Noob Question: Testing/matching linear regulators, test jig needed...?

[hymentoptera]

I've got trays and trays of LM78xx regulators here and I want build a simple jig to test them, specifically I want to mate up several pairs of L7824 regulators for stereo projects where I want precise power supply matching between multiple units. My theory is that a pair of amplifiers, in separate cases with separate supplies (such as a pair of Hairball 1176 Revision D's) with carefully matched transistors, resistors, etc, should behave nearly identically across a wide range of settings. So I'd like to match the regulators, zener, even the filter caps in the PSU, etc, in addition to all the rest of the componants in the entire build...

So, since linear regulators can vary in output (I'm thinking 5% +/- I've read somewhere?) from one device to the next, then I should be able to breadboard up a simple PSU with a socket for the TO-220 device, and simply test it's output with a multimeter, right? I can bag up all the closest pairs for my future projects.

I wanted to run the idea by you fine folks before I build it, in case maybe I'm overlooking something or overthinking this... I'm thinking 3 x 9v batteries in series would provide 27v (7824 needs at least 26v according to datasheet), a 100uF cap, device under test, a second 100uF cap, then a film cap (.1uF?) and finally a load resistor, maybe 2k4 to keep the current down near 1ma? I insert a regulator, power it up, and measure the voltage seen across the load resistor!

Is the load resistor even needed, or should I measure it floating, no load? Will the caps throw off my reading? Are the caps even needed at all? Any reason to prefer bigger or smaller caps? Is there a simpler way to do this? Am I forgetting something obvious??

I'm pretty new to all of this, so thanks for any and all comments and input! Here's basically what I had in mind:

 

[JohnRoberts]

You need to test it the way it will be used. but is this voltage critical for your design?

JR

 

[mjrippe]

You need to test it the way it will be used. but is this voltage critical for your design?

JR

^This^

If you decide it IS critical, get a length of SIP socket and break off three pins worth.  Solder into the regulator position on the PCB and plug in your regs one at a time.  Let them warm up for ~20-30 min and measure the voltage.

 

[hymentoptera]

Okay, can someone school me on why it should be tested in the same circuit I intend to use it in? I figured the main difference was from variations in the voltage reference inside, kind of like how zeners are never exactly the same, they vary by 5% or so for part to part...

Does temperature have that much of an impact? I guess I'll find out once I get a minute to build something and watch the voltage change over time. Can I just use a smaller value load resistor on a breadboard to move the current closer to .5A or so, maybe a big 25w 50 ohm, to heat the DUT up faster? Will it really have that big of an impact? I don't need precision here, just want to match them relatively closely into pairs for dual mono projects that will end up in separate cases with separate PSUs, but with the chance of being used together in stereo. I.E. 1176, etc. Nothing critical to the circuit; I'm just being anal about matching things lately.

Thanks!

 

[Andy Peters]

I don't need precision here, just want to match them relatively closely into pairs for dual mono projects that will end up in separate cases with separate PSUs, but with the chance of being used together in stereo. I.E. 1176, etc. Nothing critical to the circuit; I'm just being anal about matching things lately.

What difference does it make if one regulator measures 15.1 V and one measures 15.2 V? (Other than 100 mV, of course.)

If you think that the 100 mV makes a difference regarding headroom, do the math.

What part of the circuit requires that close matching? Divider strings for metering? Certainly not ADC or DAC references at that high a voltage.

Just curious.

 

[JohnRoberts]

Okay, can someone school me on why it should be tested in the same circuit I intend to use it in?

So the results will perfectly reflect your experience in use.

Probably the most significant variable will be load current. Next and much lesser variables will be input voltage, and temperature.

I figured the main difference was from variations in the voltage reference inside, kind of like how zeners are never exactly the same, they vary by 5% or so for part to part...

On my experience (decades ago), 95% of +15V regulators were about 15.10V, and 90% of -15V regulators were -14.87V  give or take a hair....

Does temperature have that much of an impact? I guess I'll find out once I get a minute to build something and watch the voltage change over time. Can I just use a smaller value load resistor on a breadboard to move the current closer to .5A or so, maybe a big 25w 50 ohm, to heat the DUT up faster? Will it really have that big of an impact? I don't need precision here, just want to match them relatively closely into pairs for dual mono projects that will end up in separate cases with separate PSUs, but with the chance of being used together in stereo. I.E. 1176, etc. Nothing critical to the circuit; I'm just being anal about matching things lately.

Thanks!

Methinks you are focussing on the wrong parts of the circuit...

JR

PS: I have never matched regulator voltages.