I've run into a curious situation. Please refer to the attached schematic. You can see that we are taking a portion of the AC input and feeding it into a half-wave rectifier. The output then goes into a LM337 negative voltage regulator. According to my calculations (which could be way off base) I should be looking at something around -3V output from the LM337.

I began to populate the PCB with components starting with the power distribution components. U4 gives me a nice 24 Volt output. U5 gives me a 33.5 volt output - somewhat less than I anticipated but I can live with it.

U1 is where I am having a problem. When I first powered up the board I was measuring -4.5 volts on pin 4 of U2 and U3. Well OK, perhaps I got the values for R4 and R5 wrong. But wait, low and behold the output of U1 is not holding steady. It is gradually increasing (decreasing?)!! The last voltage reading was -8.5 volts. What is going on here??

Some additional information:

• R2 is quite warm but not so hot I cant grasp and hold it.
• The voltage drop across R2 is 3.8 VAC and slowly increasing.
• U1 is in a TO-92 case
• All measurements (except AC) with respect to ground

 

What do you see across C4?

Trying to understand how current flows in your negative supply generator, which is basically a "charge pump"

On negative swings the ACINPUT node is clamped to ground by BR1, so that point can never go more than a diode drop below ground.
On positive swings, ACINPUT should be ~= +INPUT

So this supply will vary as the voltage on C1 varies, as it's basically clamped to it by BR1

IN4148's seem wimpy for a power stage?

 

1. Why do you think the U5 output is incorrect for the programming resistors you show?

2. R5 is way too large. Re-read the datasheet.

ak

 

The connections to U1 are incorrect. If you remove it, turn it 180° and refit it, it will be connected correctly. I don't know whether it will have been damaged.

 

1. Why do you think the U5 output is incorrect for the programming resistors you show?

2. R5 is way too large. Re-read the datasheet.

ak

I have no idea where my head was when I calculated those resistor values. Old age, I suppose. R5 should be 240R and R4 around 1k. I'll give it a go.

 

The connections to U1 are incorrect. If you remove it, turn it 180° and refit it, it will be connected correctly. I don't know whether it will have been damaged.

The data sheet seems to contradict that. See attached.

 

You need to double check that.

 

The data sheet seems to contradict that. See attached.
View attachment 127626 View attachment 127627

Those two pictures contradict each other. One has the middle pin as input while the other has that pin as output

Or should that be the 'muddle' pin?

 

Those two pictures contradict each other. One has the middle pin as input while the other has that pin as output

Or should that be the 'muddle' pin?

The image on the right is the "bottom" view, so they are in agreement.

 

The middle pin is the middle pin wherever you view it from. One says input, the other says output. They are not in agreement.
[Edit] My original post stands. The marking on your board (assuming that is a component side view) for U1 is reversed.

 

Where did you get that image on the right?

Check again. The middle leg is VOUT.

 

From National Semi:

 

The data sheet seems to contradict that. See attached.
View attachment 127626 View attachment 127627

Note thta your right-hand image is for a fixed voltage regulator -- not the LM337L. They are different.
Also note that the pins on the drawing for the LM337L pinout are _not_ numbered in numerical order.

 

The middle pin is the middle pin wherever you view it from. One says input, the other says output. They are not in agreement.
[Edit] My original post stands. The marking on your board (assuming that is a component side view) for U1 is reversed.

I installed a new regulator (the old one was damaged during the removal process) except I reversed the IN and ADJ pins per your suggestion. I also changed R5 to 220R and R4 to 390R. Unfortunately I am no further along than before.
Input: -10.7
Output: -10.0
Voltage drop across R2: 4.9 VAC

Is it possible that the "new" regulator was defective?

Frustrated in Western PA

 

Those two pictures contradict each other. One has the middle pin as input while the other has that pin as output

Or should that be the 'muddle' pin?

This is a follow up from my previous post. After letting this sit over night I powered it up this morning and things have changed somewhat:
OUT = -10.23 volts
IN = -7.5 volts
ADJ = -7.74 volts

Which leaves me with the following questions---
Why is the magnitude of the output voltage greater than the magnitude of the input voltage?
Why is the voltage on the ADJ pin -7.5 volts? Using the resistor values of 220R and 390R the ADJ voltage should be around -6.5 volts.

Can anyone help me out here?

Still frustrated.

 

Using the resistor values of 20R and 390R

Your picture of the board shows resistor values of 220Ω and 390Ω so is this picture not of the board you are working on?

 

Why is the magnitude of the output voltage greater than the magnitude of the input voltage?

If can't be if wired per your schematic.

 

Your picture of the board shows resistor values of 220Ω and 390Ω so is this picture not of the board you are working on?

Indeed it is exactly the same board we have been discussing all along. In post #14 I stated that the resistor values were changed to 220 and 390 respectively.

 

Connect the negative lead of your voltmeter to the right-hand end of C2 on your picture of the board. Then connect the positive lead to each pin of the LM337L and report your readings here.

 

Connect the negative lead of your voltmeter to the right-hand end of C2 on your picture of the board. Then connect the positive lead to each pin of the LM337L and report your readings here.

IN = -11.25 VDC
ADJ = -7.99 VDC
OUT = -10.5 VDC