Well to get 10 mA I got to increase my resistors:

82 Ohms
8.2 Ohms
8.2 Ohms
75 Ohms

Total: 173.4 Ohms

This gives me 10 mA at 3.4V

 

Something isn't right. That would make your Vref = 1.734V, which is way out of specifications.

What is the voltage of your supply? Do you have a couple of capacitors on the input of the LM317, say 0.1uF and 10uF?

Measure the voltage between your OUT and ADJ pins. That is where you measure Vref.

If your Vref is measuring less than 1.3v, it's likely that your LM317 is OK, but your power supply has a bad rectifier.

 

Vref is measuring 0v with my LED

Vref is measuring 1.3v with another UV LED

My power supply is 12V I have no capacitors, but I tried my power stabilizer on 3 different power supplied 12V-14V and got 10mA 3.4V with my UV LED connected.

 

Vref is measuring 0v with my LED

Then your LED is open; ie: not conducting current. It may be connected backwards, or there may not be enough voltage across it to conduct. However, if Vref is 0v, then there is no current flow.

Vref is measuring 1.3v with another UV LED

OK, then the problem is most likely that the regulator is not able to meet the regulation specifications at minimum current, which is 10mA. With such low current demand, you really should be using something like an LM317L, which has a minimum current specification of 5mA, maximum of 100mA. The problem with using the higher-rated LM317T is that you are at the ragged lower edge of what it can regulate, and the formula for calculating actual current is no longer accurate.

My power supply is 12V I have no capacitors, but I tried my power stabilizer on 3 different power supplied 12V-14V and got 10mA 3.4V with my UV LED connected.

At a minimum, you should have a 10uF cap between your LM317's INPUT pin and ground. This helps a great deal to eliminate transients on the supply.

 

Is this what I need ? Are you sure I need 10uF I see 0.1uF here http://cache.national.com/ds/LM/LM317L.pdf (page 9)

They are very large

http://www.lemona.lt/index.php?page=item&i_id=45515

Or this:

http://www.lemona.lt/index.php?page=item&i_id=26863

 

Sorry, I can't read Lithuanian, and Google can't translate it yet.

However, the reason the capacitors that you are looking at are so large is due to their higher voltage rating.

A standard "rule of thumb" for capacitor voltage rating is to look at the circuit to determine the highest voltage that will be applied at the place that the capacitor is to be used, and then double that voltage. Since your supply is 12v, then use 24v or higher rated caps. Doubling the voltage rating reduces the "leakage current" throught the capacitor, thus less heat needs to be dissipated.

You could use a smaller uF capacitor. But remember, replacing the TO-220 package LM317 with a TO-92 package will free up a good bit of space.

I have a number of 10uF 50v non-polarized electrolytic capacitors that would easily fit in the space that will be vacated by your TO220 LM317. However, I don't know what's available to you - and since I don't read Lithuanian, I'm afraid I won't be of much help searching on your vendor's websites.

 

Well, I took a stab at it, and found these:
http://www.lemona.lt/index.php?page=item&i_id=84232
They are much smaller than what you were looking at. One or two of them would probably be enough.

 

Well, I took a stab at it, and found these:
http://www.lemona.lt/index.php?page=item&i_id=84232
They are much smaller than what you were looking at. One or two of them would probably be enough.

These are ceramic capacitors and only 1uF are you sure they will work?

I'll try to order LM317L too.

The site is available in english too look at right top corner of page, click on EN.

 

Ceramics are OK. They will have a decent response to fluctuations. A few uF will likely be enough. Don't use any on the output or adjust terminals though.

An LM317L should give you much finer control over the current at the low levels you're working with.

 

Thanks for your help

I'll post back with my results.

 

Got the required parts today, just wanted to ask does it matter where I put on/off switch? Is it better on 10mA regulated line conected to LED or I need to use it on power line to the device?

 

Connect the switch in the power line to the device. If you interrupted the connection to the LED, the regulator would still be consuming some current.

 

Everything works fine I soldered 157ohm resistor 81 and 75 and got about 8mA stable output that makes my UV LED shine not very bright.

The power supply output is stable, tried it on 3 different computer PSU.

Thanks.

 

What is the voltage between the OUT and ADJ pins?
It should be right at 1.25v from your 8mA output report.

Now you can "fine tune" the current, if you wish.
For a given resistance:
Iout = 1.25/R1 (12.5 Ohms <= R1 <= 250)
For a given current:
R1 = 1.25/Iout (5mA <= Iout <= 100mA)

 

What is the voltage between the OUT and ADJ pins?
It should be right at 1.25v from your 8mA output report.

Now you can "fine tune" the current, if you wish.
For a given resistance:
Iout = 1.25/R1 (12.5 Ohms <= R1 <= 250)
For a given current:
R1 = 1.25/Iout (5mA <= Iout <= 100mA)

Hmm, since I allready used the device to gather some data I can't really change anything or I will need to measure it all again form scratch.

I think last time it was 0, but perhaps because I used 317T and it' rated at 10mA.

IT's quite hard to get precise measurement with analog multimeter now I see it about 1.2volts.

 

Ahh, OK. Never mind.

 

HI
sgt Wookie is right , but you could use an LM1084 ,This has only 1.3 v drop, it still will not run to its full potential but it will run

 

HI
sgt Wookie is right , but you could use an LM1084 ,This has only 1.3 v drop, it still will not run to its full potential but it will run

You are incorrect.

The LM1084 is a 5A regulator; it requires a minimum load of 10mA to provide guaranteed regulation, just like the LM317T does.

Our OP is supplying 8mA to his LED. The LM1084 would not be able to provide guaranteed current regulation; it is below minimums.

This required the low-current version of the LM317, which the OP has procured, and has successfully installed and has provided sufficient information to indicate that it is performing as specified in the datasheet.

If they tried changing to an LM1084, they would face similar problems as they were with the LM317T; attempting to regulate at the ragged edge of the specifications.

 

Hi, I still do not loose hope to use USB 5v to drive my power stabilizer.

I read on some website that there is a chip potential inverter 7660 that could make from 5v like 10volts. This would be enought for me. I see alot of chips here http://209.85.129.132/search?q=cach...ts_ic13.php+ICL+7660&hl=lt&ct=clnk&cd=1&gl=lt

Just dont know what I need and how to connect it?

 

You shouldn't need to boost the voltage and then use a constant current circuit.

What you need to do is accurately measure the voltage across the LED when 8mA is flowing through it. Use a DVM/DMM (digital voltmeter/digital multimeter) - they are available quite inexpensively nowadays.

Then you can calculate what size current limiting resistor you need.

For example, let's say you measure 1.23V across your LED when 8mA is flowing through it.
You want to power it from a 5v source.
Rlimit = (Vsupply - Vled) / DesiredLEDCurrent
Rlimit = (5 - 1.23) / 8mA
Rlimit = 3.77 / 0.008A
Rlimit = 471.25 Ohms
470 Ohms is a standard value. Let's see what that will do to the current through the LED.
I = E / R
I = 3.77 / 470
I = 8.02mA