Searching \ for '[OT] switching DC/DC converters (was capacitive sw' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/power/priswitch.htm?key=switching
Search entire site for: 'switching DC/DC converters (was capacitive sw'.

Exact match. Not showing close matches.
PICList Thread
'[OT] switching DC/DC converters (was capacitive sw'
1999\12\30@092630 by Byron A Jeff

face picon face
>
> Hi Byron,
>         you have not mentioned how you are getting your 12v and 5v?  I used
> inductive switching for the + rails and overwound an extra coil to get my
> -ve by stealth :)  In general I have found that inductance is not very
> critical (at least when using lm2576's :)
>
> have a look at http://members.xoom.com/andrewmuck/PSU.htm

I took a look at your page my first time through the design process. There
simply wasn't enough information to figure out the design.

As to your first question, I used a linear regulator with linear pass
transistors.

And honestly the way the datasheets specify it it seems that the inductance
is critical. Which is why I've been avoiding them like the plague.

Anyway I've decided to take a crack at the switchers using the LM2576's
for voltage inversion. Care to explain in more detail the design of the
external switch and how you determined the inductance?

Also I noticed from both the Radioshack.com and digikey catalogs that
power inductor amp ratings drop as the inductance increases. How does this
jibe with 5V @ 10A when the rated current of everything I've seen above
200 uH is 2A or less Idc Max?

I scoured the datasheets and the web for switcher information. I find it
anything but simple.

BTW how does one measure inductance anyway. I have dozens of old PC power
supplies with inductors on them. Any way to put them to use?

BAJ

1999\12\30@150645 by quozl

flavicon
face
On Thu, Dec 30, 1999 at 09:25:33AM -0500, Byron A Jeff wrote:
> BTW how does one measure inductance anyway. I have dozens of old PC power
> supplies with inductors on them. Any way to put them to use?

If I was truly desperate, I would tie them up in a tank circuit with a
transistor oscillator against a known capacitor to see what frequency
the thing runs at.

Of course, I'd need a frequency meter (which I don't have) and a
capacitance meter (which I also don't have).  I could probably measure
the capacitance with a volt meter and a known resistance if I was quick
enough or made myself a PIC interval timer.

I like the idea of buying them to save that time.

--
James Cameron   spam_OUTquozlTakeThisOuTspamus.netrek.org   http://quozl.us.netrek.org/

1999\12\30@215416 by Chris Eddy

flavicon
face
I was testing inductors by setting up a FET switch circuit ( with flyback
diode), and hitting the FET with a narrow adjustable pulse (frequently a triang
wave, pot, and comparator).  Then you measure the current flowing through the
inductor with a low value resistor and an instrument amp (use some thought
here.. BW of inst amp?).  There is a basic equation here, which is E=L(DI/DT).
Rearrange for L, and you have the ticket.  Measure the slope of the I, you know
the DC supply attached, and you are in business.  As you increase the length of
on time, the ramp rises higher and higher, and finally the inductor saturates.
This sat point is not very sharp on some core materials, and so it may be more
of a Bleeeaaaahhh...  instead of a knee.  Watch the parts at this sat point,
they tend to warm up quickly, and where there is smoke, there is..  Anyhow,
that gives you two main parameters of mystery inductors.

Then I got a used impedance meter and never looked back.  I don't get
saturation, though.

Chris Eddy

James Cameron wrote:

> On Thu, Dec 30, 1999 at 09:25:33AM -0500, Byron A Jeff wrote:
> > BTW how does one measure inductance anyway. I have dozens of old PC power
> > supplies with inductors on them. Any way to put them to use?
>
> If I was truly desperate, I would tie them up in a tank circuit with a
> transistor oscillator against a known capacitor to see what frequency
> the thing runs at.
>


'[OT] switching DC/DC converters (was capacitive sw'
2000\01\11@215459 by Andrew McMeikan
flavicon
face
Hi Byron,
   I started building my PSU before I got an inductance meter, I just
guessed, wired it up and tried to pull lots of amps (this was just for a 5v
rail at first), I got 10 Amps ok but the coil got a little warm (using 1mm
wire) so I re-did it using 1.2 mm  and a bit less turns and it was fine.  At
this stage I forked out the money for a good meter.  Instead of my 300uH I
expected, I had wound 15mH !!  The circuit did not care.  Here is my new
rule of thumb (or should it be little finger?) - find a torroid you can just
stick your little finger through and each turn is about 15 to 20uH.
Depending on your finger size and the composition of your torroid YMMV .

   Because I am using external switching transistors rather than just using
the LM2576 on its own as only a small proportion of  the total switching
current is provided by it, because the transistor sees the acition it looks
like you can do lots of 'bad things' like sloppy wiring guesswork caps,
almost anything and get away with it.  My only real comment is to use a good
diode (schotkey 16A 45V in my case) and fast transistor (2N5038).  I tried
using a slow (mje2955) with an LM2577, and that killed the LM2577 :(

The 2A limit in the data sheet is just for the LM itself, use an external
transistor for current gain and that is no longer a limitation, if you do
the maths you could even use the 2A limit multiplied up to a new limit (I am
trying to have some current limit on a new 200W design I'm working on)

The datasheet example seemed pretty simple, I just didn't take it as being
the be all and end all, ideally you should extrapolate the curves and use
their formulars, but I just grabed some ferrite torroids wrapped some heavy
wire around them and used it, the worst that can happen is efficency goes
down.  When the 15mH coil is running you can hear the (slight) sound change
at about 4~5Amps when it goes from dis-continues to continues mode, so I
would say that under 5Amps my effiecency would drop off.

I have seen lots of people using MAX787's and would be interested hearing of
any success stories with boosting their output.

I hope my site has some better details now?  I guess I should put something
in about winding an inductor.
Here is a quick and dirty to get you started.
2m of 1.2mm enammeled wire
some teflon plumbing tape
a ferrite torroid

practise with the teflon tape, get about a meter length and feed it through
half way, then wrap one half, when done wrap the other half.  this is a lot
harder than doing the wire so if that went ok you now have a nice smooth
teflon covered core that wont scratch the enamel off the wire.

put the wire halfway through, loop a turn on keeping it nice an close to the
core.  Use up all the wire.

you now have several mH of inductor.  This should let you play happily.  If
you want efficency use my 'rule of finger' above or spend some money on a
meter (worth it in the long run) I have used prewired 100uH coils but I find
the winding proccess adictive and pleasing.

you may as well try out some of those inductors in those old PC suppllies,
it really is easy to experiment with once you get started.

   cya,    Andrew...

http://members.xoom.com/andrewmuck/MP3.htm
{Original Message removed}

2000\01\11@235235 by Byron A Jeff

face picon face
>
> Hi Byron,

Hi. Glad you got back to me.

>     I started building my PSU before I got an inductance meter, I just
> guessed, wired it up and tried to pull lots of amps (this was just for a 5v
> rail at first), I got 10 Amps ok but the coil got a little warm (using 1mm
> wire) so I re-did it using 1.2 mm  and a bit less turns and it was fine.  At
> this stage I forked out the money for a good meter.  Instead of my 300uH I
> expected, I had wound 15mH !!  The circuit did not care.  Here is my new
> rule of thumb (or should it be little finger?) - find a torroid you can just
> stick your little finger through and each turn is about 15 to 20uH.
> Depending on your finger size and the composition of your torroid YMMV .

Well I tested with a 300uH coil. Got so hot it started to burn the cloth
cover. I decided to try a coil with the right amperage, so I bought a 50 uH
9.5 Amp coil. Works like a champ.

>
>     Because I am using external switching transistors rather than just using
> the LM2576 on its own as only a small proportion of  the total switching
> current is provided by it, because the transistor sees the acition it looks
> like you can do lots of 'bad things' like sloppy wiring guesswork caps,
> almost anything and get away with it.  My only real comment is to use a good
> diode (schotkey 16A 45V in my case) and fast transistor (2N5038).  I tried
> using a slow (mje2955) with an LM2577, and that killed the LM2577 :(

I used a schotkey 8A and couldn't locate a 2N5038, so I substituted a
Motorola 15003 instead. Works like a champ.

In fact I've finished the supply. I didn't cheat like you did and actually
put two more switchers for the negative voltages. I also ended up wiring
the 15V supply and regulating it with a 7812 regulator.

The machine boots and works fine. The only problem is that the sound really
bites. I may end up installing a better sound card using a horizontal
socket.

My next task is to get a keyboard and display working. This is where the
PIC comes in. I have a LED board that uses National DL1416 smart Alpha
displays (24 characters). Plan to hook it up to a PIC along with switches
and a MAX232, using the serial port as my connector.

Thanks for all the help.

BAJ

2000\01\12@062704 by Russell McMahon

picon face
>expected, I had wound 15mH !!  The circuit did not care.  Here is my new
>rule of thumb (or should it be little finger?) - find a torroid you can
just
>stick your little finger through and each turn is about 15 to 20uH.
>Depending on your finger size and the composition of your torroid YMMV .


Remember that inductance is proportional to the square of the number of
turns.
Your rule should read -

>"  Find a torroid you can just
>stick your little finger through, count number of turns, square this value
and multiply by 15 to 20uH.
>Depending on your finger size and the composition of your torroid YMWV*



* Your Mileage WILL Vary :-)


RM .

More... (looser matching)
- Last day of these posts
- In 2000 , 2001 only
- Today
- New search...