CompuColor II 17.9712 MHz crystal oscillator (Y1)

[Andrew Lynch]

Hi

I am reading the CompuColor II maintenance manual at the design of the logic board and most of the components appear to be commonly available or at least similar alternatives.

http://oldcomputers.net/Compucolor-II-Maintenance-Manual.pdf

I have not been able to find a source for the 17.9712 MHz crystal oscillator (Y1) in the lower right-hand corner of the logic board image (attached).

Y1 appears to be a regular HC-18 (possibly HC-49) style crystal oscillator.  

I have not been able to find a source or equivalent alternative to this specific frequency crystal oscillator in any of the usual places.  

I think this frequency has a specific purpose as both the source for the CPU clock and also the UART to create 9600 baud serial data rates.  

As far as I can tell, this is the only frequency oscillator used in the logic board, so it is critical to the design.  

I think it must have been common sometime in the past but not anymore.

Does anyone know of a source for this 17.9712 MHz crystal oscillator?

Any help greatly appreciated.  Thanks in advance




Merovingian: Choice is an illusion created between those with power and those without.

[Mark T]

You might try one of the programmable oscillator modules if the 8224 can accept an external clock input instead of a crystal.

https://www.digikey.ca/en/products/detail/ecs-inc/ECS-P85-B/502310

[T Gerbic]

It is close to 5X the frequency of a standard color crystal. 

Maybe an 18Mhz crystal with caps to pull it down slightly.

A dividing counter attached to a very low drift crystal should also work. 

[Mark T]

I wonder if a colour crystal could be persuaded to run at 5x overtone.

[T Gerbic]

That is a big stretch. Depending on where the crystal is mounted and space around it, some type of oscillator, PLL or divider should be possible to get the color frequency needed.  

Perhaps using a small 8-pin SMD PIC processor with a, maybe, a 50Mhz low drift SMD crystal on a small carrier board dedicated to generating the output frequency would be a much better choice.  A possible choice might be the PIC12F1571 which has an NCO module, 8-pins and a 48Mhz clock. Might be able to get it close to the frequency, but might need a non-integer crystal to hit that odd frequency. 

There are some small clock generators like the Si5351A that might work. Adafruit has a module. 

[Mark T]

That was why I sugested the module from digikey, they will program it to your requirements, so you don’t need to spend time and effort learning how to program the module.

[T Gerbic]

Looked it over and seems like a great solution. 

[Phil G]

Another suggestion from VRRRF:  https://www.vintage-radio.net/forum/showthread.php?t=224677

[Andrew Lynch]

Thank you for the responses.  I greatly appreciate your advice

I think the programmable TTL oscillator module from Digikey seems to be the approach I like the best.  Although I wonder why the designers of the CompuColor II used such an unusual frequency as critical component of their design (availability? copy protection?).   Regardless, I am interested in keeping as faithful to the original design as possible, so the programmable TTL oscillator approach is the least impactful. (HC-18 crystal -> half can TTL oscillator)

Unfortunately, I was unable to see the vintage-radio thread due to "Access from United States temporarily blocked due to AI scrapers. Try using a VPN to access."  I suspect the vintage radio people have dealt with odd crystal frequencies quite a bit due to the nature of older radios.  One of the ideas I had earlier was a custom crystal if such a thing is available.

The one thing that still is a lingering doubt to me is whether the Intel 8224 clock generator will accept a TTL oscillator input instead of attaching a crystal.  I recall from other chips like UARTs and CPUs you can substitute a TTL oscillator for a crystal by connecting only the output of the oscillator to the clock input pin of the IC.  I presume it's the same applies for the Intel 8224, but the data sheet (attached) is silent on the substitution (unlike UARTs & CPUs).  

Although the Google AI seems to indicate it will work, I remain skeptical of anything from an AI (obviously the pin assignments are wrong, that's a clue)

quote

AI Overview
Yes, it is possible to drive the Intel 8224 clock generator/driver with an external TTL oscillator instead of a crystal, but it requires careful signal injection rather than a direct, simple swap. The 8224 is designed with an internal oscillator circuit intended for a fundamental or 3rd-overtone crystal. 
Implementation Details for External TTL Input:
Injection Point: Inject the TTL oscillator signal directly into the X1 pin (Pin 16).
Unused Pin: Leave the X2 pin (Pin 17) unconnected.
Signal Level & Characteristics: The signal must be a 5V TTL-compatible square wave, ideally matched to the 8080A clock frequency requirements.
Alternative Approach: The OSC pin (Pin 12) is a buffered TTL output of the internal oscillator. While meant to drive external system timing, it can be used to synchronize other 8224s. 
Free
 +4
Key Considerations:
Frequency: The 8224 will output a CPU clock frequency that is 1/9th of the frequency provided to the X1/X2 input.
Waveform & Duty Cycle: The 8224 provides proper 8080A-compatible clock waveforms (
/ 
 levels and 33.3% duty cycle). Injecting a sharp, 50% duty cycle TTL signal directly may cause issues with the internal clock shaping logic if not managed correctly.
Internal Circuitry: You are bypassing the internal, high-gain inverter that normally drives the crystal.
Alternative IC: If flexibility in choosing the oscillator source is required, a more modern, flexible alternative is the 8284 clock generator, which is often used in newer 8086/8088/8080 designs. 
Reddit
 +3
Disclaimer: Operation outside the recommended crystal-controlled conditions may affect device reliability or performance, particularly regarding phase noise or clock duty cycle constraints. 
Electrical Engineering Stack Exchange
 +1

unquote

Again, I appreciate your help.  Thank you

--
You received this message because you are subscribed to the Google Groups "retro-comp" group.
To unsubscribe from this group and stop receiving emails from it, send an email to retro-comp+...@googlegroups.com.
To view this discussion visit https://groups.google.com/d/msgid/retro-comp/58c09a49-51fb-4dcb-95c3-745ac5560b4dn%40googlegroups.com.

[Mark T]

I wouldn’t trust AS without finding references either. AS scrapers have been causing problems for quite a few sites recently, I wonder if its deliberate to try and promote use of the services of the AS companies.

The 8224 spec says it supports third overtone, I wonder if you could persuate it to work with 5.9904MHz crystals that seem to be available.

Are you working on a replica or trying to repair a vintage computer?

[Andrew Lynch]

Hi

I am currently investigating whether a replica is feasible.  I am going through the maintenance manual and schematic reviewing the parts for the logic board to determine if I can reproduce it.

Most of the discrete components, 74LSxxx and MSI chips seem to be available or there are pin compatible equivalents.  Not a guarantee it will work but reasonable chance at least.

The design uses several small bipolar PROMs (82S123s, 82S129s, and an 82S131) as sort of "proto-PALs".  PALs were just coming out in 1978, so I suspect they weren't available during the design phase.  Instead, they used the bipolar PROMs in a similar fashion.  I believe they can be substituted for 16V8 and 22V10 type GALs (82S131 is special case).

I don't think the bipolar PROMs are practical because they are hard to get plus require special programmers.  GALs are easier to find plus can be programmed in cheap common programmers (TL866II style, etc.).  Trying to replicate with the bipolar PROMs would require a programmer like a DataIO 29B or another high-end unit most people don't have.

The trick is to replicate the CompuColor II though is getting access to the contents of the bipolar PROMs.  I don't see that anyone has dumped their contents that I've been able to find.  Getting the data in the bipolar PROMs will require someone with actual hardware willing to remove them from the logic board and dump their contents.  That will be a tough order to fill because there are few remaining units and I doubt many would be willing to extract the ICs.

That's why I am asking.  There are also some other strange aspects of the CompuColor II like the floppy disk drive and the CRT interface, but I think I have some ideas on how to handle them.

Thanks!

https://groups.google.com/d/msgid/retro-comp/e6f610fc-99ae-4b4c-ace4-006232851439n%40googlegroups.com
.