simon has seen sense and found some mains fans to use.

[Dave]

He was speaking of RMS voltage. The RMS voltage of half-wave rectified mains (which is 230V in UK) is 163V. Do the math, it checks out.

Ok, true enough but the peak is still 170.

No, the peak is 325V.

Agree, but you put the cap before the rect - this seems normal in consumer electronics with transformerless supplies, obviously at less current than 6A though.

Before a bridge rectifier, yes. Before a half-wave rectifier, absolutely not.

[Simon]

Dave

No I don't have much experience with bodges, I tend to avoid them too hence as already stated a straight 240V option is under way. I just thought it worth thrashing out every concept.

The air will be fairly clean as they are drawing air through what amounts to a guard but yes that still does not account for variations in life span and i have reiterated continually that a 240V option is best. We don't really have room for a power supply anyway.

[grumpydoc]

Ok, true enough but the peak is still 170.

No, the peak is 325V.

Oops - typo; I did say 340 earlier (assuming 240V mains, in fact as mains can be 253V then the max peak would be 357V)

Agree, but you put the cap before the rect - this seems normal in consumer electronics with transformerless supplies, obviously at less current than 6A though.

Before a bridge rectifier, yes. Before a half-wave rectifier, absolutely not.

Again, that's what I suggested - seen a single diode in crappy low current consumer stuff though.