halkoD@batman.moravian.EDU (David Halko) (02/04/91)
OK... I tried the switch... I swapped crystals on the CoCo3 with 512K to get a boost in speed... no problem, it worked most of the time <smile>... I did it a week ago over my break, so I don't remember what I was running it at, but I know I was getting a significant speed increase according to my calculations. I figure, replace my 120ns DRAMS with 80ns drams, swap a 48MHz clock crystal for what is in there to get a full 3MHz, and the machine would operate ALMOST reliably, right? (I will probably have to find faster 256x1Bit DRAMS from somewhere to make it reliable... at 3MHz, I believe I will need 60ns DRAMS, but I am sure I will be able to get by with 80ns...) Now, would there be an easy way to SYNC up the CM-8 with the high speed CoCo3? I took apart my VM2 and my CM-8 with little success... The CM-8 has a clock crystal on the motherboard of the monitor somewhere under a metal shielding case, but I can't get at it! It is secured somehow without screws or clips! The VM2 has no crystal in it at all, but it STILL wouldn't sync up with the CoCo3 after playing with horizontal and vertical hold... I figured if I exactly doubled the speed of the CoCo3, I could adjust the scan rate of the CM-8 by changing that crystal appropriately (but I can't get at it right now since I do not want to cut the shielding away RIGHT NOW with a small hack saw...) If I do change the crystal in the CM-8, would the two sync up properly??? Could I just adjust the H and V holds and play with the adjustments on the guns on the monitor??? I don't really mind if I will blow up the 6809 (I'll replace it with the CMOS equivalent) or the GIME (I'll by some stock in Tandy! <smile>) or the PIA's (I'll have a reason to socket them...) Just want some raw speed. I am sure the Eliminator will handle the speed up without a problem... any help at this point would be appreciated before I try stepping the clock rate higher (my wishing point is 5MHz... that would bring me somewhere around the speed range of an MM/1 or low line 68K Tomcat... I figured at 3-5 MHz, the CoCo would turn RiBBS into a quick BBS and give me the ability to run it on 4 or so lines without bogging down the 6809 too much... would like to run a little UUCP as well... maybe test a serial networking interface that has been in the works... the speed is a must...) When this project is complete, I will add the 1 Meg upgrade to the CoCo and I'll be ready to fly. Any help would be appreciated... Thankyou. David J. Halko
kdarling@hobbes.ncsu.edu (Kevin Darling) (02/11/91)
halkoD@batman.moravian.EDU (David Halko) writes: >OK... I tried the switch... I swapped crystals on the CoCo3 with 512K to get >a boost in speed... no problem, it worked most of the time <smile>... I did it > [...] >Now, would there be an easy way to SYNC up the CM-8 with the high speed CoCo3? > [...] >I don't really mind if I will blow up the 6809 (I'll replace it with the I think you're far more likely to blow the CM-8 with a higher horz freq! Better get/borrow a multisync monitor. >speed range of an MM/1 or low line 68K Tomcat... I figured at 3-5 MHz, the >CoCo would turn RiBBS into a quick BBS and give me the ability to run it >on 4 or so lines without bogging down the 6809 too much... would like to My thought is: it'd be a better idea to grab some old, cheap CoCo's and create EPROM paks for them which contain code to do much of the handling of the serial login ports... and then have them send data to/from the main CC3 via say, the old keyboard parallel port or ?? That should work much better than attempting to overspeed the coco board. Plus it might make a neat commercial product ;-) kevin <kdarling@catt.ncsu.edu>
zambotti@wpowz.enet.dec.com (Walter Zambotti DEC) (02/13/91)
>OK... I tried the switch... I swapped crystals on the CoCo3 with 512K to get >a boost in speed... no problem, it worked most of the time <smile>... I did it > [...] >Now, would there be an easy way to SYNC up the CM-8 with the high speed CoCo3? Ok it sounds like you changed the master crystal in the CoCo. This would speed up the entire machine including the video output. Instead of hacking your monitors to match the speed of the faster CoCo why not make the video output section of the CoCo match the original speed of the monitor. i.e. cut the trace that supplies the clock to the video circuitry and "glue" in the old crytsal (plus or minus any dividers) from the CoCo on to the cut trace/pin. This way you have fast processing and original video signals. regards, Walter
Curtis.Boyle@weyr.FIDONET.ORG (Curtis Boyle) (03/10/91)
And I was simply looking for a 32 MHz crystal to get the whole thing running at a true 2 MHz (instead of 1.78). I know Bruce Isted had his running with a 34 or 36 MHz crystal, and he had to use a Multi-sync monitor to get the video right, but barely. (By that I mean it ALMOST worked by fooling with the horizontal and vertical syncs on the CM-8/Magnavox). You wouldn't happen to know a supplier with 32 MHz crystals, would you? I haven't been able to find anyone around here that stocks them. -- Curtis Boyle - via FidoNet node 1:140/22 UUCP: ...!herald!weyr!Curtis.Boyle Domain: Curtis.Boyle@weyr.FIDONET.ORG Standard Disclaimers Apply...
zambotti@wpowz.enet.dec.com (Walter Zambotti DEC) (03/18/91)
Hi Dave, read your followup. Understand what your asking! > If I hack that trace, pop in the old crystal with a /4 circuit, will the > video still sync up to the CoCo3 bus??? I figured the coco and the GIME > might not like to chat anymore because of bus timing problems.... From this I gather you want to hack the clock trace to the GIME because you feel that this is where the video clock frequency would best be altered? I would not have done it this/that way. The 6809 and the GIME should both have the same clock chip. And I am suggesting that the video clock frequency be hacked at some point after it leaves the GIME chip. I would support your reckoning that if the GIME and the CoCo had mismatching frequencies then there would be contention on the bus. If it is not possible to hack the clock frequency after it leaves the GIME because clock frequencies are not relevant at that section of the video circuitary then you may have to alter the actual video frequency (as apposed to the clock frequency) at some point in the final stages of the video output circuitary. > If that is all that is required, then I could pop in a few chips into a > cheapo piece of perf board with the old crystal and get the whole thing > working. Could I use TTL or CMOS chips with that clock crystal for a > divider circuit? Can't say! Only someone who attempts the mod would be able to shed some light on that. And at the moment Dave it sounds like your one of the only few (if not one) who is attempting this at all. My only input on this is to : Hack the video circuitary on the CoCo sometime after it leaves the GIME, whether by clock frequency shifting or video frequency shifting. Keep the relative clock frequencies between the GIME and the 6809 the same, i.e. if you double one then double the other. Be brave. Dave it is you (and not I) at the fore front of this exciting modification. You may have to experiment until you get what you want. > In article <1991Feb13.044240.16381@tkou02.enet.dec.com>, zambotti@wpowz.enet.dec.com (Walter Zambotti DEC) writes: > > >OK... I tried the switch... I swapped crystals on the CoCo3 with 512K to get > > >a boost in speed... no problem, it worked most of the time <smile>... I did it > > > [...] Sorry but I never wrote any of this. I only replied to the above. I did write this though : > Instead of hacking your monitors to match the speed of the faster CoCo > why not make the video output section of the CoCo match the original > speed of the monitor. i.e. cut the trace that supplies the clock to the > video circuitry and "glue" in the old crytsal (plus or minus any > dividers) from the CoCo on to the cut trace/pin. regards, Walter