Where did you personally use or see LM317 applied to?

dl324

Joined Mar 30, 2015
16,996
I got a 650mA (instead of 700mA as with earlier experiment) on my PSU readout, and LM317 was not hot anymore. Warm like 5dgr over the ambient but not hot. This is good news ! But the size is enormous of that heatsink ! I dont like it.
I still didn't get 1A on the Amp meter !
I had a wonderful idea to actually measure the motor coil and it give me 0.6R (or 0.5R because was fluctuating a bit). You can simulate my motor coil using 2x1R in // to give you 0.5R and try my cct I have and see if you get the same results as I have here, and after that see if you can rise it up to 1A over this load.
Your measurements don't make sense.

You say that the motor draws 1A from your power supply, presumably set to 1.5V. That would imply an impedance of 1.5 ohms.

You only get 650mA from the LM317. What is the output voltage of the LM317?

You expect the motor to draw 1A, but you measure it's resistance as 0.6 ohms. That would try to draw 2.5A from the LM317. Even with sufficient heat sinking, the regulator would go into protection mode.

I have my workbench area torn apart for some rearranging. It will take me some time to gather parts to do any testing.

crutschow

Joined Mar 14, 2008
34,682
Can you simulate for me the Power over LM317 ? I imagine it might be a few mW like 100mW or lower. Am I right?
Not quite.
Here's the sim with the power dissipation of the LM317 plotted (red trace):
The power starts at 10W with the battery at 11V and then tapers off towards zero as the battery reaches full charge.

The dissipation curve for an actual battery would likely be different, since the charge voltage of a battery is somewhat non-linear with respect to charge state, as compared to the capacitor used in the sim.

The power would be less, of course, if a lower voltage supply was used (about 16V minimum).

dl324

Joined Mar 30, 2015
16,996
I set up an LM317 with vin=5V, vout=1.49V and a 1 ohm load resistor. I got 1.36A out and the LM317 heated up to 105F (~40C) after 5-10 minutes.

joeyd999

Joined Jun 6, 2011
5,357
I set up an LM317 with vin=5V, vout=1.49V and a 1 ohm load resistor. I got 1.36A out and the LM317 heated up to 105F (~40C) after 5-10 minutes.
View attachment 312396
For measurements like this, I usually thermally epoxy a thermocouple to the tab of the TO-220.

From there, it is easy to compute the junction temperature using θjc from the datasheet.

sghioto

Joined Dec 31, 2017
5,420
I jumped into a LM317 Voltage Regulator
Are you sure that LM317 is genuine?
I had some LM338, 5 amp regulators I thought were genuine but they would only output 3 amps max.

q12x

Joined Sep 25, 2015
1,734
I discovered something. I powered the motor again directly to 1.5V and this time it give me around 700mA instead of 1A as I got before. It was fluctuating, sometimes 800mA, sometimes 650, as I was unpower and power again and leave it for a few seconds to read the Ampmeter. So that initial 1A I got in the very begining was ... something else.
So the 700mA with the LM317 in circuit might actually show the real thing?
I start to believe this motor might be electrically aged? it is used for sure , but it might have got hot at some point in its life, and now, when Im testing it, is giving me signs of aging? Thats why the amp readings are not consistent. Im also holding with my fingers its rotor from spinning so that might also affect the current in it...? But I doubt it. Maybe the flimsy contacts might influence its current as well... I dont use sturdy contacts when experimenting... you know how it is.
So I find a solution, at least for the motor itself, I raised the voltage to 2V now, (before was 1.5V) and the amperage raised to ~1.5A when I brake completely the spindle.
The idea is that at 1A, it is powerful enough to push the hard spring in the cutting blade of the beard trimmer head. Thats why Im insisting into keeping it at least 1A.
I didn't set up LM317 to deliver 2V and 1A yet. Its the next phase. But is an interesting update.

Your measurements don't make sense.
- Im only reporting my readings.
You say that the motor draws 1A from your power supply, presumably set to 1.5V. That would imply an impedance of 1.5 ohms.
- Exactly at 1.5V, not presumably. At 1A. Yes, I agree, should give 1.5R... interesting, I didnt stay and think on this way... very smart observation.
You only get 650mA from the LM317. What is the output voltage of the LM317?
- 1.5V on Vout and powered from 5V on Vin. I measure it yesterday while communicating with another friend who asked me the same question.

You expect the motor to draw 1A, but you measure it's resistance as 0.6 ohms. That would try to draw 2.5A from the LM317. Even with sufficient heat sinking, the regulator would go into protection mode.
- After today observation, I think my motor might be old and aged electrically and most probably in its life got super hot and that weakens (usually) its internal conductivity, isolation, etc. Im only suspecting, I dont know for sure. Or I might have burn it a little bit with my experiments? But I know and remember for SURE I didnt even warm it up, nor making it hot either. So I highly doubt I damage it while it was running it cold and for short periods of time, always under my eagle blue eye.
So the fact Im getting about 700mA on Full Motor Stop/Load, this number I am reading it directly from my variable PSU. (this Amps value is usually fluctuating a lot, rarely I got it at exactly 0.70A as in the picture).

I start to believe.... that the motor itself is indeed consuming only 600mA(with its 0.6R coil) and LM317 another 100mA? so the total 700mA that we see on display is the total circuit amperage. Right?
So the question remains about why I got a 1Amp reading when I was testing the motor by itself the first time? Some kind of temporary short? Did I brake it ? lowering its internal resistance? Again, I highly doubt it. But who knows... made in china with surprises. Might have been a PSU blurp/oopsie/misreading ? The thing is that I got that 1A for a relatively long period of time while testing. But today, it is giving me these new results... hmmm - very strange.
- About your point of "the regulator would go into protection mode." - this is also another possibility while using LM317 in cct.

I set up an LM317 with vin=5V, vout=1.49V and a 1 ohm load resistor. I got 1.36A out and the LM317 heated up to 105F (~40C) after 5-10 minutes.
- I probably should do this exact same experiment that you did here to confirm if my LM317 is the real deal or not. As my american freind sghioto also mention it. Hmmmm...
Thank you for the experiment ! Very good that you make it for me.
Are you sure that LM317 is genuine?
I had some LM338, 5 amp regulators I thought were genuine but they would only output 3 amps max.
- I will do dl324 experiment to confirm !

Joined Feb 20, 2016
4,494
The motor will be generating a lot of electrical noise that can muck up your meter. One thing that can help is to add a reverse diode across the motor.
The motor start current will be higher that the run current as the back EMF has nor been generated as yet so the current will be limited by the motor resistance. When you load the motor, the current will go up.
For a QUICK test, measure the motor current with the shaft held so it cannot turn. Do not hold it stalled for long.

Do you have a reasonable capacitor on the regulator output?
Caps on the reg's inputs and outputs are recommended!

q12x

Joined Sep 25, 2015
1,734
I used these Resistor values:

You can read each rezistor value from real resistors in the image! 200R and 1k.
It started with 400mA and it increased slowly but surely to 600 and then to 700mA as seen in the image.

Do you see those red link wires at the end of the breadboard? The positive link got not hot but warm.

The yellow link that connects the positive rail to the Vin pin, also got the same warm temperature over it. You can see it under LM317 and its enormous heatsink.

The link that goes from Vout to 1R5W Load resistor, was cold, but the Load was a bit warm. The red links were more warm than the Load.
So.... I should have got 1.5A on Vout, right? instead of this increasing 400 to 700mA.
Then, its clear, my LM317 are ... busted? But they do work, they do something.
- I made a secondary experiment, since I have 6 LM317 brand new, I took out this #1 that I tested so far and change it with another brand new - never used - let's call it #2. Absolutely the same behaviour. It didnt start from 400mA as #1, but it started from 500mA and rising quite fast, I stop it at 600mA. I changed back with #1 as it was before.
So... they are all the same.
Very interesting !
I also measured the Voltage across Load and is in accordance with the amperage on the PSU , 0.50V.
I also measured the Voltage on the orange link that goes from Load Resistor leg to the other side and I got an increased voltage reading 0.6V ! (is marked with a red arrow in the image)
You can see in the image each connection is entering into 1 hole of breadboard.
Then I measured the Voltage on the leg of the 1k resistor, the next hole after the orange link, on the Vout line. Also increased to 0.7V or so.
Then I measured the Voltage on the leg of LM317 the middle one, Vout, and I got 0.8V on it.
Then I got back and measured the Voltage across Load and I got 0.52V, so I moved quickly, not to think everything was increasing as I was measuring. It was a very fast reading. I am baffled ! I believe the links and the metal nodes in the holes, act as very low resistors in themselves ? Like little shunts? The hell ?

Last edited:

Jon Chandler

Joined Jun 12, 2008
1,071
The recommended value for R1 is 120 ohms, needed to maintain the minimum current needed for voltage regulation.

This has long been a confusing point, because datasheets for LM117/LM317 show a 240 ohm resistor for R1. The LM117 needs a minimum current of 5mA to maintain regulation while the LM317 requires 10mA.

If you are drawing a greater current at all times, a higher value for R1 is ok, but the regulator will drop out of regulation if the load is removed.

From the datasheet:

q12x

Joined Sep 25, 2015
1,734
The recommended value for R1 is 120 ohms, needed to maintain the minimum current needed for voltage regulation.

This has long been a confusing point, because datasheets for LM117/LM317 show a 240 ohm resistor for R1. The LM117 needs a minimum current of 5mA to maintain regulation while the LM317 requires 10mA.

If you are drawing a greater current at all times, a higher value for R1 is ok, but the regulator will drop out of regulation if the load is removed.

From the datasheet:

View attachment 312457

View attachment 312458
Without the Load it is regulating fine to 1.5V with the existing resitor values I have right there !!!
If the mA on the ADJ were out of specifications, the regulator was not working at all !

Jon Chandler

Joined Jun 12, 2008
1,071
*shrug* Just trying to pass along information. Do with it what you like.

MrChips

Joined Oct 2, 2009
30,935
Don’t put high temperature components on your breadboard. They will destroy the breadboard.

SamR

Joined Mar 19, 2019
5,059
Use thermal paste between the component and the heat sink!

dl324

Joined Mar 30, 2015
16,996
I believe the links and the metal nodes in the holes, act as very low resistors in themselves ?
What gauge are the wires? How old is the breadboard? Have you always abused the sockets by plugging in things that are larger than the diameter of AWG22 wire?

For example, the leads of a TO-220 package are technically too large for the sockets. With repeated abuse, they could lose some of their ability to "spring back" causing poor (i.e. higher than normal resistance) contacts. I twist the leads on TO-220 90 degrees before inserting. Haven't always done that, but I do now.

"High" current, wire resistance, and resistive connections could account for some of the voltage drops you appear to be seeing.
Without the Load it is regulating fine to 1.5V with the existing resitor values I have right there !!!
If the mA on the ADJ were out of specifications, the regulator was not working at all !
LM317 aren't guaranteed to regulate without a minimum load of 10mA. Many will work with lower currents. I had one that wouldn't work until the minimum load was around 20mA.

dl324

Joined Mar 30, 2015
16,996
Some solderless breadboard data.

From Global Specialties:

From GoldTool:

I do an amp or so max and find #24 wire a bit too easy to pull out.

TO-220 case spec from National Semiconductor LM317 datasheet:

SamR

Joined Mar 19, 2019
5,059
For those breadboards mounted on a metal backplane (typically aluminum plate), I ground the backplane! Especially helpful for any oscillating circuit or preventing noise interference.

q12x

Joined Sep 25, 2015
1,734
I set up an LM317 with vin=5V, vout=1.49V and a 1 ohm load resistor. I got 1.36A out and the LM317 heated up to 105F (~40C) after 5-10 minutes.
image in #43
I had to get real close into your picture you made there and guess what resistor values you used and where everything was connected. After I read the most obvious R2 30R from ADJ to gnd, I experimentally find R1 as 150R in my case and with the imperfect breadboard I have.
1-So now I have lower resistance on the ADJ pin voltage-divider.

2-I moved closer the power clips to the pins of LM317, and not as before, circling literally around all the breadboard.

Here is the result: starting with 400mA and rising up as long as I leave it in power.

Don’t put high temperature components on your breadboard. They will destroy the breadboard.
I know exactly what you are saying ! But I did specified already, and you should believe me, that nothing is/was HOT ! It is only Warm ! And after 5min or so of functioning. If ambient temp=25dgrC the warm temp was probably at around 30-35dgrC. I didnt measure the temp, just finger it. So this temp is very safe.
Some solderless breadboard data.
Yes, I will make a note on all my breadboards: "allowable wire diameter of 0.4-0.7mm"
I usually feel when something is over sized and Im not forcing it...but sometimes I do. Rarely. But these breadboards I have are 10y old or so, and everything I build, was exclusively on them. But I have 6 breadboards so, some are newer than others since some I used rarely and others frequently... Im always keeping a very skeptical eagle eye on these breadboards.... and I never got false results because of them. All the errors I got was from my stupidity, components and wires, but I cant recall ONCE a breadboard failure... so I trust them now that they give the real result and not some distorted weirdos results.
- I believe my breadboards are fine,and the cct is weird ! Most probably LM317 is half good as a "authentic" one? Definitely not busted because it is regulating, when is under powered (in normal conditions), not over-powered as it is now.

- I also believe maybe my original formula for calculating is wrong ?!?! - maybe we should use other formulas than mine?
As a reminder this was my original formula(blue) and LM317 data(green)

P=(Vin-Vout)Iout (Iout or ILoad)
The voltage drop over LM317 is 5V-1.5V=3.5V.
the maximum Amperage the circuit was running was 700mA and LM317 was getting hot.
So the Power disipated over LM317 is
P=(5-1.5)V*0.7A= 3.5V*0.7A= 2.45W

my LM317T (what I have) << this is data from its datasheet, to ease searching, I collected it here as a convenience.
Input Voltage : 2-40V
OUTPUT CURRENT Maximum Load Current 1.5A
OUTPUT VOLTAGES adjustable 1.2 to 37V range.
Power: 2.5W @25dgrC (from a graphic) and 3.5W @0dgrC
Power Dissipation: Internally limited

Last edited:

dl324

Joined Mar 30, 2015
16,996
I had to get real close into your picture you made there and guess what resistor values you used and where everything was connected. After I read the most obvious R2 30R from ADJ to gnd, I experimentally find R1 as 150R
I knew a wire was partially obscuring the 150 ohm resistor. It's value is easy to calculate because you know the output voltage was 1.5V. That meant the drop on the 30 ohm resistor (which you could read) was 0.25V. From Ohm's Law, you know that the current was 8.33mA. So you can calculate the value of the current setting resistor as
$$\large R=\frac{V}{I}=\frac{1.25V}{8.33mA}=150\Omega$$

You could also use simple arithmetic. The 30 ohm resistor was dropping 1/6 of the voltage, so the current setting resistor needed to be 5 times larger.

At least the photo had good depth of field. I used a DSLR with an F-stop of f/25.
I believe my breadboards are fine,and the cct is weird ! Most probably LM317 is half good as a "authentic" one? Definitely not busted because it is regulating, when is under powered (in normal conditions), not over-powered as it is now.
These days, if you're not buying from a reputable place, you can't be certain what you're getting.

I suspect "erratic" results you're seeing are because the motor current will vary with load.

q12x

Joined Sep 25, 2015
1,734
mister @dl324 see if you can find another -better- formula than what I find !

dl324

Joined Mar 30, 2015
16,996
Simpler (better) than:
$$\large V_{out}=1.25V(1+\frac{R_2}{R_1})$$