AMS1117 Are they all the same?

[BoKKeR]

Hey, I have been in the market for some AMS1117 3.3 fixed output regulators(sot223). But it feels like there are just too many to choose from while they all look the same but might behave differently. The one I salvaged from a circuit home and tried outputs between 3.3-4V while being marked 3.3. The ones on aliexpress and ebay have no datasheet and are probably a gamble. Based on the datasheets I found it turns out that some need load resistors, some need tantalum capacitors, many diferent factors at play. I guess my question is. Is it worth or safe buying a ams1117 from ebay or ali because they are all the same. Or should I go with a way overpriced original IC instead? What to look out for ? thanks

[bd139]

I don't buy stuff like that of Aliexpress myself. I use a low volume so it makes sense just to buy a couple of them from a local reseller with a reputation (RS). As for load, they need a relatively heavy capacitance on the output or they oscillate. I have been using 10u 6.3v tants on output and a TDK 1uF ceramic on the input. Seems stable even with heavy RF. I have never used a load resistance, assuming that the normal load itself is a current sink.

[ogden]

Hey, I have been in the market for some AMS1117 3.3 fixed output regulators(sot223). But it feels like there are just too many to choose from while they all look the same but might behave differently.

Fact that they are manufactured in virtually every semiconductor factory and garage as well, shall discourage you to buy such on aliexpress or ebay unless you can get warranty of authenticity. In search for lowest price for such low cost, wide availability components I would just check inventory of my supplier (mouser/arrow/digikey/whatever u prefer), sort by price.

[edit] Do you find mouser 0.21$ @100 for NCP1117ST33T3G offer as overpriced? New shiny LDL1117S33R from ST is even cheaper.

[sleemanj]

The only time I encountered really bad ams1117's is a batch that had just over 10v max input and if you went over they would fail with input shorted to output... not fantastic when my 5v circuitry suddenly got 12v applied, only myself to blame because I didn't check them to make sure they actually met the specs I expected first.  After some digging I found a datasheet from some chinese manufacturer which indeed had a 10v max input, so it wasn't necessarily they were bad, just they were not specced the same as *most* 1117's. 

Anyway, that aside, I can't say I have had a problem with them other than that one time.  I use them often, I always use the ceramic+low-value-resistor trick instead of tantalum caps.

It is worth pointing out that the 1117 are not exactly efficient, of course they are linear so efficiency isn't necessarily a big consideration, but they do have quite a high quiescent current even if not in use (they will back-feed through an internal diode, and draw quite a significant current just sitting there doing nothing), so if battery drain is a consideration, the 1117 may not be a particularly good idea.

[BoKKeR]

Thanks for all the answers so far, it cleared up a lot of things. I will be using it for 5v-to-3.3v drop and it will be powered by a USB port so I don't really care about efficiency. What kind of "ceramic+low-value-resistor trick " are you talking about? What where?  And the prices on the Taiwan based LCSC IC-s are good enough and I will probably go with them. My main problem was that the only local supplier charges over 1$ a piece on top of 4-6$ shipping.

EDIT: @bd139 your solution with the tantalum worked. I am getting proper 3.3v no matter what. Thanks for the help I will probably just stock up on these capacitors.

[CJay]

That advice isn't just confined to AMS1117 chips either, I have a couple of hundred TO-220 5V regulators in tubes which now have a large, bright label that tells me that I have to use a cap on the output.

It's very clearly in the datasheet and it was a good reminder to me that not all three terminal regulators are immediately interchangeable even if the pinout is the same (that's the other gotcha, the pinout isn't always the same)

*Always* read the datasheet.

[sleemanj]

What kind of "ceramic+low-value-resistor trick " are you talking about?

See application note:
http://www.ti.com/lit/an/slva214a/slva214a.pdf

In short, when a regulator calls for a tantalum capacitor (typically on the output), you can generally use a ceramic capacitor and a low value  resistor in series instead - the tantalum is usually suggested because it has more ESR than a ceramic, by adding a series resistor to the ceramic you emulate that.

[SiliconWizard]

There are a lot of AMS1117 clones out there, sometimes with widely different specs. This combined with a rather high typical quiescent current (~5 mA) makes them unattractive IMO.

Now if you're looking to cut costs no matter what, this is another story - you can find AMS1117 clones at a little over $1 for 50 pieces - completely crazy.

[maginnovision]

I usually use a zldo1117 from diodes inc. for similar applications. Wider input voltages and mlcc compatible. I would never bother with ali or ebay. The failure rates I've had tell me to avoid at literally all costs.

[westfw]

@bd139 your solution with the tantalum worked.

What output cap did you have before?

[wraper]

Cleap voltage regulators from ebay usually are counterfeit crap or native crap with super tiny die inside. Nothing like something from decent manufacturer.

[bd139]

EDIT: @bd139 your solution with the tantalum worked. I am getting proper 3.3v no matter what. Thanks for the help I will probably just stock up on these capacitors.

Don't thank me - I just read the datasheet and the stability warning in The Art of Electronics