Searching \ for '[PIC] Oscillator stability with temperature change' 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/microchip/ios.htm?key=temperature
Search entire site for: 'Oscillator stability with temperature change'.

Exact match. Not showing close matches.
PICList Thread
'[PIC] Oscillator stability with temperature change'
2011\01\12@145717 by Gordon Williams

flavicon
face
Hi,

I'm designing a count down timer to start nordic ski racers at set intervals
over the period of an 8 hour day.  I wondering what is the best way to get
an accurate time over this period with little drift.  What I want to do is
synchronize the device to the rest of the timing system at the beginning of
the day and let it run without having to resync it later.  The accuracy that
I would like to achieve is 0.5 sec over 8 hours or roughly 17 ppm over a
temperature range of -20 C to +40 C.

The PIC device will be battery powered, have a 4 by 7 segment multiplexed
LED, voltage regulator, buzzer, a button or two and a few leds.  Not that
complicated.  It is a one off device for my own use.  I'm going to use a 1
sec pulse from a GPS to check calibration.

The question is the best way to get an accurate clock over this temperature
range.  I'm looking for something in the 4 to 10 Mhz range.  In past PIC
projects I've used in internal oscillator and don't have any experience with
crystals.

This is what I've considered so far.

My initial though was that I would have to use a temperature compensated
canned oscillator to get the accuracy over the range because a standard
crystal has a tolerance of +- 30 ppm and a stability of +- 50 ppm.

Then I thought that I could use a temperature sensor costing $0.30 or so
search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=MCP9701T-E/TTCT-ND
and then use a standard crystal and do the temperature correction.

Next I found some better crystals with +-10 ppm tolerance and +-10 ppm
stability for about 80 cents
search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=887-1265-1-ND
The problem I see here is what caps should I use?  My understanding is that
if I'm off on the cap value, I will also be off on the frequency.  As well,
the caps will have a temperature coefficient so the load capacitance with
change with temperature.  Any thoughts about the dos and don'ts about this.

Lastly, if I have a temperature sensor, could I just use the internal
oscillator to correct for temperature changes?  I would like this as it is
easy.  What I'm wondering is how repeatable it would be with temperature.
The internal osc will have a range of about +-5% over this temperature.  It
seems like a bit of a stretch to bring the accuracy from 50,000 ppm down to
17 ppm even with knowing the temperature.  Any thoughts?

Thanks,

Gordon Williams

2011\01\12@152105 by Geo

picon face
Gordon Williams wrote:

> I'm designing a count down timer to start nordic ski racers at set intervals
> over the period of an 8 hour day.  I wondering what is the best way to get
> an accurate time over this period with little drift.  What I want to do is
> synchronize the device to the rest of the timing system at the beginning of
> the day and let it run without having to resync it later.  The accuracy that
> I would like to achieve is 0.5 sec over 8 hours or roughly 17 ppm over a
> temperature range of -20 C to +40 C.
> Personally, I would use a GPS module with 1 p.p.s output but the initial sync might be a problem.

George Smith

2011\01\12@153814 by Gordon Williams

flavicon
face

----- Original Message ----- From: "Geo" <spam_OUTbuggiesmithTakeThisOuTspamgmail.com>
To: "Microcontroller discussion list - Public." <.....piclistKILLspamspam@spam@mit.edu>
Sent: Wednesday, January 12, 2011 3:20 PM
Subject: Re: [PIC] Oscillator stability with temperature change


> Gordon Williams wrote:
>
> > I'm designing a count down timer to start nordic ski racers at set
intervals
> > over the period of an 8 hour day.  I wondering what is the best way to
get
> > an accurate time over this period with little drift.  What I want to do
is
> > synchronize the device to the rest of the timing system at the beginning
of
> > the day and let it run without having to resync it later.  The accuracy
that
> > I would like to achieve is 0.5 sec over 8 hours or roughly 17 ppm over a
> > temperature range of -20 C to +40 C.
> >
> Personally, I would use a GPS module with 1 p.p.s output but the initial
> sync might be a problem.
>
> George Smith
>

The GPS will only be used for testing and calibration - not in the field.

Regards,

Gordon Williams

2011\01\12@163408 by Richard Prosser

picon face
Apart from the GPS option, how about getting a 1 second (or whatever)
tick signal from the guts of a cheap digital clock?  You should be
able to pick them up cheaply enough, and the pcbs are generally small
and powered from 1.5V or so.

RP

On 13 January 2011 09:37, Gordon Williams <gwilliamsspamKILLspamncf.ca> wrote:
>
> {Original Message removed}

2011\01\12@164834 by Scott

picon face
On Wed, Jan 12, 2011 at 3:33 PM, Richard Prosser <.....rhprosserKILLspamspam.....gmail.com> wrote:
> Apart from the GPS option, how about getting a 1 second (or whatever)
> tick signal from the guts of a cheap digital clock?  You should be
> able to pick them up cheaply enough, and the pcbs are generally small
> and powered from 1.5V or so.
>

I would imagine those cheap digital clocks are not very accurate,
especially over temperature.

I doubt adjusting the intosc based on readings from the temperature
sensor would yield the accuracy you're spec'ing. I'd suggest going
with a TCXO.

-Scott

2011\01\12@165800 by IVP

face picon face
part 1 1096 bytes content-type:text/plain; charset="iso-8859-1" (decoded quoted-printable)

> I would like to achieve is 0.5 sec over 8 hours or roughly 17 ppm
> over a temperature range of -20 C to +40 C

If you assume, not unreasonably, that the chip out of a "kitchen"
clock is good to ~ 1 minute/year, that's around 2ppm. I use them
as the timebase in many projects. You won't find better performance
cheaper

Attached is a snippet from an old application. '1' and '2' are where
the coil used to attach. These are reciprocating lines, ie 0.5Hz AC,
to reverse the polarity of the coil each second (kitchen clocks are
magnetic drive). Take one for 0.5Hz or both for 1Hz

Nominal supply is a 1.5V AA, and they use only a few 10's of nA.
I've not tried a temperature range test but that should be easy to
do. Synchronise any two clocks and put the one you want to pinch
the chip out of in the freezer

If there is a significant difference I'm sure that temperature logging
could be used for compensation. Either that or provide a small
local heater - eg resistor - for the chip or whatever oscillator you
eventually use

Joe

part 2 2873 bytes content-type:image/gif; name="kitchen_clock_pic.gif" (decode)


part 3 181 bytes content-type:text/plain; name="ATT00001.txt"
(decoded base64)

--
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
mailman.mit.edu/mailman/listinfo/piclist

2011\01\12@170641 by Dwayne Reid

flavicon
face
At 12:56 PM 1/12/2011, Gordon Williams wrote:

>The question is the best way to get an accurate clock over this temperature
>range.  I'm looking for something in the 4 to 10 Mhz range.  In past PIC
>projects I've used in internal oscillator and don't have any experience with
>crystals.

Maxim makes a 32KHz stabilized module that is relatively inexpensive.  We use them here as intermediate standards for calibrating our other 32KHz crystal-based units.

These units from Maxim are really rather good.  Sorry - don't recall the exact part number or price.

dwayne

-- Dwayne Reid   <EraseMEdwaynerspam_OUTspamTakeThisOuTplanet.eon.net>
Trinity Electronics Systems Ltd    Edmonton, AB, CANADA
(780) 489-3199 voice          (780) 487-6397 fax
http://www.trinity-electronics.com
Custom Electronics Design and Manufacturing

2011\01\12@173445 by Funny NYPD

picon face
>We use them here as intermediate standards for calibrating our other 32KHz crystal-based units.

So you change the capacitor for each crystal to calibrate the unit? Or you calibrate the crystal by software adjustment?

Funny N.
Au Group Electronics, http://www.AuElectronics.com
http://www.AuElectronics.com/products
http://augroups.blogspot.com/




________________________________
From: Dwayne Reid <dwaynerspamspam_OUTplanet.eon.net>
To: Microcontroller discussion list - Public. <@spam@piclistKILLspamspammit.edu>
Sent: Wed, January 12, 2011 5:06:39 PM
Subject: Re: [PIC] Oscillator stability with temperature change

At 12:56 PM 1/12/2011, Gordon Williams wrote:

>The question is the best way to get an accurate clock over this temperature
>range.  I'm looking for something in the 4 to 10 Mhz range.  In past PIC
>projects I've used in internal oscillator and don't have any experience with
>crystals.

Maxim makes a 32KHz stabilized module that is relatively inexpensive.  We use them here as intermediate standards for calibrating our other 32KHz crystal-based units.

These units from Maxim are really rather good.  Sorry - don't recall the exact part number or price.

dwayne

-- Dwayne Reid   <KILLspamdwaynerKILLspamspamplanet.eon.net>
Trinity Electronics Systems Ltd    Edmonton, AB, CANADA
(780) 489-3199 voice          (780) 487-6397 fax
http://www.trinity-electronics.com
Custom Electronics Design and Manufacturing

2011\01\12@180644 by Sean Breheny

face picon face
Is this a one-off or something you need to produce many of?

As others have said, I don't see why GPS could not be used
continuously rather than just for calibration.

If not GPS, then an off-the-shelf TCXO would work very well.

I doubt that a normal clock would be accurate enough because they are
intended for a temperature controlled environment (e.g., a house or a
person's wrist). I'm assuming that this device you are making will be
outdoors.

For the crystal and caps - the capacitors DO alter the frequency of
the crystal but only over a few tenths of a percent for the full range
of workable capacitance values (say a 1000% variation). So, using
standard, good quality NP0 (also called C0G) ceramic capacitors should
make their temperature variation a negligible factor. However, you
would want to do an initial calibration.

When doing your accuracy calcs, be sure to include crystal aging. I
believe that crystals age with the square root of time, so an initial
burn-in time of even a week or so may help with this.

I agree with others that this kind of accuracy probably cannot be
achieved using the internal osc, even with temperature compensation.
The most challenging thing about temperature compensation is that many
types of oscillators exhibit temperature-driven hysteresis. In other
words, the frequency variation is not a simple function of the present
temperature, but depends on the time history of temperature. Aging
would also probably be significant with the internal osc.

Finally, as an "out-there" option if this is a one-off, good used
Rubidium frequency standards can be had on eBay for about US $100. I
bought two of them and they are amazing. Biggest down side is that
they have a limited life (about 10 years of continuous operation).
However, they achieve better than part per billion accuracy. I took my
two models, purchased from different eBay vendors, and when I power
both on and connect each to a different channel of my scope, once they
achieve their own internal "lock", they have no visible drift between
each other as observed over minutes (in other words, less than about
10 deg phase change at 10MHz over 5 minutes, which works out to 10
parts per trillion difference)

Sean


On Wed, Jan 12, 2011 at 2:56 PM, Gordon Williams <RemoveMEgwilliamsTakeThisOuTspamncf.ca> wrote:
{Quote hidden}

>

2011\01\12@181834 by Dwayne Reid

flavicon
face
Our units have a trim capacitor on the crystal.

We trigger the scope from the Maxim part and monitor the unit under test by very loosely coupling the insulated scope probe tip near the OSC2 pin (no direct connection).  Its then a simple matter to adjust the trim capacitor for zero phase change over time.

Its really cool to watch the sine wave slow down, stop, then reverse direction as you adjust the trim cap.  We leave the units set as close to zero phase change as we can at room temperature.

Its really easy to see the effect of temperature changes on the unit under test.   Its also really easy to change the temperature of the Maxim part and observe that it essentially remains constant over a wide temperature range.

dwayne


At 03:34 PM 1/12/2011, Funny NYPD wrote:
{Quote hidden}

-- Dwayne Reid   <RemoveMEdwaynerspamTakeThisOuTplanet.eon.net>
Trinity Electronics Systems Ltd    Edmonton, AB, CANADA
(780) 489-3199 voice          (780) 487-6397 fax
http://www.trinity-electronics.com
Custom Electronics Design and Manufacturing

2011\01\12@182815 by RussellMc

face picon face
> ... accurate  clock over 8 hoursunder varying temperature  ...

What they all said, plus:

Budget / acceptable cost is a significant factor in decision making.
Here the application SOUNDS like it should allow reasonable cost but
your questions make it sound like low cost is important.
What is an acceptable all up cost?

There are many possible solutions, but simple "ovening" and a half
tolerable crystal to start would be a lot easier than many other
solutions.
Whether this is feasible depends on the energy availability but,
unless you need to build it utterly tiny and at minimum
You may wish to decide if you can lower the maximum temperature spec
as that allows a lower oven temperature.

eg a 1000 mAh LiIon will give you about 3.5 Watt hour or a constant
power of about 0.5 Watt over 8 hours. 2 x AA NimH 2500 mAh will give
you about 5 Wh.
Dissipate 1/2 Watt in an 0805 resistor on an eg FR4 PCB with thin
tracks. Apply finger pressure. (DTTAH) Report back as to whether
that's liable to be enough power to run a modetly insulated oven :-).

Tempertaure sensor and known clock drift is a viable option and less
brute force. More work to be sure you are right but more
"sophisticatd" and lower power.

Internal clock by itself is not viable.

A superb source of free temperature control is by strapping a device
to your wrist, inside of arm better than outside probably. May not be
feasible in your case :-). Adds substantially to stability of watch
time keeping accuracy.


Russel

2011\01\12@183611 by Gordon Williams

flavicon
face

----- Original Message ----- From: "Richard Prosser" <rhprosserEraseMEspam.....gmail.com>
To: "Microcontroller discussion list - Public." <EraseMEpiclistspammit.edu>
Sent: Wednesday, January 12, 2011 4:33 PM
Subject: Re: [PIC] Oscillator stability with temperature change


Apart from the GPS option, how about getting a 1 second (or whatever)
tick signal from the guts of a cheap digital clock?  You should be
able to pick them up cheaply enough, and the pcbs are generally small
and powered from 1.5V or so.

RP

Cheap digital clocks are likely only accurate at room temperature.  Making a
clock is not a problem - making an accurate one over the temperature range
is.  That is why I was looking for oscillator suggestions.

Gordon Williams

2011\01\12@191207 by IVP

face picon face
> Cheap digital clocks are likely only accurate at room temperature

They could be, but you can't say that for certain without testing

As an experiment I've chucked my wall clock into the freezer,
which is around -15C. I'll report how it fares compared with
the lucky clock that isn't freezing its nuts off in the dark with a
bag of peas

As you seem to have battery supply to spare (LEDs, regulator etc)
it would not seem to be too wasteful to keep the oscillator within a
narrower range of temperature with a little heating. That could be
done quite simply with a small block of polystyrene foam housing
the oscillator, a resistor and a temperature sensor

40C seems a tad warm for skiing. I'd imagine snow would be
fairly thin on the ground

Jo

2011\01\12@191841 by Jesse Lackey

flavicon
face
Hi Sean, I bought this from ebay...
Lucent Rubidium Frequency Standard 10MHz RFG-RB 160284303432
about 2 years ago and have done nothing with it, a brief search didn't yield much about the 25? pin d-sub connector on the back.  Does this happen to be the same or similar to what you bought?  If so I'd love to get any tips from you on firing it up.

Cheers,
J



Sean Breheny wrote:
{Quote hidden}

>> -

2011\01\12@193029 by Gordon Williams

flavicon
face
Was it DS32kHz that you used?  Appear available in $7-12 range.

The start up time is 1 sec so I couldn't use it as a PIC osc input.  Right?
It would have to go into one of the timers and I would use the internal osc,
which would be better as it is much faster.

Gordon Williams

{Original Message removed}

2011\01\12@201708 by Dwayne Reid

flavicon
face
At 05:29 PM 1/12/2011, Gordon Williams wrote:
>Was it DS32kHz that you used?  Appear available in $7-12 range.

Sounds right.

Maxim will give you several pieces to try for free if you wish.


>The start up time is 1 sec so I couldn't use it as a PIC osc input.  Right?
>It would have to go into one of the timers and I would use the internal osc,
>which would be better as it is much faster.

Yeah - we use ours as a transfer standard rather than for the clock in the actual units.  Those use 32 KHz watch crystals into TMR1.

We went that way because we needed battery-backed clock that was good for several weeks.  The units are used only during the winter and the PIC will actually power down if the system is not used for several weeks.  We anticipate that the cr2032 battery should be good for about 10 years because the systems only run for a few weeks every year.

I would do the same thing in your case: feed the DS32KHz into TMR1 and clock the PIC normally.

dwayne


>{Original Message removed}

2011\01\12@210753 by Isaac Marino Bavaresco

flavicon
face
Em 12/1/2011 20:06, Dwayne Reid escreveu:
> At 12:56 PM 1/12/2011, Gordon Williams wrote:
>
>> The question is the best way to get an accurate clock over this temperature
>> range.  I'm looking for something in the 4 to 10 Mhz range.  In past PIC
>> projects I've used in internal oscillator and don't have any experience with
>> crystals.
> Maxim makes a 32KHz stabilized module that is relatively
> inexpensive.  We use them here as intermediate standards for
> calibrating our other 32KHz crystal-based units.
>
> These units from Maxim are really rather good.  Sorry - don't recall
> the exact part number or price.
>
> dwayne

I use the DS3231 from Maxim. It is an I2C serial RTC with internal
crystal and temperature compensated (it uses a temperature sensor to
compensate for temperature changes, not keeping the temperature stable
like in an oven, and so it consumes very low power).
According to their data, it is stable to 2ppm.

Isaac

__________________________________________________
Fale com seus amigos  de graça com o novo Yahoo! Messenger http://br.messenger.yahoo.com

2011\01\12@213951 by Oli Glaser

flavicon
face
On 13/01/2011 02:07, Isaac Marino Bavaresco wrote:
> I use the DS3231 from Maxim. It is an I2C serial RTC with internal
> crystal and temperature compensated (it uses a temperature sensor to
> compensate for temperature changes, not keeping the temperature stable
> like in an oven, and so it consumes very low power).
> According to their data, it is stable to 2ppm.
>

Yes, Maxim make a few decent RTCs - I have used the DS3231 too and DS1307(this one not t.c), and they have others like the DS3231M (like DS3231 but uses mems resonator), DS3232, DS32B35 also with temperature compensation, accurate to 3.5ppm from -40 to 0 and 0 to 85 deg C (2ppm from 0 to 40)
All at pretty reasonable prices (IMO) and easy to use, probably worth considering - especially for a one off as you don't have to worry about availability... :-)

2011\01\13@005935 by Sean Breheny

face picon face
Unfortunately, no - the ones I bought were the LPRO units and they use
a 10-pin connector. Using them is very easy - just apply power and
ground (I think it is 24V?) and the output is a 10MHz sinusoid from
one of the other pins. I made an enclosure for one of mine which
includes a switching power supply accepting AC mains input, as well as
an LED which is controlled by the "lock good" logic output from the
unit. It also has a circuit which shapes the 1Vpp sinusoid into a 5V
TTL square wave and can drive a 50 ohm load.

The following page seems to mention your unit:

http://www.maxmcarter.com/rubidium/index.php

Also, check out the Time-Nuts mailing list:

http://leapsecond.com/time-nuts.htm

Your unit may well be mentioned in the archives.

Sean


On Wed, Jan 12, 2011 at 7:18 PM, Jesse Lackey <RemoveMEjsl-mlspam_OUTspamKILLspamcelestialaudio.com> wrote:
> Hi Sean, I bought this from ebay...
> Lucent Rubidium Frequency Standard 10MHz RFG-RB 160284303432
> about 2 years ago and have done nothing with it, a brief search didn't
> yield much about the 25? pin d-sub connector on the back.  Does this
> happen to be the same or similar to what you bought?  If so I'd love to
> get any tips from you on firing it up.
>
> Cheers,
> J
>

2011\01\13@051018 by alan.b.pearce

face picon face
> There are many possible solutions, but simple "ovening" and a half
> tolerable crystal to start would be a lot easier than many other
> solutions.

This was my thought too. Stick the whole electronics unit inside a thermos flask, with a suitable resistor for heating. When synchronising at the beginning of the event the unit could be attached to a suitable mains based power source, and the heating resistor run for long enough to get the internals to a stable temperature. This should hold the temperature to a much closer tolerance than the total temperature range originally given.
If it is possible to have the unit powered from a car battery or other 'relatively beefy' power pack (which you may need any way to run 7-segment displays for 8 hours) then a simple temperature control can be implemented to keep the unit with 10 degrees easily, without too much extra current draw.
-- Scanned by iCritical.

2011\01\13@065621 by RussellMc

face picon face
> > There are many possible solutions, but simple "ovening" and a half
> > tolerable crystal to start would be a lot easier than many other
> > solutions.
>
This was my thought too. Stick the whole electronics unit inside a
thermos flask, with a suitable resistor for heating. When
synchronising at the beginning of the event the unit could be attached
to a suitable mains based power source, and the heating resistor run
for long enough to get the internals to a stable temperature. This
should hold the temperature to a much closer tolerance than the total
temperature range originally given.

I was thinking of something "somewhat smaller".
Suitable resistor and crystal and temperature sensor in a smallish
piece of polystyrene - or for extra points, Perlite.
As noted before - a small fraction of a Watt should be enough.

Controller could be as little as a PTC thermistor - but no great
temperature accuracy - but probably an improvement on otherwise.
Or as complex as you like.
You can but commercial oven units but that is neither necessary or in
your budget.

Here's a nice enough lowish cost DIY temperature controller.

              http://www.w6pql.com/crystal_oven_controller.htm

Much detail and useful discussion.
He uses a PTC thermistor as sensor BUT a temperature controller IC may
simplify things slightly.
Just looked at Digikey pricing. PTC (or even an NTC) and comparator
look cheaper and just as good.
Bang-bang switching probably fine. Adding a proportional component
would be easy enough [tm].


Excellent Compromise

Microchip "linearised thermistor" - Digikey 30 cents US in stock
1 degree accuracy typical best versions, 2 degrees max
BUT this is across temperature so, as it is "eating its own tail" set
point temperature may not be vastly accurate BUT may well be much more
stable.
ie you can probably control at an untrimmed absolute temperature of
well < 1 C but table to much better than that medium term.
See spec sheet for details

                 http://ww1.microchip.com/downloads/en/DeviceDoc/21942e.pdf

Having said all that - a std NTC or PTC in the circuit above  would do
as well but be more difficult to predict its operating point.


Russel

2011\01\13@072123 by RussellMc

face picon face
Same man.
Same controller.
Lower power resistor load.
Smaller box.
Useful comments

    http://www.w6pql.com/vhf_ocxo.htm


2011\01\13@072625 by alan.b.pearce

face picon face
> Here's a nice enough lowish cost DIY temperature controller.
>
>                http://www.w6pql.com/crystal_oven_controller.htm

Looks like a lot of components ... ;)))))

When the marine SW went to SSB from AM, the company I worked for ovened the crystals in the transceivers to get the required frequency stability. They used a Darlington power transistor as the heating element, a PTC resistor from base to emitter, and another resistor to bias the transistor on. This gave them suitable frequency stability and ran the crystals at around 30C IIRC, in an oven similar to your description (polystyrene block around the crystals). 3 electronic components plus a copper loop to carry the heat around the crystals themselves.
-- Scanned by iCritical.

2011\01\13@093835 by Paul Hutchinson

picon face
> -----Original Message-----
> From: RemoveMEpiclist-bouncesTakeThisOuTspamspammit.edu On Behalf Of Dwayne Reid
> Sent: Wednesday, January 12, 2011 5:07 PM
>
> Maxim makes a 32KHz stabilized module that is relatively
> inexpensive.  We use them here as intermediate standards for
> calibrating our other 32KHz crystal-based units.
>
> These units from Maxim are really rather good.  Sorry - don't recall
> the exact part number or price.


The Maxim/Dallas part is DS32KHZ, it's about $8.00 in singles, $4.00 in
hundreds.

Paul Hutch


>
> dwayne
>
> --
> Dwayne Reid   <EraseMEdwaynerspamspamspamBeGoneplanet.eon.net>
> Trinity Electronics Systems Ltd    Edmonton, AB, CANADA
> (780) 489-3199 voice          (780) 487-6397 fax
> http://www.trinity-electronics.com
> Custom Electronics Design and Manufacturing

2011\01\13@103702 by Mark Rages

face picon face
On Thu, Jan 13, 2011 at 7:38 AM, Paul Hutchinson
<RemoveMEpaullhutchinsonKILLspamspamyahoo.com> wrote:
>> -----Original Message-----
>> From: piclist-bouncesSTOPspamspamspam_OUTmit.edu On Behalf Of Dwayne Reid
>> Sent: Wednesday, January 12, 2011 5:07 PM
>>
>> Maxim makes a 32KHz stabilized module that is relatively
>> inexpensive.  We use them here as intermediate standards for
>> calibrating our other 32KHz crystal-based units.
>>
>> These units from Maxim are really rather good.  Sorry - don't recall
>> the exact part number or price.
>
>
> The Maxim/Dallas part is DS32KHZ, it's about $8.00 in singles, $4.00 in
> hundreds.
>
> Paul Hutch
>

You can buy 0.5ppm TXCOs from digikey for <$6 in singles.

For a single one, you can salvage a TCXO from an old cell phone.

Regards,
Mark
markrages@gmail
-- Mark Rages, Engineer
Midwest Telecine LLC
spamBeGonemarkragesSTOPspamspamEraseMEmidwesttelecine.com

2011\01\13@122108 by Isaac Bavaresco

flavicon
face
--- Em qui, 13/1/11, Paul Hutchinson <KILLspampaullhutchinsonspamBeGonespamyahoo.com> escreveu:

> The Maxim/Dallas part is DS32KHZ, it's about $8.00 in
> singles, $4.00 in
> hundreds.

So it is more expensive than the DS3231 which is a complete RTC with an optional 32KHz output...


2011\01\13@123527 by Bob Blick

face
flavicon
face

On Thu, 13 Jan 2011 08:37:01 -0700, "Mark Rages" said:

>
> For a single one, you can salvage a TCXO from an old cell phone.

That sounds good. Please, tell me more! What do I look for? What
frequency is it likely to be?

Thanks,
Bob

-- http://www.fastmail.fm - The way an email service should be

2011\01\13@144631 by Gordon Williams

flavicon
face
I see the DS3221 at $11. The DS32KHZ is at the same price at Digikey (-40 to
85 deg 7 ppm)

Where can you get them for less?

Gordon Williams
{Original Message removed}

2011\01\13@162813 by Isaac Marino Bavaresco

flavicon
face
Em 13/1/2011 17:45, Gordon Williams escreveu:
> I see the DS3221 at $11. The DS32KHZ is at the same price at Digikey (-40 to
> 85 deg 7 ppm)
>
> Where can you get them for less?


Well, I said DS3231, not DS3221.

We didn't purchase them directly, our Korean partner assembled the
boards and did all the purchasing.
I think it was around $3 to $4.

In Maxim's site it is quoted at $3.01@1k

__________________________________________________
Fale com seus amigos  de graça com o novo Yahoo! Messenger http://br.messenger.yahoo.com

2011\01\15@000023 by IVP

face picon face
> Cheap digital clocks are likely only accurate at room temperature

Gordon,

The results of my rough temperature testing

6 hours @ -15C lost 8 minutes (about 1% slow)

6 hours @ +45C gained 10 seconds

So there ya go. Interesting. What you say is largely true then. The
typical clock is probably optimised for the habitable ambient range

Made in the many many many millions I'd expect they'd all be much
of a muchness

Jo

2011\01\15@000739 by Gordon Williams

flavicon
face
Interesting - Thanks for your work.

Gordon Williams

----- Original Message ----- From: "IVP" <EraseMEjoecolquittspamEraseMEclear.net.nz>
To: "Microcontroller discussion list - Public." <@spam@piclist@spam@spamspam_OUTmit.edu>
Sent: Friday, January 14, 2011 11:59 PM
Subject: Re: [PIC] Oscillator stability with temperature change


{Quote hidden}

> -

More... (looser matching)
- Last day of these posts
- In 2011 , 2012 only
- Today
- New search...