IBM 5110 initial info

[voidstar78]

Thank you, yes Initial is right ! I was pretty shocked to see this difference just by swapping between the two available Common and Language ROS cards. I'd be sad to get a 5110, but couldn't do anything with it without a functional disk drive (i.e. by not being able to by pass this).

I tried by-passing this in other ways - CMD-ATTN, CMD-HOLD, CMD-+, etc, no luck.

I suppose another possibility is that this other Common and Language ROS card actually is faulty. i.e. the BRANCH past this particular test happens to be corrupted? So I was curious if others had this experience with a 5110 Type 2 of not being able to by-pass the AUX IPL 013 D80 error.

(also FWIW, the two Type-2 5110 I have both only have 8 "tin cans" on their Executive ROS; contrast to the Type-1 5110 that I have which has 9 "tin cans" on its Executive ROS -- that extra tincan might be added tape-deck support, not sure yet -- that said, the "type 1" executive ROS also works in my "type 2" 5110 {and can by pass this 013 D80 error, and they're both from mid-1979}).

 

[stepleton]

And sure, maybe I'm interpreting what it is saying wrong - maybe it is just noting where the jumpers are set to, but doesn't necessarily mean just changing the jumpers makes it interchangeable - not 100% sure. But I'm curious if the Processor card from the 5100 is "forward compatible" into a 5110 (i.e. does it even have these jumpers on its board?)

Thanks for the reference. I am curious now too to know if my 5100's card has those jumpers too. I'm not curious enough to open it up and find out right now. But the next time I need to strip the machine down, I will check. My bet is that they will be present, and here's why:

There are a number of things that could be consistent with the evidence you've found. It's possible that PALM was used in other settings besides the 5100/5110/5120 series, and that those jumpers aren't specifically about 5100/5110 compatibility. (Wikipedia says without any support that "it is likely PALM was also used in other IBM products as an embedded controller.")

As the 5100/5110/5120 series and the preceding SCAMP prototype were in some ways assembled from things in IBM's parts bin, perhaps Wikipedia is correct, and perhaps the jumpers enable compatibility with some sort of hardware feature present in the 5110 (and various other IBM products) and absent in the 5100 (and various other IBM products). Or vice versa.

If you bring your 5110 to London, we can experiment

 

[voidstar78]

Note for reference: MIM page 30 (section 2-12) is what the MIM reference that led me to the 5100/5110 jumpers. Small note below the A1 board diagram that reads as follows:
"Note: See Section 050 in the Map binder for card part numbers and jumpering"



Also in the MIM 4-64 "Machine Check Jumper"

Machine Check Jumper
"This jumper connects pin J2-S07 to J2-S09 on the 5110
A 1 board. Removing this jumper allows the controller to
continue functioning When an error occurs on the
machine check line. Misleading results can be received
by running with the jumper removed. "

Except I'm not exactly sure where J2 is, but there are two jumpers on the backside of the A1 board (and it is near a label that says "J"). Won't try it now (to see if passes the 013 D80 issue with that one ROS), but someday may come back trying this out. As mentioned, the working Language ROS let me by-pass with ATTN then RESTART, with no adjustment to the A1 board.





We had scheduled for a trip to London in mid-2020 -- and well, like many other folks, we had to cancel travel plans all throughout that year. Someday perhaps will try again :D

 

[stepleton]

For negotiating wiring "coordinates" for components like these SLT boards and backplanes, you might find resources in http://www.bitsavers.org/pdf/ibm/logic/ useful. I haven't explored everything and have forgotten what all I've seen from earlier expeditions, but some things that stand out include:

Page 60 of http://www.bitsavers.org/pdf/ibm/lo...edLogicDiagrams_SLT,SLD,ASLT,MST_TO_Oct71.pdf
Layout of SLT cards and pin coordinate nomenclature for 4x4 SLT/MST "cans".

Pages 69-74 of http://www.bitsavers.org/pdf/ibm/logic/2-7047_-_SLT_Card_Design_Rules_1969.pdf
more card layout info

Unfortunately it's too late for me to keep looking through these materials, and I can't find the document I used to identify all the locations on the A1 backplane. I think there's some other useful stuff elsewhere in Bitsavers's IBM archives, too.

On the bright side, I could find the map I made from the information I gathered during an earlier tour. I've attached it to this post.

I can't remember the significance of some of these annotations and cannot vouch for their accuracy. But you might be able to reconstruct what I've forgotten. Note that some of the labels about what cards go where are 5100-specific.

 

[voidstar78]

@stepleton what a treasure of information, thanks Someday I'd like to find the address lines, and exercise the "change startup address" trick you guys did for extracting the executive ROS (just to observe it being done).
In that '69 document, there are so many diagrams, I'm very curious what software they used to prepare all that for printing - long long before WYSIYWG, I suspect a lot of it was "scripted" (something predating LaTeX or something?). Very cool.

 

[voidstar78]

Another "1978 IBM 5110" vs later 5110 difference (at least in the ones I have, I don't know if these apply universally to all "early model vs late model" 5110's) :

(#5) Here is the CRT in the 1978 5110. It has some "sticky residue" on the side where a cover on that cable has unfortunately come off. There was a little card stuck below the CRT, dated Nov 1978 and a little notice about x-ray emission...




versus the CRT in the 1979's IBM 5110's that I have:

Some of the differences are: the outer railing is more "square" than rounded, more copper behind the tube, no label on the side, there is this "large white round thing" in the back, those are the easy main things that stand out to me (obviously lots of other component differences on the circuit board).



And note that both of these have the little speaker behind them (added to the 5110 from the 5100).

Posting just as reference examples of what the CRT in a 5110 "should" look like.

 

[stepleton]

The 1978 edition has the closest resemblance to my 5100's display. Not counting the sticky residue --- I'm not sure what that could be and have never seen anything like it on a CRT before. Where did that come from?

The 1979 edition is new to me, and I wonder if it was made by a different manufacturer than Ball. The IBM-added (I assume) caution sticker suggests that it's not a replacement someone worked in there.

The "large white round thing" is the flyback transformer.

 

[voidstar78]

I don't know the history of the '78 one (with the sticky residue). The round stethoscope-diaphragm gray-rubber looking piece is also on the '79 unit at about the same location (and it appears to have a maroon colored piece below it). So my assumption is some adhesive helped keep the gray and maroon pieces together, but that failed (or was peeled off) in the '78 unit that I have, and then just collected dust over the years to end up looking rather awful. Maybe it was in some weathered storage -- I say that, because there is also some "crust" in the metal enclosure that goes around the main caps of the power supply (like possibly from water build-up, like from humidity). But aside from those two issues (that I think are just cosmetic), the rest of the system is in fair shape - but I haven't powered it on yet.

I'm still thinking some early 5110's used left over 5100 stock-parts. And I just have the impression the 5100/5110 were "hand made" by well paid IBM engineers, not cranked out in a more automated fashion. All just speculation on my part (since they were rare machines for certain professionals, not yet part of the "PC in every home" movement). No idea what the production numbers were (e.g. over 5000/year even?). But I think I recall reading somewhere that both the 5100 and 5110 were offered for sale right on up to the Datamaster/IBM 5150 era (around 1981) -- 5100 models may still have been desired, since the 5110 did have some BASIC differences (not huge differences, but enough to require a little maintenance in some older 5100 BASIC code that maybe some people didn't want to bother with)


Also, I realized the reduced size image didn't get the x-ray emission note very well, so here is the full wording on that.





The '79 one I have photo'd is consistent in both of the other two 5110's that I have, so I do think they are "OEM equipment."

Does anyone know what the white-labeled component below, near center-right (labeled DL-289 903A) might be?

EDIT: And I see there is a "green" sticker on the "far side" of this CRT - I'll see if I can get a better shot (without removing the entire CRT). That might give insight on if it is Bell related or not.

 

[stepleton]

The 51[012]0 series was manufactured in IBM's Rochester, MN plant --- one of the reasons I suppose that Ball may have been an attractive CRT supplier (it seems like it wasn't far away). I suspect that the computers weren't much more or less hand-built than many of IBM's other products at the time, whatever that means. Many of the other details about the computer seem conformant to IBM internal standards.

Based on the word "Ltd." on the green sticker you mentioned, the sticker likely lists the CRT's maker, and maybe the maker of the entire monitor assembly. "Ltd." isn't a corporation type designation in the United States, so I'm guessing this will be some company in a different country.

I think the white-labeled component is a variable inductor used to control the CRT's horizontal linearity.

The "stethoscope" part is the "anode cap", which is not usually attached to a CRT with an adhesive --- instead, there are clip-like wires that come out of the centre of the suction cup and engage a cavity in the outside of the CRT envelope. These wires carry a high voltage, so insulation is useful: that's what all the rubber is for, and sometimes a dielectric grease is also used to maintain a good seal. This grease is not an adhesive, however, and if the material in your 1978 CRT is that stuff, then whoever put it there made an uncommon mess of it!

My main hope is that there hasn't been a failure of some component elsewhere in the system that caused the mess. If something ruptured or smouldered, you could see something like this.

Within the metal enclosure you mention (by which I assume you mean the AC distribution box underneath the CRT), there is (among other things) a mains filter capacitor or set of capacitors, and depending on which version of the AC box your system has, there is some chance that this capacitor is a classic, fairly-large GE Pyranol oil-filled device (or similar from another brand). The 5110 hails from a time when the use of PCBs in things like those capacitors was being phased out owing to health hazards, but they had not been eliminated yet. PCBs yield some additionally undesirable chemicals when they undergo thermal breakdown, so if that residue could be from the filter capacitor having a very bad day, I would be pretty careful around it.

I would be cautious about powering that system up, in any case.

 

[NeXT]

Usually the sticky substance around the anode cap is corona dope, dielectric grease as mentioned or the rubber anode cap has begun to degrade due to the presence of ozone. Other than collect dust and dirt it doesn't affect anything other than being hard to clean. Silicone type anode caps don't suffer from this.

 

[voidstar78]

Within the metal enclosure you mention (by which I assume you mean the AC distribution box underneath the CRT)

Oh, actually I meant behind the CRT - I'll probably lift the CRT out eventually, to check the power input enclosure that you mentioned (directly below the CRT). The enclosure I meant is a little cover for the top half of these tall cap cylinders for the PSU. This one (in the '78 unit that I have) has a little corrosion on it. This, and the "sticky residue" on the CRT are the only visual issues I see (no corrosion elsewhere, the A1 pins and all the boards look fine).
EDIT: and this corrosion doesn't go all the way thru, the PSU and the silver canisters inside all looked ok.

I did power this up once (with the Y1 connector off the board) - nothing happened at first, so I checked the 5A fuse below the CRT. That needed to be replaced (which doesn't bode well to the history of the unit...), and after that the PSU fan at least spun up. But haven't plugged in the Y1 yet, still probing around.

I thought the CRT is powered off 12V coming through the Y1 connector (dc-only), I didn't think anything from the "power box" below the CRT actually connects to the CRT ? But then again, true, every other legacy CRT I've had needed its own AC power plug.

 

[voidstar78]

Since the CRT is working, I don't have much excuse for removing it. I'll need a smaller camera with maybe a fish-eye lens. So regarding the green sticker on the side of this 5110 CRT, all I can make out is "NEC CRT" - is this the same as the modern day NEC ?

 

[voidstar78]

I learned something new about the 5110 today - it was in a small pamphlet in the sleeve of the binders for the manuals:

To swap to lower case, you press HOLD followed by SHIFT+DOWN (or HOLD then SHIFT+UP to swap back to upper case). I haven't spotted this in the MIM, so I was never sure how to do it.

Think the addition of lower case support was new for the 5110 over the 5100. When you are in lower case mode, using SHIFT types upper case letters. (as opposed to when you are in upper case mode, SHIFT types the second label on the key).

This works in Norbert's online 5110 emulator - click HOLD, then you can click on SHIFT and then the UP/DOWN arrow as needed (sequentially).



When I ran some of the tutorial/demo software, I noticed they used lower case letters. I figured they had concatenated that using something like CHR$ (or similar) using raw character codes, but nope, looks like the system supports just typing it out like that.

 

[stepleton]

Yes, that's the NEC we know today, or as it was also known back then, Nippon Electric Company Ltd., which is what the label also says
You can find some older NEC logos on Wikipedia.

You are also correct that the 5100 does not have lowercase letters. So IBM had a little pamphlet just to tell people about typing in lowercase? Fair enough --- I knew the 5110 had the character set, but I didn't know how to get to it either.

 

[voidstar78]

Some highlights from the "BASIC Reference Handbook"






Here is the note about lower-case mode:

 

[voidstar78]

If you were ever curious what those holes are on the bottom of your IBM 5100/5110.... Two of them are purpose drilled to securely mount the machine to a table/desk (perhaps useful on a yacht! or maybe in an earthquake prone region; or at a business/bank to help avoid theft of the device). I found a fold out full-size template used to help placement to drill the holes.

I'm still curious what software was used to create posters like this, perhaps as early as 1975.

 

[stepleton]

Neat poster! The 5100 is heavy enough that I don't spend a whole lot of time looking at its underside.

I have no idea how the poster was made, but there's always a chance that it involved no software or only software in limited ways. The text might have been set electronically or via ordinary typesetting; drawings could have been made by hand and reduced (manual drafting was and remains an art; IBM could surely have afforded the best); Letraset/dry-transfers may have helped with some lines, labels, arrowheads, and so on. Then, once you had all the pieces, you could assemble them together, image the result, and make litho plates or whatever.

Publishing had been around for a long time before they made desktop publishing

 

[voidstar78]

For future reference - to add to this list of "interesting things the IBM 5110 can do", we've confirmed that multiple character sets are available in the standard Display card. The Jumpers on that Display assign which character set is the default. Aside from having different characters, these Jumpers don't seem to impact any other aspect of the system. And the Default set by these Jumpers can be overriden programmatically during runtime within BASIC or APL.

Attached is what the standard symbol set looks like (there are various APL symbols, but then obviously the first two rows is some kind of Alien communication :D jk! will have to think what kind of "art" can be made out of those - but nice to see the Greek letters represented!) - I obtained this with a small assembly program entered in the DCP. However, the key-sequence to change Display symbols doesn't work from the DCP (it is a feature provided by the Language support ROS). Attached below is what that "standard" (in my case the Display jumpers were set to US/EBCDIC) in both WHITE and BLACK background.

ADDR   CODE       PALM ASSEMBLY     COMMENTS
----   ----       -------------     ----------

2000   D501 0200  LWI R5, #$0200  ; Begin of display buffer
2004   8606       LBI R6, #$06    ; High byte of end address
2006   8740       LBI R7, #' '    ; Blank (Hex $40)
2008   7750       MOVB (R5)+, R7
200A   C567       SBSH R5, R6
200C   F005       BRA $2008

200E   D501 0200  LWI R5, #$0200  ; Let REG[R5] = Begin of display buffer 0x0200
2012   8604       LBI R6, #$04    ; High byte of end address  (256 only)
2014   8700       LBI R7, 00      ; start R7 at 0x00
2016   7750       MOVB (R5)+, R7  ; RWS[R5] = value in R7 (and increment R5)
2018   0772       INC R7,R7       ; increment R7
201A   0552       INC R5,R5       ; increment R5 (to skip a column)
201C   C567       SBSH R5, R6     ; SKIP if all bits in R6 are also set in HI(R5)
201E   F009       BRA $2016       ; 8 bytes (4 instructions) plus one extra byte
2020   0000       HALT


NOTE: I based this on Christian Corti initial assembly example that prints a character string.
NOTE: You can enter this program manually by doing CMD-ATTN near startup, do "A 2000" (Alter address $2000), type in the CODE sequence only, press ATTN, then "BR 2000" to run it)

I then put in a BASIC program to issue a similar chart, which I input using the KBD5110 project (that program is in the VUM5110 project, along with the assembly sample) - I had to modify that project to allow injecting of SHIFT number-pad keys.. The set of international (from the US) symbols is available here: (a few more yet to be added, initially I thought it was just SHIFT 0-9, but there are a few additional SHIFT-codes for a total of 15 character sets). This matches the set of symbils listed in the IBM 5110 BASIC User Reference page 83 (page 87 of the PDF).

IBM5110_SLM/CharacterDisplay at main · voidstar78/IBM5110_SLM

IBM 5110 Service Logic Manual Posters. Contribute to voidstar78/IBM5110_SLM development by creating an account on GitHub.

github.com

That set of country specific symbols from the manual is as follows:

 

[voidstar78]

I was reading more about the IBM System/360 and 370... And I noticed something:

The System/360 was a white motif, while the System/370 was a black motif. And that same difference is between the IBM 5100 and IBM 5110 models!



No idea if that was intentional - maybe IBM had identified this as a strategy to quickly be able to tell the difference between models? (either in warehouses or when servicing the systems)

Anyway, just thought it was an interesting trend.

 

[Chuck(G)]

Yabut the shells for the S/360 could be had in a wide range of colors that you wouldn't normally think suitable. But then, this was the 1960s. Yellow, orange, blue...

Never understood the revival of mid-century modern...