Hardware Hackers

[ewbray]

In the PATIUG (Philadelphia Area Texas Instruments Users Group) we had over 125 members when we had one of the largest memberships of the PACS (Philadelphia Area Computer Society) when every third Saturday of the month it would take over the unused classrooms of LaSalle University. We used to keep all of our equipment in a locked footlocker that was kept in a secure closet by the Security Department office and every meeting two guys would fetch it and bring it to our lecture hall classroom in the science building!

The "Society" back then even had a young Bill Gates as one of its speakers, who filled the university's main auditorium with standing room only and the university even piped the video feed into all of the university's other lecture halls.

In the PATIUG, Barry Traver (Publisher of Genial Travler Diskazine) & Michael Riccio (Myarc, Inc. contract programmer) were a couple of our software "guru(s)" and Alan Silverstein (Electrical Engineer for several national corporations) was the leader of our hardware "guru(s)".

 

Alan Silverstein, myself, and others would hold "hardware" sessions after the main meeting and do whatever previously arranged hardware projects {cable modifications, console video enhancements, and etc.} were scheduled.

 

I wrote about two (2) of my favorite hardware projects in the March 1988 issue of MICROpendium and April 1990 issue of MICROpendium. MICROpendium Articles.pdf

 

Looking back on my articles and my experience with their results, I would recommend that the March 1988 hardware hack was a "easier" job to perform and an "easier" job to correct if you ever had a failure of the voltage regulator, capacitors, or diodes; which the April 1990 modification did when I tried to perform my annual connection of my real TI equipment, just to exercise it the other day.

 

Some of the electrical components to our Texas Instruments hardware are rapidly becoming rare [40+ years old] and/or nearly obsolete and replacements are becoming hard to find.

Edited January 14, 2019 by ewbray

[webdeck]

Hi - Mike Riccio here. Wow - you just brought back a ton of memories.

 

I've been lurking on this forum for a couple of weeks, after I got the bug to get a retropie setup working. Spent some time over the holidays with my kids setting it up and then teaching them how to play the old games.

 

I also dug up my Geneve and tried to fire it up, but it gave me a RAM error. A lot of my old TI hardware and software hasn't survived moves and culling over the years - I think all my floppies got thrown out at some point, and I have no idea if the hard drive still works.

[+mizapf]

The Geneve had most of its chips in sockets, so maybe you could try to get replacement RAMs. If all else fails, try the Geneve on MAME.

[webdeck]

My primary goal is to get all of the data off the hard drive, so I'll need to get the Geneve resurrected to do that.

[Opry99er]

We have folks here who I bet would gladly do a hard drive recovery for you. Plenty of Geneve users too. (I am not one of them)

[+InsaneMultitasker]

The CPU RAM error may also occur if the regulator above the LED has failed, since it alone supplies power to all 16 DRAM chips. If you are careful (i.e., you don't rock the card and don't short something while testing) you can check for 5 volts at the capacitors of the top-most DRAM chips. If the chips are socketed, try pressing/reseating them carefully.

 

For reference, there are a fair number of Geneves without socketed DRAM chips. Those Geneves can be more stable than the socketed cards, thanks to cheap single-wipe sockets and poor contact between memory and socket.

[+9640News]

Hi - Mike Riccio here. Wow - you just brought back a ton of memories.

 

I've been lurking on this forum for a couple of weeks, after I got the bug to get a retropie setup working. Spent some time over the holidays with my kids setting it up and then teaching them how to play the old games.

 

I also dug up my Geneve and tried to fire it up, but it gave me a RAM error. A lot of my old TI hardware and software hasn't survived moves and culling over the years - I think all my floppies got thrown out at some point, and I have no idea if the hard drive still works.

 

Hey Mike. Glad to see you here. The source code to your program My-Term is out there both in the original code you had as well as I have updated your code and now have ANSI support in the program along with it running at 38.4K. The source code to GEME is also out there which I do not recall if I got directly from you or if it came via Lou back in the day.

 

If you want the latest code, etc, just let me know.

 

I also have a working Geneve with HFDC I would be more than willing to assist in helping you get stuff off if needed. Just let me know.

 

Beery

[ewbray]

My primary goal is to get all of the data off the hard drive, so I'll need to get the Geneve resurrected to do that.

 

Mike why don't you just get an old TI99/4a and get the old files off the hard drive that way. You can convert them to V9T9 format with TiDir99 https://1drv.ms/u/s!AjWNjAZpt3oLgYlPsOd_AXbpSajZww

[webdeck]

The CPU RAM error may also occur if the regulator above the LED has failed, since it alone supplies power to all 16 DRAM chips. If you are careful (i.e., you don't rock the card and don't short something while testing) you can check for 5 volts at the capacitors of the top-most DRAM chips. If the chips are socketed, try pressing/reseating them carefully.

 

For reference, there are a fair number of Geneves without socketed DRAM chips. Those Geneves can be more stable than the socketed cards, thanks to cheap single-wipe sockets and poor contact between memory and socket.

 

 

My Geneve does not have socketed RAM. My PEB has a modified (PC) power supply that provides regulated 5V, and all the voltage regulators on my expansion cards are jumpered to bypass them. That was done way in the past. I checked the jumpers and they are fine, so I guess I've got a problem with the soldered RAM chips - that's beyond my ability to fix at this point.

[webdeck]

 

Mike why don't you just get an old TI99/4a and get the old files off the hard drive that way. You can convert them to V9T9 format with TiDir99 https://1drv.ms/u/s!AjWNjAZpt3oLgYlPsOd_AXbpSajZww

That's my next thing to try. My eBay purchase should be arriving tomorrow.

[webdeck]

 

Hey Mike. Glad to see you here. The source code to your program My-Term is out there both in the original code you had as well as I have updated your code and now have ANSI support in the program along with it running at 38.4K. The source code to GEME is also out there which I do not recall if I got directly from you or if it came via Lou back in the day.

 

If you want the latest code, etc, just let me know.

 

I also have a working Geneve with HFDC I would be more than willing to assist in helping you get stuff off if needed. Just let me know.

 

Beery

Thanks - I may need to take you up on that kind offer.

[webdeck]

Here's a video of the boot error. It keeps looping even though I'm not pressing any keys.

 

Edited January 18, 2019 by webdeck

[+9640News]

Mike

 

Can we assume your system has no additional hardware in it since it last ran ok? Just want to make sure there are no other devices that could respond to a memory check.

That message is shown when the eprom code scans pages >00 to >3F. As I recall, it does a write and compare. Cards that could respond at page >3A or >3C without fully decoded memory with a GenMOD in place would have issues. With no GenMOD, this was not an issue.

 

I have some eprom code, and it may be possible to skip the memory check to allow an attempted boot. If that was then done, then running a program such as Ron Walter's MEMTEST program would identify if you have a single chip suggesting a bad ram chip, or maybe one of the other chips that does the mapping if multiple chips like >02, >12, >22, >32 are showing errors.

 

Of course, any memory that had the OS loaded into it, or would load a program into it, would be potentially flakey with an unknown response.

 

If you haven't already done so, the other chips (74LS138's, etc) you might want to pull them and reseat them if they are not soldered in place in the event it is not truly bad ram, but bad addressing.

 

Do you have an eprom burner? If so, I could work up an eprom that jumps over the memory test. I can also burn you an eprom if you do not.

 

Beery

[webdeck]

Mike

 

Can we assume your system has no additional hardware in it since it last ran ok? Just want to make sure there are no other devices that could respond to a memory check.

That message is shown when the eprom code scans pages >00 to >3F. As I recall, it does a write and compare. Cards that could respond at page >3A or >3C without fully decoded memory with a GenMOD in place would have issues. With no GenMOD, this was not an issue.

 

I have some eprom code, and it may be possible to skip the memory check to allow an attempted boot. If that was then done, then running a program such as Ron Walter's MEMTEST program would identify if you have a single chip suggesting a bad ram chip, or maybe one of the other chips that does the mapping if multiple chips like >02, >12, >22, >32 are showing errors.

 

Of course, any memory that had the OS loaded into it, or would load a program into it, would be potentially flakey with an unknown response.

 

If you haven't already done so, the other chips (74LS138's, etc) you might want to pull them and reseat them if they are not soldered in place in the event it is not truly bad ram, but bad addressing.

 

Do you have an eprom burner? If so, I could work up an eprom that jumps over the memory test. I can also burn you an eprom if you do not.

 

Beery

 

 

Hi Beery,

 

Nothing was plugged in other than the HFDC. I don't have a GenMOD, but I do have the memory upgrade. Here are pics of the front and back of the board: https://imgur.com/gallery/Z5WWy9E

 

I don't have an eprom burner. Thanks for your help!

 

-Mike

[+9640News]

Looks like not much was socketed on your board.

 

I have assembled an eprom image I hope to burn and test tomorrow that skips the CPU memory testing. Assuming it boots on my Geneve, I will then create a floppy disk with Ron's MEMTEST along with MDOS and LOAD/SYS programs and would mail you the eprom and disk. After that, that is all I can do.

 

Assuming you boot MDOS and can load MEMTEST, you would then need to report those results on here and hopefully someone can guide you on what chip(s) to replace. I do not know if you have the skills to desolder or not, but there may be others that can help with that process.

 

Beery

[+InsaneMultitasker]

 

 

My Geneve does not have socketed RAM. My PEB has a modified (PC) power supply that provides regulated 5V, and all the voltage regulators on my expansion cards are jumpered to bypass them. That was done way in the past. I checked the jumpers and they are fine, so I guess I've got a problem with the soldered RAM chips - that's beyond my ability to fix at this point.

If the jumper is fine, then it is most likely (1) a bad solder joint on one or more DRAM chips or (2) one or more bad DRAM or (3) your DRAM is not getting power. If you have a multimeter, check continuity across the jumpered regulator pins and then check continuity between one of the regulator pins and the 5v pin of one DRAM chip. ( It isn't clear to me how you checked the jumpers e.g., tugged on them, visual inspection, continuity. )

[webdeck]

Looks like not much was socketed on your board.

 

I have assembled an eprom image I hope to burn and test tomorrow that skips the CPU memory testing. Assuming it boots on my Geneve, I will then create a floppy disk with Ron's MEMTEST along with MDOS and LOAD/SYS programs and would mail you the eprom and disk. After that, that is all I can do.

 

Assuming you boot MDOS and can load MEMTEST, you would then need to report those results on here and hopefully someone can guide you on what chip(s) to replace. I do not know if you have the skills to desolder or not, but there may be others that can help with that process.

 

Beery

 

 

Hi Berry,

 

That is so incredibly kind of you - thank you! The floppy drive I have is a 3.5" one.

 

I used a multimeter to check continuity across the voltage regulars, and I checked the 5v to all the memory chips. I also checked continuity between the top stacked memory chip and the socket pins on the underside of the board, and they all had connectivity (except for the one leg that has a jumper), so that socketed memory seems okay.

 

I worked on "Plan B" today, which was a TI console from eBay. The video cable and HDMI adapters I was using for the Geneve wasn't giving me anything on the TI. Upon checking, I realized the adapter wasn't connecting the RCA jacks to the correct pins - I have no idea why it gave me any video output at all on the Geneve. So I fixed the wiring, and the TI booted fine. I connected it to the PEB (with a jumpered flex cable card as well as a jumper in the black box at the console end of the cable.) Now, how to see if I can get anything off the hard drive when I don't have any floppy disks with software. The HFDC manual says I can do CALL DIR("WDS1") from basic, but that just gave me errors. I tried doing CALL MDM to load the disk manager, but that just hung. I tried writing a basic program to read the directory, but that gave I/O error 02 when trying to OPEN #1:"WDS1",INPUT,RELATIVE,INTERNAL

 

So, I'm out of ideas for Plan B. It could be that the HFDC card or the hard drive don't work any more either...

 

One other data point, so I took my fixed video cable that works with the TI and used it with the Geneve. Now I get a black screen only, and a constant irritating sound on the audio. No swan and no error message. Very odd. The jumper next to the video jack is set to GND, so it should be the same composite output as the TI, right?

[+9640News]

Well, the 3.5" drive is throwing a bit of a curve into the mix. Gotta figure out how I am going to address that as I was not immediately setup to handle that drive size. Looks like I will also need to add MDM5 to the diskette as well along with GPL and an E/A module.

 

With your TI setup configuration, did you have 32K memory as well?

 

As far as the jumper next to the video jack, I have no idea. Not 100% sure mine has one and will not know its status until I open up the PEBox.

 

Beery

[webdeck]

Good point on the 32K - that's probably needed. I have a 32K expansion card - I just need to add jumpers to it and I can try adding that.

[BuckoBrand]

I like the TI. I said this cuz it's hardware, right?

 

 

 

 

 

 

 

Edited January 20, 2019 by BuckoBrand

[+9640News]

Mike,

 

I've got the boot eprom updated and tested on my system. I am going to upload it here to this group so someone else can have the binary image should they find themselves in a simular situation as yourself.

 

I'm still looking for a 3.5" drive I can connect to a system.

 

For those wanting to use the binary, one must clip the TIFILES header or the first >80 bytes from this file.

 

Beery

Geneve98NoRamCheck.BIN

[+InsaneMultitasker]

 

 

The HFDC manual says I can do CALL DIR("WDS1") from basic, but that just gave me errors. I tried doing CALL MDM to load the disk manager, but that just hung. I tried writing a basic program to read the directory, but that gave I/O error 02 when trying to OPEN #1:"WDS1",INPUT,RELATIVE,INTERNAL

Add a period after the drive number. Without it, the device name won't parse properly.

 

Does your hard drive have the terminating resistors or jumper packs in place, provided this is a single hard drive, and is the 20 pin cable connected to the right connector?

 

If the 34-pin edge connector cable is not keyed, eyeball the connector cable mating to see if the pins line up. Unfortunately, many HFDC edge connectors are cut a little too narrow, meaning there is about a half-pin worth of play. This can result in poor or no connection between the cable and connector and undue frustration when re-seating cables.

 

The 20 and 34 pin cables are easy to connect upside down. Be sure your pin 1 on the card matches pin 1 on the drive.

 

Finally, there are a few of us still around with experience recovering data from hard drives. If you think there are 'valuable' files on the drive but you have no luck getting things working, don't use MDM5 to format the drive as you will lose the data.

[webdeck]

My humble apologies to ewbray for hijacking his thread and turning it into "get Mike's Geneve working..."

 

Today was productive on "Plan B". I got the 32K card added, reseated the hard drive cables, and now I can load MDM and access the hard drive. There's definitely some corruption as I can't catalog some directories, but I can read many of the text files in MDM.

 

Now for the next question - how do I transfer the files off of the hard drive and onto a modern computer? I have a TE2 module, and I managed to get it talking to my Mac, but when I try to transfer a file (at a zippy 300 baud), it doesn't like WDS1 as a device name. I'm assuming I need a better file transfer program, but that also begs the question of how to get that program onto the TI.

 

Thanks again!

-Mike

[Opry99er]

I can send you a disk with a good transfer program if you'd like. Not sure which one would be best for your purposes, but if it's just a path name issue, couldn't you save your files to floppy and transfer from DSK1?

 

Also, I'm not sure Terminal Emulator II allows for XMODEM transfer protocol, which is likely what you will need.

[+9640News]

Mike,

 

Shoot me your address and I can send you the eprom so you can see whether you can get anywhere with the Geneve.

 

I do not have any way to get something to a 3.5" drive as I am pretty much 5.25".

 

Do you have an RS232 card? If you do, and someone can get you a copy of Telco, then you can transfer files over to your Mac. Another option is to get a Lotharek drive and connect to your HFDC so you can copy files to DSK images and transfer via a SD card. That is probably the simplest route.

 

Beery