Re: Toroidal transformer ratings?? What after being rectified?

The following is a example of a stepper motor power supply design.
For servo motors you would need a 50% larger margin of safety on
current rating...

Well...In my opinion your power supply is not to big, as i can
demonstrate...

First you should design for worst case, 3 motors moving.

This is how i would design supply...

3 motors 7amps each = 21 amps DC

67% current required for gecko = 14amps dc (.66667x21A)

The required AC current into a bridge rectifier capacitor combo is
1.8 x DC current (i think you missed this post i made). So 1.8 x 14A
DC = 25.2 amps AC from transformer......Hard to belive isn't it!!!


With that said, before i go on, i pulled a fast one. I read ahead
and looked at the link Les posted. Here is a page from it on the
amount of current through the transformer, bridge, and capacitor...

http://sound.westhost.com/power-supplies.htm#xfmr_series_r

Giving it a quick once over i see i was to conservative on 2 things.
To calculate AC current in transformer they used a number closer to
1.9 than 1.8....And in previous post i said ripple current in cap is
= to DC current...According to this article the ripple current is 1.6
x the DC current...I stand corrected....Sort of...SIGH...So many
conflicting rules...I just looked up the cap ripple current derating
in a power supply data sheet i have been using for years...It states
ripple current in cap is 2 to 3 times higher than output DC current,
not 1.6 as the web article says.....


OK so now we have to pick a capacitor value...Notice in the artical
the guy kind of side steps the issue and says 2000uF per amp...This
is because it depends on to many factors that are design
specific...Thats why we go with the gecko capacitor selection
formula...We are designing this supply to work with geckos, and the
mfg told us what cap size his product needs....

So we have a 34VDC supply (aprox) at 14ADC...
C= (80,000 x I) / E
(80,000 x 5) / 34 = 32,941 or about 33KuF MINIMUM cap size

We also need to calc our ripple current in the capacitor...I am going
to use the web article spec of 1.6 for derating...

1.6 x DC current.
1.6 x 14ADC = 22.4ADC of ripple current

Now Go to mouser and grab your ripple current charts i posted in
previous post and find a 33KuF cap at 22.4ADC ripple...MIN specs..

Here we are at mouser, looking at Vishay and Cornell Dubilier caps...
First lets narrow the search down and find a voltage range to look
at...Power supply rated at 34VDC can drift at least 15% or 39vdc or
so...

We see there are standard voltage caps at 50 and 75VDC...

In my mind i am thinking with 39VDC and regenitive braking to worry
about, that 50vdc cap is on the edge....I suspect with all the low
cost power supply designs around people are putting out, many would
use this voltage...But i would not...Even though there is 2X price
penelty to pay for security....My opinon, use the 75VDC cap...Notice
i am ignoring the surge rating here, a 50VDC cap has a surge rating
of 65VDC...I have not taken this in account because i have never seen
a spec on it...Do not know how long of a surge it can take or how it
affects life....

Next i will look at the 2 different capacitor series offered and look
for bang for the buck at 75VDC...I won't go into a lot of boring
detail of looking at different caps. But try to find a cap or
combination caps of at least 33Kuf and capable of handiling at least
22.4ADC ripple....

While selecting caps couple of things i do...I do not mix and
match...I will select a value of cap and use multiples of it. Like 4
pcs of a 10Kuf, not a 10Kuf and a 20KuF...I think the current shares
better this way...

And 1 problem you will run into...The values offered by mouser do not
always match the sheets from the mfg...So on some values you will not
have a ripple current spec...Only thing you can do is use the ripple
current value from the next highest value cap. If you are brave you
can try to interpolate the value...But be very conservative...

And finally only parallel caps, do not put them in series, unless you
really know what you are doing. Putting caps in parallel the uF adds
up and voltage rating stays the same...Series halfs the uF value and
doubles the voltage rating, and you have to play games with voltage
divider resistors on the caps...

OK, let me look at the mouser catolog now...Something in the 33Kuf,
22.4ADC range.

2 choices.

75-36DY183F075BC2A 18kUF 13.7A $18.00 2PCS 36KuF 27.4A $36.00
539CGS75V17000 17Kuf 9.8A $33.00 2pcs 34Kuf 19.4A $66.00

Of these 2 choices notice number 2 does not meet current spec...
So we use this one...

75-36DY183F075BC2A 18kUF 13.7A $18.00 2PCS 36KuF 27.4A $36.00


OK...We getting close...Still need a bridge rectifer though...So lets
go to the recton bridge rectifier pages...

We need a bridge rated at at least 14ADC and i see 15Amp
bridges...But once again this is on the edge...And it would probably
work fine with heatsink and knowing the fact that all 3 axis are not
usally used at same time...But i would not use it...As i said in a
earlier post, i usually select a current value 50 to 100%
bigger...Also i usually select a 400piv diode...Why, do i do this,
for many reasons, by going 2x on current rating i do not have to
worry about the surge current to the caps burning out the bridge.
with the high voltage rating i do not have to worry about inductive
kickback, or any kind of spike...I picked 400volt rating because it
is usually only 25 to 50 cents more. So lets make a chart of specs
and prices.

15amp 400piv 300Amp Surge $2.79
20amp 400piv 300Amp Surge $2.35
25amp 400piv 300Amp Surge $2.84
35amp 400piv 400Amp Surge $3.26

Notice the small price differance and 20 amp is less than 15 amp
(cause of ul rating)...

I would go with the 35 amp bridge, 50 cents more gets you 2x the
current you need and 8x the voltage rating...And look at the surge
current rating, 400 amps, this is also a magic number...Remember Jon
Elson's post, small transformers can put out 400 amps peak...Surge
problem solved...

Bonus information.

Wire sizes i would use for different current ratings.

10ga 30A
12ga 20A
14ga 14A
16ga 9A
14ga 6A
18ga 4A
20ga 3A
22ga signal wire only
24ga and smaller not recomended (fragile hard to work with)

More bonus information

Do not forget what the gecko power supply design manual says about
drives that are more than 18 inches from power supply. Or if you put
a fuse in the drive power line. You then need a small cap installed
directly in the drive power in terminals. A 470uF 75 Volt cap for
this example.


Wally






--- In CAD_CAM_EDM_DRO@yahoogroups.com, "turbulatordude"
<dave_mucha@y...> wrote:

>
> --- In CAD_CAM_EDM_DRO@yahoogroups.com, Jon Elson <elson@p...>

wrote:

> > R Rogers wrote:
> >
> > >Thanks to all who responded. Now, what happens to this amperage
> rating after it goes through a rectifier? I know the DC voltage

will

> increase by a factor of 1.416 due to the RMS. Does the amperage

stay

> the same or decrease due to the increased voltage?
> > >
> > >
> > This gets REALLY messy, fast. The problem is the rectifier is a
> > non-linear component,
> > and in this application the transformer is required to deliver
> short
> > pulses of high
> > current as it recharges the capacitor bank only at the voltage

peak

> of
> > the sine wave.
> > There are some rough rules of thumb, but they depend strongly on
> transformer
> > leakage inductance and the ratio of capacitor size to current

draw

> (or
> > the DC ripple
> > is another way of expressing it). Power supplies with very low
> ripple
> > are the worst,
> > because the cap is charged for the shortest duration. I've seen
> numbers
> > varying
> > from roughly 50% to 75% of the pure resistive rating for

capacitor

> input
> > filters.
> > Sometimes adding a small series resistance (in this case a

fraction

> of
> > an ohm)
> > somewhere between the transformer and cap can reduce the current
> peaks
> > without
> > seriously affecting regulation.
> >
> > Jon
>
>
> When I first started looking for components, I was thinking 21

amps.

> (3 motors, 7 amps each) but after a little contemplation, I

realized

> how the power moves.
>
> My motors are 1.7 volts, 7 amps. or roughly 12 watts. ( volts

times

> amps = watts)
>
> at 25x for the power supply, I was hoping for a 42 volts power

supply.

>
> working backwards, that is 42 divided by 1.414 or a 29 volt AC
> transformer. I had a 24 volt transformer so used that.
>
> I used a pair of the MPJA 10 amp transformers to get 20 amps

although

> I supposedly will never need more than about 13 amps (66% of full
> load for the Geckos)
>
> so... 24 volts times 20 amps = 480 watts after the transformer.
> then the loss of about 1.2 volts thru the bridge (23.8 volts) times
> 1.414 (peak voltage) = 33.6 volts.
>
> 480 watts at 24 VAC = 20 amps
> 480 watts at 33.6 volts = 14.2 amps, Quite a difference.
> and, 480 watts at 110VAC = 4.36 amps
>
> So, I needed to size wires according to the amp load at the voltage.
>
> As Jon mentioned, there are peak demands for power at the top of

the

> AC wave, but I did not considder that as part of my sizing.
>
> In reality, my machine runs one axis at a time so I could actually
> size for one motor and scale things back quite a bit. I didn't as
> the possibility I will run all three axis at some future time does
> exist.
>
> What I am unclear about is the 66% of rating for the power

supplies.

> using Gecko 201's, and a 7 amp motor, with a nameplate of 1.7

votls,

> I theoretically need 12 watts, 24 watts if both phases need to be
> included. But, at the higher voltage, the supplied wattage is

HUGE

> compared to the raw motor calculations.
>
> So the selection of correct amperage for a power supply still is
> confusing. Since it works, I can't conmplian too much, but still
> wonder if I am vastly over sized.
>
> Hope that helps.
>
> Dave