Wanted. 12 digit bubble LED displays

[Specmaster]

A long shot I know, but I thought I'd try it anyway, I find myself now with 2 of the older style red bubble LED display, commodore calculators that would benefit from the fitting of 2 new display modules as some segments have given the ghost and popped their clogs.

The model numbers that I have are SR4912 and also 4120D, both are scientific models, but it seems that all of their special purpose calculators which includes, financial, statistician and navigation types I think employ the same displays.

Photos of the type below.

Can anyone help me out here, or know where I could look for these at all?

[MegaVolt]

Old Soviet indicators АЛС318 (Cyrillic) for 8 digits. They can be filed and put two side by side.

https://www.ebay.com/itm/255641303727

[Specmaster]

Thanks for the link and the idea, though I'm not at all sure that would work (apart from the size, to be checked later) but the display would have to be cut and filed as you say, to make up the 12 digits  that I need, but in doing so, the connections to the digits and segments on the display to be modified are also going to be chopped and the signals decoded to route the required signals over the second display module.

I need to think very carefully before adopting such a radical approach I think.

[MegaVolt]

Thanks for the link and the idea, though I'm not at all sure that would work (apart from the size, to be checked later) but the display would have to be cut and filed as you say, to make up the 12 digits  that I need, but in doing so, the connections to the digits and segments on the display to be modified are also going to be chopped and the signals decoded to route the required signals over the second display module.

I need to think very carefully before adopting such a radical approach I think.

The lens can be easily removed and processed separately.
The printed circuit board is a little more complicated.
To merge, you need to merge the lines that control the segments. If your driver is powerful enough, this is not difficult.

Here is a picture of what is under the lens.

https://cs14.pikabu.ru/post_img/2022/03/11/11/1647022706134822755.jpg

Somewhere there were the same ones with an additional lens of a stretching figure in height .... Here it is drawn with a dotted line. This is some kind of additional letter after the name.

https://arbatex.ru/upload/resize_cache/iblock/701/650_380_15462ec0b699c9953a0e709557c3c2f4a/701c1895acac828a7a3531d1485e5631.jpg

Judging by the description and weight, these are the letters "В" or "Г".

http://msevm.ru/data/rd/als318/index.htm

[AndyBeez]

Are you 100% sure the segments are dead and it is not the multiplexer that is wavering?

Maybe it is possible to farm displays from retro Texas, Texet or Rockwell calculators from the mid 1970s? A quick search on ebay for 'vintage calculators led" reveals a few likely donors.

[Specmaster]

Yes, it is the display at fault, I have another Commodore that uses the same display, and I have transplanted that screen, and it works perfectly in those that I'm seeking to replace.

I already have a search set up in eBay for vintage LED calculators and also for bubble LED displays, trouble being with the calculator route is that most of them are only 9 digits. Those are directly compatible, the cost plus postage is high and most of them are also working (allegedly) calculators, and in some cases the very same models that I'm attempting to repair 

[austfox]

I found these in my parts drawer, probably purchased in the late 80's? Not sure of the colour.

They appear to be 12 digits, but looks like your ones might have an extra 2 bubbles for the '- sign'?

EDIT: Red, common anode. Measures 60mm x 18mm.

[Specmaster]

The extra 2 apparent digits on mine, under a microscope, just seem to be nothing more than solder pads. The measurements and the pin spacings appear to be just fine, but the final connections seem to be different to mine. For example, yours have tracks disappearing under the display from terminations 4,6,12,14 and 18. Mine have tracks doing the same from terminations 4,10,14 and 16.

The PCB's must be multilayer ones to connect with the pads beneath the lens plastic. So it could be that the actual incoming terminations would be the same order as each other?  This certainly is the case in the case of my calculators, as the display board is the same in each case, but the calculator models are all different. One is financial one which displays 11 digits plus the negative sign, the other 2 or both scientific models, but different models which uses the first 9 digits, the first being the negative (if required) and then 8 digits. Digit 10 is not used at all, and digits 11 and 12 display the mantissa.

So these might be worth a try, what are you looking for them?

[austfox]

I'll draw you a quick diagram later on today so you can see how to wire them up. I have 4, so can offer you 2 at no charge.

Postage to the UK from Australia is not cheap though. These things only weigh 5g each, but the minimum 50g parcel will still be AUD $15.30.

I can send as a letter, which is AUD $3.70, but it needs to stay under 5mm thickness (maybe wrapped in a couple of sheets of A4 paper but no additional padding protection). International letters are only for printed matter, so there is a chance they get confiscated.

[Specmaster]

Well, that's very kind of you. They could be, perhaps, taped to a couple of pieces of stiff cardboard and popped into a jiffy envelope, either way it is certainly doable as long as I can (if needed) rearrange the connections to make them work OK.

I await you drawing later and take it from there. 

[austfox]

Actually simpler than I thought. It is only a double-sided board. There are 8 tracks on the top side that run horizontally the full length of all digits (4 above the digits and 4 below) and every segment's anode has fine bond wires to each.

Therefore, the segments are made from Pins 4, 6, 8, 10, 12, 14, 16, and 18.

Each digit's cathode is all the other pins (and in order from left to right) 1, 2, 3, 5, 7, 9, 11, 13, 15, 17, 19, 20.

Edit: The clear plastic bubble part measures 57mm x 13mm. Each bubble is about 5mm in height.

[Specmaster]

Well dimensional it sounds to be perfect. Can I ask how you determined the pin connections? Did you use a meter in diode mode or what?

The last thing I want to do is to use the wrong method of checking mine and risk damaging further segments by burning them out  by using too higher voltage output on a meter in ohms range?

[austfox]

2.1V from my power supply in current limit mode.

[Specmaster]

So far so good, they seem to be 100% compatible even though the PCB is completely different. I suppose the only thing to do now is to confirm that the actual segments agree as well. This is what mine is displaying by pin, left to right:

4 = C,  6 = DP, 8 = A, 10 = E, 12 = D, 14 = G, 16 = B and 18 = F.

[austfox]

Yep, exact pinout match.

[Specmaster]

Brilliant, I'll PM you.

[austfox]

Two displays on the way to Specmaster. Feel free to post a couple of pictures of the restored Commodore calculators once fixed...

[Specmaster]

I shall do that gladly, looking forward to getting them up and running once again with a display that you have confidence in the results again  .

[Specmaster]

At last, I have actually got round to replacing the 12 digit LED displays in 2 of my Commodore calculators which was enabled thanks to the generosity of a fellow member austfox from Australia who had a couple of spare displays in his collection and the postman finally delivered them today.

The display was fitted without any drama at all, but the second one did cause me concern when I went to solder the ribbon cable onto the display, shock horror, I discovered that I had one core unconnected at the end 

Back I go and double-checked all the freshly soldered connections and yep, not a single one was missing  The display has 20 connections, the other 2 calculators also only had 20 connections and all cables were correctly located and soldered and were working. So what was happening here, then, was I going mad, had the excessive heat (33C) in my lab finally sent me round the twist, had I become dehydrated and seeing things. Nope, I counted the cores and there was 21 of them, panic began to set in, how could I make such a mistake 

Switched on my trusty Ti1906 bench meter, put it into continuity mode and checked the connections through, 1 to 20 were correct so what was 21 doing and where was I going to connect it, then it struck me (working blind), no service data available I checked to see if connection 21 appeared anywhere else on the PCB of the main board, I mean it was certainly soldered onto connection 21 of the display board, so it had to go somewhere right? Wrong, turned out it had no connections on the main board so begs the question, why was it there, why did they go to the trouble of having special boards made for the main and display, just to terminate a core that did nothing, why not just rip that core away from the ribbon 

Anyway, here is a shot of all 3 of them showing every digit and all the correct segments works 

Many thanks to austfox, without his kind help, these 2 would be confined to the repair queue for a long time before donors could be found, or worse still, scrapped.

[Brumby]

I have also had the pleasure of austfox's generosity - even to the point of dropping stuff off at my doorstep because he was in the area.  Muchly appreciated. 

This sort of support from members always makes me smile and very appreciative of having found the EEVblog.

[Specmaster]

Thank you everyone for your kind support in the restoration of these lovely vintage calculators, I have more coming up, currently work in progress is a Texas TI-59 programmable in bits at the moment, looking for other donor machines to appear in the future, hopefully. Also have another Commodore inbound, a 4148R, status unknown, sold as not working.

In the photos below there are left to right Commodores SR4120D, 4902, SR4148R, SR4912 and then Commodore PR100 and Texas Ti58 programmable. All are fully operational, although 2 of the Commodores have had replacement displays fitted because of failed segments, and worthy of note is that although the SR4120D and the SR4148R look the same, they are not, the SR4148R has rechargeable batteries and an extra 2 digits on the display.

So far I'm happy to say that not a single calculator has been sacrificed in order to keep the others working, and the PR100, Ti58 and SR4148R have had their internal rechargeable batteries replaced and slightly modified to keep them working as intended.