Kismet

I'm not sure I'd call those done right (three revisions a piece), but a licensed clone is at least held responsible for problems. Unlicensed clones, are simply made to scam people out of their money, anyone selling them, assumes 100% of the risk, as the (chip) manufacturer is not going to be in business very long. Which is why Hyperkin (Retron) and Cybergadget (Retrofreak) both switched to SoC's and then stole open source emulators to run them. This is where the AVS, NT Mini and Super NT simply beat the ASIC clones at their own game. Sure they're not as cheap, but the FPGA implementations can actually fix bugs.

Some (presumably chinese) chip fab out there produces the same chip for all the clones. That's why every single ASIC clone is the same, and has the same problems. The build quality of most, if not all cheap clones is also abysmal with sloppy solder, jumper wires, and glue everywhere. The chips you see on SNES clones are TCT-975(APU), TCT-976 (PPU's), TCT-978 (CPU), there's an entire thread from 2015 on it over http://www.sega-16.com/forum/showthread.php?30576-Yobo-FC-Twin-SNES-clones-chipsets-questionsif you want to see what chipsets there are. There's basically 3 generations of chipsets. With the Retron 3 basically having the last version. Suffice it to say, any ASIC isn't going to get a revision. When they run out, they produce more. It's more likely The RetroFreak/Retron 5 is what we will be seeing more of since it's cheaper to buy a cheap ARM SoC than it is to produce more ASIC's. If you look on eBay you will see plenty fake NES Mini Classic and SNES Mini Classic's with those 9-pin controller ports on them.

Load the 240p suite if you have a SD2SNES or Everdrive and use the color bars ore the Gray ramp to find out. If it's set wrong, you'll not see the bars on the end correctly (eg you'll see two whites and two blacks) or if it's set the wrong way in the other direction, what is supposed to be black will be grey,

It's probably the scaler in the monitor/tv. 1080p to 2160p is a 2X simple scale, but 720p to 2160 is a 3X simple scale. If it's slightly blurry at 720p, then it's also adding a linear filter to the scaler, and that would basically be a second step.

This appears to be a SD2SNES compatibility issue. Looks like the SD2SNES hooks got turned back on and that's what breaks the game. (I should have clued into that when the clock was reset)

So in picking random things to see if it works, I found one that doesn't make sense. In the smokemonster list, there is a title called "Brandish 2 - The Planet Buster" under JP , it does not get past the KOEI screen on the SD2SNES on the Super NT 4.3 firmware. When I put it in my SNES, it ALSO didn't get past the KOEI screen, the first time. But I hit reset and it went right to menu screen in the game. "Brandish 2 Expert mode" also does the same thing just a black screen after the KOEI logo. Same with the fan-translation. I was only able to get the game past this screen once, on the real SNES.

The problem is that the issues with the JB are separate from issues with real carts, and issues with using the Everdrive/SD2SNES, some which don't happen on the real hardware, and some that do. The ED/SD2SNES has different performance characteristics from real cartridges, plus they have a "program" that runs before the games do, so that means the default state (superCIC, hooks, etc) of the SD2SNES can also impact the games. Hence why anything perceived broken has to be tested with the real cart, and then has to be tested in a real SNES to eliminate the ED/SD2SNES from being the issue. eg Sailor Moon S (Fighting game), appears to have a scan line at the top and bottom that is broken. But... It actually shows that on the SNES too. What we need is a spreadsheet checklist of something like:

On 4.1 I had to enter the house 75 times before it crashed with the real cart, and the Japanese version on the SD2SNES. The (smokemonster) version on the SD2SNES only took like like a handful of tries before it crashed or rebooted on 4.1 On 4.3 I haven't been able to replicate the crash yet, but I'm going to assume it's fixed if I can't crash it within 5 minutes like before.

The file browser is in the "Load Custom Pattern" menu. The SD card is used to save pattern, so the speed makes sense since the card can be used. So no surprise at all.

It's more likely that they just captured the video since they are equal brightness. To give you an example: This is the raw capture from the Super NT with the SD2SNES running Mario All Stars. This is the raw capture from the Super NES with the real Mario All Stars cart. (HD Retrovision cable) This is via S-Video, same SNES, same capture card. Please note that the capture card still uses BT709 for S-Video 720x480i This is via Composite, same SNES, same capture card. This is what it looks like running at the same time (The 4K is on the left running the capture full screen via HD Retrovision cable, the SuperNT is on the right.) On a CRT, you'd see the aperture grille. The more likely thing is that the capture device or camera used was not aimed at a CRT but at a LCD running composite.

Yup the capture card correctly sees the Colorimetry and color range. Super Metroid RGB32

That's certainly a testable scenario (input latency), but i somehow doubt it's more than one frame. I ordered a SGB2 a while back and it's still in transit.

You're not getting it. If it was kevtris, that would likely put him in some breech of contract. NDA means NDA, and since the amount of people who know the design of the Super NT is exactly one, that would be a problem. You're not going to see additional cores "hacked into" it, because they don't exist in the firmware. All the evidence needed to suggest someone else dumped the firmware is there. There is a JTAG port, there is a USB port, even the cartridge port. However given the time frame, it had to be someone with a review unit. If kevtris actually was the one who handed it to smokemonster, I'm pretty darn sure he wouldn't hand a broken firmware to him. Broken features is the trademark of someone who hasn't actually tested it. Seeing as no "fixed' firmware came out, that also tells us that whoever released it, does not have the skill to fix it.

I assumed NT was for "New Technology". At any rate there is at least one person in the thread that is obsessed with naming things not yet produced, so let's not go there. I just want them to avoid the insane naming scheme current devices are which is basically "same as the last name, plus buzzword", because nothing confuses people more than devices that use two brands, or model numbers that don't go in order. Beanie Babies are not consumer electronics, a "better one" never comes along. Once they're gone, they're gone, and replaced with the exact same thing next year in new hat, like Barbie. My opinion is that the NT Mini's production is done, there won't be another Aluminum case run, and if there is another run at all, it will likely be on the Super NT more powerful design with the analog output as an accessory. Yes, they could do another run of the PCB, but you have to consider the ability to source the parts. They'd obviously have to cut the 8bitdo controllers from any future run, and from the manual since they've removed those from sale (or simply renamed them.) More to the point though, the AVS already fills that niche if the NT Mini is out of production. So it would be Analogue's best interest to focus on another product, and come back with a multi-cartridge design (eg Sega MD, Super NES, Nintendo NES/FC) that includes the NES/FC support. Who knows, maybe Analogue will just sell nothing but Super NT's this year and then the next thing that comes out is based on the Cyclone 10 which might be cheaper in quantities by then.

While I certainly wouldn't have minded getting the NT Mini when it was last in stock, my main interest was in the SNES first. If the Super NT gets any additional cores/pin converters I'd rather go that route. My opinion, is that they likely retooled for the Super NT production so it's probably done. There's no point doing another production run if it's more expensive than the Super NT. If they do another run, the next thing I expect is a 3-slot Super NT, which has the cart slots for both the FC/NES and SFC/SNES, and uses the SNES controller ports. Or... Have Analogue gun for the Retron5 and produce a system that can do all of that, and not be hideously ugly.

There's more than one way http://www.nirsoft.net/utils/dump_edid.html and https://developer.amd.com/amd-edid-utility/ and several others. You usually need them for screen-replication type of setups.

The NT Mini doesn't have a physical reset button does it?

It will depend heavily on the screen. My 4K monitor seems to be fine regardless of the input resolution as long as the input resolution is 720p, 1080p or 2160p. If it switches to 480p I have to go into the menu and change the aspect control and overscan settings, otherwise it seems to default to "fill the screen" That said, the scaler in this monitor seems to have a slight anti-aliased look to it at 720p, where as the 1080p looks like it's straight integer scaled. So if anti-aliasing is being invoked, it's likely has slightly more latency than straight integer scale. (Integer scale just duplicates the next pixel, where as anti-aliasing has to mathematically take into account the surrounding 3 scan lines at least.)

What I would like is a separate monitor/device and cartridge profile. eg the Super NT remembers what monitor it's plugged into and that becomes the "base" profile, and then the cartridge profiles (eg read the cart checksum) override the base profile. Though it might not really be possible to do that with plugged in carts.

They probably don't know where the screws are.

It's probably the memory bandwidth. Like, the SD2SNES has a XC3S400-PQ208 FPGA, with MT45W8MW16BGX-701WT/IT as the memory chip which is 104MHz PSRAM not the 144Mhz and 166Mhz SDRAM in the Super NT.

I think the problem here is the question. I don't know what the question was asked, but it was likely "could the Super NT make chip games work with my everdrive" or something to that idea.

If you learned how to do something on X hardware, and along comes Y hardware that is a port of the same game. Try playing any fighting, shump, or platformer on the arcade hardware, then try to train on the home hardware. When you throw software emulation on top, that home port becomes even worse. Today, there's even crappy mobile ports that throw the entire idea of accuracy out the window. My point is that a FPGA implementation of the 8 bit and 16 bit systems is something worth doing, especially against a HDMI target, but many of those home computer systems did not have the kinds of games that arcade machines had, and many home computer versions of games only targeted the slowest version of the system.

It's probably not "easy" to RE, it's just unlikely, given the timing, the brokenness, and the source. If I were kevtris, and I was under contract to provide firmware updates to Analogue, and not to make it possible to run games from the SD card, then I would take the bare minimum required to do that, while not making it painfully difficult to troubleshoot problems. You're not going to keep making parallel versions of firmware, no you simply change your build pipeline so things like extra cores aren't included, but that doesn't mean the firmware didn't have "run from sd card" in it already, just hidden. Which is easier to do? Maintain two versions of code, one with all this other stuff in it, or maintain one version of the code, and just have the optimizer or pre-processor "no-op" the functions out. I don't actually know if someone could pull the compiled FPGA core out of the firmware and produce their own clone with it, that would require producing a clone of the hardware that is exactly the same as the Super Nt, and thus would cost about the same. Even if someone produced their own firmware and removed the cartridge slot to save a few pennies. To that end, we're just going to keep speculating who, and why. Sure, it's entirely possible it could be kevtris, who slipped the source code on a usb drive under smokemonster's door anonoymously, and all he had to do was hit compile. But that's pretty stupid. It's more likely that the firmware was downloaded via the JTAG, and the decryption key/algorithm was pulled directly off the device. It's been done. http://www.zdnet.com/article/cryogenically-frozen-ram-bypasses-all-disk-encryption-methods/ It's also possible that someone just desoldered the RAM and the flash from the board and dumped it that way. We don't know, and that's an awful lot of effort. But someone who got an early release of the Super NT certainly could have done that. kevtris mentioned before that the Super NT won't allow the Mini NT's firmware to be flashed onto it, so there's clearly it does some kind of check, but it could also just allow arbitrary firmware blobs where it identifies itself as SuperNT firmware. In other hardware, they usually prevent you from downgrading. The Super NT doesn't. So the bootloader isn't dumb.

I don't think software emulation is any better until you get to hardware that is off-the-shelf, because off-the-shelf hardware is well documented. Basically anything that has a BIOS and Operating system becomes a much easier target for dynamic recompilation efforts. That said, a FPGA Amiga, Atari ST, MSX, X68000, PC88, Tandy 1000, C64, or Apple II has a lot of parts that can't be software emulated accurately, but many of the software titles were ported to these machines from another machine originally, so accuracy is not generally thought of as being as critical as say the GB/NES/SNES/SMS/MD/GG/PCE systems. Computer systems, except for the Amiga, didn't have capable hardware, and console/arcade ports were often thought of as being terrible. The MiST project really needed someone like Analogue to actually produce a viable low-cost hardware box that anyone can use (eg with a USB keyboard, USB mouse and HDMI screen) and that's really the downfall here. However the larger obstacle to that was the fact that it's not possible to produce any kind of "SD-card"-only emulation box, software or hardware that isn't meant to infringe copyrights with impunity. So for a PC/Atari/Amiga/Mac/AppleII/etc all of these have a BIOS, and some version of DOS or basic OS that are also needed to boot the machine. You can't simply plug MiST into an original floppy drive or hard drive. To that end, you can legally acquire the software needed for the Amiga under "Amiga Forever" cheaply, and that makes MiST/MiSTer have access to two legal sources of machines that could justify mass production. PC and Amiga.