Ldo regulators

"Ldo" wrote in news:455a4d54$0$7634$ snipped-for-privacy@reader1.news.tin.it:

Something about this post makes me want to find a blanket to bite.

Ok, if you have a regulated 3.3V supply, and a regulator on a board that would do the same thing, why not just bypass the onboard regulator? Whatever you do, you're not going to get a zero volt drop from a voltage regulator, unless it's a power converter. Even so, there's no point in this case. Just check your 3.3V supply carefully, and if it's good, bypass the onboard regulator.

A piece of wood may be better. Less fuzz between the teeth after letting go of it :-)

Plus with all luck it might begin to oscillate. Many LDOs will easily start to sing. Some have an undocumented failure mode where they choke up upon seeing or thinking to see a rising source impedance. Then the current rises and rises. Bzzzt - pop - POOF.

--
Regards, Joerg

http://www.analogconsultants.com

Joerg wrote in news:yvJ6h.7451$ snipped-for-privacy@newssvr11.news.prodigy.com:

I must practise this. If my crowns will let me. >:)

Nice thought. I tend to avoid LDO regulators, preferring to reduce the drop at the rectifier, but I might have to use one some time. Any advice on which types, or makes, are generally well behaved?

Just got a black eye with the TPS71550. It seems to become unstable with too much input impedance, resulting in self destruction (they don't have an overtemp limit) and occasionally taking everything connected into the grave with it. TI hinted that there might be an issue with source impedance but insisted on scope plots before they'd investigate further. It is hard to take scope plots if it blows after a few hundred cycles and you don't know when. So, I guess we'll ditch that regulator now. I had sent them my rather simple schematic and they could have thrown that on the simulator. Nope, not going to happen, they said :-(

Thing is, I can't simulate because they won't release the innards of it. Ok, I had a 7uA to 2mA load change. Muffled it by easing it on over a millisecond. They said that might still make it choke up if there was a high input impedance. My take is that this stuff needs to be mentioned in datasheets. Either that or good enough SPICE models.

Then there was a National part, LM29xx or something like that with the same issue. At least they vowed to change the data sheet and mention that problem.

Then there were a whole lot of others. Forgot the P/Ns and various mfgs because these weren't my designs. I did the re-design and flung all the LDOs out the front door immediately.

Anyhow, I won't use LDOs unless there is absolutely no other way. Usually there is.

--
Regards, Joerg

http://www.analogconsultants.com

Joerg wrote in news:pAL6h.7475$ snipped-for-privacy@newssvr11.news.prodigy.com:

If they won't accept your proof at that basic level, they won't accept anything more elaborate. You made the right call, don't waste your time.

I've never used a SPICE model in my life. I'm not at all a high-powered tech though, I guess I don't need to. Even so, it's just a model. Plenty of people who ARE high=powered techs and better will say similar things. As with laser diodes, it's far better if the company will push the envelope on four samples from different batches and post that data in an addition to the data sheet. It's easier for them, and it means caveat emptor for the buyer, who at least has something real, not a model or a guess, to go on.

I like their parts, generally. That news is good. I buy my LM317's from them, in (very) modest bulk. :) I'd try their LDO's so long as they tell it like it is.

Yes. I think if the overhead voltage is pared down to the point where there is not, it's a bad design anyway.

Sorry, i was in a hurry and i wrote the post before going to bed. And sorry, i also took the only blanket left (it was pretty cold and i couldn't get to sleep).

There are 2 boards involved.

- Board A provides a regulated 3.3V

- Board B gets its power from board A

- Board B has the LDO regulator (3.3V) and i guess it was designed to be used with supplies greater than 3.3V.

I cannot cut the piece of board containg the regulator because it's not me who made the boards. Maybe i could try to bite the regulator and see if it breaks.

Now i would like to know whether applying 3.3V at Vin i get 3.3V at Vout anyway (regulated or not) If i can't, i cannot use the board B, unless i find 5V somewhere.

this

That was pretty much verbatim my thought process after the last phone call with them. TI is a really good company but I was a bit taken aback by their unwillingness to throw it onto the simulator. It would have taken a guy there maybe 30 minutes in total and they might have discovered something serious with their own product.

BTW I had a similar issue on Monday, totally non-technical. Well, maybe not totally. The property tax bill looked iffy. So I called. After a bit of run-around I got the auditor's office on the line and they had me fax it in. They called back, profoundly apologizing. The final number was ok (they said...) but the line item display was pretty hosed. Nobody had noticed that the computer was doing this since who knows when...

If the models are good SPICE is ok. Once I discovered a reliability issue that was impossible to measure with the usual scope technology available in the early 90's.

Depends. When you have to use three AA cells and the uC needs at least

3.6V you are between a rock and a hard spot. But I have wised up and use switchers now, mostly sans PWM chip for cost and other reasons. Wasn't really possible in the TPS case though. That will require some nifty discrete linear design with 0201 parts and stuff. Guess I'll need better reading glasses now.

--
Regards, Joerg

http://www.analogconsultants.com

"Ldo" wrote in news:455b995a$0$13758$ snipped-for-privacy@reader3.news.tin.it:

No worries, Joerg's suggestion of wood should do nicely.

Mhmm. I'd stick to blankets. And pitch pine.

There is a voltage drop in a linear regulator that applies no matter what goes in, so you won't get your full 3.3V out. I haven't tried, but I think you might be ok to short the input and output of the regulator on board B if it's a standard three terminal type. Even if this made it try to draw more power it can't destroy itself when there is no way that current can actually go through it, as it's shorted. It would be simpler than removing it, and easy to undo later if you get a different supply.

According to the data sheet for the REG102NA-3.3, it states the minimum input voltage for normal operation is Vin + 0.4V. You probably won't be able to get this part to act correctly with a 3.3V input.

Can you unsolder the REG102NA-3.3 part and add a jumper to allow the external regulated 3.3V into the board?

Dave

I haven't bought the Board B yet, but i guess it should deploy the ldo in a "small" package, the Board itself is simple but not so cheap. I'm not very good in soldering, let's say i'm not very patient in jobs that are not creative (someone could say soldering is an art, i agree but that's not my case). That said i'm afraid that if i try to do the "short" i could end up in biting my hands this time!

I have to build a small adapter board, i hoped i could get away with just connectors. Now i should use charge pump ICs to raise 3.3V to 5V and the feed it to the ldo (it's sad i know). MAX619 might be it, but i'm afraid that it doesn't provide enough current (50 mA). Perhaps ICs from the same family have more power.

that

case,

I know about +0.4V. But i was surprised to see a home made PIC programmer working with 5V when it actually had a 7805 regulator inside. 7805 should require 5V + 3V at least if i remember. I thought "if it works with 7805 maybe...."

Board B is not very cheap so i wouldn't put to the test my skills in soldering and unsoldering!! I'm investigating on charge pumps instead.

That can be a bit much for many charge pumps. You could also use a little boost converter chip from National or TI. You'll need an inductor but it's small. If you want to take the easy route I believe TI makes complete modules that can simply be dropped in.

--
Regards, Joerg

http://www.analogconsultants.com

just

same

I also found TPS60130 (300 mA). The power required by the chip on board B is

240 mW. So i guess it should do. But its package is too small for my patience, i was hoping for a DIP.