Kudos to TI...

[Jeff_Birt]

I recently found a TI-92 on fleaBay complete with box and manuals for a good price. It was an educational package not a retail which was interesting. I will say that the rumors of how terrible the TI-92 display is, are true. I have heard the TI-92Plus are much better.

This prompted me to get out my 20-year-old TI Voyage 200, the last of the TI-89/92 lineage, which I bought new. Put batteries in and it sprang to life. The formulas and notes I had put in were still there (no doubt in flash) and it 'just worked'.

Then a column of the LCD went out in front of my eyes. Then a second one, but it came back. This is a common failure of TIs as the conductive adhesive strips used on the LCD FFCs comes unstuck over time. Looking at pictures it seems the FFC for the columns folds UNDER the LCD while the one for the rows comes out the right side. This means you have to peal up the row FFC to try and fix the missing columns. I think for one missing column I'll leave well enough alone for new. If I can find a junk V200 with a bum LCD I'll experiment on it.

On the positive side I searched for the TI Connect software and found TI still hosts it on their website, but it was last updated for Win8. I downloaded it and installed it without incident on Win10 and proceeded to back up my V200.

So, kudos (sarcastically) to TI for making the LCD nearly impossible to repair.

Also, kudos (genuine) to TI for still having the PC interconnect software available for a 20-year-old calculator that 'just works' on modern operating systems.

[Jeff_Birt]

I just found the TI BASIC program editor I wrote in VB.NET 1.0. Looks like I updated it last with Visual Studio 2013. It built and ran fine in VS 2019. Now to find all the programs I wrote for it back then.

[blackjetrock]

(11-26-2023 09:42 PM)Jeff_Birt Wrote:  Then a column of the LCD went out in front of my eyes. Then a second one, but it came back. This is a common failure of TIs as the conductive adhesive strips used on the LCD FFCs comes unstuck over time. Looking at pictures it seems the FFC for the columns folds UNDER the LCD while the one for the rows comes out the right side. This means you have to peal up the row FFC to try and fix the missing columns. I think for one missing column I'll leave well enough alone for new. If I can find a junk V200 with a bum LCD I'll experiment on it.

As an experiment I had some new heat seal connectors made for the Sharp PC-G850VS series of pocket computers. They were not cheap, but I had a small number made. I had fairly good success applying them, but a couple of columns were still bad. i think I may eventually be able to take the old connectors off and put the new ones on. Alignment is a major problem.
So you may be able to fix things like this (with the application of money) using new parts. I think once you peel these connectors off then they are done for, they won't go back on. Applying heat to fix the connections does work, I've found, but only a few times, then that doesn't work either. I also suspect that all heat seal connectors will eventually fall off, which isn't good.

Andrew

[Jeff_Birt]

(11-27-2023 08:29 AM)blackjetrock Wrote:  As an experiment I had some new heat seal connectors made for the Sharp PC-G850VS series of pocket computers. They were not cheap, but I had a small number made. I had fairly good success applying them, but a couple of columns were still bad.

I suspect that if an LCD is run for an extended period of time with missing segments it may be permanently damaged. I had new zebra strips made for the TRS-80 Model 100 LCDs and while this fixes many problems I have still run into the issue of a missing row. Since these LCDs are symmetrical you can rotate them 180 and the same row, now flipped top to bottom is still missing. This may also be from some damage to the conductive coating on the glass. I don't have the equipment to try and get a good view of the conductive pads on the glass.

I have had good success resealing FFCs that have not moved. A few I have had to try a few tries at it. As you say alignment is a bear. I have noticed that several sellers on Aliexpress are now selling ends of rolls of the Anisotropic Conductive tape. Perhaps only a meter or so per roll and supposedly they are not yet pat their expiration date. I will order some to try when I get a good repair candidate I don't mind sacrificing to the gods of learning and knowledge.

In the TRS-80 PC-2 service manual they show an interesting technique for reattaching the FFC on the cassette/printer interface. (Providing you can detach it without damage.) They show offsetting it ~1mm from the edge of the pad. This gives an exposed bit of copper to touch your iron to solder the FFC to the PCB.

[Jeff_Birt]

I uploaded 'TIEdit' to GitHub. Looking at the project this morning I found it had been updated to .NET Framework 2.0 at some point. I was new to the idea of object-oriented programming at the time and chose VB .NET as I had recently started at the university and needed to learn it since they were switching over from VB6 for student use.

I can remember being excited when I found a way to do function pointers in VB which made it a much easier task.

You should be able to build this with the free version of Visual Studio. At some point I'll try to add an installer project and add a proper release to GitHub.

[blackjetrock]

(11-27-2023 01:32 PM)Jeff_Birt Wrote:  I suspect that if an LCD is run for an extended period of time with missing segments it may be permanently damaged. I had new zebra strips made for the TRS-80 Model 100 LCDs and while this fixes many problems I have still run into the issue of a missing row. Since these LCDs are symmetrical you can rotate them 180 and the same row, now flipped top to bottom is still missing. This may also be from some damage to the conductive coating on the glass. I don't have the equipment to try and get a good view of the conductive pads on the glass.

I don't think a disconnected LCD row would suffer from anything, if you apply DC it's not happy, but applying nothing? More likely LCD coating, as you say.

Quote:I have had good success resealing FFCs that have not moved. A few I have had to try a few tries at it. As you say alignment is a bear. I have noticed that several sellers on Aliexpress are now selling ends of rolls of the Anisotropic Conductive tape. Perhaps only a meter or so per roll and supposedly they are not yet pat their expiration date. I will order some to try when I get a good repair candidate I don't mind sacrificing to the gods of learning and knowledge.

If you mean 3M 9703, then yes, you can get bits of it. I have tried it and it does work, but the cells are too big for the PC-G850 connector pitch, so it didn't work for me.

Quote:In the TRS-80 PC-2 service manual they show an interesting technique for reattaching the FFC on the cassette/printer interface. (Providing you can detach it without damage.) They show offsetting it ~1mm from the edge of the pad. This gives an exposed bit of copper to touch your iron to solder the FFC to the PCB.

Can you solder to heat seal cables? The ones I have seen are carbon traces, as opposed to the FFC which is a copper trace.

[Jeff_Birt]

(11-27-2023 02:04 PM)blackjetrock Wrote:  I don't think a disconnected LCD row would suffer from anything, if you apply DC it's not happy, but applying nothing? More likely LCD coating, as you say.

Can you solder to heat seal cables? The ones I have seen are carbon traces, as opposed to the FFC which is a copper trace.

W.R.T. LCD damage. My line of thinking is that if a common row driver is disconnected the vertical drivers are still biasing the pixels. If each LCD segment is not equally biased (50% of time current flowing one direction, 50% other direction) it will fail fairly rapidly. It seems that pad damage on the glass is more likely though.

W.R.T. soldering FFC. Yes, you can only do this with copper. Not carbon printed ones.

[Jeff_Birt]

Here is a link I had previously saved. There is a note in the description saying that the AC-2000 series is correct tape for FFC to PCB bonding.

https://www.aliexpress.us/item/3256802972756543.html?

Here is some of the 2000 series I found:

https://www.aliexpress.us/item/3256801165737220.html?

Some searching for 'AC-2056R-35' revels it may be a Hitachi product and this datasheet:

https://bonding-machine.com/wp-content/u...-Sheet.pdf

[blackjetrock]

(11-27-2023 03:10 PM)Jeff_Birt Wrote:  Here is a link I had previously saved. There is a note in the description saying that the AC-2000 series is correct tape for FFC to PCB bonding.

https://www.aliexpress.us/item/3256802972756543.html?

Here is some of the 2000 series I found:

https://www.aliexpress.us/item/3256801165737220.html?

Some searching for 'AC-2056R-35' revels it may be a Hitachi product and this datasheet:

https://bonding-machine.com/wp-content/u...-Sheet.pdf

That is very interesting. I'm going to have to try some of these films. I think if it works you could get cheap flex circuits made up an duse the tape to fix the ends to the PCB and the LCD. Before I had the heat seal connectors made up for the PC-G850 I had flexi circuits made up and tried the 3M tape with those. that didn't work as the 3M tape cells are too big. This tape is much better on that parameter, so it might just work.
Thanks for the information.

Andrew

[polbit]

Is the Ti-92 Plus aflicted by the same LCD connector issue as the Voyage 200?

[Jeff_Birt]

(11-27-2023 11:19 PM)polbit Wrote:  Is the Ti-92 Plus aflicted by the same LCD connector issue as the Voyage 200?

Yes it is. The 80 series TI calcs can also have this issue.

[Jeff_Birt]

(11-27-2023 04:29 PM)blackjetrock Wrote:  This tape is much better on that parameter, so it might just work.

I have just ordered some to try and found a cheap TI-92 to practice on. Will report back and/or make a video about it.

[polbit]

You just can’t win with these! From HP 48 and the deteriorating foam, to 41 and the posts, to the cool TIs and their LCD connectors. Always a major design flaw that is not easy to fix cleanly (at least for an average user). Thank goodness there are people like you with much more skill then me making them work again…

[Bushcat]

If this film is Anisolm, it has two apparently contradictory characteristics: on the one hand, its expiry date is 6 or 7 months after manufacture which suggests it ages rapidly. On the other hand, then-Hitachi had a repair note which says that imperfect bonds can be repaired i nthe field by heat and pressure, simply by using something like a broad-head soldering iron and some kind of buffer material to protect whatever substrate you're pressing down on. I've done this successfully with a couple of Casios in the past.

[blackjetrock]

(11-29-2023 05:08 AM)Bushcat Wrote:  If this film is Anisolm, it has two apparently contradictory characteristics: on the one hand, its expiry date is 6 or 7 months after manufacture which suggests it ages rapidly. On the other hand, then-Hitachi had a repair note which says that imperfect bonds can be repaired i nthe field by heat and pressure, simply by using something like a broad-head soldering iron and some kind of buffer material to protect whatever substrate you're pressing down on. I've done this successfully with a couple of Casios in the past.

Yes, lot of material has over-enthusiastic expiry dates. For DIY it usually doesn't matter too much.
The repair procedure you describe is one that I have used. I use a wide 40mm or so bit on my temperature controlled iron and some silicon sheet as a buffer. I have improved quite a few heat seals like this, but after 2 or 3 'reworks' improvement tends to stop for me.

Andrew

[blackjetrock]

(11-28-2023 01:50 PM)Jeff_Birt Wrote:  

(11-27-2023 04:29 PM)blackjetrock Wrote:  This tape is much better on that parameter, so it might just work.

I have just ordered some to try and found a cheap TI-92 to practice on. Will report back and/or make a video about it.

So have I.
I have ordered this:

https://www.ebay.co.uk/itm/223536976654?...2854493156

I'll also report back.

Andrew

[Jeff_Birt]

I'm reading through the V200 'guidebook' again and have to say it is far inferior to HP's manuals. A few months ago, I read the HP-41CX user's manual Part #1. It was a very good introduction to the machine giving you an overview and a high level look at how to use it. Part#2 is more in depth about each aspect of the machine, i.e. written as more of a reference.

The V200 'guidebook' is written sort of like a half-assed textbook with no real rhyme nor reason to order things are presented. Lots of detail about minor aspects are presented before you even grasp the larger picture. It reminds me of reading the 'Circuits 1' textbook that my university used for years (because it was written by a faculty member). About 1/3 the way though it dove into great detail about something rather obscure and I remember thinking, "Why are they talking about this then they have not even mentioned this other very fundamental and important concept?" I carefully read back through previous sections and finally found the 'fundamental and important concept' mention almost as an afterthought. I mentioned this to a graduate student friend who had previously taught the class and he just shook his head and said, 'Yeah, there is a LOT of that!'.

Still like the V200 but the docs are sub-par.

[John Garza (3665)]

(11-28-2023 11:04 PM)polbit Wrote:  You just can’t win with these! From HP 48 and the deteriorating foam, to 41 and the posts, to the cool TIs and their LCD connectors. Always a major design flaw that is not easy to fix cleanly (at least for an average user). Thank goodness there are people like you with much more skill then me making them work again…

The root cause of all this is engineering decisions were made to reduce costs while sacrificing long term reliability. You see the same with HP machines. Everything was downhill quality-wise after the Classics.

-J

[Jeff_Birt]

(11-30-2023 06:18 AM)John Garza (3665) Wrote:  The root cause of all this is engineering decisions were made to reduce costs while sacrificing long term reliability. You see the same with HP machines. Everything was downhill quality-wise after the Classics.

I have mixed feelings about it to be honest. On one hand, I understand that these sorts of devices were not designed to last 50 years. It was not a design requirement and if it were, they would have cost much more. I understand why the decisions were made. On the other hand, I like playing with old calculators and want to keep them working.

I have a book about the history of the Allis Chalmers company. They originally started out making steam engines, the gigantic things used to power whole factories via a system of rotating shafts, pulleys and belt driven machinery. Material science being what it was around 1900 the engines were vastly overbuilt. While they required a lot of maintenance in terms of lubrication, adjustments and tweaking of controls they did not really break. Almost no spare parts were sold.

Given the limited market for huge steam engines and advancements in technology leading to electrification they put themselves out of the steam engine business. The huge stationary steam engines were scrapped even though they were in working order as electrification was more efficient. The effect was that the owners of the steam engines overpaid. They bought machinery which outlasted its usefulness by a large margin.

Over-engineering is as big a sin as under-engineering. Yet, though my mind understands this my heart wants my old calculators to function perfectly.

[John Keith]

(11-30-2023 06:18 AM)John Garza (3665) Wrote:  The root cause of all this is engineering decisions were made to reduce costs while sacrificing long term reliability. You see the same with HP machines. Everything was downhill quality-wise after the Classics.

Partly true, but graphic LCD's require a much larger number of connections than segment type displays. I don't think that one should expect such display connections to be as reliable as older, simpler designs.

Also, not everything after the Classics went downhill- the Voyagers are very durable and reliable.