Another Variable power supply

[PEU]

is this circuit correct?

I assembled it from design examples in the LM350 datasheet and other resources, the goal is to have a regulated power supply without the 1.2v lower range imposed by the LM350. Thanks

Pablo

 

[georges80]

Your bridge rectifier is wrong - you have the -ve going to the LM350 and the +ve going to the LM7910.

I think an output capacitor on the LM350 wouldn't hurt.
Also a cap on the input and output of the LM7910 would be a good idea.

You'll want some substantial heatsinking on the LM350 if you pull lots of current at the lower output voltage settings.

george.

 

[PEU]

the anodes (+) go to the lm350 and the catodes (-) to the 7910 that is not correct?

yes capacitors can be added to both parts of the circuit.

My main concern is if the design is correct, here is the revised schematic:

Pablo

 

[georges80]

No, the cathodes go to the lm350 and the anodes to the lm7910. The cathodes of the diodes will be where the +ve connection is. I think a couple of uF is plenty on the input of the LM7910 - since it's just feeding low current to the reference diode.

With a big cap on the out of the LM350 - you may need to implement the diode protection stuff around it as per the LM350 datasheet.

george.

 

[PEU]

Ok, I just corrected the previous schematic as per your suggestion Georges, what do you think ? /ubbthreads/images/graemlins/blush.gif

Should it work as expected?


Pablo

 

[georges80]

I'd make the input cap to the LM7910 bigger as per one of the datasheets I saw, at least a couple of uF. Check the LM7910 datasheet and see if they recommend a minimum output capacitor value - some regulators require a certain about of capacitance to be stable.

I would also size up the input cap to the LM350 - you had about 220uF I think before. You need a pretty big cap there to smooth out the rectified AC - the size of the cap will be dependent on how much current you plan to pull from the LM350. The input voltage (across the input capacitor) will sag at higher current draw when the AC is cycling to to zero volts - the input cap is there to 'fill' in that part of the AC waveform.

With a larger input capacitor on the LM350 - make sure your bridge rectifier is rated to handle the current surge as the input capacitor is first charged up.

george.

 

[rdshores]

Another similar circuit found here....
http://www3.telus.net/chemelec/Projects/Power-Supply/Power-Supply.htm

 

[PEU]

revised again as per your suggestion Georges !!!, I checked the lm79xx and the LM350 datasheets and these are the suggested values for the capacitors, here is the revised schematic (I removed the previous ones to avoid confusion)

Thanks

Pablo

 

[jtr1962]

Bigger caps, especially on the LM350 output. This will cut down on the output ripple. How much current can the transformer continuously deliver? I can suggest appropriate values for C2 based on that. I'd say C3 is OK as is, but maybe I'd use something larger, like 100 uF, just for a safety margin, and to provide a lower ESR. A smaller ceramic or tantalum cap, around maybe 1uF, in parallel with the larger electrolytics wouldn't hurt either. The reason for this is to filter high-frequency EMI. And I might also suggest putting a 1uF tantalum or ceramic between the potentiometer wiper and ground. Any noise picked up on that line will filter through to the output.

 

[dat2zip]

You can use this general rule of thumb formula for computing C2.

1000uF = 1V/A.

What this means is, if you had a 1000uF capacitor and your load on the capacitor was 1A you would get 1V of AC ripple on the capacitor. 200uF is 1/5 that, so the AC ripple with 1A would be 5X more.

Remember, to subtract the AC ripple component from your rectified DC component and ensure that the lowest peak value is still higher than the Vin min of the regulator.

As an example: Assume your max Vout is 5V. I'll assume the LM350 Vin to Vout minimum is 2V. Thus, you would need 5V + 2V on the input capacitor at all times to keep the linear regulator happy. If the AC rectified voltage was 7V, and you hooked up a 1A load, you would have 1V of ripple. That means the AC component of the Capacitor will dip down ~1V or 7V - 1V and would be below the 7V computed above as the minimum input. In this case you could raise the DC component on the capacitor (different transformer), raise the capacitance to lower the AC ripple or a combination of the two above.

The higher the AC ripple is also induces heat build up in the capacitor. In general, try reducing this value as much as practical.

Wayne

 

[MrAl]

Hi there Pablo and other CPF members,

In addition to the other great posts i'd like to
add the following....


Pablo:
There are a few problems :-)

First, the input voltage is too high. The max differential
voltage the ic can take is 35 volts. This means when you
adjust the output near zero volts the input will see the
full peak output from the rectifier diodes, which can be
very high. A better limit for input rms voltage is 20
volts. This will give a nice safety margin with high
input line also.

Second, the main cap, C2, needs to be around 10,000uf to
avoid output ripple at high output voltages. Since peak
current will be quite high in the caps, a good idea would
be to use ten 1,000uf caps in parallel. This is common
practice in these kinds of power supplies. Remember also
that if you dont have enough filtering you'll get ripple
output and if you're not looking at it with a scope
you wont know it's happening and it could interfere
with the testing of some device you've got connected--
might make it look better or worse than it really is
just because of the ripple.

Third, the negative regulator has to put out almost 20ma
so the input cap for that (C3) will have to be about 220uf.

Fourth, the output cap could be aluminum if you increase
it to 25uf. You'll need to add a protection diode across
the LM350 if you do this also.

Fifth, there's no mention of the part number of the rectifier
diodes, which have to handle a significant current flow.
I'd use 3 amp diodes and never 1 amp devices for 3 amps out.
Of course the diodes for the negative regulator can still
be 1 amp devices.

Sixth, the 680 ohm resistor is too high of a value to
properly bias the LM385 when the pot is adjusted to
zero ohms (for zero volts out). This is a mistake on
the LM350 data sheet and a better value is 470 ohms.

Seventh, there's also no mention of the heat sink surface
area, nor the package type of the LM350. The TO-3 package
is better, but more expensive, and to handle the power
at low output level and full current i would use a heat
sink with an area of at least 30 square inches, but 45
even better.

Take care,
Al

 

[vicbin]

Hi all, /ubbthreads/images/graemlins/smile.gif

Mr.Al, if you're still remember in my other Adj. PS thread, with max. load approx. 3-4 Amps with 35VAC tranformer output, do I still need a huge cap to reduce the ripple (fyi, this PS uses buck switcher as pregulator with linear as final) ? I used 3 x 6800uF/100V caps, is this enough ?

Sorry Pablo, I don't mean to hijack this thread but I think it is still related.

Regard,
Vic

 

[PEU]

no prob Vic, just put a link to your PS thread for our (my) enlightment /ubbthreads/images/graemlins/smile.gif

Edited Later: Vic, your project is better than mine!!! and is not much difficult, do you have the final circuit you used? I'm also reading the forum where the circuit is hosted, nice resource!


Pablo

 

[vicbin]

Hi Pablo,

Yes, I got working one but with a lot-lot of modifications made on the PCB which is looks like a huge mess, especially on the voltage reference module.

Please read through all the thread since we came out by adding a switching preregulator module at the original circuit to reduce disspation on the linear part (yes, it makes alot of difference, from a truck sized heatsink without the pregulator to small CPU heatsink).

I'm planning to make a new one soon since the prototy....er..... /ubbthreads/images/graemlins/grin.gif I mean the 1st one already gone as a gift to my friend an as a return he promised me an Art Deco wooden box for my second one. /ubbthreads/images/graemlins/blush.gif

Of course there is still a lot of "consultation" needed to finalize the better one. All of this depends on my great sensei Mr.Al. /ubbthreads/images/graemlins/grin.gif

PS Circuit with Adjustable Constant Current Mode

Regards,
Vic

 

[MrAl]

Hi again,

Pablo:
Are you still going to build the previous power supply
with the LM350 or no?

Vic:
Do you have a link to the entire power supply yet, like
with the switcher too? I was wondering how the final
circuit looked too now since we talked about this so
long ago. Thanks :-)

Take care,
Al

 

[PEU]

Hi, Al

GOOD QUESTION, I built the one with the switching pc supply, it works just fine!, but I miss the other features of my bigger and older one, so after looking around for more parts I found a 30+30/220 (3amps) transformer, a voltmeter using intersil 3161/3162 and tough of building the one in my schematic.

But Vicbin kind of changed my mind with this design/link, I readed the thread where you discussed with him and it sounded just a great supply to build.

In that thread I saw how it was improved and liked it, so, to answer your question, if you have the schematic completelly debugged, I would like to build it, the most expensive parts are already at hand, and the others arent that expensive /ubbthreads/images/graemlins/smile.gif

if this is the case, lets drop this thread and continue in the Vicbin one.

in any case, Thanks to all for helping me out AGAIN /ubbthreads/images/graemlins/happy14.gif

Pablo

 

[vicbin]

Hi,

Mr.Al,

I'm really sorry since the 1st one is no longer with me (gone as a gift) with my friend and unfortunately he doesn't have any digicam. I can tell you the final part looks like a big mess especially on the reference part and interconnect to the switcher preregulator.

Anyway, I'm really enthusiast to post my modified schematic and pics of my current preparation for the 2nd one and currently I'm looking for place to host for them. Where do you host most of your schematic pics ? Is it free ?


Pablo,

This is great, /ubbthreads/images/graemlins/smile.gif now I got other CPRer who also interest in building the PS together. Honestly, if you're looking at that thread, it just only 2 active participants which is me and our sensei Mr.Al. /ubbthreads/images/graemlins/frown.gif

I can see you're also planning to use a digital display, was planning to use cheap 2 dollars LCD DMM for Volt and Current display and it turned out to be very ugly and complicated front panel construction, and while toying with LEDs in less ambient light, LCD display is really difficult to view from any angle.

For the upcoming one, I will use Intersil ICL7107 which can drives LED display (not LCD) and with built in 8 segment display driver. Fyi, ICL7107 based DVM is less complicated to build and it needs only external passive components (resistor, cap and LED of course) to build a 200mv, 3.5 digits DVM. If you can find it easily, I suggest you should consider 7107 rather than 3161/3162 which needs external BCD -> 8 segment decoder and more complicated PCB layout.

Regards,
Vic

 

[PEU]

Hi Vic,

I have the 3161/2 voltmeter already assembled & tested, from a project I did about 18 years ago (yes 18 years!!).

If maxim-ic.com is willing to send samples to Argentina I'll consider building one or 2, but in the meantime, lets do it with what I have /ubbthreads/images/graemlins/smile.gif

I hope you can draw a schematic of the revised board, I used the free demo version of microcap8 as a schematic drawing program, its really easy.


Pablo

 

[MrAl]

Hi Vic and Pablo,

Vic:
What i was using for a lot of schematics about power supplies
and LED circuits was some web space Pablo had generously
let me use for the EE course and related stuff.
If you ask him if it's ok i'll gladly post your schematic
there too and link here. I really need to take a good
look at it again because it's been quite a while since
we talked about it, and this way Pablo can take a look too.
It's a nice power supply for sure, and i had been working on a smaller version myself, so that makes THREE of us
working on this now! /ubbthreads/images/graemlins/smile.gif
Once i take a fresh look i'll be able to advise a little
better, but i think offhand the output of the switcher
might not need as much cap as it has now, and yes,
if the min switch on period isnt very short you'll need
some min load, but it should work to load the switcher
itself.

Pablo:
Welcome aboard /ubbthreads/images/graemlins/smile.gif
Now there's three of us working with this power supply.
Mines a lower current version, but it will still
benefit from a switching pre regulator like Vic's.
My advice to you is if you only need 3 amps out and
you can get by with 20 volt output then stick with the
LM350 design because it's much simpler. If not, you
can follow Vic's layout once we get the schem up.
If you dont mind i'll post it to your web site if
Vic sends me the whole schem?

Take care for now,
Al

 

[PEU]

Hi Al,

One more time Al, space is yours to use /ubbthreads/images/graemlins/smile.gif /ubbthreads/images/graemlins/smile.gif no need to ask /ubbthreads/images/graemlins/happy14.gif, there is PLENTY of space available.

I'm in no hurry for building this one, so I'll wait for the debugged schematic and if my knowledge allows, help with the PCB board.

I do want to take advantage of the nice 30+30 transformer, but no need to use it at 3amps, since the transformer is rated 100VA total

for your reference, the coin at the base of the transformer is a US quarter.


Pablo