May 2016
I am a mechanical engineering consultant by profession. I design the mechanical and hardware/software for my client‘s ideas and projects. I then 3D print the mechanical components, fabricate the PCBs, put everything together, and provide my clients with working production-intent prototypes. I love my job!
One of my elderly friends (early 80s) came to me with an antiquated product called “The Electronic Rolodex 128k.” He asked me how he could get the data onto his computer because the LCD on this device is starting to fail and he is no longer able to read the names and phone numbers. I dabble enough in this area where I said that I am very capable of figuring this out. So, I took on the challenge.
The only data I/O on the device is an IR transmitter on the left side and an IR receiver on the right. There are NO serial ports or any other connectors that would enable me to connect to a computer. The purpose of the left and right transmitter and receiver is so that if you have two identical devices, you can transmit the data from one device (on the right) to the other device (on the left).
I borrowed an IR to USB adapter from a friend and opened up a Tera Term window. I also used my signal analyzer to determine the appropriate baud rate (4800). I can see the data reception and it seems that I am receiving complete garbage characters. I have played with my signal analyzer and have inverted the signal, swapped LSB and MSB, and many other tweaks.
When that didn’t work, I opened up the device and soldered leads directly to the TX and GND pins at the IR transmitter. I then connected the leads to my signal analyzer and also my oscilloscope. I verified the baud rate and also noticed that the voltage levels were TTL, not RS-232. I borrowed a TTL to USB adapter, but this did not work either.
I came to terms with the fact that the data is compressed and/or encrypted. So, I searched far and wide for some software for this device and I found it ... well at least a similar device. I installed and ran it in a variety of configurations and still cannot get receive data that makes any sense. I have done enough testing and playing around with this product where I can almost write a book, but I still have no success in transferring the data to another similar device or to a computer. I have searched for this device on Google, eBay, Amazon, etc., and they are extinct. I’ve purchased a slightly newer generation of this product, but I do not believe that they are compatible. I have called the manufacturer, but they no longer support the device. So, I am at a total dead end. I have read your magazine for many years and thought that this would be the perfect venue for a Nuts & Volts case study/project for DIYers and the like. I would greatly appreciate any advice that you might have in doing what I originally expected to be a relatively simple problem to solve.
Please log in to post an answer.
I just had a thought, but it may be too late since the display is getting unreliable. Photograph each record, download the photos to a PC and import those in to Microsoft OneNote. OneNote has the ability to OCR a photo or scanned document.
However if the display resolution or quality is poor this may not work anymore.
“Google is your friend”, as I keep telling my girlfriend.
A 10 minute search on Google revealed a couple of interesting tidbits:
Apparently, RadioShack had, at one point, sold a computer interface for this device, though given the manual on their web site, I’m guessing it was 20 to 30 years ago.
Also, I saw references in other places that it’s possible to get the Rolodex to transmit a single record. If you really want to decode it, what I would suggest is to set up to record the signature of the IR signal (Digital Storage Oscilloscope?), and have it send a known [short] entry. (Maybe do so a few times in case it’s using something that changes each transmission). Then change one letter in the known entry, and send it again. With some patience, you should be able to break the code.
(I really doubt that they got very heroic about security on this gadget, especially as it’s quite old.)
Of course, if you really want a challenge, you _could_ decode the LCD code. (Remember that LCDs require a balanced signal, so go both positive and negative to the actual display).
It sounds like you have successfully tapped into the data stream in a couple of ways so that is not the problem. What you need is an algorithm for interpreting that data. I suspect your best bet is still the OEM. You have tried calling. Perhaps a politely worded letter to them, explaining what you are trying to do, would find it's way to someone who has been around there long enough to know and with enough time to actually answer. Perhaps you could arrange to have that letter delivered with an order for several hot pizzas and drinks around lunch time.
(I'm not familiar with this device, but I did a lot of similar stuff when this was probably built).
In the early days, the only real serial standard was RS232 which used +/-15V. These 'new pretenders' could not supply this, so many used 0-5V only. This just fell within the spec so that a standard port could talk to it (usually).
There's also the pain of getting handshaking right on adopters etc. You don't have a real port to deal with, having tapped into the IR part, but it could help explain some of the difficulties.
Although there's no RS232 port, data coming via IR may not be reliable without software handshaking (eg XON/XOFF with ASCII), but ASCII is not the only kid on the block. It could be EBSDIC or more likely as memory was always at a premium, encoded with reduced bits (think teletype close with shift-in and shift-out).
If you're determined to do it in hardware, bare the above in mind, and that most interfaces only like standard ASCII.
Or, try to tap the LCD feed? But, as the display is still just working, why not just re-key all the data - surely that would be quicker?
What a challenge! Good Luck!
Wow, lots of work so far! Could you stick a chip clip on the memory chip and monitor it while trying a transfer (to force it to dump the DB)? Maybe it isn’t encrypted until the IR transfer, that is, unencrypted coming out of the memory chip.... obviously, assuming there is one.
Opinion: there is some kind of handshaking protocol that needs to take place that doesn’t. In your post on the Microchip board (found while searching for software), you say 700 contacts at guesstimate 50 characters per contact. That should take 58 seconds at 4800 baud. Yet you only see a 1/2 second data burst. I think the unit doesn’t see the handshake, gives up and goes away. Success might depend on getting a similar unit to handshake with.
P.S. I see you are also in Hollywood. Look me up under KM4SJN and I’ll buy you a beer & swap war stories.
Have you opened the device and identified its processor? The data through the IR ports may be compressed, but the data sent to the display isn’t compressed, and maybe easier to capture as LCD digits and segments.
Then you can convert that back to ASCII and send to a new computer.
The next problem is how to attach to the LCD digit and segment signals without doing any damage to the original device. The scan frequency is necessary to sync the signals at the right moment when being read. The processor ID will give you the correct signal pins and help reading through a second device.
The last solution is to read the data on the LCD one-by-one and type the information manually to the newer computer. It may be the only non-invasive solution.
