flashjazzcat

Wasn't one supplied with the board?

OK: let us know what happens.

Well, I got my first ever YouTube payment the other week: £78 (which took a year to earn), so I guess I am being paid for it. Patreon is also open (thanks to the person who signed up with a monthly pledge already, BTW). But yeah: if people want written instructions provided at the point of sale, this is a vendor problem. I am perfectly capable of producing such material if enlisted to do so, but the more I see of the vendor's installation philosophy, it appears completely divergent from how I like to do things anyway. As for videos vs. written material: one literally CANNOT WIN here. I write a sixty page user manual, people ask for a video instead or say the documentation is too in-depth. I make dozens of videos and people want written material. Anyway: I don't sell the product so any useful contributions should be regarded as bonus material.

See tf_hh's comment earlier in the thread. The entire system runs at a slightly different clock speed depending on whether the machine is NTSC or PAL, and VBLANK runs at a different frequency. The peripheral has to sync with the host machine; 1050s of either standard are apparently able to do so, but the XF551 cannot and has to have the drive firmware matched to the host.

That's fine, but you came in hot putting "extensive user manual" in quote marks and giving the impression that the information you were looking for was nowhere to be found. Of course I meant to provide supplementary, illustrated written material as a companion to the video on my website, but that's just one of the things which never gets off the to-do list (largely since I'm primarily concerned with documenting the firmware I wrote for the thing). In point of fact, I prefer a written reference myself and I have had my own video freeze-framed for two days as I work on a pair of 800XL U1MB installations. I wonder to myself why I didn't simply screen-grab the video or hunt around for the original graphics, but whatever; it wasn't a massive headache to fire the video up. Well, while the video is supposed to be instructional, it's not aimed at someone who wants to get the device installed in five minutes. If a video is needed (or rather, written instructions) on how to shorten or build your own ribbon cables, I can produce something. Sure: I've filmed the process before, but it's doubtless buried in this or one of the other U1MB install videos. The video is aimed at someone who wants to do a neat install, has the dexterity and tool ability to do so, and is prepared to research and learn how to do new things. YouTube's chapters feature is pretty good now; I keep meaning to go through older videos and add descriptive jump points. Anyway: I take on board the comments regarding written instructions and even agree to a large extent, but when I wasn't actually asked by the vendor to produce installation guides in the first place (instead doing so off my own back since I was doing the work anyway), said videos then get opportunistically linked on the vendor's product page (without me asking for this to happen), and then people complain that the installation instructions aren't good enough, it is possible to end up in a double facepalm moment. For reference: the grey connectors are called 'Harting' connectors and can be a little difficult to source. I do have some spares, but attempt to re-use the existing connectors in 100 per cent of cases. This requires the deft use of a jeweller's screwdriver and takes practice before you can get away with not breaking them. The IDC connectors at the other ends of the ribbon cables are pretty easy to source, meanwhile. They're not intended for re-use either, but you can get away with doing so if you're handy with a craft knife. The initial U1MBs (produced by Candle O'Sin) shipped with bespoke ribbon cables tailored to the host machine and quite nicely documented on Candle's website (all of this material could have been reproduced and still up on the Internet in a centralised location, but it wasn't and it isn't). This required the customer to decide which model of machine they were going to install the board in before they ordered it, but did avoid all this cable hacking. I just happen to think Candle's original installation philosophy was spot on and a good reference for others, but many people prefer to go their own way and hacking up the ribbon cable, etc, is a risky proposition for many. Anyway: the videos document my approach to installing stuff in customer machines, and if they inspire others to do a decent job, they served their purpose.

The blobbed ones are all connected together by traces anyway. The rust is right between GND and the 5V trace running around the outside of the board, though, so an intermittent dead short is quite possible, especially since the power LED didn't initially come on.

My 800XL installation video has diagrams with labelled arrows pointing to the four vias. There is really no need to run around in circles looking for that information. Shortening the ribbon cables needs a pair of scissors, a jeweller's screwdriver and a desk vice, but the cable lengths are described in the video too. The five minute install described in Lotharek's perfunctory instructions will give you exactly what you typically end up with: board hanging off a screw, four wires soldered directly to the CPU legs, and a concertina of redundant ribbon cable. Even that will take significantly longer than five minutes, owing to the socketing work which may be necessary. If you started with the video and actually watched it, though, you have the via locations to start with. It usually takes me at least an hour to put U1MB in, including adding or replacing sockets, drilling, shortening cables, crimping, etc. Selling the thing on a 'five minute install' implies a pre-socketed machine and a naff looking installation. As for a clear, step by step installation guide: as the videos show, you can take a horse to water, but you can't make it drink.

So it should, since the 1050 is TV standard agnostic, unlike the XF.

It is in fact a requirement for U1MB.

Try swapping out the delay line IC as well (one with missing legs). They sometimes die and render the machine stone dead.

If you want to test the extended RAM, XRAM will do. Or is something else acting up?

It almost certainly won't happen on an LCD display. I'm not an electronics guru, but I guess the luminance intensity changes are messing with the geometry of the CRT raster. I've seen this happen on many old CRT monitors, and it's a thing which is nicely absent on LCDs. Perhaps the CRTs would benefit from recapping, etc, but I don't know enough about that technology to offer informed insights; just what I've observed with CRTs in the past.

That's a good question, although Jurgen reminds us that this is only worth worrying about with the XF (although there are others; my Rana2000 won't work with PAL Ataris).

Are the drives and computer all NTSC or PAL?

It has nothing to do with the cartridge or the software. It's a CRT problem.

Get SIDE3.

Thanks - I will add that to the list of things to do when I get some free time. I think I remember that episode.

Ouch. It was not, fortunately. I'm now at the stage where I'm not sure if I made some critical, invisible error with a passive component or whether I'm somehow managing to always have at least one destroyed IC installed.

Oh crap... Herb just reminded me of my magnum opus mad skillz moment. Last April or so, I accidentally got the12V and 5V regulators on my 1088XLD swapped and pumped 12V through every chip on the board when I first turned it on. Fortunately I left the U1MB out for the first power-on test. Michael had labelled the regulators as I recall, so I have no idea what the Hell happened here (although I blame the UK government for terrorizing and pyschologically torturing everyone for the past eleven months). Still haven't been able to figure out what blew (I appear to have a complete set of spare discreet logic ICs), and having rectified the error, the system boots to a PMG stripe down the screen with U1MB installed. Maybe I can diagnose it when I get a scope, which will be shortly.

It's an H743ZI with a maximum frequency of 400MHz. Electrotrains lent it to me for UNO cart hacking purposes. Do you think it's good enough for this application?

A professional recently installed VBXE in the GTIA socket of a 130XE, couldn't understand why it wouldn't work, started a thread about it with photos, and it took two dozen posts before any of us noticed what he'd done. Sadly the thread got deleted, which is rather regrettable since no-one gave him a hard time about it. That was the best one by far of recent times, anyway.

I did this once years ago when I used to employ an unused gate on the CD4050 to buffer the sync signal on VBXE machines. I forget which machine it was (one of my own, fortunately), but one of the spare inputs or outputs was tied to GND, so I cut the trace before wiring things up. It was only months later that I noticed I'd completely screwed up the legacy video output. When I looked more closely, I'd completely (and completely unexpectedly) isolated the GND pin of the CD4050 by cutting that trace, leaving the whole thing floating.

No: this would be in XL/XE mode using the inbuilt PBI math pack overlay ROM. It already has an implementation of Hias' HSIO driver with SIO2BT support added. So, the need for OS patching is totally done away with. CPLD for logic, but everything procedural is done on the 6502. I still have a Nucleo development board here (borrowed), which I guess would work. Looking forward to the full details!

Heh... what Bruce said there. Story of my Atari life.

This is extremely interesting. If the source isn't too big, I could attempt to implement a driver in the U1MB/Incognito PBI BIOS.