Another "New TRS-80 Model I with issues" thread

[Eudimorphodon]

In a rare fit of impulse buying I BIN'ed an untested storage-locker-find Model I keyboard unit (only) the other day, which just arrived a few hours ago. (My motivation was somewhere between thinking it's be nice to have a parts unit for my beloved other TRS-80 and thinking it might also be nice to have a test mule to use for... building things.) I have zero stars for the seller when it comes to packaging, he charged $40 for $13 worth of postage and the worst packing job I've seen for quite a while, but miraculously the unit arrived not looking any worse than it did in the listing photos. (Severely dusty but no major plastic cracks or anything.) I probably should have opened it up and looked for anything obviously bulging or scorched, but I was terminally impatient so I connected it to the monitor and PSU for my other Model I and powered it up.

Good news: It powered and spat out a "Ready" prompt. (Which actually seemed vaguely wrong in a way I couldn't put my finger on yet.) I tried typing a few lines of BASIC but it had the worst case of keybounce I've ever seen, which I figured was probably par for the course. When a mistyped line produced "What?" instead of a syntax error I realized that one of the mysterious switches on the back of the unit:



Is a switch to go between Level I and Level II BASIC. (The machine came with a buffered cable still attached to it and it has the DIN RAS/CAS/MUX modification so I know it was attached to an expansion interface.)

Anyway, bad news: the screen display wasn't particularly stable, being slightly jittery. I started messing with the switches to see what they did and (I think coincidentally) it started getting very bad, flickering side to side and eventually rolling vertically. (For the record the red pushbutton toggles between the two BASICs, I think the white toggle is for a non-Tandy lowercase modification, and the blue switch is either momentary or broken and resets the machine when poked. Wondering if they merely extended the reset button or if there's more going on, will find out when I open it.) After messing with the vertical hold on the monitor I was able to mostly stabilize it in that orientation, but the horizontal wavering continued... until it didn't. Eventually it stabilized and looked almost fine. Which it stays for a minute or two before going all wavy. Wash/rinse/repeat.

Here's a couple shots of the output. "Stable":


And unstable: (Actually taken a couple minutes before the "stable" shot.)


(I apologize for the not great contrast, I had some backlighting and the monitor wasn't adjusted so well. And, well, it's the original Radio Shack monitor, set expectations accordingly.)

To rule out issues with the monitor I pulled out my other keyboard unit to test, and it looks like it's fine. Wondering if anyone has a good top-of-mind idea what could cause this sort of cyclic instability?

(There may be other issues internally; at one point when I was typing I was getting what appeared to be a vibration-related issue with the character generator or another aspect of the video output chain. I didn't see that again after messing with the switches so I'm wondering if it may have been related to poor contact on the lowercase mod switch. Obviously job one is cleaning the whole thing up.)

 

[Eudimorphodon]

Minor update: I've started cleaning the system up externally but I haven't opened it yet. After the first external cleaning I decided to see if I could better characterize the symptoms by firing it up and letting it run for a while. Here's about the size of it:

On the initial cold power-up I had some "interesting" behavior that I think may have been due to having the Level I/Level II switch in the Level I position at the same time the lowercase mode was enabled. Moving them Level II/No Lowercase and hitting reset let the machine boot. (Also I was apparently wrong above, the red pushbutton is the lowercase toggle, white toggle is BASIC version. Still haven't found anything the blue toggle does besides reset the machine.) Here's the relevant part:

1: After the initial powerup I had a pretty stable display. I won't say "rock solid", there is a small degree of both horizontal and vertical shake, but it's readable and the letters weren't noticeably distorted.

2: Shortly after powerup the display started horizontally "weaving" like in the picture in the previous post. It started off gradually but within a minute or two it was very difficult to read the display. (I was typing in a simple character set display program through the noise.)

3: Eventually (maybe within four or five minutes of initial power-on) the weaving subsided, fairly suddenly, returning the display to its initial state of being "mostly clean" (IE, vertical lines straight, not tilted/weaving) but still a little shaky compared to my "good" TRS-80.

4: The really bad behavior seems to be resolved once it's fully warmed up; if I turn it off and only leave it a minute before turning it back on it doesn't go through the wavy song and dance again. I'll have to experiment to see how "cold" it needs to get to show symptoms.

This is starting to feel like it could just be an old capacitor problem, although that wavy behavior seems very specific and might mean there's something else sketchy in the video section; I plead ignorance as to what that might be that would be so affected by temperature. If anyone has a suggestion where to start that would be awesome.

 

[Eudimorphodon]

FWIW/if it matters, the specific model number of this unit is a "26-1006D"; if the references I can find are correct the "1006" indicates it was sold as a 16K Level II unit, I don't know the significance of the D. Serial number is in the 55,000s. This serial number is higher than my other unit (44,000-something), which claims to be a 26-1006-no-D. My other unit has a lowercase mod with descenders, this one appears to still have the orginal MCM6674 with the weird graphics characters in the first 32 positions instead of another copy of the uppercase letters and lowercase letters shifted up.

 

[Eudimorphodon]

... And one more update: Letting the machine cool down to "room temperature", as determined by a hand on the rear vent area, is enough for the shakes to resume after power up. (Figure I had it off for... 20 minutes? within 10 feet of a ceiling AC vent.) Figure 8-10 minutes before the shakes subside "mostly completely".

 

[ldkraemer]

It appears to me that the Model 1 is working correctly, you just need the 8046673 Character Generator installed.

I've attached the "FOR" text in the program showing what a 8046673 would display.


Larry

Address 0x06 = "F"
Address 0x0F = "O"
Address 0x12= "R"

 

[Eudimorphodon]

I’m not worried about the character generator, I just thought it was interesting because the two previous TRS-80s I’ve owned both had the one with descenders so I’d never seen the original character set in the flesh.

I guess I’ll see if I can get a video of this warmup issue/instability, it’s very hard to get a still photograph of that’s not just a blurry mess. (And once it’s warmed up a fast shutter can make it look like it’s perfectly stable when it’s still jittering to the human eye.)

 

[Eudimorphodon]

No updates on chasing down my new Model I's warmup problem, real life has been too busy. I did dig up my stash of 5-pin DIN plugs so I can see about soldering up an AV cable to use a monitor that's a little less... idiosyncratic, than my original TRS-80 monitor for testing.

Sort of as an aside, after I BIN'ed the keyboard unit off eBay I noticed the same seller had the matching Expansion Interface listed separately. Their asking price was fairly dear (twice as much as the computer) and my goal for this machine was either as parts or an experimental mule so I passed on it. Over the next few days the seller gradually hacked down their BIN price; when it hit $99 I went ahead and bit. So now I have the expansion box, just arrived today. I'll be honest the main reason I went for it is I'd blithely assumed that it'd be pretty trivial to make a PSU brick for one of these things if it turned out it was good enough to want to keep whole, but, man, unless you have a box of different value AC transformers lying around that has ones you can use in it you're going to spend about $60 just in parts to make a new one from scratch. The Expansion Interface has both PSUs in it, and it also came with a printer cable, so, yeah, I guess it was worth getting the band back together. Big disappointments: no double density board or serial card. The serial number on the EI is in the 6000s, which is much lower than the one for my other TRS-80. It has the DIN/buffered cable mod.

The expansion unit looks like it's physically okay but it's extremely dirty, and of course the edge connectors look completely dingy. Does anyone have a favorite method for cleaning one of these up without hurting the notoriously cr***y tin plating? Are DeoxIT or other contact cleaners a good place to start? Back in the day I remember doing the pencil eraser thing, but I kind of feel like we're beyond that.

Has anyone a recipe in their pocket for getting a Flashfloppy-modified Gotek running on a single-density Model I? (disk image preparation, wiring gotchyas, etc.) I will grant that thinking about that now is probably jumping the gun.

 

[Eudimorphodon]

Been making slower progress on this than I hoped…

Yesterday I cleaned the exterior of the keyboard unit, including pulling off all the keys and deep scrubbing them, because I was frankly worried about catching anthrax or hauntavirus from touching the system in the state it was in. It actually cleaned up surprisingly well, under all that grime is a TRS-80 that’s got some scratches but a lot less paint rubbed off than the other unit. Keyboard only has cracks in about a quarter of the keystems instead of basically all of them, another good thing. Did not have time to open it up.

Today I cleaned and disassembled the Expansion Interface. Also looked a lot better clean. Kind of an oddity is it only has 16k, the second row of sockets is empty. In the first row there are seven ancient ceramic RAM chips and one newer plastic one, kind of surprising the kind of owner that would hack the keyboard as much as they did and possibly replaced a bad RAM chip didn’t max it out. The PCB has the solder mask flaking off in spots near the PSU section. I’ve seen this on older PCBs before, hopefully it’s not worth worrying about.

Anyway, point of this post: The two big filter capacitors in the EI look pretty sketchy, especially the 10,000uf 16v one. (Crud/white scale around the contact post at one end. No massive leak but a definite “fuzz”.) I’m going to guess the equivalent parts in the keyboard are in similar shape, could this cause the warmup/shaking issue with the video by any chance? I’m going to replace the ones in the EI before I smoke test it just in case.

 

[Eudimorphodon]

Since I have for now at least fixed my video instability issues that were actually voltage regulator problems I decided to start digging into if I could get the expansion interface to work. (Also the can of contact cleaner I ordered finally arrived; I don't know if there are global shortages of this stuff or what, but the hardware store didn't have any both times I looked, despite the inventory on the website claiming otherwise, and my Amazon order for it was delayed like three weeks.)

This morning I soaked all the edge connectors in contact cleaner and scrubbed them with swabs before letting them thoroughly dry out. Then, a couple hours later, before hooking the EI to the keyboard unit I took the bottom off it and and checked voltages at the empty 4116 sockets (strangely this EI only has 16k installed), and when those checked out I fired it all up.

First time up it didn't register that it was connected at all; no boot-screen garbage and PRINT MEM was still 15572. Then I remembered I hadn't done anything with the DIN MUX connector and, yeah, it was terrible. Actually rusty inside. I took the EI back outside and hit that with contact cleaner and a brush. And while that was drying out I popped the lid off the buffered cable (It had already been pried apart at some previous juncture, one of the four plastic pegs was broken off) and examined it to figure out how it was powered so I could check voltages in it and, presumably, work my way up to clamping the logic probe and/or oscilloscope onto it to see if signals were at least getting that far. But I guess I didn't need to bother, after trying it again with the DIN plug cleaned up:



(And I had to hold break to get that far so maaaaybe the disk controller is working, that's the next thing to poke at.) Tried loading a cassette game that supposedly requires 32k or more and it was fine so, yeah, guess it's working.)

Dumb question I think I've worked the answer out to: Even if DRAM chips are not installed in the EI sockets the decoding circuitry is still going to assert a buffer on reads from the memory space associated with those sockets, so if, say, I wanted to slap together a simple RAM expansion using a SRAM chip I'd have to do something to disable the EI memory is I still wanted to use the EI's disk controller and whatnot, correct?

Dumb question the second: I found a PDF of the manual for these older EIs (serial number 6751) but it doesn't have annotations covering what was changed for the buffered cable/MUX mods. Anyone have links to docs about that?