console speedup mod

[marc.hull]

You will almost certainly have to use a 4MHz-rated version of the 9901 when you swap out the 9900, and as CC Clarke and Atrax27407 already mentioned, timing is going to be an issue with a lot of software. Geneve users had similar issues when running stuff made for the /4A at full Geneve speed. Sometimes things worked well, sometimes sprites did crazy things, and sometimes a program wouldn't work well at all. Timing will cause issues. So long as you have the capability to switch between standard speed, 4MHz speed, and overclocked 4MHz speed, you should be able to keep things running fine. PEB cards don't have issues with the higher speeds--as Geneve owners (and 99/8 owners with Armadillo Interfaces) can attest--although the RS-232 card will do strange things because the BAUD rate tables won't be right. I may have to read out the DSR that I have for the 99/8-modified RS-232 card that I have and compare it to the standard DSR, as it is using a DSR designed for the /8 (the card is otherwise identical to a standard RS-232 card).

I thought the Geneve had the same clock as the 99-4a. Don't have a 99-8 but doesn't it also have a 12 mhz clock ? I think all three machines share the same 3 mhz state cycle am I off base ?

[+mizapf]

PEB cards don't have issues with the higher speeds--as Geneve owners (and 99/8 owners with Armadillo Interfaces) can attest

 

but only because the TMS9995 offers a CLKOUT from the input rate divided by 4, and the 99/8 users get even less speed (because of the 10.7 MHz clock at the 9995).

[MueThor]

Dear Ksarul,

 

Thank you for your thoroughly outlined insight.

Sometimes things worked well, sometimes sprites did crazy things, and sometimes a program wouldn't work well at all.

Has the problem with the sprites sometimes occuring probably got something to do with the VDP or with the communication between CPU and VDP? Would it be possible to resolve this problem, by modifying the communication between VDP and CPU respectively adding a special controller to manage their communication? Or would this problem only have been resolved, if there had existed a VDP for higher working frequency?

 

BTW, what is irritating me very much is following: Many TI ICs are sold by companies in China and I have the suspicion that there is a lot of product counterfeiting in the pie ...

 

 

Regards

Edited December 11, 2016 by MueThor

[CC Clarke]

BTW, what is irritating me very much is following: Many TI ICs are sold by companies in China and I have the suspicion that there is a lot of product counterfeiting in the pie ...

 

Your concerns are certainly justified. The Chinese tend to imitate, rather than innovate (thus far. . . )

 

While TI has a Chinese presence brought on through the acquisition of chip-makers struggling in China and elsewhere --at bargain prices, to be nearer their intended markets without high import tariffs that cut profit, the bulk of their high-technology chip fabrication is based in the US. US export laws are intended to minimize high-technology transfers that can be used in cutting edge weapons research and development. This of course doesn't hinder their desire to gain proprietary advanced materials and production techniques via espionage. The reported theft of F-35 design data from Lockheed is one such example.

[+OLD CS1]

Yeah. I just bought 20 SN76489s and five SN76494s. I have concerns.

[+InsaneMultitasker]

... the RS-232 card will do strange things because the BAUD rate tables won't be right.

The same holds true for most terminal emulators and other programs that directly access the RS232 serial port. They usually have their own tables or settings that do not rely upon the DSR; sadly, a modified EPROM will do nothing to fix those problems.

[atrax27407]

For epromming (and other purposes), I consider EVERY chip that I get from a Chinese source as counterfeit. Some work quite well, others do not. You just have to be selective once you find a "good" chip source from China.

[MueThor]

Hi folks,

 

As CC Clarke had written in an earlier post for experimenting with hardware socketing is recommended. In this context I have now many questions. Which technical data has to be fulfilled by a socket to be soldered onto the TI board? There exist different DIP sockets for IC with different specifications for pitch, row width, contact resistance, contact material and contact plating. I will have to buy DIP IC sockets with rectangular pitches, but with which row width (0.3" or 0.6")? Is a maximum resistance for the contact resistance of 20mΩ okay? Are copper alloy as contact material and tin for the contact plating enough? Please, can anyone of you guys answer my questions, also with regard to very long lasting and stable socket solutions?

 

 

Regards

Edited March 7, 2017 by MueThor

[Swim]

First impressions and review of my TI-99\4a version 40

 

Stuck home from work due to health issues so I got bored and picked this project out of the pile on the work bench. I have been experimenting over the years with speeding up the stock 4a console. The result of those experiments created the console that has been reliably running the Hidden Reef BBS for the past few decades. That particular console has the 32k on the 16 bit buss and runs overclocked to 3.58Mhz. Back when I was playing around I think I tested out at least three versions of available designs that put the 32k on the 16 bit buss until finally settling on the design from System Ninety Nine User Group of Germany. The SNUG design required a daughter board which was produced by the group but was the most reliable of the designs I tried and added a couple of features handy for programmers and could be setup to toggle on and off so the console could still be used at stock speed for gamers. Back at that time I had also experimented with running at 4.0Mhz but found that 4.0 broke too much software and wouldn't boot with certain 3rd party hardware in the P-Box.

 

All that said, I had chose a 4a motherboard with the NEC -3 VDP memory which is 150 nano second instead of some of the other 4a MB's with 200 nano second or slower VDP memory as the base for this project. If you are lucky you might find a 4a MB with NEC -5 VDP chips which are 100 nano seconds. First thing I did was remover the stock clock crystal and install a new 4.0Mhz clock crystal. Sure enough, it did boot but was very flaky and showed dropped characters and wrong colors as has happened in the past when I tried a 4.0Mhz crystal in a stock MB. I next removed the stock TMS9901 and installed a TMS9901-40 and also began replacing the 74LSXXX logic in the memory management circuit with 74HCTXXX logic and replaced the stock electrolytic capacitors with new capacitors. At this point, I rigged up the MB and booted it to make sure of my work to this point. It booted nicely and was helped by the TMS9901-40 being installed. Did some things like run the size command while in X-Basic, access DSK1 just to see what would work and what wouldn't. I then removed the TMS9900 from the MB and installed a TMS9900-40. Booted up the MB to check my work and it did boot and ran as expected. In the photos you can see I installed a RAM disk in the P-Box and formatted it and loaded software to it. I also made a log-in on the Hidden Reef using TELCO at 4800 baud and to my surprise the screen didn't show any dropped characters and I didn't experience any other problems I was expecting from past experience. It was late and I had lost energy so I left the system running the EKMEL Fish Tank software and went to bed. Now 14 hours later the system is still running nicely without having crashed. Tonight I hope to finish changing the remaining 74LSXXX logic with 74HCTXXX logic and also perform the SNUG 32k on the 16 bit buss mod to bring this MB up to running at 16 bit speed at 4.0Mhz. If I'm successful I will install this MB into the Hidden Reef system and make it the first S&T BBS running at 16 bit, 4.0Mhz speed.

 

[atrax27407]

Were you using the RS232 code that I sent to you to do your BBS tests?

[Swim]

Were you using the RS232 code that I sent to you to do your BBS tests?

 

The RS232 Rom is the one you sent. But I only made the one log-in with TELCO so I wouldn't really call it a test at this point. Time will tell.

[atrax27407]

Looks like the clock speed detect works. Hopefully the baud rate tables are correct as well. SKYPE if you can.

Edited May 16, 2019 by atrax27407