Author Topic: Help with my LM317 limiter?  (Read 9067 times)

bokeb

  • Guest
Re: Help with my LM317 limiter?
« Reply #30 on: July 20, 2021, 04:53:20 AM »
OK, I had to reread that 50ma->40ma 365Ohm statement a couple times to understand what you meant (it confused me at first).

You were saying that you STILL had the 20.4Ohm resistor between Vout and Adjust but put a load consisting of the LED and a variable resistor between Adjust and the positive lead of the ammeter with the negative lead going to ground.

OK, so you increased the load resistor (LED current limiting resistor - a.k.a. a ballast resistor) to 365 Ohms and at some point started to see the current drop.
All exactly correct. I then increased the load resistor, and stopped as exactly as I could when the multimeter started reading a solid 0.04A without wavering between 0.04 and 0.05. Then I pulled the multimeter out of series and used it to read the voltage across the load resistor (20.4v), and then pulled the load resistor out of the circuit and read it at 365 Ohms.
Quote
LED (figure 2.5V drop if its not white or blue and about 3.5V drop if it is - blue and white LEDs are different technology).
well what I've got are red and yellow. I'll attach a picture with em. Don't understand much at all about LEDs other than that I think I have saved some pots and my milliammeter in the past by having a led in series in the circuit. have blown a few leds tho yeah.
Quote
40ma through a 365Ohm resistor is 14.6V drop plus 2.5V is 17.1V.  Remember how I said the LM317 needs about 3V across it to function?  17.1V + 3V = 20.1V plus the 1.25V there should be across the 20.4Ohm resistor is 21.35V.  You have 27 unless you've already drained your 9V batteries (they're not very high mah rating - like 650ma for alkaline cells). So 27 minus 21.35 = 5.65. Unless you're battery stack is no longer reading 27V, you shouldn't see any drop in current from 50ma.  To see the current start to drop off, you'd need to have the resistance around 405Ohms though thats not THAT far off 365, your battery stack may not be reading 27V anymore and your meter, if it can only resolve 10ma, you won't need to get to 40ma exactly for it to show 40ma - just <something> less than 50ms so its possible.
Awesome, good to know the math *can* check out (still learning this). I would agree with the ammeter rounding theory - whatever internal logic + sensitivity the 10mA resolution ammeter has, it has to decide somewhere to call it one integer and not another.
I have experienced how quickly these 9v batteries can drain, but I did barely use these last night (and immediately fully disconnected after)-
and just read the voltage again just now, now actually reading 27.4v. So it must be the ammeter. About to find out.
Quote
Also, just as an aside, its rather unwise to push 50ma (or 40 even) through an LED. Usually these things are spec'd at a max current between 20 and 30ma.  With the efficiency of these things these days, even 3ma for a lot of them is painful to look at.
Yeah, this is actually why I put the LED in - to save the load-simulating-pot and have the led blow instead if something were. Though I don't know, maybe the leds I have can handle more than the pot I used could. I definitely have blown a couple pots this way. I am pretty sure once a couple weeks ago I blew my first milliammeter fuse by putting somewhere around this much current through a circuit without putting a LED in it.. so since then I've been putting LEDs in if it seemed like a lot of current. Anyway sorry this is kind of off track
Quote
Yeah, I'd love to know whats up with that power supply.

I'm using a laptop charger/supply that puts out 19.3V at I think 3.5amps (or maybe even 5amps).  Laptop chargers aren't light-weights.  This is a good, name brand one though.
Yeah, me too. Sending a picture with the culprit. The first half (that plugs into the wall and the transformer) is mine from a long time ago, have used it interchangeably for my laptops in the past no problem, though not in a little while. The second half (starts with the transformer and ends with the plug for the laptop) I bought from a tiny electric crunk shop that just had that half-a-charger sitting around, obviously used. Didn't want to strip the end of my charger (though I have figured how to use these now without stripping the end). Took it home, stripped it, plugged it into the first half of my charger, tested its voltage - got 19.6v ... just assumed after that that it was good to use. IDK!

*** Today my job was - after buying a new milliammeter - running to a couple cafes with my circuit and my old (culprit) power supply and testing electricity in a couple other locations. R1=125 Ohms, R2 removed. Got the same results (0.16a and 0.19a depending on LM317T model). Knowing then that it must be the cord, I went to the local laptop-repair store, and tested it with a new charger they were selling and the milliammeter I had just bought - I got a reading of about 9 point something (forget) milliamps (I think i had the GKDAT model installed, didnt get the time to try both). I remembered your 10 milliamp figure for the 125 Ohm resistor. Hallelujah. Brought home the new cord - about to run tests with load-simulator again using this and my milliammeter.

Quote
Those are both ST microelectronics parts. The logo on the right one is just a little crude (it happens sometimes - they crank these things out by the boatload so the faster the packaging line moves, the better and we both know that speed is inversely proportional to quality - wink).

ST is a VERY reputable manufacturer. They're up there with the likes of Texas Instruments,...  Definitely not "chicom" based.
Good to know. Though I have to say the the logo on the right one I pictured was one of the CCOV3's ... I have 4 of them, and the logo on all 4 of them look exactly the same as each other. and the logo does look clean cut and quite intentional. Idk. It does look similar to the ST logo at first glance but on closer inspection it seems very clearly to be intended to look the way that it does - a single character, not 2, maybe a character from another alphabet (Hebrew? Sanskrit?).. or maybe just some fanciful emblem.
I will definitely keep testing both of these models going forward. It is a peculiarity (I think) with them giving different numbers as they have done). Perhaps I will wind up favoring the ST versions, knowing that they are a reputable manufacturer.
Quote
Um? Be careful with that new meter (wink).  It might be prudent to take the back off (they might even tell you in the manual) and figure out what the exact type and rating of the fuses are and buy a spare set or two.  Finding fuses is easy (sometimes they'd need to be ordered off the web - stores don't normally stock electronics fuses - lower current, fast blow usually) but it shouldn't be hard to find a source.
Lmao, yes, good advice. Clearly I need it. I think I have been learning a lot in the past week though and have learned my lesson. *knock on wood*
Good call, I will order a few 200mA fuses online. In my last run to the city my successes included 3 250mA fuses, i think they were the only, smallest fuses for individual sale in this half of the country. Hopefully I won't have to find out whether the meter can actually handle the 250mA.
Quote
Make sure the circuit is working properly at 10ma first.  Then start cranking the current down to see how far you can go before it gets unstable and never go below that current (for safety's sake, stay at least 1ma above it).

Keep us posted.
Yeah, I'll definitely keep R1 at 125 Ohms now what with a working milliammeter. Only needed the higher amperage last night to work with my busted one.


I'll do some tests with my new cord and milliammeter now and post the results in a bit.
« Last Edit: July 20, 2021, 05:02:50 AM by bokeb »

bokeb

  • Guest
Re: Help with my LM317 limiter?
« Reply #31 on: July 20, 2021, 10:07:23 AM »
All right.
Fresh test.
New power supply: 20v 2A
R2= 10 Ohm minimum, 1668 Ohm maximum potentiometer (load-simulation-pot). (no LEDs)
Ammeter in series between R2 and ground.
*I tried to report data for columns 2 and 3 for when IOut dropped by about 0.1mA

                          | IOut when                | R2 when IOut              | VR2 when IOut    |
                          | R2 = 10 Ohms         | begins to lower*           | begins to lower*  |
-------------------------------------------------------------------------------------------------------------------------------------------
R1 = 468 Ohms | GKDAT = 2.70mA    | GKDAT > 1668 Ohms | GKDAT > 4.43v    |  **so far have been unable to achieve lowering of IOut when
                          | CCOV3 = 2.72mA    | CCOV3 > 1662 Ohms | CCOV3 > 4.46v   |      R1 = 468 Ohms. I stopped testing at ~1660 Ohms of R2.
                          |                                  |                                    |                              |
R1 = 98 Ohms   | GKDAT = 12.81mA  | GKDAT = 1369 Ohms | GKDAT = 17.08v  |
                          | CCOV3 = 12.87mA  | CCOV3 = 1351 Ohms | CCOV3 = 16.91v  |
                          |                                  |                                    |                              |
R1 = 47 Ohms   | GKDAT = 26.9mA    | GKDAT = 664 Ohms   | GKDAT = 17.05v  |
                     *** | CCOV3 ~ 15.5mA    | CCOV3 = 672 Ohms   | CCOV3 = 17.00v  |
                                                                                                                                       
***CCOV3 began overheating and giving unstable IOut figures when R1 was at 47 Ohms and R2 was at 10 Ohms. As R2 was increased from 10 Ohms, IOut rose steadily until R2 was at some point between 50 and 100 Ohms, at which point IOut stabilized at 27.0mA. I tested this with all 4 CCOV3 models and got similar results.


This is where I will end this test for now. Tomorrow possibly I will try it with R1 around 250 Ohms, and get R2 with multiple resistors up to the point where IOut stops to drop.

Regardless, conclusions for now:
1) The circuit is working, the power supply is working. thank god.
2) I've got about a maximum of 17 volts to work with across the cell. I was expecting 15.75 volts (20v - 3v -1.25v, right?) ... IDK.
3) While I am not getting the wild current differences across the board with the 2 different LM317T models that I was getting with a bad power supply, I am still getting some anomalies at low resistance levels with the CCOV3 model (the one that I think may not be made by ST). I will probably start using the GKDAT models only.
4) I need to learn what the expected resistance ranges are for my setup, so I know what pots to put in my board when I solder it soon. I will research this on the board tomorrow as well as probably create an expected cell situation with the right sodium carbonate mixture and try to measure its resistance. So far what I know is that I will be using a 1oz bullion anode and understand from Kephra that I can run up to 20mA or so doing so. Therefore I would like to design a board that can output a range of 5mA-25mA, or thereabouts. Any advice in this area much appreciated.
5) I understand from some youtube LM317 videos that this circuit can be improved somewhat by adding in a couple of capacitors in parallel in a couple spots. I want to figure out what specifications of capacitors I should be looking to use. My knowledge in this area is very limited - any advice here appreciated as well.
« Last Edit: July 20, 2021, 10:13:04 AM by bokeb »

Offline Gene

  • Expert
  • Participant
  • ***
  • Posts: 1757
  • Likes: 190
Re: Help with my LM317 limiter?
« Reply #32 on: July 20, 2021, 08:28:52 PM »
If you have 17V usable with those CCOV3 regulators that means that there's 20-17-1.25V (1.75V across the regulator, right?

1.75V * 0.0155 = 27.125mw heat dissipation in the regulator (thats 0.027125 watts).  It shouldn't even get warm.

Something is obviously wrong with those CCOV3 regulators.  What you got with the other one makes perfect sense.

Remember I said in electronics "exact" is a word we never use.  You're in the right range though with those true ST regulators.

Perhaps you're right and the others are some cheap chicom knock-off that have serious limitations or perhaps even 2nds someone got their hands on from a third party manufacturer for next to nothing and branded them to look like an ST. Fraud IS a big issue in the semi industry and much less so with the big, complex parts like processors and more so with the "jelly bean" parts.  Thats why, as a company making electronics devices, you always buy from reputable distributors where there's traceability directly to the manufacturer who made the part. Costs a little more but in the long run saves you money for fewer returns, dealing with issues that might even have you recalling something, ...

Try to lower the current at the top end (voltage across load at 17V) and see how low you can go before things start getting squirrely.  I bet you'd hit 5ma.

Well, if you want 25ma max, that'd be R=1.25/0.025 = 50 Ohms assuming your pot can achieve 0 Ohms. If not, figure out what it can do at best and then subtract that from 50Ohms and go with that. On the low current end, 5ma would be a total resistance of 1.25/0.005 = 250Ohms but if you can go lower in current, take that current and divide it into 1.25 and go with a pot around that.  I realize the general availability is 250, 500, 1000,... At 500Ohms you'd be able to hit 1.25/(500+50) or 2.27ma (probably won't get there with that regulator but who knows?).  At 250, 1.25/(250+50) so 4.167ma.  The math is simple.

Remember, the regulator maintains 1.25V across the resistor+pot so the equation is simply

R = 1.25/<current in amps>
<current(amps)> = 1.25/R

17V with 20V input I'd consider rather good.  That means the real headroom of the regulator in this particular topology is only 1.75V which seems really good.  For sure as the current increases the regulator is going to require more headroom but at these lower currents perhaps not. All good.


bokeb

  • Guest
Re: Help with my LM317 limiter?
« Reply #33 on: July 21, 2021, 01:29:56 PM »
Hmmm.. Ok, so I re-created the circuit the same way with the ST limiter, R1=47 Ohms, and R2 a 1668 max Ohm pot (again in series, simulating load).

If you have 17V usable with those CCOV3 regulators that means that there's 20-17-1.25V (1.75V across the regulator, right?
Raised R2 from low until the very first signs of IOut beginning to drop (less than 0.1mA drop). Measured VR2... 16.75V. Honestly not sure where to get a proper reading across the regulator. Measured voltage across VOut and Adjust... 1.21V (I know this is the one supposed to be 1.25v right). Measured voltage across VIn and Vout... 1.65V. Measured voltage across VIn and Adjust... 2.88V but a bit unsteady (was this config interfering with the limiter's logic?)
Anyway it would've been my guess that the proper way to read voltage across the regulator would be from VIn to VOut, right? With a 16.75 VR2 figure, I was expecting a 2V regulator figure... I guess 1.65V is close? 0.35V missing somewhere? I measured voltage across the milliammeter (I am using my earlier, part-busted multimeter to measure voltage/resistances now, while leaving the ammeter in series)... 0.04V.
Well anyway we're mostly on target here.
Quote
1.75V * 0.0155 = 27.125mw heat dissipation in the regulator (thats 0.027125 watts).  It shouldn't even get warm.

Something is obviously wrong with those CCOV3 regulators.  What you got with the other one makes perfect sense.

Remember I said in electronics "exact" is a word we never use.  You're in the right range though with those true ST regulators.

Perhaps you're right and the others are some cheap chicom knock-off that have serious limitations or perhaps even 2nds someone got their hands on from a third party manufacturer for next to nothing and branded them to look like an ST. Fraud IS a big issue in the semi industry and much less so with the big, complex parts like processors and more so with the "jelly bean" parts.  Thats why, as a company making electronics devices, you always buy from reputable distributors where there's traceability directly to the manufacturer who made the part. Costs a little more but in the long run saves you money for fewer returns, dealing with issues that might even have you recalling something, ...
mm, makes sense. Well, I know how to identify the ST models now and what store I got them from. Will no longer use the others.
Quote
Try to lower the current at the top end (voltage across load at 17V) and see how low you can go before things start getting squirrely.  I bet you'd hit 5ma.
Ok, if I'm understanding this correctly, you're saying increase R2 until VR2 = 17V (or whatever causes IOut to begin dropping), then incrementally increase R1 in order to decrease stable IOut, each time re-increasing R2 in order to reattain VR2 = 17V at that IOut level, then watch IOut for a minute each time, looking for a point where the regulator is no longer maintaining a stable IOut reading.

Alright, following this procedure, I got IOut (stable) down to 2.51mA, which took 7110 Ohms of R2 to start making IOut and VR2 start jumping around. I then pulled R2 back to 7030 Ohms, which brought IOut back to a stable 2.51mA, and VR2 to 17.15V (I noticed that as I kept installing more resistors and pots to achieve higher R2's at lower IOut's, the VR2 that would cause IOut to destabilize kept creeping up by small amounts). At this point I have installed just about every pot and resistor I have to achieve these levels and cannot test any further down (need to pick up some 5k resistors) - but indications are that (if I am doing this right) it could keep performing lower.
Attaching a shot of my setup for this test for posterity.

I really am getting a stable 2.5mA IOut when R2 (load-simulator) is anywhere between 1k-7k Ohms. If I am testing this correctly, this is the test to determine what minimum levels my LM317T can regulate current down to in this current-limiting config, right? Is 2.5mA stable an unreasonable level to believe this LM317T is performing at? Maybe this particular ST component is exceeding its minimum expected performance by quite a bit? What would be an unreasonable level to expect an LM317T to possibly perform to?
Regardless of all of this, if I understand Kephra's directions correctly, I'm basically never going to want to run at 2.5mA anyway. Just trying to understand this circuit at this point. Now that I have this knowledge I may like to build more of them possibly at various specifications in the future.

Quote
Well, if you want 25ma max, that'd be R=1.25/0.025 = 50 Ohms assuming your pot can achieve 0 Ohms. If not, figure out what it can do at best and then subtract that from 50Ohms and go with that. On the low current end, 5ma would be a total resistance of 1.25/0.005 = 250Ohms but if you can go lower in current, take that current and divide it into 1.25 and go with a pot around that.  I realize the general availability is 250, 500, 1000,... At 500Ohms you'd be able to hit 1.25/(500+50) or 2.27ma (probably won't get there with that regulator but who knows?).  At 250, 1.25/(250+50) so 4.167ma.  The math is simple.

Remember, the regulator maintains 1.25V across the resistor+pot so the equation is simply

R = 1.25/<current in amps>
<current(amps)> = 1.25/R
Right. Duh. I've been so deep in the weeds just trying to get this circuit to work the last couple weeks that I guess I forgot that I don't really have to take my load resistance into account when designing this circuit... because that's already been calculated with Kephra's formula (a solution of 99.9% distilled water and 0.1% of 1 Molar Sodium Carbonate solution) ... right? I guess that produces an always-the-same Load resistance at the start of a run, that will only go down by a (small-ish?) amount by the end of the run with the solution now containing some silver nanoparticles.

Anyway, so OK, all I really need to keep in mind is the measurements across VOut-Adj. The voltage between the 2 always being 1.25, the resistance between the 2 always being the adjustment pot + safety resistor, and the current between the 2 being the resulting desired IOut into the cell.
Don't really have to do the calculations for the whole circuit, other than making sure I have plenty of voltage across the cell (sounds like with 15V you're in pretty good shape with a sodium carbonate electrolyte). Doesn't stop me from being curious what the typical resistance in this cell is. I did try setting up the cell and measuring it - got an extremely unstable reading somewhere in the hundreds of thousands of Ohms. That can't be right, I must not be measuring that correctly.

Got it. Just have to decide what parameters I will choose. I have considered I may hammer out my bullion bar into a higher-surface-area anode, with which apparently I can use up to 40mA or so - I may try building a circuit that can handle that.
I managed to find a couple of the tiny 10 turn blue square trimmer pots that you have mentioned. They're 500 Ohms. I like them, I'm intending on using one of them. I'm reading 2 Ohms at their minimum before the reading gets squirrely. So if I go for the 40mA max, I suppose I will try to make the safety resistor about 23 Ohms.
Do you have an understanding on what happens inside these 10 turn trimmer pots when you turn them all the way to one end of their range, but keep turning? Normal 1 turn pots have a maximum turn range and then just won't turn farther. I want to make sure once I solder this circuit together, that I don't do permanent damage to the pot each time I try and max out the amperage.
Quote
17V with 20V input I'd consider rather good.  That means the real headroom of the regulator in this particular topology is only 1.75V which seems really good.  For sure as the current increases the regulator is going to require more headroom but at these lower currents perhaps not. All good.
Nice. I'm feeling good about it.
To be fair, I only measured one of my 3 ST-built LM317's for what maximum voltage it could push across the simulated cell with my 20V input. Perhaps testing this on my other 2 tomorrow will be my final test before going forward.

Thanks for all your assistance here, Gene. It has been both a good exercise for the old noggin, as well as a great crash course on electricity and electronics that's gotten me a working product.

The main item for tomorrow I think will be to set everything up to make my first batch of basic colloidal silver - straight through the breadboard. Had all the reagents and sundries for a 20ppm uncapped for a little while now.

If all goes well, (color looks good and all), I will look to figure out adding a couple capacitors to the circuit (still want to figure that out) and solder it together.
Also looking at making the double transistor circuit for extra credit since I mostly get everything now and have pretty much everything I need I think. Have a few 2N3904 TO-92's that the clerk assured me would work the same as 2N2222's. idk if that's true or not.
Finally if successful it is my intention to make a step by step thread named "The Absolute-Complete-Idiot's guide to building your own current limiter". Maybe we can help more people learn to make true colloidal silver for the cheap cost it aught to be - I get the impression most people here are shelling out for the lab bench power supplies.

Offline kephra

  • The older I get, the better I was
  • Administrator
  • Participant
  • *****
  • Posts: 8883
  • Likes: 286
  • Illegitimi Non Carborundum
    • My World As I See It
Re: Help with my LM317 limiter?
« Reply #34 on: July 21, 2021, 02:01:10 PM »
Quote
Kephra's formula (a solution of 99.9% distilled water and 0.1% of 1 Molar Sodium Carbonate solution) ... right? I guess that produces an always-the-same Load resistance at the start of a run, that will only go down by a (small-ish?) amount by the end of the run with the solution now containing some silver nanoparticles.
The cell resistance depends on the electrode sizes (wetted area), distance between electrodes, temperature, and amount of electrolyte.
There is the unknown and the unknowable.  It's a wise man who knows the difference.

Offline Gene

  • Expert
  • Participant
  • ***
  • Posts: 1757
  • Likes: 190
Re: Help with my LM317 limiter?
« Reply #35 on: July 21, 2021, 09:55:15 PM »
Yup, Kephra's correct (as always).  Thats why you have to tweak the cathode depth to adjust the cell voltage high enough at the beginning of the run. Change anything (even a drop more or a drop less of electrolyte) and things change.

I run a 14 gauge piece of silver wire, 10ma, 150F and over the course of a run to make a quart (don't have any 1L containers) of 120PPM notice that the cell voltage drops about a volt as more and more silver is accumulated in the water which lowers the resistance of the water.  As per Ohms law, if the current is fixed and the resistance is lower, so is the voltage.

Even stirring vs. not changes the cell resistance.  Stirring increases the length of the path the electrons have to follow so they have to move through a longer distance in the cell to get to the cathode and as such, that longer path increases cell resistance and for a given current your cell voltage increases.

No I've never tried it as I don't have a heated stir plate but as I recall, the above is what Kephra explained to me a good while ago and it does make sense.

I just set the starting voltage to 12-13V (about as high as I can get it) and let it run. Its never close enough to 10V at the end of a 3 hour run to the point I only check it maybe once an hour, if that.

If you're running lower current (in the say 3-5ma range), you may not even be able to use the full 20 drops of electrolyte per liter.  I found that out early on when I was running 3ma making cold. Others I recall said the same thing - they had to drop down to 16-17 drops to get the cell voltage up over 10V.  At 8-10ma I don't have an issue anymore and just use the full 20 drops up front.

Its all a big balancing act - just like everything else in life.  You get used to it after a while and find the sweet spot that works best for you.

I'm glad you finally figured out what was wrong with that current limiter but I'm still baffled as to exactly WHAT that old power supply was doing to make it happen.  We may never know.

Offline kephra

  • The older I get, the better I was
  • Administrator
  • Participant
  • *****
  • Posts: 8883
  • Likes: 286
  • Illegitimi Non Carborundum
    • My World As I See It
Re: Help with my LM317 limiter?
« Reply #36 on: July 21, 2021, 11:51:05 PM »
Quote
I'm glad you finally figured out what was wrong with that current limiter but I'm still baffled as to exactly WHAT that old power supply was doing to make it happen.  We may never know.
What would happen if the input power polarity was reversed?  Thats one thing I would surely check.
There is the unknown and the unknowable.  It's a wise man who knows the difference.

bokeb

  • Guest
Re: Help with my LM317 limiter?
« Reply #37 on: July 22, 2021, 07:09:30 AM »
Jesus crimeny... Kephra you're right.

The cord has a black and a white wire. One of the very first things I did like a month ago was look up which color is positive and which is negative. I got a few results stating that for DC, black is positive and white is negative. I measured the voltage across it with my multimeter around then, but wasn't really sure if I was using the meter correctly to measure which direction was positive.
Well now I've measured this new cord and definitely know that if you connect the positive voltmeter lead to the positive (hot) wire, and the negative lead to the negative wire, you will get a positive voltage reading.
...And it turns out the white wire is the positive wire and the black one is the negative wire.

...jiminy.

Offline kephra

  • The older I get, the better I was
  • Administrator
  • Participant
  • *****
  • Posts: 8883
  • Likes: 286
  • Illegitimi Non Carborundum
    • My World As I See It
Re: Help with my LM317 limiter?
« Reply #38 on: July 22, 2021, 01:43:54 PM »
Mystery solved, and now you have two good power supplies :)
There is the unknown and the unknowable.  It's a wise man who knows the difference.

Offline Gene

  • Expert
  • Participant
  • ***
  • Posts: 1757
  • Likes: 190
Re: Help with my LM317 limiter?
« Reply #39 on: July 23, 2021, 05:54:41 AM »
AND... for house wiring its the exact opposite. White is ground (neutral) and black is positive(hot). No clue why but it sure is.

In electronics, black is almost always ground.

Usually though for laptop supplies there's a symbol on the label that tells you which pin on the connector is which. Did you cut the connector off or did it not have one?

But even still, there's something squirrely with those LM317's that have the funny "ST" symbol on them. They're not acting properly.

I never even thought to ask if you checked power supply polarity. Geez!

No, the LM317 has no reverse polarity protection and no I can't tell you what it'll do if you power it backwards though at least for this we kind of know now (sigh).

I'm glad though that Kephra caught this. Now there's no doubt as to what was really wrong anymore.

indocomp

  • Guest
Re: Help with my LM317 limiter?
« Reply #40 on: August 04, 2021, 02:07:16 AM »
Does LM317 both T or Z basically producing constant current by reducing voltage?
I use only potentiometer 50K to reduce from initial 30v 5mA to 2mA and it voltage drop to around 12v 2mA,
i am using 125 mm silver wire with 2 mm diameter as anode
thanks

Offline Gene

  • Expert
  • Participant
  • ***
  • Posts: 1757
  • Likes: 190
Re: Help with my LM317 limiter?
« Reply #41 on: August 04, 2021, 08:44:32 AM »
The LM317 current limiter works by keeping the voltage across the sense resistor (between VOut and Adj pins) constant at 1.25V.  If the voltage across a resistor is constant, the current through it is constant also.

You adjust the current setpoint by changing the value of the resistor which, given the constant 1.25V across it, will change the constant current through it.

Unlike a pot, the LM317 adjusts automatically and constantly to changing loads.

As the conductivity of the cell changes due to accumulation of silver in the water, the LM317 adapts to keep the current constant.

As long as the cell voltage never rises above VIn to the regulator minus about 4V, the limiter will continue to function flawlessly.  Set it and forget it basically.

The issue with the LM317 is that the TO-220 version is only guaranteed for a minimum current of 10ma. You may well be able to get it down to 5ma but I doubt much lower.

The LZ version shows a minimum current of 2.5ma but I bet you can get it below this though I can't say how low.  Thats why I recommend the LZ version.  Its good to 100ma (NOBODY runs that high - usually 5-10-20ma which this thing can easily do).

The whole thing is only 2 or 3 parts. The LM317, a current setting resistor (fixed) or if you want adjustability, a pot and a fixed resistor in series with it.  The reason for the fixed resistor is that those little blue 25 turn trimmer potentiometers can achieve almost exactly zero Ohms on one end and you do want to limit the top-end current out of the LM317 so put a 50Ohm resistor in series with it between VOut and Adj.  I=1.25/50 = 25ma top end limit. On the low end you'll be able to go as low as the LM317LZ you have can go (something less than 2.5ma). I usually recommend a 1K Ohm potentiometer. That'll get you down to 1.25V or around there somewhere or as close to that as the part you have can go.

indocomp

  • Guest
Re: Help with my LM317 limiter?
« Reply #42 on: August 04, 2021, 10:52:55 AM »
Hi thank you for your explanation, i am thinking to limit the current at 53V because im using 53V adaptor so i need 3 item which is:
1. Potentiometer 1K Ohm
2. 500 Ohm 1/4 watt resistor
3. LM 317LZ

I hope i am not wrong
Another question is Do we need to connect all the potentiometer leg and connected like below picture i modified from your diagram
Or just use 2 legs which is input and signal and leave the ground alone unconnected
No 1 is from the LM317Z output
No 2 is delivering positive current to anode
no 3 Which is ground, im not sure about it
Please check it, thanks
So sorry i only know a little basic knowledge about electrical


Offline Gene

  • Expert
  • Participant
  • ***
  • Posts: 1757
  • Likes: 190
Re: Help with my LM317 limiter?
« Reply #43 on: August 04, 2021, 08:28:13 PM »
First off, did you really mean 53V?

The datasheet for the LM317LZ I have states a max Vin of 35V and the TO220 version states 40V.

I'd suggest you stay around maybe 32V just for a little safe area because I haven't checked all the various manufacturer's datasheets (everyone makes these parts and they're not all carbon copies of each other though they all meet the specs they state in their datasheets) so its always good to err a little on the side of caution.

With electrolyte and things adjusted properly you'll have 10-12V across the cell  The limiter will work famously in this scenario even off a 19.3V laptop charger (what I use).

Remember, max cell voltage plus about 4v the current limiter needs to continue to function properly plus 2-3V above this for a bit of wiggle room and you're good.  More is better if you can but stay below 35v for the reasons stated above.

Yeah, its kind of hard to find a wall cube that puts out more than about 15V these days and most over 12V are difficult to find.  Wall cubes are cheap.  Power supplies aren't.  HVAC and maybe bell transformers are usually 24V but they're AC, not DC. You need DC so that won't work.

The series connection of the pot and resistor are what the regulator maintains 1.25V across. Remember, for any given fixed resistance, the current through it, given the voltage is fixed at 1.25V by the regulator, will be maintained constantly at whatever you set it to.  In other words, the regulator maintains 1.25V across the regulator output (VOut) to what you labeled [2].

The equation to compute the current is

I = 1.25/R

where I is in amps and R is in Ohms.

So with the example you provided, the maximum current would be

1.25/12.5 = 0.1 (a.k.a. 100ma) which is excessive for a Colloidal Silver cell.

This is because at one end of the potentiometer setting, its end to end resistance will be zero or damn near zero so the only thing in circuit would be R1.

The minimum current (depending on how low the regulator can go and still regulate) would be

1.25/2012.5 = 0.0062 (a.k.a. 0.62ma)

You'll never get there, perhaps not even much below 2ma with an LM317LZ and maybe closer to 5ma with the TO220 version.

This is why I recommend a 50Ohm resistor and a 1K pot.

With these the max current would be about 25ma and the min would be about 1.2ma which gives you the maximum range.

Also, remember that those little blue trimmer pots are 25 turns.  The wider the current range you make them move through (0-25ma is 1/4 0-100ma) the less adjustment resolution you get.

In other words, if your range is 0-100ma and the pot is 25 turns that means for each single turn you're adjusting through a 4ma range. If its 0-25ma, now 1 turn is 1ma so you get 4 times the adjustable resolution making it much easier to hit your target without splitting hairs.

See the difference?

Normal panel mount pots are 1 turn. If you go this route you're going to need 2 pots in series - a coarse adjust and a fine adjust to be able to set your current accurately and thats just plain annoying though it may well allow you to use one turn (inexpensive) panel mount potentiometers.

Just an example to explain this.  You have 2 pots.  A 1K followed by a 100Ohm still with the 50Ohm in series for safety and to limit range for more accuracy.

So now, if you set the 100Ohm to its mid-point and adjust the 1K to get you to about the current you want (lets say 5ma which would be a resistance of 200Ohms), now you get to adjust the fine adjust pot +/- 50Ohms which gives you a one turn adjustability range of resistance 150-250 (about 4.2ma-6.25ma so call it 2ma so +/-1 ma above/below what you have the other pot set to with your fine adjust set to midpoint).

See how it works.

IF you can get a 50Ohm pot (no I'm not sure if it exists but I'm pretty sure 100Ohms does) your range would be about +/-0.5ma so 1ma total for the whole range of the 50Ohm pot.

Rotating a 1 turn pot through 25ma and trying to hit something to at least 1 decimal place in current is going to be rather tough.  With the 2 pot solution, it gets a helluva lot easier but its more parts and more cost but then you only need to build 1 of them, ever.

I've never seen a panel mount 25 turn though I guess its possible.  They're usually (high res ones - read ain't cheap) 10 turns and if you go this route, choosing to put everything in a little enclosure simply with a current adjustment knob on the outside, at 0-25ma, one turn would adjust through 2.5ma.

Using 2 panel mounts - the 1K and 100 gets you effectively a 10 turn pot. Using a 1K and a 50Ohm effectively gets you a 20 turn pot - just with 2 knobs (wink).

The LM317 current limiter is a series device. It has no ground BUT you have to connect the ground of the cell (cathode) to your power supply ground or negative to complete the circuit.  Current flows out of your power supply positive through the limiter, through your cell and back into the negative power supply terminal. And this is shown in your pictorial.  Build it exactly as show - just change the value of the pot and resistor as I've explained above and if you go panel mount for the potentiometers, seriously consider the fine and coarse adjust pair. You'll be glad you did OR bite the bullet and buy a 10 turn 1K.

Your choice.

Keep us posted of your progress.

indocomp

  • Guest
Re: Help with my LM317 limiter?
« Reply #44 on: August 05, 2021, 02:46:30 AM »
Ok thank you and i will keep it posted
I actually recently bought dc power supply for this colloidal silver generator stuff with current limiting but it only has 30 volts maximum output
Does this device actually solved current limiting problem or do we still need custom made circuit using LM137 LZ?
Because from what i experimented before when the current hit the maximum limit , the voltage will dropped

thanks