annoying Power Supply and LM317 problem

Thread Starter

Amann

Joined Apr 6, 2011
3
Hey everyone,

I've just started out doing the physical prototyping of a stepper motor driver I've been theorizing for a while. But since I'm a bit of an electronics newb, I quickly found myself stuck on an initial step I thought wouldn't need much planning. It's driving me INSANE. I was wondering if anyone here could provide me with insight:

I have a LM317t circuit setup like the "Adjustable Regulator with Improved Ripple Rejection" from the typical applications section of http://www.national.com/ds/LM/LM117.pdf using a 220 ohm resistor as R1 and 5k trimpot as R2 set to around 660 ohms to ideally provide a 5v logic supply.

The IC has an aluminum heatsink mounted on it and I've also placed a few resistors in series from the circuit's output to ground, ensuring it has over 10mA of load on it.

It provides 5v perfectly after connecting a cheap 10v dc wall wart (without an earth ground connector).

HOWEVER when If I feed it the + and - terminals from my 35v large toroid based linear unregulated power supply (that I plan to build my project around), the circuit outputs a constant value around 9v, regardless of whether or not I ground the - DC rail to the chassis (that the incoming AC ground is tied to).

If I tune the 5k trimpot while the circuit is powered up, the output voltage forms a bit of a parabolic curve with a minimum at around 9v: If I turn the resistor too low past a certain point, the voltage actually begins to steadily INCREASE. I avoided turning it in a direction that made the output more than 15v.

STRANGELY, if I leave the trimpot at 660 ohms (for the intended 5v output) the proper regulation begins to kick in AFTER I switch off the AC line to the power supply and the 2x22000 uF caps start to drain into the circuit. The LM317 output quickly drops to and stays at 5v for around 20-30 seconds before dropping slowly to 0.

I started using a multimeter to poke around the PCB of my power supply's bridge rectifier module in order to see if I was misunderstanding the outputs or something. There is a terminal labelled "c" (assuming it stands for "common") between the + and - terminals that I deduced as a floating line to a point halfway in between two equivalent ~1.6k bleed resistors that lie in series between the positive and negative DC output rails. While on, the measured voltage between the - terminal and "c" is ~-17v and between + and "c" it is ~+17v. Just for kicks I tried using "c" as the negative rail for the LM317 circuit and it actually worked perfectly. But since the current is limited by resistors, I don't plan on actually using it. Also, in another just for kicks moment I assumed "c" meant "tie me to chassis" so I tied it to the chassis/ground. Then BAM!! absolutely nothing changed.

So in summary: the voltage regulator circuit works if the voltage is 10v from a wall wart, 17v from the +/c of the large toroid PSU, or the 35v->0v transient voltage at +/- after powering down the PSU. It doesn't work right if it's connected in the same way while on.

I've been trying to figure this one out for a couple of days and I have the feeling I've overlooked something really stupid/simple. Are there some characteristics of my PSU that are incompatible with the LM317 circuit? Is there too much ripple in the PSU's output? Does the LM317 behave this way only if it's operating in some sort of failure mode? Is there some grounding issue? Did I overlook something on the LM317 datasheet? I'm very stumped and I guess slightly afraid of electrocuting myself if I continue blindly probing for the source of this problem.
 

#12

Joined Nov 30, 2010
18,224
You're over feeding it. Too much voltage across the regulator puts it into self protect.

Is this enough information to get you running?
 

Thread Starter

Amann

Joined Apr 6, 2011
3
Isn't lm317 supposed to support up to 40v differential voltage? or am i interpreting the datasheet wrong?
 

#12

Joined Nov 30, 2010
18,224
It's been a very long time since I read that datasheet, but I think it's called "safe area of operation".

Look at figure 4. As the difference voltage across the regulator goes up, available current goes down. That's because the chip heats up and goes into thermal safety.

I know this isn't the exact right datasheet, but it is close. The T package in a variable regulator.

Ummm...you have the right datasheet. It doesn't explicitly explain this property. I musta read it 30 years ago. Experience counts for a lot.
 

Attachments

Last edited:

Jaguarjoe

Joined Apr 7, 2010
767
The 220Ω resistor between Vout & ADJ is too large. An LM317 is not the same as an LM117 in some regards. The 117 only requires 5ma of output current to run. The 317 requires as much as 10ma. therefor that resistor should be 120Ω to ensure proper operation.
 

Thread Starter

Amann

Joined Apr 6, 2011
3
yeah thanks for the tips guys, I'm probably just going to use a separate, lower voltage supply to derive the 5v's. going from 35v down to 5v is kind of inefficient anyways
 

nbw

Joined May 8, 2011
36
yeah thanks for the tips guys, I'm probably just going to use a separate, lower voltage supply to derive the 5v's. going from 35v down to 5v is kind of inefficient anyways
It sure is. Can you put something useful in between, such as high-power LEDs or a fan?
 

Jaguarjoe

Joined Apr 7, 2010
767
How about a switcher? Somewhere around here I saw a 3 pin TO-220 like switcher that pin for pin mimicked the 7805. I think it was called a 78xxx05.
 

THE_RB

Joined Feb 11, 2008
5,438
I think you should use a smaller transformer with a lower voltage for your 5v supply.

The "35v large toroid supply" probably puts out about 45 to 50v into the PSU cap at light loads. That is not a good way to get 5v. :)

You could buy one of those little $5 9v transformers that makes about 12v DC at light loads, or you could even put an over-wind coil on the toroid to give you another transformer secondary at about 10v to 12v DC.
 
Top