[vcf-midatlantic] examples of non-crystal based clocks in digital computing?

Dave Wade dave.g4ugm at gmail.com
Sun Mar 13 06:30:37 EDT 2016


> -----Original Message-----
> From: Ben Greenfield [mailto:ben at cogs.com]
> Sent: 13 March 2016 09:33
> To: vcf-midatlantic <vcf-midatlantic at lists.vintagecomputerfederation.org>
> Subject: examples of non-crystal based clocks in digital computing?
> 
> Hey All,
> 
> I have one simple specific question, without a simple answer and general call
> for examples.
> 
> I’m researching timing clocks used in general purpose digital computing. I’m
> starting to think that Ace/Pilot Ace by Alan Turing is the start of clock based
> general purpose computers. I’m starting to conclude the earlier computers
> were all basically clocked by the media holding the instructions.

It depends on how you look at it. The Pilot Ace used mercury delay lines. The speed of propagation of waves in mercury is fixed. 
So for a given length of mercury, and a given number of bits to be stored, there is a fixed bit rate that’s required. 

So I would argue that for any machine using delay line store the clock cycle is fixed by the mechanical properties of the delay line and the word length.

The Manchester SSEM (and I assume the Manchester MK1 as it was derived from the SSEM) has a master clock that can have its frequency changed by a knob. In fact the clock speed was tweaked several times early in the development cycle. So machines that use Storage Tubes for displays can be built as true variable speed machines. 

Even the Pegasus which derives its clock from the drum has a cheat. It uses the tracks on the drum in pairs so the data can be clocked off at twice the bit rate of the drum. 

The real break through came with Computers like the Atlas (and I assume Stretch) where you start to get asynchronous processing and the various parts of the machine are no longer all clocked together...

> 
> Is this correct?
> 
> I’m currently looking for examples of general purpose computers that do not
> use crystal based clocks.
> 
> Any thoughts appreciated.
> 
> Ben
> 

Dave





More information about the vcf-midatlantic mailing list