LM338: LM338T Parallel Configuration

Hi Christian,

You can parallel LDO's and we have recently published comprehensive literature detailing how to do this with ballast resistors (those links are below if you are interested).  This application circuit uses a combination of parallel LDO's using op-amps and ballast resistors (which is also fine).  The main consideration here is stability analysis of the op-amp and selecting the ballast resistors.  But before we do any of that, have you considered the thermal management of this system? If each LDO is load sharing equally, nearly 50W will be dissipated across each LDO and even 10% of this may be too high for linear regulators.  If you have the thermal management solved then we can discuss in more depth how to design this in, otherwise I would strongly recommend a switching converter (we have lots of options here and I can route this E2E thread to the switching converter team if you would like).  If you need the performance of an LDO in your system (and cannot tolerance the switching converter noise) then you can also buck the 28V down to around 15V (or enough to meet the dropout and headroom requirements) and we can then parallel the LDO's to give you clean, well regulated output voltage at whatever current you need.

New Parallel LDO white papers (using ballast resistors only)

https://www.ti.com/lit/wp/sbva093/sbva093.pdf

https://www.ti.com/lit/wp/sbva100/sbva100.pdf

https://www.ti.com/tool/PARALLEL-LDO-CALC

Thanks,

Stephen

Hi Stephen,

Yes the thermal issue is figured out. (Fans along with heat sinks--although the 15V intermediate regulator is an interesting idea).

However, I was wondering for the above circuit, how the Opamp would set the ADJ (voltage) for the top 2 regulators, and which/if any, opamps you would recommend given that LM307 is discontinued.

I was thinking something like the LT1494 from Analog.

Hi Christian,

Okay thanks for the confirmation.  Let's proceed with the parallel LDO design.

We have op-amp options in our portfolio that will be excellent candidates for you.  I can reach out to the op-amp team and ask for a modern replacement to the LM307.  I don't have the white papers written on parallel LDO's using op-amps yet but it is a problem we are working on and I'll help with guidance in your design if you need the op-amp. In my work I've noticed that an op-amp configured as an integrator (so, a single capacitor from Vin- to OUT) will sufficiently compensate the new feedforward loop.  You may need a series Riso between the op-amp and ADJ pin if there is parasitic capacitance loading down the op-amp.

What tolerance do you need on the output?  Do you have any specific key requirements other than voltage and amps (such as PSRR, noise, etc)?  Do you expect transients or is the system relatively DC?  It's possible that you don't need the op-amp and ballast resistors will be enough but that depends on your other system requirements.

Thanks,

Stephen

Hi Stephen,

No other hard requirements.

The tolerance is likely +/- 500mV or so.

I have the above circuit setup to test with the analog opamp and it actually does look good.

I cannot test the high current at the moment, but I just wanted to let you know.