I did not think carefully about using an LM317 for producing 3V3 from 5V0. While @paulvdh 's comment about low current appears to be correct I would probably not design that way. One point is that when a regulator operates close to dropout it’s performance will (in many situations) not be as good. But don’t shy away from using "LDO"s (Low Dropout Linear Regulators). There are plenty of good ones. But the classic definition of an LDO is that it works with an input-output differential of < 1 volt, so the LT1117 is sort of in “no man’s land” because it does not quite meet that.
That photo makes me wonder:
What is the destination of those components being removed and sorted?
How much is this person doing the work being paid?
What is the precise location and what are the other circumstances surrounding this work (politically sensitive.)?
Having said all of that, it would be great if there were a good approach for electronics recycling.
Each manufacturer a little different data and it is not only the question of pictures. The ICs are little different. In my BOM I accepted ONS LM317L (SO8) but not TI LM317L (SO8)
You should not assume we have here 1.7V:
what is the tolerance of USB 5V,
what is voltage drop at contacts (I have read USB specification many years ago, but remember that it was specified there),
what is voltage drop at used cable.
And working at drop-out limit means IC is no longer a stabilizer but rather like two diodes in serie. Much worse damping of voltage fluctuations at input, and much worse reaction to load current change.
LM317 parameters are specified for voltage drop 3V or higher.
Do anybody knows about what current load we speak in this thread ?
Seeing no radiator I can assume small, but …
I have seen publication assuming that you can stabilize 2V/3A from 35V using LM350 as 35V and 3A are in LM350 rated specs. And it was in electronic magazine. And when I wrote them I got into stupid conversation with one of their designers. He even wrote that mathematical calculations do not necessarily have to be believed. If even mathematics did not appeal to him I send the whole conversation to Editor in Chief and then got info that those one with whom I was speaking is now ashamed of everything he wrote.
May be even more then I when I was teaching students at Technical University. My salary those time was about $13 per month. But I suppose that she gets less.
The datasheet I looked at was 25 year old one from National.
But you’re absolutely right, it’s not very wise to go to the limits with specifications like this.
Tolerance on USB is also quite wide. I would not be suprised if it can be below 5V and still be within spec, and then there are bad quality cables and such.
ESP07 module is mentioned. According to the datasheet it needs a power supply “>500mA”
But average current (No WiFi) will be well below 100mA, and in the uA ranges when sleeping.
Yesterday there was a screenshot of a PCB in the form of a ghost, but I don’t see it anymore.
I guess it also has some LED’s but not much more.
Sure! Don’t we all live and work in a liquid nitrogen bath?
That comment reminds me of a magazine article I wrote years ago regarding an “ideal diode” circuit based upon a MOSFET and op amp. The application circuit really needed an op amp, but an editor (who did not understand the circuit) changed it to “op amp or comparator”. Fortunately that was not the final edit, and the editor listened to my complaint…
im going to solder some new components on a new board and make sure continuity is good but its all simple components 3x led,two N.O. buttons,and a oled screen so i doubt any of those will be the coulpret.
when i tested the power pads to the mcu it was regulating correctly at 3.3v from power line and ground was tested correctly but once i tested the same pads with the board fully powered it was giving me 0.23 from the multimeter so i thing there might be a short or something fishy
so i think officially it is a board/component issue,and did you want to see the board file or the physical board?
this was the brand not sure how reputable but based on the previous tests its not the regulator any longer that is the point of concern its the board continuity
Mostly the physical board, just to judge the components, the soldering quality or maybe some other visible issues. Of course you’d need a good macro camera or microscope, otherwise there’s probably not much visible at these scales. Of course if you want to send the board file, we could look at issues there, too.
okay i got it i re-soldered everything to a new board and used a different power supply,i believe the first power supply i used the amps were to high and fried the regulator everything on the board is working as it should.
thanks to everyone for your help (moderator edit: did your autocomplete slip after that, or something? You can continue metadiscussion politely if you want to.)
Not really how it works, the Amp rating of a power supply is just the maximum (assuming it delivers a constant voltage) so it shouldn’t deliver more than the regulator “needs” or requests (by controlling its resistance).
Of course if it works with some other power supply, there still could be an issue, like maybe a very unclean power output or incorrect (way too high) voltage.
AMS1117 is NOT pin compatible with LMxxx devices. Look at the datasheet for your device and ensure that your schematic pinout matches the datasheet.
Both input and output need 10ufd AND .01ufd caps. The device has very high gain bandwidth and may be oscillating at up to 20MHz. Check with a scope, but be sure to add the bypass .01 caps.
There needs to be a minimum load, add a R to draw 10mA.
Agree 100% re needing some small caps (read the datasheet carefully to see if they recommend anything specific)… I always put 100nf (=0.1uF) within 1cm of any voltage ref, input and output.
(interestingly, one device spec sheet just mentions the 22uF on the output, saying “The addition of 22µF solid tantalum on the output will ensure stability for all operating conditions”. Another spec sheet says 10uF on input, 10uF on output. If you don’t ground the ground / adjust pin, you’ll need a cap alongside the R)
(Several jobs ago, they had a washing machine control panel, with a 7805 on… good news: they had 100nF caps on the voltage reg; bad news: they were about 6 inches away, down a thin pcb track!!. I demonstrated that putting one adjacent made a “non-working board” into a working one. They just measured with a multimeter…)
Adding 100nF decoupling capacitors to both input and output of voltage regulators, and to each pair of power pins of each IC is good practice, and quite often good enough.But when you tun into the domain of high frequencies (Radio stuff, FPGA’s, bigger microprocessors at higher clock frequencies) the default 100nF capacitors are not good enough anymore and you have to fine-tune the decoupling to the application.
Yup, my stuff is audio, so 100n works just fine. However, I sometimes need to put a 1nF on the feedback of my input op-amp, just to stop it singing supersonically! So, yes, smaller caps for higher frequency circuits!
Help us help you. Make a photo of your exact board with the exact part on it. Zoom in so we can read the part. Make multiple photos if necessary.
Show us the exact markings on your regulator part.
Show us the KiCAD PCB with the view on which the part is mounted. Don’t make us have to think backwards. Show us your schematic.
Upload it some where so we could in principle open it ourselves.
Linear regulators become unstable with out capacitors on the input pins to ground and the output pins to ground. Read the manufacturers specifications for capacitor values.
Here is screen shot of an OnSemi data sheet where the input and output capacitors are explained.
Your layout is unnecessarily poor because you have placed your input cap near your output pin and your output cap near your input pin. You should make the layout look like that bit of schematic in the data sheet. Also you should connect the ground pin to the large ground tab. (Caution if you ever use the adjustable version that tab is not ground.)
You might ask, Are these small differenced important?
To which I must reply, Is the regulator regulating important? If yes then you must use the part as the manufacturer explains how to use it.
To test solder the capacitors of the correct values directly on the pins of the regulator with the shortest leads you can get to reach.
The older application notes from National Semiconductor for LM317 give general insights on how to use these kind of parts and what can go wrong, but you must also refer to the exact data sheet for your part.
The LM317 is not a low drop out voltage part so that general insight of which I wrote cannot be fully generalized to the NCP1117 which is low drop out.
I would advise you place a series resistor between the USB power connection and the input to the regulator so that you can measure the current. The current is the voltage across the resistor devided by the resistance value. You may have to use a fraction of an Ohm.
You wrote, "only 5v 1amp is being supplied "
How did you measure this?
I suspect you did not but are quoting a supply rating.
If 5V at one amp then your board is presenting a 5 Ohm resistance as a load. Quite the coincidence and I suspect you need to think through what is happening for more insight.
You need to measure the current your assembly is drawing from the supply. Thus the series resistor of which I wrote.
You could cut up a USB cable and insert a series resistor on the power wire to make your test too.
Old regulators models are old, they are difficult to use. I would buy a Buck/boost regulator board from Ebay and try to clone that. For a little higher input voltages, I use switching regulator like these R-78E3.3-0.5 https://www.digikey.fi/fi/products/detail/recom-power/R-78E3-3-0-5/3593412
There are plenty of manufacturers and models and they are easy to use.
I use LM5017 to step-down directly from 48V to 3V3 (with simple choke). If you take care of limits on Ton and Toff it can work with 100V at input and 3V3 at output.
