Re: One Second Pulse-Width, Once Per Hour

Whichever way you do it, you need a power on reset circuit so the thing doesn't require an hour to straighten itself out.

Tam

yup, pic12c508 eight pin dip internal 4mhz clk, less than a buck each. But the old timers'll never go for it...

Remove "HeadFromButt", before replying by email.

I don't have any in my junk box. CK722's, though ;-)

But I'm open to education... what do I need to buy to learn this "new-fangled" thingy... is it anything like a tooob ?:-)

...Jim Thompson

Its an internal 4 MHz RC oscillator with programmable calibration. here is the link to the datasheet.

Remove "HeadFromButt", before replying by email.

Thanks for the reminder. The existing R/C oscillator in the PICs is not very stable.

I designed a replacement that will be in future PICs. It's similar to "CMOS-Osc-NoClip.pdf" on the S.E.D/Schematics page of my website, except that the R-tap approach isn't used, since that node does not have an external connection.

But my approach still isn't anything near what you can get with a crystal.

...Jim Thompson

Octal base if you like. ;-)

Unadjusted accuracy is +/-7% over temperature @5V Vdd, which isn't too bad, but there are better ones.

The similar 8-pin PIC12F629/75 is within +/-2% over Vdd and temperature, and within +/-1% at 3.5V and 25°C. Of course that's because it's calibrated at the factory. $1.11 in 100's.

Best regards, Spehro Pefhany

Even a resonator is pretty good compared to anything made with ordinary RC components. +/-0.5% initial tolerance and very low drift. Cheap; only one added part.

Best regards, Spehro Pefhany

I imagine you modified the input-protection circuit to allow beyond- the-rail swings, thereby stopping any undesired input-current draw? Very nice, this reduces any errors from power-supply voltage drift, etc. Nevertheless, surely Microchip's in-circuit post-calibration approach is still necessary, given the usual poor capacitor accuracy?

Thanks, - Win

whill_at_picovolt-dot-com

Their in-circuit calibration scheme talks about 1% overall accuracy.

Thanks, - Win

whill_at_picovolt-dot-com

There is no input protection needed... the gate doesn't go to a pin.

The main inaccuracy is poly R's, NOT the (fairly) accurate C's.

...Jim Thompson

Is there any large company you *haven't* designed for? :-)

Al

So, how do you solve that? External R? Or is it back to requiring a factory cal? What *uncalibrated* accuracy are we talking about here, 1% 2% 5%?

Thanks, - Win

whill_at_picovolt-dot-com

I have 20-30% in mind for unadjusted resistor accuracy.. don't know if that's still true, have things improved, Jim? Some of the specialist linear IC makers have much more accurate resistors but I suspect they are trimmed metal film.

Best regards, Spehro Pefhany

Sheesh...You can lead an old timer to water, but you can't make em drink. (The water that is, tequila is an other story)

Remove "HeadFromButt", before replying by email.

These particular ones are poly, and there are analog switches to "calibrate".

...Jim Thompson