Jump to content

Falonn

Members
  • Posts

    215
  • Joined

  • Last visited

Recent Profile Visitors

915 profile views

Falonn's Achievements

Chopper Commander

Chopper Commander (4/9)

164

Reputation

  1. The chip on the left in your photo is 2MB of flash storage. And the chip on the right is a CPLD that is probably handling the more advanced bank switching that opcode mentioned earlier. So I'm guessing the benefit an SGC cart brings is lots more storage (vs. a mega cart) along with being able to write back to the flash (for features like saving).
  2. I think it's a jack for a standard TRRS video cable (like this one). The components on that board (notably the LM318) are just the through-hole versions of the AV mod in your other recent post.
  3. I don't have anything very useful to add to the good advice that's already been given, but wow! I see reflow work like that in my nightmares. 😬 All of the SMD stuff looks like it was absolutely swimming in paste which then wasn't heated enough. And the wires (especially VIN and VCC) look like cold solder joints that are hanging on by a thread. I'd be curious to learn the backstory on that board. Maybe Luke had a young relative visiting, was teaching them how to do surface mount work, and somehow that board got mixed in with the rest of his stock. 😅
  4. This looks like it will be a great resource for the community, thanks! As someone that's been thinking of dipping my toes in Colecovision programming for a while now, I was just browsing the code and spotted this curious double call to vdp_no_interrupt. Is there some quirk of the hardware that requires it to be called twice? Is it in case there is an interrupt during the first call? Does the VDP interrupt flag need to be cleared by the first call before the interrupt is actually disabled in the second call? Thanks! (And, by way of introduction since I don't believe we've ever spoken directly before: I was just appreciating your winning IOCCC submissions last week. All of your work is very impressive!)
  5. This is a great tool, @Tursi, thanks for sharing it with the community! Two colors per 8 horizontal pixels in bitmap mode is a harsh limitation. I wonder if there would be any value in trying to minimize error by performing the dither eight separate times, shifting the input image one pixel horizontally each time? You could keep track of some total error metric for each version and choose the one that did the best. It would also be neat to be able to offer up some variable number of sprites to spot-correct the highest error areas remaining in the image after the initial conversion! I tried to build it from source, but it appears there is some dependency on a closed-source library that doesn't seem to exist anymore? Outside of IS40_BrightnessHistogramEqualizeImage, it looks like it wouldn't take too much effort to migrate from ImgSource over to the stb_image* single-header/public-domain libraries.
  6. Those are great and that battleramblog site is quite a resource. That last yellow knock-off version is something special! hehe. I suppose this whole mystery gets a lot easier to unravel if we allow talking between the INTV and CV folks. If the Intellivision artist just grabbed their own photo of the toy from a low-angle and based their artwork on that... and then eventually Mattel passed that game artwork along to the team doing the CV port, it explains away all the similarities in one fell swoop. The timeline is still curious. Do you happen to know roughly when that unreleased Intellivision IV shot was created? It might have gone the other way around: the CV artist could have drawn it first (which would explain the big paper version) and then the INTV4 artist could have adapted it to the more capable hardware along with restoring/cleaning up the more stylized bits. One thing is certain: we know that the castle in the original Intellivision game did not participate in the source of this artwork...
  7. I don't know, the Filmation version seems quite different to me. After a lot more digging, at this point I'm convinced both dev teams were given the same low-angle photo or mock-up/painting of the Castle Grayskull toy; probably a color guide given to the overseas factory for the paint masks (and then subsequently ignored for cost-cutting measures when they were told to paint it free-hand 😂 ). The most readily available assets that Mattel would have at their disposal wouldn't be frames from the Filmation series, but factory prototype pictures. Going from the assumption that the INTV4 and CV versions were drawn/digitized by separate art teams from the same source material, it's interesting to compare/contrast the creative decisions each artist made. The biggest differences I see are: The lower-jaw design: CV uses a more stylized and idealized set of teeth + tongue. The left-side tower in the background: again, the CV version is a little more stylized and less perspective-correct. The right window isn't drawn quite as straight-on in the CV version. Beyond that, the similarities are everywhere. Here is a side-by-side comparison after a little aspect-ratio stretching: Using the annotations I added to the CV version, here are some details that both artists chose to mimic almost identically: Extra-wide, puffy (almost two-layer) "cheekbones". Upper canine teeth that form the top half of the door-frame. A concentric repeating tunnel pattern. A little side feature that juts out at the right (and strongly resembles the roof and building profile you might see if the castle toy were partially open). In both versions, these areas use a distinctly different color from the rest of the castle, which again, feels like the decision you'd make if you were looking at a factory color guide. This "X" shape (which is actually just a rock pattern if you look closely at the toy). The pattern on the top of the "helmet". These wooden posts under the window (and above the skull). They get a lot of attention, especially in the CV version with a dramatic color choice. These thin blue/purple stripes along very particular tower features also feel like a detail that would be present in a color guide (but didn't quite make their way to the toy because of the extra work it would have taken). If you compare the toy's lower jaw to the INTV4 image (where the artist appears to have followed the source material more closely than the stylized CV version), you can actually count/match the exact tooth pattern. In 2013, Mattel showed a new version of the castle at a toy fair. It appears that this version was inspired by the original mock-up: In some ways it resembles the videogame images more than the original 1980's toy castle does: the distinct coloring at the bottom (#5), the high-contrast wood posts (#8), and the side-roof (#4) are all rather prominent, even if it reimagined-away a few of the other details. The left side tower and feature #3 still make me think this was a technical drawing or mock-up more than a literal photo. (That would also account for the "sad" eyes we don't see in the final toy.) But whatever it was, it was used by both game artists. That said, the Filmation version doesn't have any of those features (or even a lower jaw), so I'm guessing it wasn't used by either game artist. Thank you for coming to my TED talk. 😅🤣
  8. Those are way too close to be a coincidence! (The third image down in this topic shows the pixel-grid sketch that the CV version was based on.) The most curious thing for me is that this isn't quite the same perspective for the castle as seen during the beginning/end credits in the cartoon. I skimmed a half-dozen episodes to see if this particular shot is used anywhere but didn't find anything. This has got to be some key art that was used somewhere before. Google Image Search wasn't very helpful either. A lot of comic book covers feature the castle, but none that I spotted used that particular vantage point. The closest I found was this promotional image for the 2021 MOTU Revelation series, which looks like it was based on the same original key art, only a little more loosely: The perspective there is close and I'd be willing to bet the artist for this piece started with the other one as a reference. One of the most distinct features in the older images (not seen in the Revelation version) are the "sad" skull eyes. Every other picture of the castle I could find has the more traditional eye socket look like the one above, but both the CV and Intellivision IV images are quite unique. Does anyone know the source for the original key art used for these games?
  9. If it's limited, I haven't found the limit yet. It seems like you can press the button anytime you like and use it as long as you like. Those non-flying screens seem very easy. In the video I intentionally let a couple of the fireballs hit me to demonstrate the knock-back effect, but otherwise picking the top or bottom of the screen and just walking forward seems like the way to go. In the videos of the Atari and Intellivision versions I've watched since then, the effect on that screen doesn't look like an energy field around your character. Instead, it looks like he's just holding a shield out in front of him. The only thing I haven't tried out on those screens is getting hit by the dagger (sword?) that Skeletor throws periodically. On the CV version it blends in with the fireballs and is easy to miss, but on the other platforms the art for the sword is much more distinct. I'm guessing that can't be stopped by the shield and presumably on harder difficulties he throws more of them?
  10. This is great work, thanks! I like that you can re-map the controls right inside the app. I was just showing my kids a few Colecovision games yesterday and was lamenting that the controls in most CV emulators (that I've used, anyway) are buried in a configuration file. JollyCV solves that problem: if I can't remember the controls, I just set them again! One request: I was looking for a way to reset, checked through all the menus, read the README, and only just saw your F1, F2, etc. instructions in the top-post on my way back here to ask if it was possible. Including that list in the documentation or (even better) on-screen while the Tab menu is open would be great! Thanks again.
  11. How about a video? 😃 (YouTube kind of butchered the quality.) EDIT: Hmm, it looks like YouTube made it only 30fps so you can't see the 60Hz ship-damaged effect in the 2nd flying level. Here is the original (100MB) MP4 in full quality, captured at 60fps.
  12. Hmm, in the two years since this reply I seem to have forgotten I'd already asked here, so I just re-asked it over in the 2600 forum a couple hours ago before I noticed this one again. But the answers here (that the -5V supply pulls 0mA and that pin 8 on the controllers would be at -2.5V) both seem at odds with the schematic: (Here it is, live in the simulator.) The intersection between R1, ZD1, and ZD2 is definitely at -2.5V, but the pin 8 controller GNDs are just a hair below 0V at -63mV. And instead of 0mA, there should be a steady 43mA pouring mostly through ZD1 and R1. So, if it is for Coleco controller compatibility, it's a tiny <0.1V tweak that comes at the expense of ~200mW constantly being burned. For Atari joysticks, it seems like the -5V rail could safely be omitted. But I still feel like I'm missing something. Maybe I should get my Expansion Module 1 back on the bench and test the power draw, too. I wonder if fieroluke's was behaving out of the ordinary?
  13. I'm also curious about this ~21 year old question. Poking around in @ChildOfCv's excellent schematics (and comparing/contrasting with the original 2600 using some 2600A Rev 16 schematics I found on the RetroSix wiki; the last image on that page), it looks like both the +12V and -5V lines from the Colecovision expansion port are being used for something. 12V The +12V line is pretty straightforward. It looks like the color adjust pin on the TIA has an input range just a little wider than 5V. Here is a side-by-side comparison of that part of each, simulated in Falstad. The +9V on the 2600 side is from the unregulated power supply input. The +5V in both come from a 7805-like regulator (even though Expansion #1 already has a regulated 5V supply coming in, it still includes another regulator from 12V to 5V). So the trim pot will swing from 0V to 5.9V (instead of maxing out at 5V if they'd just used the 5V supply directly). Without adding a small boost converter someplace, this one looks like it's actually required. -5V This one is a little weirder. There's nothing analogous to it in the original 2600, so there isn't anything to compare it to. The only place it looks like it's being used in Expansion module #1 is for the controller ports: The 1N5222B's there are 2.5V Zener diodes. And pin 8 of the controller ports are supposed to be GND. Simulating this in Falstad, instead of getting 0V on both pin 8's, they're seeing something like -60mV, with 18ohms of impedance to ground, which seems strange to me. Worse, between ZD1 and R1, this circuit is always burning ~200mW for nothing. If you just disconnect the -5V supply, you save those 200mW and the ground pins on the controllers actually see 0V (granted, it's through R33 which isn't great). Am I missing a reason why you'd want the controller GNDs to be slightly negative? I know the hardware designers back then were analog wizards, so I just assume everything is more clever than it seems, but seeing a redundant 5V regulator makes me wonder a little.
  14. The RF board isn't required. (I did most of my testing during the TMS-RGB's development with the RF board removed.) So, at a minimum you can rule that out as being part of the problem.
  15. Without checking my notes, this is close to what I remember. I think one of the smaller ICs might be able to be left out, too, because it mostly only had to do with giving the sync signal a sharper, stiffer output than the sync stripper was capable of. I was even wondering if the BOM couldn't be collapsed another line or two by using a THS7314 (or '16) on both the input and output. But then you'd need to find a separate solution for the sample-and-hold, defeating most of the benefit. I'm still putting myself through a kind of electronics curriculum--reading textbooks in order (E&M, circuits, and now ~60% done with electronics) and doing all the exercises along the way. (The basics finally don't scare me. I think the most important thing I learned is that it's Ohm's law all the way down; all of the more advanced analysis techniques seem to boil down to just applying it repeatedly until every problem is solved.) 😂 The component version of the board is still on my radar, but my self-learning progress is creeping along a little slower than I'd planned. After the general books, I have a few others that are more specific to analog video (HF noise, Poynton, etc.) that I'm going to skim explicitly for TMS-Component info. And then I think I'll finally be ready to do a nice job.
×
×
  • Create New...