Failing TPS54331 Supplies

Can i see schematics and waveforms?  If you are building up these designs on bare boards with only the supply circuits populated, then I have to beleive it is a layout issue.  Did you follow the layout guidlines?  The most critical issue is to have the input bypass capacitor as close to the IC as possible.

OK, how do I do that?

Jim

You should be able to attach files. Use the "paper clip" icon second from right  above

Please see attachments. Input cap is C8.

Try removing, R4, R5, R24 and R25.  let me know if that works.  Otherwise I will need to see waveforms.

That has already been tried.  Nothing I can do with Ven makes any difference.  What I find most odd is that as I crank up the input voltage, one supply seems to regulate for awhile, then as the voltage increases, it goes out of regulation, while the other one starts regulating.  I can't find an input voltage that makes both circuits stable, while the design software indicates they should both be very stable across a wide range of input voltages.

Jim

What are the *.TSC files?  I can't seem to open them.  Also can you provide any waveforms?  I would like to see PH and Vout for sure and any other waveforms that demonstrate your problem.  Are you measuring the low output voltages with a meter or are you observing the values with a scope?  It may be that you are in low headroom drop out AND eco-mode at the same time.  In this case you may be getting a dc output with around 2.1 V of ripple that will make the output appear to be low when measured with a volt meter.  The circuit will exhibit breif periods of switching followed by extended sleep time.  From your decription, this sounds like it could be the issue, but I will have to see waveforms to be sure.

Moderation reply.

The files are Tina-TI files.  After loading up the reference design, modifying parameters and running a simulation, Im not sure Im any closer to an answer.  Please see the two attached files.  Both of these circuits worked fine with no warnings in SwitcherPro Desktop, but they don't work in simulation.  I played with one of the circuits and got it to work somewhat at full load, but it fell apart at light load. Can you please help me to determine why these circuits don't work?

Thanks,

Jim

I can tell you that switcherpro does not take into account any of the esoteric operating modes of the TPS454331.  That includes low headroom drop out and eco-mode.  What I think you are seeing is eco-mode and low headroom drop out occurirng at the same time.  I f ai can see some wave forms < I may be able to make a call on that,

John,

The scope waveforms show no hints.  They are very flat with very little noise.  Certainly no evidence of instability or stepping.

Probing the outputs, the 1.8V supply seems to regulate at input voltages just above 4V, but the 1.2V supply shows no signs of regulating at all.  The output voltage varies as the input voltage varies.  In other words it just looks like a resistor.  But still the output looks very flat.

Thanks,
Jim

so, no switching at all on the PH node?  I will try to look at your sims tomorrow.

John,

Any luck? I'm building a 2nd proto board now and I really need at least the 1.8V supply working.  I'm rewiring the input so it will run from 5V this time.

Thanks,

Jim

I was expecting you to provide waveforms.  Do you have any that show your start up issue?  I ran teh tina files and they do not work w/ 5 V input, but seem to work fine with Vin = 12 V.  I will have to see if I can find out who did the model and look inside.  Getting teh model to work is one thing, but I am ore concerned w/ your circuit.

Unfortunately I do not have any waveforms worth sharing.  As mentioned before outputs are pretty quiet and at DC, just the wrong voltage.

Also unfortunately I don't have a voltage higher than 5V in my system.  I guess what I'm most interested in is how a supply rated for 3.5V input won't even work at 5V input levels.  I think you are right  to want to look inside the model.  But since the model seems to somewhat validate what I'm seeing in real life, I'd like to know what if any "undocumented" constraints there are on getting this silicon to work as a voltage regulator.  Was it mischaracterized?

Sorry if I seem short but that's the way time is running on this issue.

Thanks,

Jim

i think that is a model issue not a device issue.  I have passed that on to the modeling team and they are baffled for now.  They are going to run it in teh unlocked master where they can access the internal nodes.  So far as I know teh actual silicon should work w/ 5 V input, especially if you do not have any resistors connected to EN to set external UVLO.  The tolerances on teh external UVLO can possibly cause teh EN threshold to be too high at low voltages.  If I have a board handy I wil try it in teh lab.

I checked a baord.  the start threshold for VIN with no UVLO resistors on EN is 2.95 V.  Exactly what I would expect.

John,

OK, thanks.  I should have my 2nd board modified and ready to check out tomorrow afternoon.  I'll let you know what I find.

Thanks,

Jim

John,

I have both supplies working now.  Apparently the EN input doesn't work quite as documented.  With no resistors the design works.  I haven't checked across a range of loads yet but so far so good.

Thanks for all your help!

Jim