1980 -81 Tandy 'TRS80 Model II' Strewth!

[Wozzer1]

I remember carrying out a service call to Staffordshire Ambulance headquarters - I was to fit extra memory to the Boss Man's Compaq.
From memory, It was a square "daughter board" which plugged in - Set the DIP switches and away you go.
I recall it was 4MB and cost just over £35 per MB - That was back in the 80's.
I started out on CPM, with 8" floppies, and taught myself DOS when the company gave me an IBM with a single 5.25 floppy & no hard drive to play with - Oh yeah, and a daisy wheel printer!

[majex45]

The first computer I programmed (at school) was an Olivetti Programmable calculator, It could work to 22 digits if you combined two 11 digit numbers. The command used was "A split V" if I remember correctly. We had great fun(?) programming it to calculate such as a+ax+ax2+ax3 (squared, cubed etc) and calculate prime numbers to 22 digits. These were often left running overnight as they took so long. Prime numers started very quickly but soon slowed down to one a mnute and even per hour. It used magnetic cards to store the program. This all from a somewhat old and rusty memory.
Somewhat later I used a Varian Data Machines mini computer (19" by about 5 Units high) to control a microwave dish.

[duncanlowe]

Pretty sure our first school computer was a TRS80 II. Prior to that there was a 'staff only' terminal via an acoustic coupler, terminal for the OU mainframe in MK. There was a loan PET between those, from the local Polytechnic, courtesy of my Dad. That meant me and a mate of mine were the only pupils allowed near the new TRS80. And I think we broke it. Were working on a simple game program Poking to the video RAM. And calculated the number of screen locations wrong, so POKEd off the end of the video RAM. I've never proved it, but when we did this it seems we POKEd into ROM and broke it. Certainly it stopped working straight after. The local Tandy did replace it though.

[SiriusHardware]

There are several machines which have a legendary 'killer poke', where, through user action at the keyboard and due to bad hardware design, it is possible to cause two or more elements of the hardware to fight against each other on the buses.

As an example, the design of the address decoding for the ROMs may not limit actions in the address range occupied by the ROMs to read-only, as it should: It may enable the ROMs when the micro tries to write to that address range as well, in which case the ROMs and the micro will both attempt to place data on the databus when the micro goes to write to an address in that range: This would be a potentially damaging situation, with the data outputs from one IC attempting to overpower the data outputs of another.

I wasn't aware of the TSR80-II being one of the machines with a potential 'killer poke', but I don't know the Tandy machines at all well.

[Radio Wrangler]

The same law which requires volumes to be reported in olympic swimming pools, lengths in London busses, and areas in football pitches requires computing power to be reported in Lunar Landers.,

David

[ortek_service]

Originally Posted by SiriusHardware

There are several machines which have a legendary 'killer poke', where, through user action at the keyboard and due to bad hardware design, it is possible to cause two or more elements of the hardware to fight against each other on the buses.

As an example, the design of the address decoding for the ROMs may not limit actions in the address range occupied by the ROMs to read-only, as it should: It may enable the ROMs when the micro tries to write to that address range as well, in which case the ROMs and the micro will both attempt to place data on the databus when the micro goes to write to an address in that range: This would be a potentially damaging situation, with the data outputs from one IC attempting to overpower the data outputs of another.

I wasn't aware of the TSR80-II being one of the machines with a potential 'killer poke', but I don't know the Tandy machines at all well.

I would expect many computers back then had little h/w protection against you trying to write to ROM Memory-space. The Beeb had Sideways-ROM and later RAM, so the area was also writeable - although hopefully the nOE of these sockets would be connected to nRD / nR/W. But maybe some some simply enabled the ROM inc. output whenever that space was being addressed, irrespective of read / write status.
However, I wouldn't expect that to be cause instant death of IC's, as they will normally have quite limited current sourcing on the data lines, even with buffers on some of the better designs, so limiting current into a sinking set-low output.
There was also the problem that the CPU's didn't have illegal instruction traps, and could end-up trying to execute data if there was a s/w bug etc.
So some of the 'illegal' (undocumented officially, as they didn't work) instructions, like the Z80's legendary Halt and Catch Fire' could end-up getting run.

But I thought it was POKE's to video h/w, that might try and make a monitor run at wrong frequency, may be more-likely to cause damage.

[Mark1960]

I think halt and catch fire was 6502, puts the 6502 in a state that requires reset, possibly even power off.

[ortek_service]

Originally Posted by Mark1960

I think halt and catch fire was 6502, puts the 6502 in a state that requires reset, possibly even power off.

Well it seems true HCF's might not have existed, and they just halted.

And 6800, 6502 & Z80 etc were all affected under certain conditions
- Although with the Z80, an NMI would escape out of the endless-loop condition: https://en.wikipedia.org/wiki/Halt_a...ire_(computing)

[TonyDuell]

There is often some confusion about the early TRS-80s.

The model 1 was a computer-in-the-keyboard with a separate video monitor and cassette recorder. The ROMs were either 4K 'level 1 BASIC' which was very cut-down (no maths functions, 2 string variables but no string functions, 1 array variable, no PEEK/POKE, etc) or 12K 'level 2 BASIC' which was normal Microsoft BASIC. These ROMs were always in the memory map, starting at location 0. Because of the ROM, video RAM and memory-mapped I/O, this machine was limited to 48K RAM

The model 2 was a unit containing the mointor, computer electronics and an 8" floppy drive. This had just a bootstrap ROM which was disabled after the boot so you could have 64K RAM.

The confusion is that a number of people seem to think a Model 1 with level 2 BASIC is called a Model 2. It is not!

From a quick look at the service manual, on the Model 1, there are buffers between the ROM data lines and the rest of the machine which are only enabled on reading. The video timing is a hardwired chain of TTL chips so prograamming can't affect that.

The Model 2 does use a 6845 chip for the video timing. But on that machine the video RAM does not normally appear in the CPU memory map, so it seems unlike you'd be directly poking it from BASIC.

My guess is you had a Model 1 Level 2 machine and were just unlucky.

[Phil__G]

I did a complex RTTY and morse-code transceive program for the Model1 which had its own split-screen video routines - there is no problem writing to video ram - its no different to how the TRS80 does it itself As Tony says, sounds like an unlucky coincidence.

[TonyDuell]

There is no problem writing to video RAM on the Model or Model 3 (or on Model 4 in Model 3 mode) machines. It's just a 1K block in the CPU address space starting at 0x3C00 I seem to remember.

But on a Model 2/12/16 or on a Model 4 in true Model 4 mode, the video RAM does not normally appear the CPU address space. You have to fiddle with some I/O ports to disable part of the main RAM and replace it with the video RAM. This is not something you can do from BASIC, it needs a machine code program and even then it's tricky (better to use the OS routines....)

Fortunately there are excellent technical manuals for all TRS-80 machines if you want to play such tricks.

[G6fylneil]

I've still got a TRS80 model 1 in the loft with its B&W monitor. I bought it from a friend who bought it new. I've got a "MS typing tutor" (on a cassette) for it. It was impressive at the time.