voltage regulator chips in parallel?

[HiTech]

I was servicing a 12vDC to 120vAC inverter/charger and to my surprise the mfgr. has two LM317K (TO-3 case) regulator chips connected in parallel, apparently to increase current handling. I didn't know you could effectively connect them in parallel.(?)

 

[eng1]

A more reliable solution is to add power bjt(s) around a single LM317.

 

[audioguru]

Mickey Mouse connects regulators in parallel. Sometimes he adds an equalizing resistor in series with the output of each one. Then they share the load but the voltage regulation is lousy.

They won't be exactly the same (unless they are separately adjusted to have exactly the same output voltage) so the one with the highest voltage will power the entire load until its voltage drops down (maybe when it gets too hot?) to where the second one begins to work. The first one will be very hot and the second one will be cool. Not very good.

 

[Andy1845c]

Is there a reason they would do it this way? I did some reading on voltage regulators yesterday for my LED project, and if I understand correctly, the LM350 will handle up to 3A and the LM338 upto 5A with adequate heatsinks. Why not use one of these insted of two of the 317? Are they that much cheaper?

 

[audioguru]

All three adjustable voltage regulators can handle the same small amount of heat before shutting down.
All three reduce their output current when there is more than 10V or 15V across them.

So maybe it is best to add a current-boosting power transistor to an LM317 circuit.

 

[Roff]

If they are configured as current sources instead of voltage regulators, they can be connected in parallel. Can you post a partial schematic?

 

[Optikon]

I was servicing a 12vDC to 120vAC inverter/charger and to my surprise the mfgr. has two LM317K (TO-3 case) regulator chips connected in parallel, apparently to increase current handling. I didn't know you could effectively connect them in parallel.(?)

Because you used the word "effectively" then as pointed out, the short answer is NO if they are truly parallel. I prefer to use an outboard power transistor to provide load current beyond what the regulator can do safely.

 

[HiTech]

I didn't get a chance to draw out the diagram, but this device is a "professional", high quality unit. The construction is near mil-spec, and a majority of the crucial components are custom mfgr'd. The company name is Wilmore. Oh, and the dang thing weighs well near 50 pounds with quite a large torroidal xformer in it!

 

[Optikon]

I didn't get a chance to draw out the diagram, but this device is a "professional", high quality unit. The construction is near mil-spec, and a majority of the crucial components are custom mfgr'd. The company name is Wilmore. Oh, and the dang thing weighs well near 50 pounds with quite a large torroidal xformer in it!

If you could draw out how this section is wired up, I'm sure the audience will provide good feedback/suggestions for improvement.

 

[Hero999]

It's probably a constant current battery charger so there's nothing wrong with the design.

 

[HiTech]

from a preliminary look at it. I think you're right. It is to recharge a deep cycle battery and there is no supporting circuitry to provide a low current maintenance charge. Almost sounds like a device that would get used in an environment where power failures occur regularly with the battery getting cycled quite often.

 

[Rio71]

hi,
yes, you can easyly paralleling linarregulators like lm317.
i have twelve L78S12CV parallel in a diy 25 ampere powersupply.
this is good 10years old an works perfectly.
i have the parts all three pins connectet "hard" together, no resistor between..

only one is really matters:
watch out that your regulators are absolutely from the same batch --> exactly same printing on it, especially the second line !!!
so differences from lot to lot are not present.

 

[Roff]

hi,
yes, you can easyly paralleling linarregulators like lm317.
i have twelve L78S12CV parallel in a diy 25 ampere powersupply.
this is good 10years old an works perfectly.
i have the parts all three pins connectet "hard" together, no resistor between..

only one is really matters:
watch out that your regulators are absolutely from the same batch --> exactly same printing on it, especially the second line !!!
so differences from lot to lot are not present.

According to the datasheet, the output voltage temperature coefficient is -1mV/C. If all parallel units start out sharing current approximately equally (as they should if they are all from the same lot, as Rio71 points out), then the negative tempco makes them inherently stable, temperature-wise. If one gets a little hotter than the rest, its output voltage will attempt to drop, pushing some of the current load to the other(s). This would not work with LM317s unless they all shared one feedback network (voltage divider). According to the Temperature Stability graph in the National datasheet, the reference tempco is also negative (above 25C), so they should be stable. Personally, I would never try this. If one device fails, the rest will fail (probably just shut down) in rapid succession under a heavy load.