LM2575T regulator still not working!!
FlyingFishFinger
Posts: 461
Hi!
Is it normal for a LM2575 to deliver ~10V AC in addition to 5V DC or am I doing something wrong?
Can I put this into the Stamp?
Rafael
Post Edited (flyingfishfinger) : 11/21/2005 8:16:47 PM GMT
Is it normal for a LM2575 to deliver ~10V AC in addition to 5V DC or am I doing something wrong?
Can I put this into the Stamp?
Rafael
Post Edited (flyingfishfinger) : 11/21/2005 8:16:47 PM GMT
Comments
10 vac -- what's that all about?· Ripple?
Try placing a load resistor on the output, like 220 ohms.· See how it is with that.· Switching regulators require a minimum load, they do funny things when they don't have anything to do.·
And, y'know, with these switching regulators: you·must concentrate all of the Grounded points together, you can't run wires hither & yon,·point-to-point fashion, to join them (that's bad form.)· That's emphasized in the data-sheet/s.
·
Could it have anything to do with the coil I'm using?
Rafael
Post Edited (flyingfishfinger) : 11/20/2005 1:04:17 AM GMT
Another question then: What's the input (look) like?
Maybe I should go and calculate the right values again.
My input was, just for testing purposes, a 9V battery, but when I'm done I'd like to be safe with up to 24V
Rafael
I just don't know what's wrong
The DC is reading pretty close to 5V (5.08, if that's close enough)
Rafael
[noparse][[/noparse] "Calculations": Oh, me! ]
Are you sure that you're not actually measuring 10mv ac?
Lookit --·on the data-sheet there·are circuits provided (esp. a simple 5V config).·
Do you have a pdf of the datasheet?· Please U/L it and indicate (by page number) the circuit which you have constructed.· If yours isn't exactly one of theirs, then note where yours differs.
Look on P. 12 for the fixed output version.
I read on( 2 pages further) and consulted the graphs to determine my inductor and capacitor values for my desired input range. I used a 47uF 25V on the input and a 220uF 25V on the out, I have the diode and a 330 uH inductor.
Rafael
· N.B. :
· Linear design and implemantation is much easier than switching design.
I did this with ~15 V and it nearly triggered the thermal shutoff.
What inductor value do you recommend?
I used the diode with the part number from the text, not the diagram.
If I don't have tantalum caps, CAN electrolytics be used? My nearest good supplier for parts is half an hour from where I'm at, and in addition I'm dependant on my parents for getting there. Having gone twice in the past week, I think I've exhausted their good will.
Mail order takes far too long when I only have a week off school, half of which I;m not here.
Rafael
In answer to your question about the efficacy of electrolytics in this design:· No, electrolytics cannot be substituted for tantalums, because the ESR (electrical series resistance) of electrolytics is much greater than that of tantalums.· You're dealing with high-frequency and they're necessary.
If you want to get it going, the only way to do it is within the manufacturer's guidelines.· That's the approach that I'd take, somewhat as you are now:
for my part, instead of trying to construct all that (which [noparse][[/noparse]glance at schedule] may take another month or two) I might as well get this:
http://www.alldatasheet.com/datasheet-pdf/pdf/26742/TI/78ST100.html
But who sells it???
Do I have to order directly from them???
Rafael
Post Edited (flyingfishfinger) : 11/22/2005 3:34:38 AM GMT
Steve
···I've nothing more to add than: truly, the mind boggles.
·· Later.
Lots of heat, too. Generally an unhappy device.
(Amplifiers oscillate, oscillators won't. Basic electronic theory, that [noparse]:)[/noparse]
Steve
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Chris Savage
Parallax Tech Support
csavage@parallax.com
I'm lurking this thread since a while...
Since Rafael made his first post, a lot of guesswork, like "try this", "try that", "what about the coil?", etc. is piling up here. Why? Because Rafael did not give enough information about his problem, and nobody else has asked for it. Too bad to see how much time has been wasted on both sides so far - Raphael still does not have a solution for his problem, and others have spent their time posting possible solutions, guesses, and questions to provide help somehow.
Raphael - assuming that your LM2575 design is not "top secret", so why don't you simply post your circuit's schematic together with some information, how it is powered?
If you had done this together with your inital post, I'm pretty sure that you would already have a solution, instead of still fishing muddy water now, 16 posts later.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Greetings from Germany,
G
So, except for using the (wrong) caps, I have breadboarded the exact circuit from the datasheet, and for Chris, there is nothing before the regulator at this stage except a 9V battery and there probably will never be. Direct primary regulating.
Here is the circuit I was using, but I used capacitor values and the diode obtained from the text and the graphs following the diagram.
Rafael
I'm using the LM 2574 (similar device with 500 mA max current) with great success - over 300 units installed so far w/o any trouble, and my circuit is exactly the same as you have shown. I don't use tantalum caps, just "regular" aluminum caps. Maybe, your breadboard setup causes trouble due to excessive long leads, or stray capacities.
IMO, the most crtitical part is the Shottky diode - it actually must be a fast-switching diode. I'm using a standard 300µH inductor as it is available from almost all distributors - Digikey should be a good source in the US.
I have designed a small 2-sided PCB (see the attached picture), and it works w/o problems.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Greetings from Germany,
Günther
OK, ich habe einen 330uH inductor und eine Diode mit der Bezeichnung 1N5818. Das war eine der in der Tabelle angegebenen Teile.
Ich werde noch einmal versuchen, die verbindungen so kurz wie moeglich zu halten, dann sehen wir mal weiter.
GERMAN OFF
Let's see...I'll try my setup again and then we'll see
Rafael
On a breadboard? Plugblock type breadboard? This is why I asked you to post a photo. Build methods matter massively on switchers.
The odds on getting this to work on a breadboard are low. Too much stray capacitance, inductance, resistance. Switchers are, as I'm sure I mentioned earler, far fussier than 7805s.
If signals couple into the feedback node (pin 4), it's going to behave pretty randomly. The currents and edge rates make it very, very easy to capacitively couple signals from node to node. Doing it on a breadboard makes it orders of magnitude harder.
Steve
instead of a 1N5818, I'm using SB150 diodes (avaliable in Germany from Reichelt) but I think the 1N5818 should work fine as well.
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
Greetings from Germany,
G
Rafael
If I was forced to do this, I'd contemplate either cutting copper tracks into a piece of PCB (cutting gaps to make tracks, if you see what I mean), and soldering the parts down. Building something on square-pad board, where you can join adjacent pads with wire / solder to make tracks would also do, but you do need to lay out the components and tracks exceedingly carefully, to make sure that currents go where you want them to.
If this is your first project, then I'm afraid that I just wouldn't recommend a switcher... Either try to get away with a linear regulator, or get a switch mode power supply as a module, one way or another.
It's possible that you could get it to work on a breadboard, but with your lack of test equipment and experience, I suspect your energy would be better spent on other parts of the project, and return to it when you've got better access to test gear and a bit more experience of high speed / high current electronics...
Linear Technology (http://www.linear.com) wrote a very good introduction to sitch mode power supplies, for people having a hard time - very much like you. It's here:
http://www.linear.com/pc/downloadDocument.do;jsessionid=DEu6zviL6rpI9l3GHCB3RHPx00eeScLKb6ixOrsYuPREz2H781c5!-239514596?navId=H0,C1,C1154,C1009,C1021,P1222,D4120
Hmm - I'd put money on that link not working. Just go to http://www.linear.com and search for 'poets' - it's an25 you're after. an47 is another classic, for high speed stuff.
Hope that helps a bit,
Steve
Or I manage to get that TI chip I mentioned earlier, but I haven't found anybody who·has it yet. I need up to 25V in, 5V out.
Otherwise, I'll just have to try 78xx with big heatsinks...
Thanks for your help
Rafael
No, no, not me; it's Herr Daubach.· He must have a spare.· C'mon, GÜNTER, out of the shadows with you --·we all know you're out there.
Size?· Take a look at the picture of his board, you should be able to judge the size from it.· It looks like appx 1 in. x 1 in. (25.4 mm x 25.4mm).
Specs?· He says it's good for 500mA.·
Well, if he's going to be shy, then send him a Private Message.
-- PJA --
I'll give him a day or two and then I'll PM (Threat, threat..
Thanks
Rafael
The fact that you get 5.08 V average out, suggests to me you may have a measurment artifact, not a real AC signal....
The LM 2575 switches at 52 Khz, are you seeing ripple from the regulator, or 60Hz AC, or something completely unrelated?
Trying to run it off a 9V battery is a little iffy, the impedance of typical 9V batteries is high, unless brand new alkaline or fully charged NiCD or NiMH. You definitely want to go overboard with the input cap size in this case. I wouldn't expect to see problems though unless you had a significant load on the regulator, like a couple of hundred milliamps.
Mike C.