lm2596 buck converter

[dannyf]

where did you get yours

ebay, a couple years back.

I have another that's more recent (over the last 6 months). I will test it when I find out where it is,

[MrAl]

Hello again,

Thanks for the pictures and notes from you all.

I see now that there is a lot of variation in the product here.  There are at least three different boards i have seen now.  Mine does not have the compensation capacitor, while the ones on eBay in the link previously in this thread does have that capacitor.  The capacitor is for compensation for certain operating conditions.

I see the waveform too posted by RadioFlash (thanks), and mine is not like that.  It is much slower, so i guess the chip i have is a fake, that's the only explanation.  It is marked LM2596 and JM24RP, but it cant be that chip.

[Arcamax]

:)i hope this one i ordered is like those green board pictures and not that blue board.

There must be something to it.   When someone would call up with a car problem they would always ask what color the car was.  As a manager once told me....The optics of a project are everything.

If your saying i should have checked/asked what it was, then yeah i should have done/thought that. wasn't thinking lol.
http://www.ebay.ca/itm/DC-DC-LM2596-Step-Down-Adjustable-Buck-Converter-Power-Supply-Module1-5V-35V-/360672163864?pt=LH_DefaultDomain_2&hash=item53f9bc7818
5 out of 6 pictures are actually a green board but one blue, so it should be good

[electrophiliate]

I have several of these cheap buck converters with the IC marked "LM2596S".

I got these from here about a year ago:

http://www.ebay.com/itm/3pcs-LM2596-1-23V-30V-Output-DC-DC-Buck-Converter-Step-Down-Module-Power-Supply-/400516985374

So I did a brief test to compare with the results in the post by [radioFlash].

Input: roughly 13V from SLA battery.
Output: 5.00V into 100R and 10R loads.
Both channels are AC coupled.
Ch1 (yellow): pin 2 of the alleged "LM2596S".
Ch2 (blue): output ripple.

Mine are generally only running at about 56kHz, the ripple is comparatively much worse, and there is ringing on pin 2 (presumably indicative of discontinuous mode).

See attached images below.

I also have a few of the much smaller buck modules mentioned by [jarrodhroberson], from the same supplier as in my previous link:

http://www.ebay.com/itm/2pcs-DC-DC-4-5V-28V-to-0-8V-20V-Step-Down-Module-Power-Supply-Brand-New-/370965365951

I did not want to spend too much time on all this but those performed much better than the bigger modules.

With a 10R load at 5V output the ripple was under 50mV and the waveform suggested a switching frequency of 1MHz as claimed in the link posted by [jarrodhroberson].

[akis]

opps sorry this  http://www.ebay.ca/itm/1-Pcs-DC-DC-Buck-Converter-Step-Down-Module-LM2596-Power-Output-1-23V-30V-EK-/201126191001?pt=LH_DefaultDomain_0&hash=item2ed40e1f99   was this miss labeled when it said Input current: 3A(maximum). Isn't it supposed to be the output maximum 3A

Be very very very very careful with these. I have bitter experiences with the bigger brother of the LM2596 the LM2679. When it blows, and it does, and I have blown a few, 50% of the time it goes closed circuit and lets the full input voltage to the output. Then you also fry everything downstream.

In particular with these ebay LM2596 boards, I have used a lot, and have fried a lot, and those fried were mounted and working happily inside a closed box, and one day they decided they had enough. I did not draw anywhere near 3A from the output, more like 0.5A if that.

[electrophiliate]

Quick update. Another capture of the "LM2596" (LM2576?) based buck converter modules described in the previous post.

This time with a 2A load, which resulted in a minor voltage drop (I did not look at the transient response just a DMM), and quickly heated up the non-heatsinked IC.

See attached image. Looks like it in now in continuous mode so the previous ringing is gone. Ripple is about the same as with the 10R load (500mA).

[edit 1: It did not like that test and no longer wants to put out 2A @ 5V but still seems to work OK for lighter loads.]

[edit 2: It seemed to recover overnight and behaved better today and took at least 10 minutes of punishment but became hot.]

[edit 3: Updated the attached capture.]

[Arcamax]

opps sorry this  http://www.ebay.ca/itm/1-Pcs-DC-DC-Buck-Converter-Step-Down-Module-LM2596-Power-Output-1-23V-30V-EK-/201126191001?pt=LH_DefaultDomain_0&hash=item2ed40e1f99   was this miss labeled when it said Input current: 3A(maximum). Isn't it supposed to be the output maximum 3A

Be very very very very careful with these. I have bitter experiences with the bigger brother of the LM2596 the LM2679. When it blows, and it does, and I have blown a few, 50% of the time it goes closed circuit and lets the full input voltage to the output. Then you also fry everything downstream.

In particular with these ebay LM2596 boards, I have used a lot, and have fried a lot, and those fried were mounted and working happily inside a closed box, and one day they decided they had enough. I did not draw anywhere near 3A from the output, more like 0.5A if that.

Quick update. Another capture of the "LM2596" (LM2576?) based buck converter modules described in the previous post.

This time with a 2A load, which resulted in a minor voltage drop (I did not look at the transient response just a DMM), and quickly heated up the non-heatsinked IC.

See attached image. Looks like it in now in continuous mode so the previous ringing is gone. Ripple is somewhat better than with the 10R load (500mA).

[edit: It did not like that test and no longer wants to put out 2A @ 5V but still seems to work OK for lighter loads]

  ok so then should these good for 5volt and ~500mA output?

[Kjelt]

Use them for critical applications in a two stage design with failsafe. So for instance you have 18V input and want 5V output, let the 2596 do the 18 to 8V stepping down followed by a traditional 7805 do the 8 to 5V. Monitor the 7805 input if its higher then 10V shut the ps down.

[electrophiliate]

ok so then should these good for 5volt and ~500mA output?

Yeah, that should be OK, but I would be reluctant to use them for critical applications, but [Kjelt] suggested a failsafe workaround in the above post. I have a handful of different cheap modules from Ebay but have never really used them much or left them running for long periods of time, so my experience with them is limited.

Today for a test I let it run at 5.1V 510mA for 30 minutes and it seemed OK and the IC was only warm to the touch.

I think I might order some genuine LM2596-ADJ and LM2576-ADJ ICs to build a few circuits and compare the performances with these cheap modules.

[MrAl]

Hi,

I am now reading around the web that other people have had problems with these regulator boards too.  Not just this one, but with the LM2596HVS where the input is allowed to go up to something like 60 volts.  The genuine chip takes 60 volts input no problem, but the fake HV version chip blows out at around 45 volts input.  So there is clearly a difference and some fake chips floating around.  Where they came from and how they worked their way into the parts stream is unknown at this time.

I have had very good luck with the LM2576 TO220 package, building the entire circuit from scratch myself.  I've built them on a regular PC board etched just for that circuit, and also on those little 2x3 perf boards with the dip patterns etched on them from sources like Radio Shack, wired by hand with small 22 gauge jumper wire.  They all work just as expected.  One had a 15v 1 amp output for use with a flat bed scanner.

Someone suggested that i treat one of my 'fake' boards as a faulty circuit and just start replacing parts to see if the problem goes away with brand new parts.  I have to wonder if a cheap inductor would cause this too if it saturates.  It's not as likely now though, now that i have seen the fake HV version blow up.  I guess it is worth a shot though.  Swap out the inductor with a better quality unit first.

[Seekonk]

I think many of the problems relate to relying on the on board capacitors alone.  These are for localized use only to prevent the circuit from oscillating.  They are of dubious quality chosen on cost and unlikely to handle ripple currents long term.  Always parallel other capacitors with these for reliability.  If that capacitor opens up, the chip will likely destroy itself in short order.

[Zepnat]

These findings are the same as mine. The fake IC is total garbage for anything other than light work.
Please test a genuine lm2576. I blew (exploded) 20 fakes while the genuine one isstill working fine.

[Saronni]

I'm adding my LM2596 module story to this thread for the benefit of anyone searching the archives for these modules.

I too picked up several modules with suspected "fake" ICs in June 2016 from this ebay seller http://www.ebay.com.au/usr/shieldsfans. While the IC is clearly marked LM2596 it is only switching at 54kHz. In the scope picture below the yellow trace is the AC coupled ripple on the module output and the blue trace is pin 2 of the "LM2596" (also AC coupled).