LM317 Voltage Reference not constant

Hello,
This is my first post on the forum. I've built a laser before but now I'm on summer break from school so I have time for a more in-depth project.

I've ordered a SF-AW210 Blu-Ray diode which is a few days away so I've just been working on the driver using a test load in the mean time. My driver is basically the one from the "It Can Be Done" page, except with 25 ohm pot, three 10 ohm resistors in parallel, and a 47 uF capacitor. And I am simply using a 9V battery as my power source.

For my test load I'm using the 6 1N4001 diodes as is needed to represent the blu-ray diode. I couldn't find any 1 ohm resistors so to measure the current I've used two methods:
1) three 10 ohm resistors in parallel in place of the 1 ohm to give me a 3.3 ohm eq resistance (you may be noticing a pattern here, I have a crapload of 10 ohm's)
2) removed all diodes and just completed the circuit through my DMM set to measure current.

Now here, when I started taking measurements, is where I noticed things didnt seem to be adding up correctly. With my three 10 ohm resistors (parallel -> Req = 3.3 ohm) in series with my 25 ohm pot across the constant voltage reference of the LM317 I calculated beforehand that my current range should be something like this:

Min ~ 1.25/(25 + 3.3) = 44mA

Max ~ 1.25/(0.2 + 3.3) = 350mA (give or take)

This was the goal from the beginning, getting a nice range where I could easily get to about 200 - 250mA. But, when I started taking measurements at my load I found that while my Min current was behaving beautifully, giving me right around 40 - 45mA, there was something wrong with my max current. It was increasing as I turned my pot down, but not as fast as it should have been. The max current I was getting with my pot all the way down was about 145mA which I knew was nowhere near as much as it should have been.

So what I did was went back to my driver and started taking measurements there. First I checked to make sure that the resistances were correct. It swept from about 3.5 ohms to about 29 ohms so I knew that the problem wasn't in the resistors of the pot, they were functioning perfectly.

So then I rehooked up the power and started measuring the voltage across the constant reference pins. I started with my pot turned up and the voltage across there seemed good, about 1.19-1.22V. BUT THEN, as I started decreasing the resistance between the pins the voltage started decreasing with it!!!! With the pot turned all the way down I was getting only about .5V or less across the pins that were supposed to remain at a constant 1.25V. Thus, comes my max current of about 140mA (.5/3.5 = 142mA).

I'm just wondering has anybody encountered this before? I have no idea what's going on. Just to test it and be sure, I replaced the resistors with a really large one just to see if the voltage would stay at 1.25V or change. When I did this the voltage across the pins shot up with it, I believe all the way up to 4.5 or 5V. So clearly something isn't working properly with the LM317, the "constant" voltage seems to be changing with the resistance rather than staying fixed.

Does anybody have any advice on this? Should I just get a new LM317 and go from there or is it something else I'm doing? And does anyone have any additional pointers for building a laser from a SF-AW210 blu-ray diode?

Thanks

Here's a picture of my circuit on my protoboard if you'd like to see. Sorry it's kind of tough to see some things (also note that not everything in the picture is exactly how I described it above. I took this before it was finished):

Link

I would guess its your battery. Those rectangle 9V don't produce enough current. Try 6 AAs. Or better a bench PSU.

hmmm interesting, that never even dawned on me. I thought the LM317's only required a certain voltage to operate properly. I'll give the AA's a try, thanks

Kevlar said:

I would guess its your battery. Those rectangle 9V don't produce enough current. Try 6 AAs. Or better a bench PSU.

Click to expand...

Ok, I gave that a try... well sorta, I already had a 4 AAA battery holder and a 2 AA battery holder so I hooked those together. The source voltage was pretty decent, about 8.5V, but once again when I hooked it up to the circuit the current flow was crap and the voltage across the reference pins was all over the place. It was even worse than with the 9V battery.

Was it because I mixed in some AAA's with the AA's? Do you have any other suggestions? Is it possible theres just something wrong with my LM317?

From your picture... I can't see how your LM317 is wired to your
Test load...

BTW... since you "have a crapload of 10 Ohms" using 10pc 10 Ohm
resistors in parallel for your Test Load will give you a 1 Ohm resistor
making it simpler to measure current flow of your driver... 1mV=1mA...

Jerry

lasersbee said:

From your picture... I can't see how your LM317 is wired to your
Test load...

BTW... since you "have a crapload of 10 Ohms" using 10pc 10 Ohm
resistors in parallel for your Test Load will give you a 1 Ohm resistor
making it simpler to measure current flow of your driver... 1mV=1mA...

Jerry

Click to expand...

Sorry about the picture, here's a closer up shot:



Link There are other pics of the same thing in there as well.

The orange wire is connected to the positive terminal of my source. You can see the three resistors in parallel connected to the middle pin of the LM317. The other side of those resistors is then connected to dark green wire which goes into my pot. the other green wire is coming from my pot and meets the red wire (which is connected to the left pin of the LM317), capacitor, and diode with the lime green wire as the output bus going to the test load.

I realize that a bunch of 10 ohms will give me a 1 ohm eventually but I figured using 3 would be enough. I just get a higher voltage and divide it by 3.3 to get my mA. At worst this extra resistance should only add a little bit of current to my reading. But that's not the problem, the problem is in my driver. I still cant seem to figure out whats going wrong here. Does everything else look okay? Am I missing something obvious here?

Thanks

Maybe you're getting shaky contacts from your breadboard.
Try soldering everything together.
I've built a couple of LM317 drivers and the only time the ref voltage was lower than 1.2x
was when there was not enough input voltage.

Like when I was trying to drive a ~3V red diode from a single LiIon cell (~4V) for example.

Have you been to this page already?
http://www.rog8811.com/laserdriver.htm

anselm said:

Maybe you're getting shaky contacts from your breadboard.
Try soldering everything together.
I've built a couple of LM317 drivers and the only time the ref voltage was lower than 1.2x
was when there was not enough input voltage.

Like when I was trying to drive a ~3V red diode from a single LiIon cell (~4V) for example.

Have you been to this page already?
Laser driver - It can be done

Click to expand...

I'm pretty sure my breadboard is solid, I use it for school. And I've been measuring the input voltage right on the pin with reference to the ground and it's been around 8-8.5V everytime. That should be more than enough for the LM317 to work.

The only thing that I can think of is that maybe my LM317 is bad. I've had it for years and I used it to experiment on my other laser so I'm thinking that perhaps at some point in the past I crossed my wires and fried something. The LM317 seems pretty robust so that wasn't the first thing that entered my mind but it's really the only thing I can think of at this point. I'm gonna go pick up another and see if it fixes my problems, I hope so.

Thanks

Ok,
I think I fixed the problem. Looks like you had the right idea Kevlar, thank you so much. When I first tried your fix it was actually worse than the beginning. But then I tried expanding upon it by putting about 11V to the input and BAM! everything started working like a charm immediately. I'm not sure why it took so much voltage but everything seems in order. I've got the 9V running in series with 2 AA's and that seems to be working great. I might look into a better battery in the future because it doesn't seem like I should need so much voltage to make it work, but this definitely works for now. It's sweeping beautifully from 42mA to 350mA every time.

I actually have two power switches, one a standard switch, and then another illuminated push button so I'm gonna see if I can wire it to dissipate some that extra voltage in the light. If you have any advice in this area please let me know. Thanks guys

It's a dropout problem, common in linear circuits.

You are using a BR diode (needs 5.5 to 6 V, depend from the FV of the diode) ..... plus a LM317 regulator (that have an internal dropout of approx. 2.5V), and the RSET (that drop another 1.25V), so, you need at the input of the circuit a voltage that must be AT LEAST 0.5 to 1 V MORE than (6+2.5+1.25) 9.75V ..... this means 10.25 to 10.75 V, for the chip work correctly as regulator.

Thanks HIMNL9,
after going over each part of the driver and load with my DMM I started seeing how there just wasn't enough voltage there to work. I was just so thrown off for a while because my understanding of the "It Can Be Done" page was that 9V was all that would be needed and any extra voltage would be dissipated as heat. How did he get that 9V, is that calculation correct with a better (rechargeable perhaps?) power source than what I'm using?

If you use two 9V bricks in parallel, you will get enough current delivery to satisfy the lm317, gor about 1 minute and a half. Then you'll have toswitch off and wait for about two to three minutes. if the bat cans get warm, switch batts or add another one in parallel. They'll work, but they are a pain. Burgess (Dura-cell) used to sell a high current delivery pack battery, but it was 3"x5"x3/4" and expensive. look into stacking LiPo type batts (cr2's, cr123's, and on & on).

Well, a 9V battery can work good for a red diode, cause they have a FV of 2.8 / 3V max, but for a BR is not enough, cause they go from 5.5V to 6V (i had a PHR with 6.5V FV, but it was a single case).

Also, if you have the space, try to always use 100 or 220 nF ceramic capacitors on the in and out of the regulators (in parallel with electrolitic ones, if the electrolitics are more than 10uF) ..... is difficult that this happens, but time by time, linear regulators go in self-oscillation, both if used without capacitors, or with high value ones, and these ceramic capacitors prevent this.

Ahh ok, so it was more of a power issue. I'm gonna look into upgrading my power source but for right now what I've got should work well enough without getting too hot or anything.

I'm running out of room quickly but I think I'm gonna move up to a bigger project box to house everything. My breadboard took up a little more space than I was expecting. I think I should have room for the ceramic capacitors. How exact do the values need to be? I think I've got a couple 470 nF caps on hand.

.47 Mikes should work fine on the input side, but you need a little more on the output. I think that the input cap has more to do with quelling the oscillations and only has to be a small value. It iw a small issue with a battery supply and more of a concern with a.c to d.c. conversion with ripple on a loaded supply, or one without sufficient filtering. It is caused by the error comparator in the IC continually trying to adjust the output to keep the reference voltage set. I don't think (but could be wrong) that the output cap has much to do in a system with a relatively fixed load (no constant switching in the load circuit, causing voltage or current requirement fluctuations). There seems to be a school of thought here that the big output is needed to protect the laser diode from transients (don't understand it much, seems like a big cap would cause more problems than it would solve, but the more experienced guys here do it, it works, so copy-cat it).

Ok thanks... Yea I've got a 47 uF on my output so I figured that should be big enough. The .47 micro I was talking about was just for my input cap. I wasn't even aware that caps were also used on the input as well before you guys mentioned it so I wasn't sure about my values, but it sounds like I'm good with what I've got.