3.6V to Precise 5V Reference IC required

[carl]

Hi There,

Anyone any ideas on the best way i can get a Precise 5V from a 3.6V source? The best 'step-up' DC/DC converter i can find is AS1321. This gives a +/-0.125V (Ta = -40C to +85C). I cannot find any better??

It needs to be low power, and also allow a complete shutdown control pin.

The only way I can think of achieving this is in two stages, using two DC/DC converters:

3.3V to 6V+
and then 6V+ to a lower precise 5V source.

I cannot see another alternative to get a precise 'step-up' 5V from 3.6V.

Any suggestions most appreciated.
Carl

[SherpaDoug]

Another option might be a small boost converter where you insert your own reference in the feedback. I don't think you are going to find a single chip solution.
_________________
The search for better is endless. Instead simply find very good and get the job done.

[temtronic]

Silly idea...certainly cost effective and should meet your requirements.
I've done this a few times in the past 30 years..

Take a spare I/O pin and 'toggle it' at some reasonable rate.Feed the output into a voltage doubler (2 caps, two diodes), add a 3rd cap for 'storage', feed this (about 6-7 volts) into a 5 volt reference chip. Done.
Ref chips don't draw a lot of current, so PIC pin is fine.

Set the pin low for 'shutdown.

As for the 'toggle rate'. It just has to be fast enough to charge up the caps and if the 3rd 'filter' is say 100mfd, it'd keep the ref chip 'alive' for quite awhile.

As I said, silly but it does work.

[newguy]

[carl]

Thankyou all for your replies - much appreciated.

Temtronic, your idea sounds interesting. How much time though would be needed each time the pin is toggled to generate enough power through the doubler into the 5V ref chip? Assuming the toggle speed was appropriately set to some level.

would a burst of 100-300ms be adequate?
and then how long would it take to turn off the ref chip, when the pin is kept low (between points 5 and 6 below)?

so:

1) Turn on pic
2) Strart pin tiggle
3) After 100ms, 5V is now present out of the ref source.
4) Do something (which will take 200ms).
5) turn off toggle and set the pin low
6) go into sleep
7) after 500ms start the process again

Thanks for all your help
Carl

[carl]

Actually, just thought but this idea may not be suitable depending on the total amount of capacitance in the circuit.

It can't be more than a few hundred uF.
The reason being is that the circuit will be I.S.

[temtronic]

Hay Carl

This is one of those try it and see.I used to drive a 78L05 no problem with 1N4148s and 22mfd caps.Nowadays, there are real nice reference chips(low power, ultra precise) that may only need a few ms burst with 1mfd caps.
In my original circuit the 'burst' was only 100ms.

Think of this as a poor man's MAX232 'power' section

... if you use a tripler you can get almost 10 volts from 3.5
... if you invert the diodes you can get a -ve output

"The reason being is that the circuit will be I.S."
hmmm... what does I.S. mean ?

[RF_Developer]

I.S. : is that Intrinsicly Safe, which, amongst many other things, depending on the IS class places limits on the energy stored in a circuit...

...or not... maybe.

[carl]

Thanks for your replies

I think it is going to be one of those 'try and see' solutions. RF_Developer your quite correct - it is intrinsically safe. some of the designs I do are I.S. which means that capacitance, inductance, voltage and power have to be as low as possible - they are all inter-related.

mF caps are way to large though.

I had a look at the 'toggle' idea (its on microchip web) and it looks a valid idea - but I have to compare the power and noise (especially noise because I will be driving a load cell and ADC) this idea will generate compared to using some sort of dedictaed boost or voltage doubler I.C.

nice to see the old '7805' reg and '4148' diode reference - been around for years and everyone has them in there stock!!

If I did go with the 'toggle' plan - what would be the best way of setting this up - using the CLKOUT pin?

Anyway I will keep looking at the possibilities.

Thanks
Carl

[carl]

I will try using pin RA6 (PIC16F628A):

RA6/OSC2/CLKOUT RA6 ST
CLKOUT — CMOS In RC/INTOSC mode, OSC2 pin can output
CLKOUT, which has 1/4 the frequency of
OSC1.

Carl

[temtronic]

Using CLKOUT means it's on all the time, no 'shutdown' feature.

I'd just use any 'spare' I/O pin.
2 modes of operation

[carl]

Thanks Temtronic,

Thats a great help - I will definetly consider it.
The only reason I thought of using the CLKOUT, is because maybe I could use it in conjunction with the 'sleep()' function and wake up with the 'wdt'. (I cannot use an external wake up event). This in essence would therefore start and stop CLKOUT as well as conserving power all in the same instruction.

I have never used sleep or wdt before, and am sceptical about its function. All the info I know about the wdt is that it is inaccurate and annoying. I checked the datasheet and all I can find is the variance with no prescaler. But I will need the variance 'with prescaler', to assess its accuracy.

31 TWDT Watchdog Timer Time out Period
(No Prescaler)
7* 18 33* ms VDD = 5V, -40°C to +85°C

To test, I tried setting the fuse# to include the prescaler but it doesn't like it e.g. WDT128 or WDT1024.

If I was to use this method... How do you control (turn on, turn off, set the prescaler) the WDT and where do you put it in the code.

What would your opinion on this idea be??

Thanks for your help
Carl

[temtronic]

I look at the overall 'picture' a bit differently.
While it's nice to try to get 'sub-micro-amp ' performance from any system there is a huge cost in time and effort to achieve it ($$$). Trimming code, optimizing routines, etc. as well as parts values all play their part. Back in '85 when I developed our remote energy control systems the WHOLE system had to be battery backed up. Remotes ended up getting Gates sealed lead acid batteries, same as in US subs. Today's batteries pack a LOT more energy in a small package. Nowadays a coincell battery would do the job ! My point is that it's far easier, quicker and cheaper to put a larger capacity battery into the system than to tweak it every other way. There's a misconception that 'smaller is better' but over the decades I've found that 'bigger is better' and more reliable.

[carl]

Hi Temtronic,

Totally understand your point.

Thanks for all your help
Carl

[carl]

Found the WDT spec's now in the graph at he back of the datasheet.
It's not too bad for what I need - only varies by 30ms over the full temperature range at 3.6V.

So that will do - I will try to write the code for WDT and sleep

Carl