alexGS

Thanks, Thorsten! 1. yes, 2. yes, and 3. yes - It is relatively easy to find a 1994-2004 laptop (or desktop) that can boot into DOS (a bootable USB drive can be made with Rufus). TCLOGO is a small .COM that seems to run easily in DOS with trivial memory requirement, but not in DosBox when I tried it; “out of space” was the message I saw. I imagine some kind of stack for its own variables. Most laptops of suitable age have parallel and serial ports as well, so I am happy to entertain all options; I just thought that the PCMCIA slot was closest to the original ISA bus used by the 9771 card; and this seems to be the way we might achieve I/O at port 925 without any additional software tricks. But we’d have to get past Card Services (the evil gatekeeper who will deny our logic!) Your data sent to/from the Arduino, being a single set of eight bits for outputs and inputs, seems to be more promising for original software compatibility than the parallel port cable I’ve made; that requires the use of three separate ports, one for outputs, one for inputs, and one for power output; this would be difficult to forward/translate to port 925. In summary, I like your serial port output via Arduino translation, the hardware is easier to make than my PCMCIA idea, compatible with a wide range of machines. Thank you for explaining the need for the USB-TTL adapter, I have one lying around, and an Arduino Nano for that matter, so I shall now try to make your circuit :) Now imagine if we could redirect that serial I/O, so that the TC-LOGO program can use it. This was what I was thinking a TSR program could do (like the soundcard example above) but I was worried how it would redirect inputs as well as outputs. I’ve just had a better idea for a logically-simpler solution. How about using a tool like Ghidra to diassemble the TC-LOGO program (quite small, only 34KB!) and look for where port 925 is used (39Dh) and change it to… 3F8h? The COM1 address :) Since the 9771 card has a jumper to change the I/O address, I guessed there must be a way to change the address used by the DOS version of TC-LOGO, but it is undocumented and I can’t find it by poking around so far. Imagine if it was simply a /p926 command switch…! It feels like there are many potential ways to solve this problem, I look forward to trying Ghidra tomorrow. I don’t know if it can make an edit and reassemble… Cheers -Alex

Hello Thorsten, It’s ironic that you post this today, just as I was replying to your new thread (relatively new anyway) which I have only just seen today while doing Google searches. First, well done on assembling this information in one place, I think it is useful. Second, thank you for a solid attempt at an answer to my question - which I am still struggling with :D I’ve spent about five hours today dredging through ancient PCMCIA specifications. And yes, I read all about that 24-bit IO card, which still can’t quite do the job because it can’t mix input and output bits in the same latched port (remembering that the LEGO software can only talk to one port). To clarify, my idea was to attach the latches/address logic of 9771 to the PCMCIA slot, which appears to have all the suitable address and control lines as well as the data bus (as found on the ISA bus). Sadly, between the ISA bus and the PCMCIA, there is both a hardware layer (the Intel 82365SL) and a software layer (‘socket services’, ‘card services’) that need to cooperate. Both could be programmed to allow the access as port 925, and the PCMCIA slot is electrically and timing compatible with ISA. The configuration via card services looks like it needs this programming in the form of memory registers attached to the PCMCIA device - ie. the interface would need an EPROM, or something, to supply its configuration data. I don’t know how to make that work. http://elinux.org/Flameman/pcmcia gives a summary of the layers/device definitions. It still seems possible. The objective is to run the original LEGO software on a slightly wider range of hardware, being laptops as well as desktops (it is generally difficult to find anything with an ISA slot, but there are more devices out there with PCMCIA and able to run DOS, such as from booting DOS off a USB drive). I have two other left-field ideas in mind, but neither would offer the LEGO software compatibility that I would like to see. Those left-field ideas are - the Velleman USB interface (electrically suitable, six output bits and two input bits etc.), or the use of a RaspberryPi - actually, that’s miserable with the 3.3V, so perhaps the Arduino instead, but with either of these - like the printer port idea - it seems everything has to be done from scratch. The previous Arduino solutions haven’t actually shown any use beyond turning motors on and displaying inputs - there has to be a method of actually programming it :) I hope your Arduino implementation works out differently :p Is it possible to read and write those bits with a single byte written/read to/from serial? And if so, could some kind of device driver (a TSR perhaps) make that available to the LEGO DOS-mode TCLOGO? A possible lead on that last idea; these TSR programs written to redirect sound card access to an ISA or LPT port. I think the task we face is similar (instead of port 205h, 388h, etc. we could redirect 39Dh to the serial or parallel port, though I wonder how inputs would be handled - perhaps the program can intercept inputs as well as outputs. It’s written in assembly language that I know almost nothing about - http://github.com/JKnipperts/TNDY I’m definitely interested, even though seeing holes drilled in LEGO and tubes ground off the underside of a tile has made my heart skip a beat ;) It also confused me when you said a USB-to-serial converter would be needed for a modern PC, considering the Arduino has its own micro-USB built-in. I guess all will be revealed in your next instalment. Thanks again -Alex

It’s always quite exciting when topics are discussed at the very time that I am thinking about them - i.e. within a day or two (for something that I last considered at least a year ago) :) I am wondering whether the 9771 card logic (which I had found previously at the other link) could be interfaced to a PCMCIA slot as found on old laptops? Would the LEGO software then ‘talk to’ port 925 and would those signals appear at the PCMCIA connector? PCMCIA pin out http://www.hardwarebook.info/PCMCIA - seems to have OR, OW, and EN? And address/data of course. I’m not clear whether a system with a PCMCIA slot has an intervening ‘card services’ in BIOS that only configures ‘approved’ devices, or only enables certain address space. Or does it simply expose all the ISA signals at the PCMCIA slot? The wiring of the LEGO card would presumably configure it at a particular address - as I recall, there is a jumper to select “adress” (!) 925 or 926? The only other practical solution I can think of, to use the LEGO software on a system that doesn’t have the ISA card installed, is to somehow modify the software to redirect output to the printer port. I presume via disassembly the port address could be changed from 925 to 378h? I guess it would be very difficult to match the 9771 port characteristics in terms of output bits and input bits being at the same port address. Thanks for any insights in solving this conundrum -Alex I just realised that I have hijacked your thread in replying to Thorsten on his topic (which I landed on from a Google search). Sorry about that. I am glad you have found the software you were looking for - I shall download it too, in case I get one of those little blue cameras for my Mindstorms :) -Alex

That’s fabulous - thank you :) I will watch the complete video shortly I note that on their Macintosh LCII, the Dacta interface doesn’t fill the modest 640x480 screen properly. I guess that’s why I tend to use even older models like the Powerbook 100! I’ll try it on a Classic too… The advantage of the Windows version, which seems to have come out after that video was made (they’re showing the DOS version), is that the windows can be re-sized and multiple pages displayed at once. Again, wonderful find - thanks for sharing your viewer’s resource here! -Alex

Ah right, I understand now :) Nice work! Love the reference to 2400 baud as ‘light speed’ :D I also wondered whether a Psion Organiser could be used for controlling some of this stuff. It has a serial or a parallel port and a programming language built in. Just another piece of period 1980s tech, but small and handheld :) Cheers -Alex

Evan put it there ;) I’m guessing the software from the link you shared is the same, but maybe I should check. There’s separate versions of Control Lab for DOS and for Windows 95/98. I bought the Macintosh version and Evan put that there too for me. -Alex

Hello Thorsten, I think you need a 9771 card for that wonderful PC - you’ve got that into great shape. Then you need 9750 Interface A, and the BASIC code that appeared in the 9750 guide booklet, complete with typo which I enjoyed discovering and fixing ;) The PC is too old for use with the RCX or Scout :P The Toshiba laptop sounds perfect! The LEGO software was highly annoying, but can the Toshiba run Brixcc/NQC? -Alex

Hello Rob, There has been much information in this thread above - which I revived recently to add the Macintosh software. However, I think the overall easiest solution to use the Control Lab is the original Windows 95/98 software which you can download from https://archive.org/details/vintagelegorobotics along with the manuals (reference guide etc.) as example files are included. I think having old hardware is easier than trying to work with alternative software and partitioned/emulated operating systems. For the correct look and ‘feel’, I bought a 1998 Toshiba laptop last year (came with Windows 2000) and I put Windows 98 on it. This cost me $50. I suggest if you search for ‘Windows 98 laptop’, on your local marketplace website, you’ll find something suitable. You need a 9-pin serial cable of the ‘null modem’ type (perhaps you have the original cable with the interface, but if not, they can still be purchased new from good electronics stores). The Macintosh software option may cost you a bit more due to the rarity of the hardware, but if you go down this path, Mac OS 9 or earlier will work. The machine needs to have round serial ports, so the iMac is ruled out - the last machine to have serial ports is an early 1998 Powerbook G3. The software does not expand to fill the screen, so something older with a smaller screen will look best. PowerBook 5300, 520c are quite expensive and fragile. There are many other monochrome-only PowerBooks which are cheaper (e.g. 140, 150, 520). Any desktop Mac before the iMac will work. I use an accelerated Amiga (which I happen to have) to emulate a colour Mac. I also have a PowerBook 100 which runs Control Lab fine as well (monochrome). Brickfactory.info has the instructions for set 9701 if you’re missing them. Click ‘By Number’, then use the drop-down list to find 9700-9799. Yes, a global effort, I take my setup to LEGO shows here in New Zealand… Have fun! -Alex This is a great summary, many thanks. My iPhone translated into English for me with no problems. In particular I see the Control Center-based 9702 set instructions there, which I have not seen anywhere else. -Alex

Thanks, Thorsten - I’ve posted a couple of videos on your other thread which you’ll probably like :) We’ll keep to the Control Lab stuff over here, as this thread was useful to me in understanding the Control Lab software options. -Alex

Hello Thorsten, thanks for your comment on the Control Lab thread :) ZX Spectrum pre-dates the RCX by many years and for that reason, I choose to team it with early 80s Technic in the style of the book, ‘Make and Program Your Own Robots’. I have built a modern replica of the RoboTEK interface and I can sell you one if you like, though it’s not a cheap option at around 80 Euros (I had the circuit board specially made). Here is the three-motor arm project from the book; really crude programming compared to yours. The inputs and outputs of the RoboTEK are all over the place, so the inputs are not valued at 1, 2, 4 as you would expect, but instead 4, 1, 32… I could have easily rectified this and could have made improvements with double-pole relays etc. but I wanted to keep it completely compatible with the original projects in the book. And here is a look at my part of New Zealand’s national brick show: -Alex

Hehe - 512x342 and 512x384, so it will be screen-filling for those two :D It might save you some puzzlement if I point out that to select objects, you shift-click them. I fumbled for some time. The most subtle feature of the lot is that a motor in the setup page can be associated with an Angle sensor, which will mean OnFor runs for units of angle, not the default time units. Feel free to discuss how-tos here. -Alex

You’re most welcome :) I think the Windows version is technically better (multiple, resizable windows, faster code?) - but the Macintosh version has a nice look to it. I’ve always liked the small font used on the Mac. -Alex

Hello Chris, please see https://archive.org/details/macintosh-control-lab-1.0 Cheers -Alex

Finally uploaded to https://archive.org/details/macintosh-control-lab-1.0 Sorry for the wait! I managed to get the Scanner example debugged and working reasonably well, though not quite as well as the Windows PC version. Results a bit fuzzy… may need tuning of the parameters. A complex program that I haven’t had time to fully understand. Have fun and let me know how you get on! Incidentally if running with an Amiga 500+/Shapeshifter, you’ll most likely be limited to only a few colours due to the ECS chipset. With AGA, it runs well in 16 colours but I suspect with the A500+ the bitplane conversions might be too intensive. In fact, the Colour version can still be used in a black-and-white Mac environment (looks identical to the B&W version; the latter is just smaller to fit on an 800K disk). Cheers -Alex

Cheers for your enthusiasm, I feel bad for not seeing this post sooner. Eurobricks had an old email address that I can’t see any more, so I missed the notifications. A progress report at last - the package finally arrived a few days ago after nearly two months in transit. I will upload the unmodified disk images tomorrow, and also a modified disk that I have been working on - see below. Both disks (B&W and colour versions of Control Lab 1.0 for Macintosh) read perfectly in my Powerbook 100 (a story in itself repairing that thing but at least the floppy drive, which is external, was fine). Both versions display identically on a B&W Mac, so I think the only reason for the B&W version is that it fits on an 800K disk. The colour version is 1MB and supplied on a 1.4MB disk. As my PC can read the 1.4MB disk into an image using HFV Explorer, I’ve chosen to work with that version. I cannot get the 800K disk image into the PC without corrupting it, though I can image a 1.4MB Mac disk with the 800K image file copied onto it. Hopefully, anyone trying to use a Mac Plus will also have another Mac to read the 1.4MB image and write the 800K disk. The colour version on the 1.4MB disk appeared to include some project files (Untitled, Untitledt, FAST CAR), but none had any procedures (code) or interfaces (pages); only the setup page. It became clear that these are not original files; uploading that disk image as-is would have been confusing and useless. The B&W version had no project files at all; just the program. The Windows 95 Control Lab (downloaded from the archive) includes four interesting sample files (Conveyer Belt, Wheelchair, Robot arm, Scanner) whose code and interface pages are not described in the written materials and not found on the Macintosh disks. Therefore, I am transcribing these into the Macintosh version. Even though that may be confusing, it will not be useless. There were subtle bugs to solve, small differences. One of the samples (for the robot arm) in the Windows Control Lab is far more ambitious than it first appears. It is a program that records program code into a text box, which the program can then run. When you click the buttons to move the arm, the program measures the angular distance moved OR the time moved (as applicable for the two motors) and records that into the text. The “parse” function is a mindbender, and so is the error trapping provided to run the recorded code which the user can also modify in the text box (hence, error messages are passed through to a temporary text box). Crazy stuff! The “scanner” program promises to be even more exciting, as it modifies a picture shape on the fly… The other crazy aspect is the development platform I’m using. Obviously, the monochrome Powerbook 100 was limited, so instead I have installed Shapeshifter on my Amiga 1200, using two CompactFlash cards, one for PC-card file transfer and the other for internal ‘hard drive’. Finally the 40MHz 68040 accelerator card with 32MB RAM has come in useful - it gives me the speed of a Macintosh Quadra 650, except for the graphics, which have to be emulated and are relatively slow (0.8x the speed in 16 colours, 0.48x speed in 256 colours). I got Mac OS 7.6 installed and the Control Lab software runs perfectly, with the interface connected to the Amiga’s serial port (9-25pin adapter). One obvious limitation, which became particularly apparent at a larger screen resolution, is that the Macintosh version does not allow the procedures/interface window to be resized - it is fixed to 512x342 (suitable for a Mac Plus), though the small command window and tool palette can be moved/resized outside this. I have the Powerbook G3 which has just arrived, but that isn’t yet as easy to work with as the Amiga, due to a lack of any way to transfer files in/out of the machine. No USB etc. Even a simple thing like the picture background of the samples was a bit of a challenge - I took a screenshot on the Windows 98 PC, cleaned it up in Paint, saved a bitmap file, copied it via a USB drive to a Windows 10 PC that can write it to a CF card for the Amiga to read, copying it to the Amiga’s hard drive from where it can be copied to the Macintosh partition (only when Shapeshifter is running), so that Graphicconverter can open the BMP and then it can be copied and pasted into Control Lab… LOL. I did something similar for the procedure text, stripping out the CR/LF etc. I forgot to mention that the data files from the Windows version obviously aren’t compatible with the Mac version - I did try that first, setting creator type etc. but they locked up the application trying to open them. Just a couple more of the four interesting samples to do tomorrow - then I will put the resulting disk image into the Internet Archive, along with the original disk image with no files. I will leave the creation of the ‘standard’ project files (i.e. those created by following steps documented in the written materials) for others to replicate, since that can be done without the Windows version running. -Alex

A few weeks ago, I managed to buy the original Macintosh Control Lab software - two 3.5” disks, black-and-white and colour versions - from a Bricklink seller. I was unreasonably excited about this, especially as I’m not really a Mac user :D just want the experience… I’d been looking for some time… I’d even seen the request in Macintosh Garden, and the Robolab software offered there. I have Robolab 2.5 on the way as well. While waiting for the disks to make their way to me in New Zealand (one more week I hope), I’ve also bought a PowerBook G3 (early ‘Wallstreet’) which should be ideal with its serial port. It’s probably still a month away. However there isn’t a floppy drive, so my recently-acquired PowerBook 100 curio will join in the fun to read the disk and make an image file etc. which will be uploaded to Archive.org in due course :) Cheers -Alex See post above - I have the software on its way to me and I will share soon :)

Certainly WD40 or even vegetable oil will work, but vegetable derivatives become sticky and messy, and with any mineral oils you run the risk of damaging the plastic (it will crack). My suggestion would be silicone grease - as the safest grease for plastic parts. It may be a little hard to find this where you live - try an engineering shop, or try a model shop (whichever is easier) - it is sold for small R/C car gearboxes at huge cost (but you don't need much). Don't try to use silicone heatsink grease (white) from electronics shops - similar, but NOT the same! The grease you want is translucent, and melts slightly with temperature. You only need the slightest smear to achieve the result you want. I used to use silicone grease with my Technic models, and it made a huge difference to the smooth running. If you can't get silicone grease, I would expect silicone spray (auto parts store) to be almost as effective (and still safe to use), though it probably won't last as long and application may be messy. -Alex

Good results there; after only five days! I find that it takes at least a week and usually two. It would be a good idea to check the state of printed elements every day or two, I now believe they are OK after a week, but get a little faded after two weeks. Paper-based stickers can be peeled off first and carefully kept... -Alex

It's been a little while since I visited, but this new colour scheme plays crazy games with my eyes... I don't think the buttons appear properly either? Please... who wants it back the way it was? Thanks, -Alex

Good ideas. My main item is my airport (I used to make tremendous runways with a hundred LEDs, fourteen years ago now.) I have a fair number of roadplates, and there will also be a race circuit somewhere. I was careful to have a short piece of 'double track' suitable for my 7866 crossing, which is hopefully on the way from Germany. The bridge could perhaps be a giant viaduct on the diagonal section, since that may be the part that 'falls off' the layout table. My layout table used to be a large cork noticeboard (in unused condition) which fitted on a queen-size bed. I've recently acquired a pair of doors 2m high x 90cm wide. All three things together might be enough space. I must have another go with Track Designer to see if I can draw-in bounding boxes (of specified size). Thanks for the input! -Alex

Hello all, I've been busy with Track Designer (and eBay: buying more track!), and this is what I've come up with. The only current problem is that I've just realised that the 'table' I was going to use (two large doors) is only 2m x 1.8m, so the layout won't fit :-D One main criteria was that I wanted no dead-ends - I wanted it to be possible for a train to run uninterrupted and traverse the whole layout. I started with an 'endless loop' (a loop that went out through the crossings and re-joined itself in three increasing circles) but it became a little complicated by the need to include sidings. Which means that now, a train cannot traverse the whole layout at once - ah well... I am fortunate enough to have four trains (with a fifth on the way) so there should also be good potential for collisions ;-) Does anyone have similar layouts to share? Is there anywhere to find .tdl files on the 'net? Thanks for looking! -Alex

Hi, thanks for the response! :) Yes, the bleach does work - the results are best on white LEGO. I have had amazing success with other white parts (please look on the LegoFan forum for examples, I can't seem to post pictures here unless they're truly tiny - only a few Kb!) You can bleach light grey (or other colours) but unfortunately the colour gets lightened as well. For these colours, I recommend leaving outside in bright sunlight (ironically!) as this seems to return the colour about 80% of the way to normal. It's proving hard to do this here in NZ, because it rains every day and we only get a few hours of sunshine every couple of days. Ah well. I would start with about 1 third bleach:2 thirds water. Leave parts for 24 hours, then check on progress. For the really yellowed (browrange) pieces, up to one week is useful. I find that after a few days, the bleach solution becomes less effective, and this needs to be allowed for (i.e. one week is about twice as effective as two days). As for using hot water, I would expect that to speed things up, but how long is it going to keep hot for? ;) Bleach does not affect the printing on LEGO elements (to my great relief...) I don't believe that bleach affects the integrity of the plastic. I have had problems with split LEGO in the past, without using bleach. The pieces that I have bleached still seem to behave perfectly. Cheers, -Alex

ha - yes, I think I understand the reference! You must be on both forums... :-D Right - well, my conclusion so far: 1. White bricks can be made white again, 50% bleach solution, up to a week. Works great. 2. Light grey will always end up lighter once the yellowing is removed. Time period up to 24 hours. 3. Points MUST NOT be bleached (even after four hours, corrosion was pitting the surface of the rails. I didn't leave it overnight...) 4. Conductor rails CAN be bleached - for some reason, the corrosion does not seem to occur (better grade of stainless steel?) 5. It would be lovely to buy new parts - but the PRICE... During my 'day job', I get about US$7/hour. It is easy to blow through a week's work just buying a few items off eBay. Today, I spent a day's work on some train stickers :) Six pieces of track generally sell for about US$20. Better to save what I have, I think. I would like to buy a MISB 7740 set. That will be about US$500 (months of savings!), but at least the pieces won't be yellowed. Cheers, -Alex

As the title suggests, I'm trying to improve the appearance of old 12V track. It's been a nerve-wracking process! Over the last week, I've had some white LEGO pieces sitting in a 50% bleach-water solution. It has been pretty effective, turning the yellowed (and in some cases browned/oranged) pieces back into white. *sweet* So, I put some grey tracks in a 50% bleach solution as well. After 24 hours, I took them out. They looked great! Encouraged, I put in some more track - and took that out after about 18 hours. It still looked a little yellowed... ...but by this time, the first set of track had continued to lighten even after removing & rinsing... as it dried, it turned into the track you see in the top straight in the first picture below! :-/ The bottom track shows how the middle and top tracks looked before any treatment. Horrible, I'm sure you'll agree. The middle track is what came out today, after 18 hours. Initially still a bit yellowed, but as you can see, it's now a little too light. This is compared with the little-used point on the left of the picture. Here is a crossing and signal rail. You can see the need for lightening, compared with the similarly-aged but original-condition items alongside. So also last night, I put a point in with the crossing and some conductor rails. This was perhaps a little too adventurous. After 18 hours, there were the most picturesque orange 'mushrooms' growing on the rails of the point. And when I took the point out and cleaned it, I found that the corrosion had eaten right through the rails in two places :'-( The third picture shows the results (the crossing, the rail, and the point). As you can see, the light grey is a little closer, but the dark grey of the point has also lightened too much (this happened after rinsing & drying). I have some more track in there. And another point. It's been four hours. Can I bring myself to leave them overnight? X-O -Alex

The price - looks like a bargain compared with any 12V trains that I try to buy on eBay :) Just the 7760 Diesel Shunter locomotive sold for 225 pounds - that's, what, US$270? And I'm trying to find a 7740 Inter-city set - most are around 250 Euro, but I did find a couple that were Mint In Sealed Box - for US$500. They both sold, too. -Alex