Continue to Site

Welcome to EDAboard.com

Welcome to our site! EDAboard.com is an international Electronics Discussion Forum focused on EDA software, circuits, schematics, books, theory, papers, asic, pld, 8051, DSP, Network, RF, Analog Design, PCB, Service Manuals... and a whole lot more! To participate you need to register. Registration is free. Click here to register now.

Register Log in

Super precise 32768Hz Oscilator

Status
Not open for further replies.
simce

simce

Full Member level 4
Full Member level 4
Joined
Jan 9, 2004
Messages
220
Helped
12
Reputation
24
Reaction score
12
Trophy points
1,298
Activity points
2,536
rxcontrol.free.fr

Hi every one,
I need solution for super accurate oscillators fo 32kHz for RTC.
I need to have almose exact time of two distance stations (it is more important that their RTC times are close one to each other than they are close to apsolute time in earth). I have tried DS32KHZ from dallas/maxim but for one month they can deviate one from another for about 10-15. This can be due to temperature differences, because i tested 5pcs of these oscillators on room temperature and for about 10 months maximum deviation was 1sec, that is very good, but when variable tempearture is applied to them they start deviating....
It can be used also higher frequency oscillator but it has to be with good ppm accuracy.....
 

32768hz receiver

I think the reason it deviated because you are using a sample product, try to buy one and you will see the difference, sampled products were tested about their reliability but not on their stability.
 

32768hz accurate

How about a chamber with controlled heating and therefore constant temperature.
This is done if very high accuracy is desired.


Mik
 

32768hz oscilator

Can you afford GPS receivers? These will give you the time at both ends and they will always be identical. Another expensive option is to buy from one of the major test equipment manufacturers stand alone high precision frequency references. These are usually 1 part per 10^8 accuracy.

Another option is to have receivers locked to the VLF time signals like WWVB. You could also have your own synchronizing transmitter which your units receive. Or one of the units can also have the synchronizing transmitter.
 

think the reason it deviated because you are using a sample product, try to buy one and you will see the difference, sampled products were tested about their reliability but not on their stability.
Click to expand...
No i'm not using samples, i have about 100 pcs of DS32KHZ and i noticed that some of them are more cross acurate than others. This way i made board with 10 DS32KHZ and rtc's and i left them for one moth and then i measured deviance and then i selected pairs that have closest times...

How about a chamber with controlled heating and therefore constant temperature.
This is done if very high accuracy is desired.
Click to expand...
I was also thinking of this but as you can see i've made one board with 10 oscilators and they are all in same environmental coditions. Even they are in same conditions they are diviating. Also thia application has to be for low power conditions...
Can you afford GPS receivers?
Click to expand...
Unfortunately for this application i can no afford GPS because of low price of the application. At first i though to use some kind of transmitter/receiver system so i can synchronize them at least one a day and this way i can have low power consumption and high precision of RTC's.
Also i was thinking of zeropower RTC's from ST, and i need information if you have any experience using them?
 

Your low cost solution is to "burn in" the oscillators. Run them for several months before you trim their frequencies by adjusting one of the capacitors in the feedback network. This will get most of the long term drift out of them.
 

Your low cost solution is to "burn in" the oscillators. Run them for several months before you trim their frequencies by adjusting one of the capacitors in the feedback network. This will get most of the long term drift out of them.
Click to expand...
Yes this is one possible solution. Because i can not find accurate enough oscilators with price that is acceptable i'll use DS32KHZ, but i'll choose pairs as i've done untill now.
I've found TCXO's that are for higher frequncy but it has the same ppm factor. This tell's me that even if i use them with frequency dividers i'll have final frequency with same deviation as with DS32KHZ. Can enyone tell me what is the meaning of ppm. What will be frequency deviation if i have oscilator of 100Hz with 1ppm?
Does it will be 100Hzx(1e-6)=+/-0.0001Hz?
 

simce said:
... Can enyone tell me what is the meaning of ppm. What will be frequency deviation if i have oscilator of 100Hz with 1ppm?
Does it will be 100Hzx(1e-6)=+/-0.0001Hz?
Click to expand...

ppm = parts per million
so your calculation is right

Mik
 

If you can receive one of the low frequency time normals like DCF77 in central Europe you could use this as your synchronised time. A more complicated solution is using one of the time normals that are available on the internet.

best regards
 

If you can receive one of the low frequency time normals like DCF77 in central Europe
Click to expand...
Where can i find more information regarding this matter?
 


I do not better (cheeper) way to get the enough time information as DCF77. If you want to have 1 sec deviation in 10 months over range of temperatures you need accuracy of 0.04ppm. Prices of oscillators with such accuracy are in range of kE or k$, whatever your favorite currency is, since they must be some kind of OCxO. Plain TCxO you can not find better as 0.2 ppm. If you try to build your OCxO you should first characterize each crystal resonator, "burn it" (as said) and than tune it on desired frequency and keep on constant temperature.
And remeber, although you by resonators that are specified up to some tolerance in entire temperature range, they all have different behavior and just the limits are the same.

I hope it can help
D.
 

simce, how much accuracy do you need, and over what period of time?
What is an acceptable price?
How many do you need? Is this just a one-time project?
 

Status
Not open for further replies.

Similar threads

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top