Hi!
I'm using a LM317 voltage regulator. Instead of connecting two resistances, one between output and adjust and one between adjust and ground, I would prefer using a trimmer connected in this way:



In this way the trimmer acts in a ratiometric manner, so I prevent voltage drifts due to temperature effect.
But I think it could be a problem: when the wiper reaches the upper bound, the output and the adjust are shorted. So a big current flows in the pot and I think it could be a problem for it.
Are these considerations right? Or does the LM317 have any protections and the current can't exceed a limit?

(I found a similar post with a similar title but it doesn't answers may question)
Thanks!

 

What voltage drift problem do you think you're preventing? My advice is to use the LM317 as intended. Use a separate resistor to set the adjust current and use the pot to adjust voltage.

EDIT: I forgot to mention that using the pot in that manner will give an "interesting" voltage adjustment. As you change the resistance between output and adjust, you'll change the current in the other part of the pot. When you set the pot to it's maximum, you'll get 1.25V and higher voltages for other settings; assuming you don't short the output and adjust terminals.

 

What voltage drift problem do you think you're preventing?

If I use a resistor characterized by a high temperature coefficient of resistance, the effect of temperature could cause a change in the fixed voltage. Using the pot configuration like in the figure I minimize voltage drifts due to temperature because I am acting in a ratiometric manner and both resistances, the upper and the lower, have the same effect of temperature.

 

Read my first response. I edited it to include some additional information about the time you responded...

Temperature drift for resistors is measured in ppm. On a K ohm type resistances, it will be small.

 

If you use the arrangement recommended by the manufacturers you will obtain the best overall performance. The ADJ terminal expects a particular resistance to the output terminal (240 ohms from memory) for the best temperature-related performance.

Yes, the LM317 has an internal current limit circuit as well as over-temperature protection. You have to be pretty clever to destroy an LM317

 

assuming you don't short the output and adjust terminals.

This is the problem I mean. In this case I risk to damage the pot.
Maybe I was wrong when I said that acting in a ratiometric manner I minimize voltage drifts due to temperature. The voltage between Output and Adjust is always the same, so variations of resistance due to temperature aren't minimized: temperature behavior of pot's lower resistance is the same of a fixed resistance! Right?

Temperature drift for resistors is measured in ppm. On a K ohm type resistances, it will be small.

Yes, but i would to fix an output voltage of 1,65 V so the lower resistance should be 40 ohm.

If you use the arrangement recommended by the manufacturers you will obtain the best overall performance. The ADJ terminal expects a particular resistance to the output terminal (240 ohms from memory) for the best temperature-related performance.

125 ohm to satisfy the max. minimum load current.

 

Temperature drift in resistors is: change in ohms per total ohms. Every place you adjust the pot to, there will be a different number of total ohms in each end. The 317 chip is guaranteed to 1% maximum drift for all temperatures it can survive. It's hard to find a resistor that is anywhere near that sloppy! You want to stop temperature drift? Choose a different chip, not a special resistor configuration.

 

If you use the arrangement recommended by the manufacturers you will obtain the best overall performance. The ADJ terminal expects a particular resistance to the output terminal (240 ohms from memory) for the best temperature-related performance.

Yes, the LM317 has an internal current limit circuit as well as over-temperature protection. You have to be pretty clever to destroy an LM317

Commodore Business Machines must have been pretty clever then....................

For a while I had a sideline of salvaging 5V switchers from a place that scrapped old computers and casing them up with the 9-0-9V iron cored transformer rescued from the scrap CBM PSU. It didn't take long to give up replacing the 7805s - there was no guarantee it'd last any longer than the one CBM put there!

 

Your awareness about minimizing the effects of thermal tracking between the two resistors is good, and mandatory in high precision instrumentation circuits. But the 317 is only a 1% part. If you use two metal film fixed resistors and a CERMET trimpot, the temperature coefficients will track closely enough to be a trivial source of error. Also, unless the circuit has to hold a tight adjusted value over a wide temperature range like +25C to +75C, there won't be enough drift to matter no matter what kinds of resistors you use.

ak

 

Your awareness about minimizing the effects of thermal tracking between the two resistors is good, and mandatory in high precision instrumentation circuits. But the 317 is only a 1% part. If you use two metal film fixed resistors and a CERMET trimpot, the temperature coefficients will track closely enough to be a trivial source of error. Also, unless the circuit has to hold a tight adjusted value over a wide temperature range like +25C to +75C, there won't be enough drift to matter no matter what kinds of resistors you use.

ak

I wouldn't chance it with carbon composition resistors - they still exist, the RF guys use them.