Power Supply Rectification and Smoothing

[[email protected]]

I'm building a power supply for a 36V, 500W DC motor to drive a small
lathe.

I have scrounged a transformer from an old microwave oven and removed
the 2000V secondary windings leaving the 240V primary windings intact
and undamaged (I will test with a megger before powering up.)

I will wind some 30A automotive flex around the secondary core and
test until I have around 36V (the rated voltage of motor and speed
controller)

I have found a bridge rectifier to use for the rectification of this
secondary 50Hz current. It is rated at 400V and 35A. Will this be
suitable? Is there a downside to using too large a BR?
This is about the only one that would appear to handle the rated 18A
for the motor. Is 35A rating enough for the BR?

Now to smoothing. I have calculated that I will need something like a
56,000 microFarad capacitor to reduce my ripple to 10%.

I'm not even sure if this will be smooth enough for my motor
controller which is a Chinese scooter controller here:

http://secure.oatleyelectronics.com/files/spc536notes.pdf

Anyway, I can't find a cap this size in Australia (I'm sure they exist
- they do in Digikey - very expensive but unavailable singly) but I
can find 50V electrolytics of 4700 microF for around 75c (US) each. If
I strung a dozen or so of these between the DC output terminals, would
I have any major problems? Yes, I would prefer 100V caps but the 50s
are all that are offered by this surplus goods vendor.

Otherwise, I was thinking of using 3 cheap 12V car batteries across
the DC terminals. I've heard these have ginormous capacitance, and
hence smoothing effect. They might be short-lived however, and be
false economy.

As these caps are so cheap, would it be of advantage to hook several
dozen up in series and parallel to increase capacitance and voltage
handling?

I'm learning furiously, but as a newbie, some things that are obvious
to the experts are rather hard to find out about, especially when you
are not quite sure of anything. jack

 

[[email protected]]

How did you calculate that ?

"Smoothing capacitor for 10% ripple,
C = (5 × Io) / (Vs × f)

C = smoothing capacitance in farads (F)
Io = output current from the supply in amps (A)
Vs = supply voltage in volts (V), this is the peak value of the
unsmoothed DC
f = frequency of the AC supply in hertz (Hz), 50Hz in the UK "

from:
C:\Documents and Settings\Administrator\Desktop\Power Supplies.mht

Your load is ~ 15A.

More like 20, isn't it?

on a 50Hz supply the period during which the cap is being
discharged is typicall ~ 7.5ms.

from delta V= I.delta t / C we get 15*7.5.10^-3 / 56.10^-3 = 2V *peak to peak*
of ripple. That'll have an RMS value of < 1V.

Try something like 22,000 uF.

That makes it cheaper, thanks

Any opinions of the cheap caps I can get?

Also, what do you think about that 35A 400V bridge rectifier?

jack

 

[[email protected]]

36V rms will be wrong. When rectified it will provide ~ 49V less what the
circuit loading and ripple reduces it to.

Thanks Graham, for simplicity, I had skipped the nitty gritty here
where I would adjust the secondary turnings to give me 36 or so Volts
of smoothed DC

I also doubt you'll be able to wind enough flex around the core to do that in
the first place. I suggest you try 'magnet wire'.

How many turns were you assuming? I have drawn the window to scale and
can easily fit 42 turns of this stuff in. As there are around 200
primary turns, a 30 something VAC output would require around one
seventh of this so about 30 turns.

Twelve gauge magnet wire seems to be made of unobtainium here, and I
think it might be difficult to wind into an already assembled
transformer core without damage. And anyway, it gives me an excuse to
buy 100 metres of some handy 30V auto wire

Try targeting around 28V. Rectification results in a DC supply voltage equal to
the *peak* value of the AC voltage you see (less rectifier voltage drop, loading
and ripple effects).

Thanks for that, Graham, makes the winding even easier. jack

 

[Eeyore]

I'm building a power supply for a 36V, 500W DC motor to drive a small
lathe.

I have scrounged a transformer from an old microwave oven and removed
the 2000V secondary windings leaving the 240V primary windings intact
and undamaged (I will test with a megger before powering up.)

I will wind some 30A automotive flex around the secondary core and
test until I have around 36V (the rated voltage of motor and speed
controller)

36V rms will be wrong. When rectified it will provide ~ 49V less what the
circuit loading and ripple reduces it to.

I also doubt you'll be able to wind enough flex around the core to do that in
the first place. I suggest you try 'magnet wire'.

Try targeting around 28V. Rectification results in a DC supply voltage equal to
the *peak* value of the AC voltage you see (less rectifier voltage drop, loading
and ripple effects).

Graham

 

[Eeyore]

Now to smoothing. I have calculated that I will need something like a
56,000 microFarad capacitor to reduce my ripple to 10%.

How did you calculate that ?

Your load is ~ 15A. on a 50Hz supply the period during which the cap is being
discharged is typicall ~ 7.5ms.

from delta V= I.delta t / C we get 15*7.5.10^-3 / 56.10^-3 = 2V *peak to peak*
of ripple. That'll have an RMS value of < 1V.

Try something like 22,000 uF.

Graham

 

[Eeyore]

"Smoothing capacitor for 10% ripple,
C = (5 × Io) / (Vs × f)

C = smoothing capacitance in farads (F)
Io = output current from the supply in amps (A)
Vs = supply voltage in volts (V), this is the peak value of the
unsmoothed DC
f = frequency of the AC supply in hertz (Hz), 50Hz in the UK "

from:
C:\Documents and Settings\Administrator\Desktop\Power Supplies.mht

A curious 'cookbook' equation. Mine is from electronics first principles.

More like 20, isn't it?

20A x 36V = 720 W !

15A x 36V = 480W.

That makes it cheaper, thanks

Any opinions of the cheap caps I can get?

I'd look further. Can't Dick Smith ? Electronics offer these ?

Also, what do you think about that 35A 400V bridge rectifier?

Bridge rectifiers are rated according to the average forward current which is your
load current. Therefore it should be just fine.

Graham

 

[Phil Allison]

Any opinions of the cheap caps I can get?

** WES Components at Ashfield have very nice 15,000 uF @ 100V bolt
terminal caps for $ 28.

With a trade order.




....... Phil

 

[[email protected]]

A curious 'cookbook' equation. Mine is from electronics first principles.

Well, yes, I assumed that. but is it not useful? It's within an order
of magnitude. What would be the downside of using twice the
capacitance here?

20A x 36V = 720 W !

That's about right. The motor is not perfect and its 500W output will
require something like that flat out.

15A x 36V = 480W.



I'd look further. Can't Dick Smith ? Electronics offer these ?

Yes, but at around $5 each. But he doesn't have these. Dick's a bit of
a waste of space for many things. Altronics have exactly the same spec
for $3.90 each. What is the problem you see with the cheapies?
When I find how smooth the current needs to be for my DC motor
controller, I will know exactly how many of these I will need.
Is the 50V spec a bit small? Could I run these caps in series and
parallel to increase the voltage handling?

Bridge rectifiers are rated according to the average forward current which is your
load current. Therefore it should be just fine.

Thanks, I wasn't sure how much safety factor was needed here. There
will surely be a current surge when the whole shebang gets turned on.
Perhaps if I turn on the mains before I energise the motor. In fact, I
thing the controller won't start until the speed control is set to
zero first. Thanks for your help, here, Graham, regards, jack

 

[PhattyMo]

"Smoothing capacitor for 10% ripple,
C = (5 × Io) / (Vs × f)

C = smoothing capacitance in farads (F)
Io = output current from the supply in amps (A)
Vs = supply voltage in volts (V), this is the peak value of the
unsmoothed DC
f = frequency of the AC supply in hertz (Hz), 50Hz in the UK "

from:
C:\Documents and Settings\Administrator\Desktop\Power Supplies.mht



More like 20, isn't it?


That makes it cheaper, thanks

Any opinions of the cheap caps I can get?

Also, what do you think about that 35A 400V bridge rectifier?

jack

Rectifier should be fine.It will need to be mounted on a heatsink.
50V rated caps should be fine,just make sure your raw DC voltage doesn't
go above their ratings.
Paralleling a bunch of smaller caps might actually be even better than
one large cap,This tends to have a lower ESR than a single cap of the
equivalent value. (Having a low ESR might not be a big issue in your
application,but it can't hurt.)

I managed to *barely* fix 2x windings of 10AWG stranded wire on my MOT
core,for dual 12V outputs.
http://www.users.qwest.net/~ptaylor/Electronics/MOTpowersupply/MOTPSpage.html
It's ugly,and there have been some changes since I wrote that page
up,but the basics are all there.

 

[[email protected]]

** WES Components at Ashfield have very nice 15,000 uF @ 100V bolt
terminal caps for $ 28.

With a trade order.

Thanks Phil. I might need two, depending on how smooth these cheapie
DC motor speed controllers demand their input. That's getting a bit
pricey, when markup and freight are added. jack

 

[Hammy]

Yes, but at around $5 each. But he doesn't have these. Dick's a bit of
a waste of space for many things. Altronics have exactly the same spec
for $3.90 each. What is the problem you see with the cheapies?
When I find how smooth the current needs to be for my DC motor
controller, I will know exactly how many of these I will need.
Is the 50V spec a bit small? Could I run these caps in series and
parallel to increase the voltage handling?

You may want to determine the Line ripple current (100 or 120Hz)!
Unless you want to watch your cap cook and spew electrolyte all over.
I doubt a cheap $5 cap will handle the rms ripple current at your 20A
output you'll be looking at large can screw terminal types.

 

[[email protected]]

You may want to determine the Line ripple current (100 or 120Hz)!
Unless you want to watch your cap cook and spew electrolyte all over.
I doubt a cheap $5 cap will handle the rms ripple current at your 20A
output you'll be looking at large can screw terminal types.

Thanks, Hammy, this is the sort of thing I need to know.

If I had say 24 of these 75c 4700uF 50V electrolytics arranged across
the rectifier output as 12 sets of series pairs, with a fan directed
towards them would this $20 outlay have any chance of success?
I could keep an eye on the temps with my handy radiation thermometer.

jack

 

[[email protected]]

I took a chinese scooter speed controller apart... little plastic box
with a couple of to220 mosfets in it.... I dont think that thing cares
about the quality of the voltage going to the motor... it just chops
it. A couple caps with a big L in series will smooth out the volts...
call it a filter choke if you want... maybe just the secondary of a
transformer would work?

Yes, there are many differnt models of controllers.
I already have several. A couple of the cheap 100W, and a couple of
the much better 350W jobs. The one I will be using on the 500W motor
is the Rolls Royce with reverse and all.

I understand they operate with PWM and current limiting. They are
designed to run off batteries, and as they use minimum spec
components, I wondered whether any ripple would stress anything.
The motors can withstand gross overloading for short periods, but what
lets out the magic smoke first is the controller if you take it much
above the rated voltage. jack

 

[Phil Allison]

<[email protected]>



** Why don't you just.......

**** OFF TROLL !!

Before you electrocute your stupid self.







......... Phil

 

[Daniel Mandic]

I'd look further. Can't Dick Smith ? Electronics offer these ?

Graham

Hi Graham!


Nice to see here are helpful ppl on the work.

By the way. My former School I showed you, is from 1907. I was far away
that.



Kind regards,

Daniel Mandic

 

[Phil Allison]

"Daniel Manic Moron "



** Piss off - FOOL.




....... Phil

 

[Daniel Mandic]

"Daniel Manic Moron "

are you talking about the past? That isn't present, indeed.
Manic, ok, but

** Piss off - FOOL.

you deserve nothing Phil, as you serve for nothin' and nobody. Nobody.

...... Phil

don't make me laugh...



Best regards,

Daniel Mandic

 

[BobG]

I took a chinese scooter speed controller apart... little plastic box
with a couple of to220 mosfets in it.... I dont think that thing cares
about the quality of the voltage going to the motor... it just chops
it. A couple caps with a big L in series will smooth out the volts...
call it a filter choke if you want... maybe just the secondary of a
transformer would work?

 

[[email protected]]

If I had say 24 of these 75c 4700uF 50V electrolytics arranged across
the rectifier output as 12 sets of series pairs, with a fan directed
towards them would this $20 outlay have any chance of success?
I could keep an eye on the temps with my handy radiation thermometer.

Just to elaborate and check my calculations, I'll draw an ascii
diagram of my propsed circuit.

+36V-----------------------------------------------------------------------
| | | | | | | | | | | |
= = = = = = = = = = = = | |
| | | | | | | | | | = = =
= = = = = = = = =
| | | | | | | | | | | |
0V---------------------------------------------------------------------------

Each cap is 4700 uF so each pair will be 2350 uF across the output
leads. Twelve of such pairs will be a total of 28,200 uF across the
output. Will each cap be exposed to only 18V? And the ripple current
shared between the 12 pairs equally?

Does this make sense? jack

 

[Eeyore]

Well, yes, I assumed that. but is it not useful? It's within an order
of magnitude. What would be the downside of using twice the
capacitance here?

Unnecessary cost for one !

Graham