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.

does input clock frequency pin have to come from crystal oscillator?

Status
Not open for further replies.

dl09

Full Member level 4
Full Member level 4
Joined
Feb 12, 2020
Messages
226
Helped
2
Reputation
4
Reaction score
1
Trophy points
18
Activity points
1,706
i have studied the datasheet of the ov7075, 1 pin is input clock frequency. does the input to input clock frequency have to come from a crystal oscillator? what if the input came from a 555 timer? the minimum input clock frequency of ov7075 is 10 megahertz.
 

You are asking about OV7670? Typical input frequency is 24 rather 10 MHz (for 30 fps). The chip might principally work with an inaccurate RC oscillator, but I doubt that it's useful for a regular video application.

How do you process the video output?
 

I'm not sure what an OV7075 is but I'm guessing its a camera module.

If it is, the scanning frequency which is derived from the crystal frequency has to be fairly precise, within a few Hz so any other kind of oscillator probably isn't suitable. A 555 timer can't produce frequencies as high as 10MHz but even if it could, its stability would be far from acceptable.

Brian.
 

if i use any device that is not a crystal oscillator to provide input to the system clock pin of the ov7670, it's stability would be inadequate for this application?
 

Correct - not only stability but the frequency has to be exact as well. The video scan rate and data output rate have to be exact or it will not synchronize the picture on a normal monitor.

Brian.
 

so there is no way around this problem, without using a crystal oscillator?
 

Linear Technology produces resistor seteable MEMS oscillators, which are used to replace crystals in certain applications.

But if you know anything about LT’s products, is that they are very high performance but not cheap. What is your concern with crystals? They are ubiquitous and relatively inexpensive.
 

i am worried i won't get them working. i have not tried using a load capacitor. i have bought some crystal oscilllators, before.
 

Better to buy an oscillator, rather than trying to make one with a crystal.

Crystal oscillators are not hard, if you are into that sort of thing. If you are not....buy a clock! Lots of places sell ones for a $1 or so. They will start up, without problems, and stay on frequency, under a wide range of conditions.

Now, if you want to learn about making a crystal oscillator...........yeah, that is a subject guys like me can help you with. But, in the interim..............buy a clock. Much easier and certainly faster.

One last point.............$1 clocks are pretty much a commodity, these days. Don't invest a lot of time comparing specs on competing ones. Buy what you can find, at the right price, from wherever you buy your parts from. The differences, in this case, don't really exist.
 

There might be an existing and suitable clock resource
in the system of the end application, but for characterizing
or playing around maybe this doesn't help.

Some bench instruments provide clock references (e.g.
spectrum analyzers, AWGs). A decent pulse generator
or AWG could be a "good enough" clock source, won't
know until you either hook it up or understand all the
specs. Now the 10MHz reference output of spectrum
analyzers while usually a pretty good stable source,
is low level 50 ohm RF and would need squared up and
level shifted to be any use. And it can't be varied.

I'm all for the clock-in-a-can-for-cheap approach,
unless you need an excuse to buy test equipment.
Such as some desire to vary the clock under some
control. You could consider a VCO or NCO based
PLL eval board as an interesting play, one that has
an on board reference oscillator would be cute.
Such a critter would probably come with a PC interface
and demo CD. Might cost you if you don't smell like
volume sales, but maybe you can get something
from the various vendors of such products for not
too much money.

I have been collecting those FeelTech AWGs
off eBay when I find one going stupid cheap. The
new ones are out and the old ones are going cheaper.
They'll do >20MHz pulse if you get the right model.
 

Better to buy an oscillator, rather than trying to make one with a crystal.

Crystal oscillators are not hard, if you are into that sort of thing. If you are not....buy a clock! Lots of places sell ones for a $1 or so. They will start up, without problems, and stay on frequency, under a wide range of conditions.

Now, if you want to learn about making a crystal oscillator...........yeah, that is a subject guys like me can help you with. But, in the interim..............buy a clock. Much easier and certainly faster.

One last point.............$1 clocks are pretty much a commodity, these days. Don't invest a lot of time comparing specs on competing ones. Buy what you can find, at the right price, from wherever you buy your parts from. The differences, in this case, don't really exist.

you said "clock", what is that?
 

you said "clock", what is that?
It's sloppy speech, the technical term is crystal oscillator or crystal clock oscillator. A crystal with an oscillator circuit in a small package. You apply 3.3V or 5V power and get a square wave at a fixed frequency.

E.g. https://abracon.com/Oscillators/ASA.pdf
 

the crystal clock oscillator does not require a load capacitor?
 

Generally - but there may be exceptions:
A device with a built in oscillator circuit needs something connecting to it to set the frequency, normally a quartz crystal and with loading capacitors at each end.
A clock module is a ready built oscillator with the crystal and all other components inside a metal can. All you provide is power and it gives you the frequency out.

When a device needs a steady reference frequency, you can use it's in-built oscillator if you provide the crystal and capacitors -or- if it doesn't have a built-in oscillator you have to provide the clock from a module or equivalent external circuit. Some devices give you the option to use an internal oscillator or to use an outside frequency source, this is sometimes useful if you need to exactly synchronize several devices by driving them from one source.

The significance of the quartz crystal as a reference is it will produce a very stable fixed frequency. It really is a crystal made from quartz and it vibrates when an electrical stimulus is applied across it. The vibration is at a very precise rate which is decided by the physical size and shape of the quartz. This means you have to have them made specially for the frequency you need. Some commonly used frequencies are mass produced and you can buy them 'off the shelf' for others you have to contact a crystal manufacturer and tell them what you need.

Brian.
 

It would work, but it is an overkill, as the Si5351 is a frequency synthesizer.
Good if you require a frequency-agile clock source.

If all that is required is a single, fixed frequency, use the low cost clock modules suggested above.
 

do the clock modules require a load capacitor?
 

No, read the data sheets.

A crystal clock oscillator has a no other pins than power supply and buffered digital output. A load capacitor has no effect on the operation of the internal crystal.
 

can you provide links to stores that sell crystal clock oscillators?
 

Newark,
Farnell,
Digikey,
Mouser,
RS Components,
Rapid Electronics.

Just do a search for "crystal oscillator modules" and you will find hundreds of sources.

Brian.
 

Status
Not open for further replies.

Part and Inventory Search

Welcome to EDABoard.com

Sponsor

Back
Top