915 terminal

[Jimhearne]

Hi all,

Has anybody got anything on the TI 915 terminal, i can't find any mention of it at all on the net.

Seems to be more like a TI 770 Intelligent terminal but with a different part number.

It seems to have all the LEDs for upgrade to a DS990/1

 

Jim

 

[jbdigriz]

I have several scans of 770, 771, 990/1, and 915 maintenance standards, FCO's, training materials, etc. that  I haven't put up yet. And some items that have to wait untill I get a large format scanner for foldouts. This is stuff Guntis sent me  way back when. The ones that are done  need some post-processing, trying to get to that, but here's the one for the 915. Far as I can tell it's basically the the same as the 990/1, but I haven't really delved into it.  Maybe different firmware?

2268696-9701_Maintenance_Standard_For_915_Remote_Terminal.pdf

Edited September 6, 2022 by jbdigriz
excuse the chronic editing. Sticking keys, fumble fingers, and network lag.

[Jimhearne]

Thank you very much for scanning that, it's the first mention I've seen of the 915 terminal.

I think it could be upgraded to a 990/1 , almost certainly with different firmware, but i don't think mine has all the cards for running the floppy drives etc.

And the odds of finding the extra cards wouldn't be very high.

 

I'd be interested in anything else you get a chance to scan in.

 

Many thanks, 

 

Jim

[jbdigriz]

Ah, ok, I see now. Missing the file processor board and the disk controller, which would be in slots 2(red) and 3(orange), respectively. And there's only one RAM board (slot 5), probably the low 32K. For a DS990/1,  the boot loader ROM would be either on a hi 32K board  in slot 7 or a 64K board in slot 5, depending on model and whether it was an upgraded 770 or 771. So, looks like the 915 is strictly a terminal with no local storage. Assuming all the boards are there, which seems fairly safe.

 

From front to rear, I see:

 

VDU controller

Low 32K, or maybe 64K, hard to tell, but it will likely be marked.

CPU board

Async board #1

Async board #2 Can't remember which one is for the serial  data line and which is for a printer.

 

I will post some pics shortly of the boards in a 990/1. The rest of the machine is in bad enough shape that I'll save that for another day.

[jbdigriz]

Sorry, TI called what I referred to as a "File Processor" the "I/O Processor". BTW, that's a strange looking chip on your VDU controller. Nothing like it here.

[Jimhearne]

I will take some pictures of the boards.

I think it is just the terminal version.

 

Jim

 

[+FarmerPotato]

I have two floppy controllers but no IO interface. You’re welcome to one of them. doubt I will ever assemble  a full DS990/1 either. 
 

my photos:

 

 

[Jimhearne]

Here are dumps of the 2 roms from the 915 terminal memory board (aka 770 terminal i think).

 

Interestingly the board is labelled as a 770 memory 2263555-0006, even gives a part number for the schematic, 2263557

 

915 rom Memory U37 2267683-1.BIN915 rom Memory U41 2267683-2.BIN

 

What was initially strange was that there are 8 roms sockets, they are connected in 2 banks of 4 with common address lines and separate data lines for the 2 banks.

But, in each bank, all 4 of the sockets have all 24 pins connected together, including the 2 chip select lines.

After scratching my head for a while, i realised they must be using the programable options in the TMS2732 for selecting the active state on the 2 chip select pins.

So the 2 Roms in my system have both chip select 1 and 2 active low.  

But for the other 3 pairs of roms they must have the roms programmed with the other 3 combinations of the 2 chip select lines.

 

I just used a simple adaptor to swap pins 18, 20 & 21 around to convert the TMS4732 to a TMS2732A pinout to read them in the minipro. 

This of course only works for these roms which need both chip selects low.

 

Jim

 

 

 

[Jimhearne]

I knew the keyboard was bent but didn't realise quite how much.

The key switches are fitted into a metal frame which is what is bent and the PCB is following the shape.

 

The switches are 4 contact Hall effect switches so at least there shouldn't be any dirty contacts, lets just hope they all still work.

I think the only way i'm going to safely straighten the metal frame is to unsolder all the switches and remove the PCB from the frame.

That will give the desoldering gun a good workout.

I will try a couple and see how easily they desolder.

 

Jim

 

[Jimhearne]

Well, i think i got away with it, only 2 of the hall effect pcb came out of the switches, that's not a major problem, you can just put them back in with a dab of hot melt.

 

Now to get all the switches out of the framework.

I did find some damage to the tracks around one of the LED's , look like somebody had tried to remove it from the bottom of the board only and ripped a track off the top side.

 

With hindsight I do wish I'd tested the keyboard before taking it apart, if any of the TTL is duff it would mean taking the keyboard apart again to test them.

 

Jim

 

[Jimhearne]

Well , that's a pain, at least 50% of the switches on this keyboard are duff.

I had been reading on the net that these switches don't age very well. 

These are hall effect switches, they have 4 pins, power (5v), ground and 2 open collector outputs.

There are different output configurations.

On all the main switches the 2 outputs should both give a low pulse (about 50us) when the switch is pressed.

The ones that don't work are a variety of not taking any current, or only 1 output working , or the output not pulling fully down to 0v

I had the outputs connected to a scope and in some cases you could see the output pulse fail (loose amplitude) over a dozen or so presses of the button.

 

I guess i will have to start looking around the keyboard forums on the net and see if i can find replacement switches though i'd be worried about more failing in the future.

That or make a replacement keyboard with normal switches (and a suitable interface) and hope i can find some switches that will take these keycaps.

 

Jim

 

 

 

[Jimhearne]

It's strange, the keyboard has a date of 1977 on it even though the system has a date of 1981 inside.

The keyboard is also serial number 000019 , that's very low. 

I wonder if it's been swapped out at some point.

 

Jim

 

[jbdigriz]

Haven't dismantled the keyboard  in my DS1 yet, (Apparently that was the common short name for the 990/1. At least I've seen it several time in the docs I have.) but I know I will be, so your experience with the 915 will be helpful here as well. Interested to know what the failure mechanism(s) for the switches are, and to what extent they're rebuildable.

[Jimhearne]

The switches use a little Hall effect IC on a ceramic PCB

It's just the bare chip under a little blob of sealant.

It seems from what i've read so far that the silicone isn't very well sealed and over the years it can absorb moisture and degrade, i guess it may well depend on how it was stored.

Judging from the rust on the rear of mine and the dust inside i think it was probably in a loft or garage for quite a while.

 

The switches are Micro Switch (that's the make, not helpful for searches !), mostly part number 1001sd1a3b

 

Jim

 

Edited September 29, 2022 by Jimhearne

[jbdigriz]

That is most helpful. Presumably the moisture leads to corrosion of the solder then. Perhaps the IC's are still good. Worth checking unless cheap replacement switches are available. Pot in epoxy rather than silicone, maybe. Bears investigation anyway.

Edited September 29, 2022 by jbdigriz

[jbdigriz]

I'm thinking that maybe some caution will be warranted using the I/O processor you're getting from Erik until we can pin down exactly what machine it's for. 2270080 is not a part number I'm finding anywhere in this list of switch and jumper setting for 770, 771, DS1, & 915 terminals (also variously called by TI "Distributed Processing Systems" and "Small Business Systems") You'll at least want to trace and compare the pinouts, I'm sure. I don't see that PN in any superseding documents I have, either. It may be from a later period, or possibly earlier, say '77-'78. Still looking. 

Alert_Memo_43-79-0071.pdf

[jbdigriz]

Hmm.  I think I will rescan that bi-tonal, should clean up better. Was in better shape when I got it long ago, but that can't be changed now. Figured it would be useful though when swapping boards etc.

Edited September 30, 2022 by jbdigriz

[Jimhearne]

On 9/30/2022 at 12:07 AM, jbdigriz said:

That is most helpful. Presumably the moisture leads to corrosion of the solder then. Perhaps the IC's are still good. Worth checking unless cheap replacement switches are available. Pot in epoxy rather than silicone, maybe. Bears investigation anyway.

Sorry, i assumed the forum would send me alerts that somebody had replied since i started this thread but it didn't.

I suppose it wouldn't hurt to try reflowing the faulty switch boards, not sure if the silicone would be soldered down (can you solder to silicone) or if it would be some sort of conductive glue.

 

I've had no joy finding replacement switches so far, though I've not resorted to asking on the keyboard forums yet.

There is one of these keyboards on Ebay in the US but they want silly money for it given it has keycaps missing.

https://www.ebay.com/itm/265757155891

With shipping and Duty that would be more than i paid for the whole terminal.

 

At the moment i have some bought some SC70 packaged hall effect sensors, if they work i will combine them with a PIC micro (to generate the 2 pulse outputs from the single on/off output from the hall sensor) on a replacement thin PCB to fit in the switch housing. 

Visually they would look the same.

 

Jim

 

 

[Jimhearne]

On 9/30/2022 at 4:49 PM, jbdigriz said:

I'm thinking that maybe some caution will be warranted using the I/O processor you're getting from Erik until we can pin down exactly what machine it's for. 2270080 is not a part number I'm finding anywhere in this list of switch and jumper setting for 770, 771, DS1, & 915 terminals (also variously called by TI "Distributed Processing Systems" and "Small Business Systems") You'll at least want to trace and compare the pinouts, I'm sure. I don't see that PN in any superseding documents I have, either. It may be from a later period, or possibly earlier, say '77-'78. Still looking. 

Alert_Memo_43-79-0071.pdf 7.31 MB · 2 downloads

Thank you, that does look useful.   

Getting the keyboard repaired is the first task.

 

Jim

 

[jbdigriz]

5 hours ago, Jimhearne said:

Sorry, i assumed the forum would send me alerts that somebody had replied since i started this thread but it didn't.

I suppose it wouldn't hurt to try reflowing the faulty switch boards, not sure if the silicone would be soldered down (can you solder to silicone) or if it would be some sort of conductive glue.

 

I've had no joy finding replacement switches so far, though I've not resorted to asking on the keyboard forums yet.

There is one of these keyboards on Ebay in the US but they want silly money for it given it has keycaps missing.

https://www.ebay.com/itm/265757155891

With shipping and Duty that would be more than i paid for the whole terminal.

 

At the moment i have some bought some SC70 packaged hall effect sensors, if they work i will combine them with a PIC micro (to generate the 2 pulse outputs from the single on/off output from the hall sensor) on a replacement thin PCB to fit in the switch housing. 

Visually they would look the same.

 

Jim

 

 

Unfortunately the only machine likely to have these particular switches would be another SBS or maybe something like a Silent 700 model 733. The keyboard hounds have already got to all those they can find. I couldn't justify robbing either for switches unless they were total junk otherwise, couldn't afford what the keyboard ghouls have driven the market up to, and don't want to see any more perfectly good terminals parted from their bespoke keyboards and then scrapped, as seems to happen way too often.

 

Cool on your plan, please do let us know how it works out. I've seen some repro switches that MAY work, have to do some more checking.

[Jimhearne]

I have come across these switches before, about 20 years ago i acquired a keyboard that used them but it had the version of the switches with the constant outputs, both outputs went low as long as the button was pressed.

TI only use those for the CTRL and SHIFT keys.

For the main switches TI use a different switch where the 2 outputs only pulse low and then go high again each time the switch is pressed.

Hence the need for something (the 8 pin PIC micro) to look at the single output from the Hall effect sensor and generate the 2 pulse outputs.

The existing switch is open collector but i can simulate that on the PIC by switching the pins between a low output and being an input.

 

If this plan doesn't work out, the next plan is to try and find a normal contact switch that can fit in the same steel frame (maybe with a 3D printed adaptor) and take the TI keycaps (again maybe with an adaptor) and make a complete new PCB with whatever circuitry needed to generate the same interface to the rest of the terminal.

Tonight's task is to hook up one of these hall effect sensors and see if the magnet in the switch will operate it.

 

Jim

 

[Jimhearne]

Well, the Hall effect sensor i bought worked, but too well, it has to be almost outside the switch to stop it triggering all the time.

Luckily there seem to be less sensitive versions so i've ordered a couple of other versions for tomorrow.

It would be nice to put the Hall chip in the same place on the PCB.

The Hall chip is a SOT553 package, 1.7mm  x 1.25mm , that's going to be fun to hand solder.

The PIC chip is a huge 3mm x 2mm compared with that.

 

Jim

[jbdigriz]

Not as much fun as iPhone or Android motherboard repair, probably. Is that potting material rubbery or hard like cured resin? Contractor's solvent might loosen it up either way. Might be difficult to explain "doll" repair solvent, which is otherwise cheap for a small bottle on Ebay. 🙂 I think that stuff is there for mechanical support as much as anything else. Don't see how it would be conductive, and I think I see solder pads under the IC. 

Edited October 6, 2022 by jbdigriz

[jbdigriz]

Hmm. Have to be careful there. What is PCB material but reinforce, cured resin? Note to self: Have to test, and only apply to the potting.

[Jimhearne]

The switch PCB material is ceramic so pretty much nothing is going to damage that.

The material over the silicone chip is soft, a bit like silicone sealant.

Some Acetone softens it more and it came off.

Here are a couple of closeups of afterwards.

If the Silicone is soldered down there isn't a lot of it.