Searching \ for 'tapping xtal oscillator' in subject line. ()
Make payments with PayPal - it's fast, free and secure! Help us get a faster server
FAQ page: www.piclist.com/techref/timers.htm?key=oscillator
Search entire site for: 'tapping xtal oscillator'.

Truncated match.
PICList Thread
'tapping xtal oscillator'
1999\08\27@203436 by Anne Ogborn

flavicon
face
I'm using a xtal for my project, as I need an accurate
timebase, and I have to be cheap (production device).
So no can oscillator.

I'm wanting a stable 1MHtz source to drive the ISD chip as well,
one locked to the CPU clock.

The obvious thing to do is to tap into the xtal output.
So, can somebody tell me what sort of electronic wizardry I have
to do to make the 4MHtz crystal drive some sort of gate so I can
divide by 4 and drive the ISD chip???


--
Anniepoo
Need loco motors?
http://www.idiom.com/~anniepoo/depot/motors.html

1999\08\27@214434 by David Covick

flavicon
face
Anne,

By using a CMOS 4024 Ripple Counter you can clock 4 MHz in to pin 1.  Pin 11
is a divide by 4 and outputs 1 MHz.

Single chip solution.

David


{Original Message removed}

1999\08\27@223034 by Anne Ogborn

flavicon
face
David Covick wrote:
>
> Anne,
>
> By using a CMOS 4024 Ripple Counter you can clock 4 MHz in to pin 1.  Pin 11
> is a divide by 4 and outputs 1 MHz.
>
> Single chip solution.
>
> David


How do I tie pin 1 to the crystal?  Just tie it to OSC1 or OSC2?

--
Anniepoo
Need loco motors?
http://www.idiom.com/~anniepoo/depot/motors.html

1999\08\27@232103 by Bob Drzyzgula

flavicon
face
On Fri, Aug 27, 1999 at 07:27:43PM -0700, Anne Ogborn wrote:
>
> How do I tie pin 1 to the crystal?  Just tie it to OSC1 or OSC2?

I think that, on the PIC, OSC1 is nominally the input,
while OSC2 is nominally the output. However, I don't think
(watch everyone else say different) that it's recommended
that you use those as a clock source for other chips if
you're using the built-in oscillator with an external
crystal; I'm not sure that the PIC's oscillator has
sufficient drive for that (I know that it's been discussed
before, but it may come down to one of those questions
where the answer depends on how much of a problem it
would be for you if it didn't...)  If you forgo the PIC's
oscillator and make your own (mostly you need some more
resistors and an inverting buffer like the 74x04), then
the inverting buffer will have plenty of oomph to drive
both the PIC and other devices. uChip shows how to do this
in most data sheets, e.g. Figure 13-6 in the PIC16C6X
sheet. Where it says "To Other Devices", that's where
you'd pick up the signal to feed the 4024. Alternatively,
I suppose you could simply feed the 74x04 output straight
into the 4024, drive the PIC off pin 12 of the 4024 and the
ISD off pin 11.

BTW, the 4024 is also available in HCT; uChip shows
using an AS device feeding CLKIN... does it matter
if the CLKIN is fed CMOS or TTL levels?

--Bob

--
============================================================
Bob Drzyzgula                             It's not a problem
spam_OUTbobTakeThisOuTspamdrzyzgula.org                until something bad happens
============================================================

1999\08\28@002518 by tec

flavicon
face
I've used the OSC2/CLKOUT to drive two loads a 16C750 and a 16C552. No
problems.

Cypress has some nice clock products (probably to much money for this
project).

Todd Conard

Bob Drzyzgula wrote:

{Quote hidden}

1999\08\28@044606 by Anne Ogborn

flavicon
face
Thanks.

I'll try OSC2 - it just has to drive the ripple counter,
so it's worth it.

dern - my chip count just went up 8o/

1999\08\28@093109 by Dave VanHorn

flavicon
face
> The obvious thing to do is to tap into the xtal output.
> So, can somebody tell me what sort of electronic wizardry I have
> to do to make the 4MHtz crystal drive some sort of gate so I can
> divide by 4 and drive the ISD chip???


Careful.

Check the spec, there should be an output drive spec for that pin. If not,
you're in shaky territory, no matter how many hobby projects or protos "just
worked".  The most interesting part of a data sheet is often what is NOT in
it.

I had a similar case, where I disapproved of this arraingement because there
was no output drive spec, and the eng mgr told me that all the protos
worked, and "don't worry.."   Guess who got to fly to kaohsiung christmas
day to debug an assembly line full of product that didn't run... :(

Startup time on the osc may be affected too, but only you can answer if that
matters to this application.

In many cases, that output is designed to only drive a crystal.

Also, remember to check the effect of that gate capacitance on the osc
circuit.

1999\08\28@135046 by Morgan Olsson

picon face
Dave VanHorn wrote:

> > The obvious thing to do is to tap into the xtal output.
> > So, can somebody tell me what sort of electronic wizardry I have
> > to do to make the 4MHtz crystal drive some sort of gate so I can
> > divide by 4 and drive the ISD chip???
>
>
>Careful.
>
>Check the spec, there should be an output drive spec for that pin.

Concidentally, I plan to divide down a PIC driven 4,096 MHz Xtal to 1,024 MHz to feed a A/D (MAX110), for maximum 50Hz reject.

I plan to use 1/4 HC393, which is a nice little 16 pin device containing two /2 and two /8 small ripple counters. (its sister 390 has /2 and /5 counters)

Looking at the spec (Motorola) I see that the *maximum* input capacitance is 10pF.

>In many cases, that output is designed to only drive a crystal.

And a 30pF (typ) capacitor!

>Also, remember to check the effect of that gate capacitance on the osc
>circuit.

Yep, just reduce the cap you would normally use on that pin, with the nominal capacitande of the clock input you are driving. (something 7pF ? )

Being a CMOS device, it has negligible high DC resistance.  (do not use TTL...)

We also have to take in count the average voltage on thet pin; should we use HCT or HC?
Need to measure, or it it stated somewhere?
Anyone?

Regards
/Morgan
Morgans Reglerteknik, HŠllekŒs, 277 35 KIVIK
 tel 0414-446620, fax 70331, mrtspamKILLspaminame.com

1999\08\28@152313 by Morgan Olsson

picon face
Hej Anne Ogborn. Tack fšr ditt meddelande 08:38 1999-08-28 -0700 enligt nedan:

...

>In RC mode, OSC2 pin outputs CLKOUT which has 1/4 the
>frequency of OSC1, and denotes the instruction cycle time.

It is OK to drive the OSC1 pin with an external cirquit in Xtal configuration, but in RC mode that pin is both input and output, which will probably cause some EMI and more power consumption.  I think however that it is possible and will be OK to drive it througt a resistor approx 1kohm, although it is pretty ugly.

>OK - if I can find a reasonable price can oscillator, I don't need the divide by 4
>chip. That might justify some/all of the can's price.

The HC393 in my previous letter is a cheap one.  (and have extra free counters i need, too) Of course there are a lot of cheap counters to select from.  And I believe the design is safe.  Will test it next week.

/Morgan
Morgans Reglerteknik, HŠllekŒs, 277 35 KIVIK, SWEDEN
  tel +46(0)414-446620, fax -70331,   .....mrtKILLspamspam.....iname.com


'tapping xtal oscillator'
1999\09\08@185732 by p.cousens
flavicon
face
How about using a simple single transistor crystal oscillator circuit,
with the output going to osc in.
If the chip fuses are set to rc,
this will give you 1/4 clock out on osc out

Anne Ogborn wrote:
{Quote hidden}

1999\09\08@205243 by Anne Ogborn

flavicon
face
hpc wrote:
>
> How about using a simple single transistor crystal oscillator circuit,
> with the output going to osc in.
> If the chip fuses are set to rc,
> this will give you 1/4 clock out on osc out


though I'm called off right now by other emergencies, my
fix is probalby going to be a can oscillator and the
RC solution.

--
Anniepoo
Need loco motors?
http://www.idiom.com/~anniepoo/depot/motors.html

1999\09\09@090018 by Sam Laur

flavicon
face
> How about using a simple single transistor crystal oscillator circuit,
> with the output going to osc in.
> If the chip fuses are set to rc,
> this will give you 1/4 clock out on osc out

Except that you can't drive the osc1/clkin input with an external clock
in RC mode...

A cheap oscillator would also be a couple inverters, no can oscillators needed.
Of course it will take more board space than transistors.

1999\09\09@191103 by AppTech

picon face
>> How about using a simple single transistor crystal oscillator circuit,
>> with the output going to osc in.
>> If the chip fuses are set to rc,
>> this will give you 1/4 clock out on osc out
>
>Except that you can't drive the osc1/clkin input with an external clock
>in RC mode...


This shouldn't be true. I haven't tried this with a PIC but it is simply
implementing an oscillator in which the frequency determining element is the
RC time constant on the xtal in pin. A bit of playing with circuitry should
allow one to "fool" the PIC oscillator . The data sheet shows the circuit as
an inverter with the input timing capacitor charged high by the timing
resistor. The gate resets the capacitor to zero when a threshold is reached.

A little hands on would help here but it seems that driving the pin from an
external clock via a series resistor plus perhaps some capacitance on the
pin would allow one to "fool" it. In Annie's case (she started this thread)
this would save her a package of gates. A crystal oscillator can be made
with a single transistor. One would have top be fairly pressed economically
to justify this approach but a reliable design should be possible.


RM

More... (looser matching)
- Last day of these posts
- In 1999 , 2000 only
- Today
- New search...