LM21305: VOUT Range not in datasheet

CHRISTOPHER YECKEL

Part Number: LM21305
Other Parts Discussed in Thread: LM73605, TPS54424, TPS54824

Hello,

I have integrated an LM21305 into my design. The idea is that I use a digital potentiometer to adjust the voltage into a resistive filament. The idea is as the filament warms and becomes more resitive I can increase the voltage to apply more power.

I am doing this by replacing RFB2 with a 1Kto10K digital potentiometer. I put this regulator in this configuration on my board and it does not work.

I purchased an evaluation board and with the same potentiomenter I can run all the way from 1V up to 11V with perfect 4A of output current in all cases.

Clearly I have a layout issue on my board I need to resolve, which I am able to do.

The question I have is am I wasting my time? The datasheet and worksheets make it sound like the device really only wants up to 5V VOUT and what I am trying to do (up to 11V) is not recommended? I don't want to spend all the time doing this if it is just going to break.

I have been running 11V/4A out of the evaluation board for an hour now and the device is staying pretty cool. Do you think I am ok running the device like this or am I going to have problems? The datasheet is not clear on the recommended output voltage range, I assume as a buck regulator it is ok up to its VIN right?

Thanks,

-Chris

over 6 years ago

0 JohnTucker over 6 years ago

TI__Guru**** 191445 points

Chris,

i can't think of any reason why it should not work. It is a buck converter so obviously the input voltage must be above the maximum output voltage with some headroom. Your maximum duty cycle will be limited by the minimum off time as lited in the datasheet.

0 JohnTucker over 6 years ago in reply to JohnTucker

TI__Guru**** 191445 points

Chris,

Also let me know if you need layout help.

0 CHRISTOPHER YECKEL over 6 years ago in reply to JohnTucker

Prodigy 155 points

Thanks for the quick response John!

Well these devices are not the easiest to replace so we were worried we blew it up because of too much voltage fluctuation on the FB pin possibly because of too high a VOUT.

My first layout was not perfect... but I thought it was good enough for it to work...

So in your opinion based on the limited information I have provided that I am having a layout issue, since if I take my components off and hang the evaluation board off of my board it works fine.

I plan on using the layout I have found in the evaluation board datasheet and basically copying it. If you have a better reference I could use for a perfect layout I would be very happy. Otherwise I might have to do some guessing. Thanks Again!

-Chris

0 JohnTucker over 6 years ago in reply to CHRISTOPHER YECKEL

TI__Guru**** 191445 points

Chris,
The EVM layout is a known good design. If you want to post your preliminary layout, I can take a look at it before you spin PCBs.

0 CHRISTOPHER YECKEL over 6 years ago in reply to JohnTucker

Prodigy 155 points

John,

I think since I am doing it all again i am going to move over to the LM73605. It seems more robust and the layout is very clearly defined in the datasheet (page 42) and I have some room for it. It also has curves up to 12Vout which eases my mind.

Its also a new product, which means it will be supported longer?

Thanks again!

-Chris

0 Anthony F over 6 years ago in reply to CHRISTOPHER YECKEL

TI__Mastermind 35115 points

If you are thinking of changing the parts I suggest that you also consider the TPS54424 (or TPS54824 if 4A is not enough). These newer parts have very similar features to the LM21305 so I'm thinking they are close to what you need in this application. Also since this design needs to support a wide range of output voltages the flexibility of external compensation may be helpful.

0 CHRISTOPHER YECKEL over 6 years ago in reply to Anthony F

Prodigy 155 points

Hello again,

I have decided to take your advice and use the TPS54824. Question: what would you recommend for the minimum resistances for RFBT and RFBB? I would like to use a 1K resitor in series with a 10K linear digital potentiometer (MAX5483EUD+) as my RFBB and a 20K fixed resistor for my RFBT. This would give me a theoretical range of to 1.7V to 11.5V. I don't care much about the efficiency at light loads, Is 1K too low of a resistance for RFBB to operate?

I would put a 50K linear pot on RFBT, but I can't get as good of a range playing with that side of the divider.

Thanks,

-Chris