What determines the speed of a Quartz Oscillator?

name9902

Member
Feb 12, 2005
48
0
0
I was curious as to what was it that determines the speed of a Quartz Oscillator...?

Since they seem to provide the "beat" that controls the CLK wire, in turn controlling the speed of the entire comp

so what determines this speed?
 

Fullmetal Chocobo

Moderator<br>Distributed Computing
Moderator
May 13, 2003
13,704
7
81
I believe it is inherent to the crystal itself. Different types of crystals pulse at different frequencies. Hence you have multipliers and what not to reach the frequency you need for the computer.
Tas.
 

name9902

Member
Feb 12, 2005
48
0
0
hmm, i figured...people speak of 5-10MHz quartz...so you need a multiplier to reach enough to be able to pulse the CLK wire and be efficient and able to "keep up" with the Clock Speed...?
 

Fullmetal Chocobo

Moderator<br>Distributed Computing
Moderator
May 13, 2003
13,704
7
81
I don't know so much about being efficient as being effective, but yes. Instead of creating or finding crystals to pulse at higher speeds, you just put a multiplier on it to reach your higher clock speeds.
Tas.
 

name9902

Member
Feb 12, 2005
48
0
0
mmm, dumb question here, don't know much about how multipliers work, but i suppose you can't put a multiplier on a the clock speed, which is why it sets the speed for everything else?
 

Fullmetal Chocobo

Moderator<br>Distributed Computing
Moderator
May 13, 2003
13,704
7
81
Well, there is no need. Since current technology works in synchronous transmissions, it all has to be clocked for everything to work together. Once you get into asychronous data transmission, there is no need for a clock, asa it just doesn't matterr at that point. But it wouldn't really do much to increase the core clock speed (the crystal). It is much easier to increase the output, or atleast how it is received.
Tas.

EDIT: I'd be willing to say that you could probably put this conversation in HT (Highly Technical) and get really good answers. As it stands, I'm waiting for someone to from over there to come over here and pwn my ass about all that I've been saying. :)
 

name9902

Member
Feb 12, 2005
48
0
0
haha, i'll throw it up in there i guess, tho i worry about not understanding half the sh!t they say...still don't get why it is that speeds wouldn't be faster with CPUs if you can have multipliers with em...but like i said, i understand not as much as i'd like of this, just looking for an answer that'll satisfy me for now so i can move on lawl

I tend to have a problem with moving on with studying this stuff w/o knowing every single detail about the prior subject...i can't just skip over stuff i have trouble understanding heh
 

Fullmetal Chocobo

Moderator<br>Distributed Computing
Moderator
May 13, 2003
13,704
7
81
Oh, why don't JUST the processors have the multiplier, instead of having it on the mobo? Well, your memory, north bridge, south bridge, and everything inbetween all have to talk at the same frequency... The core speed (crystal), along with the multipliers and other things sets the speed of which EVERYTHING works together. That is why you have to watch your AGP bus while overclocking, because you can set it too fast, and the video card just can't keep up.
Tas.
 

name9902

Member
Feb 12, 2005
48
0
0
ye i got that, i'm just asking like...theoretically, why can't there be a multiplier on the CPU + multipliers on the Mobo busses, increasing the multipliers on mobo busses, so that "theoretically" there's no roof on CPU speed if you just put multipliers everywhere....

*same disclaimer applies as above* =P
 

Fullmetal Chocobo

Moderator<br>Distributed Computing
Moderator
May 13, 2003
13,704
7
81
Clock speeds aren't the limiting factor. And that is basically what the multiplier does (for the processor). They have taken chips to 4+ GHz now with overclocking. The limiting factor is primarily heat. As you rev up clock speeds, the processor produces more heat. They have reached amazing processor speeds by cooling with liquid nitrogen and the like...
Tas.
 

name9902

Member
Feb 12, 2005
48
0
0
mmm i see, so even if some kind of internal multiplier existed...the heat would be way too high for the CPU to not be engulfed in flames...i got it...

I still wonder if it would be possible to have an internal multiplier of the clock speed on a CPU tho or are ya saying that that already exists for the CPU allowing them to make it to 4+GHz?
 

Fullmetal Chocobo

Moderator<br>Distributed Computing
Moderator
May 13, 2003
13,704
7
81
Well, I haven't OCed in so many years (my Xeon mobo doesn't do it at all), but aren't there like 3 things you set? The core speed (66MHz, 100MHz, 133Mhz last time I checked), the multiplier, and then the voltage? Something like that? Well, you can crank them up to whatever you want, but your HSF have to be able to compensate for the extra wattage that is generated.
Tas.
 

name9902

Member
Feb 12, 2005
48
0
0
ye i got it, there's a multiplier on CPUs, but the limit is based on heat...makes me think that the ceiling on CPU speed gets fixed by fixing other setttings to make them faster and more efficient
 

Fullmetal Chocobo

Moderator<br>Distributed Computing
Moderator
May 13, 2003
13,704
7
81
It (as we are referring to here) is a small silvery box on your mobo (there are most likely several) that pulse at a specific frequency, and as such, determine the base of clock speeds for your computer. If you have a quartz watch, it operates on the same principal (IE every 6,000 pulses, add one second kind of thing. [random number picked...]).
Tas.
 

name9902

Member
Feb 12, 2005
48
0
0
well, i'm only focusing on the system crystal that sets the beat for the CPU...i think there's only one...?

the others i'll worry bout later
 

Fullmetal Chocobo

Moderator<br>Distributed Computing
Moderator
May 13, 2003
13,704
7
81
Originally posted by: name9902
well, i'm only focusing on the system crystal that sets the beat for the CPU...i think there's only one...?

the others i'll worry bout later

I'm not quite sure. Just depends upon the mobo manufacturer really. I've never had the luck to see wiring schematics for a motherboard, although I would absolutely love to.
Tas.