kevtris

He is right, I never had a genesis. I had a SNES though. Friend and I swapped systems for a month or two; I got his genesis+sonic and he got my SNES+super mario world. Up until this year, about the only things I played on it were sonic games.

hah. I try my best!

They sure do, I've witnessed it on the discord chat so far. I still am trying to figure out what I ever did to them to result in such negativity and hate from them.

I didn't post anything because I just assumed it would've been deleted with extreme prejudice. I had plenty I wanted to say, but I would've just been labelled a troll and a hater, and didn't want "competition" <cough>. Obviously none of that is true. I ended up talking to bmack on dischord and he was nice about it.

They got peoples' money, no need to talk about it any more. It will remain to be seen if they produce anything in the next 5-6 months.

Interestingly The Opcode guy thanked me in a private message for the info and I'm glad he got something out of it and saw it before it got nuked. Without constructive criticism, projects get stuck in a hug-box echo chamber and then people can lose sight of the big picture. Sure, trolling is not a good thing and isn't constructive, but calling actual criticism "trolls" when it is not helps absolutely no one. Once your project gets a wider audience than 1 sub forum, the narrative cannot be controlled. I saw lots of complaints about youtube comments/videos in that thread. If those people only got to see the hate that I receive all the time from youtube and other sources, they'd blanch. I just take it in stride and don't worry about it. I am always happy to receive well thought out constructive criticism, and I do not need to be protected from wrongthink. There was a lot of mister talk in this thread awhile back but I didn't say anything or complain about it because people want to talk about it. It's like going to a Coke convention and then talking about how Pepsi is so much better, and you shouldn't drink Coke. I bet if I was talking about the virtues of my 1080p HDMI Coleco core in the Phoenix thread, it would've been nuked from orbit in literally seconds.

Did my post on that get deleted? I didn't notice. If so, that was a douchey thing to do, because I was just being honest and trying to save them a huge headache down the line when they were getting hundreds of parts only to find out half of them don't even work, because they were pulls sold as new, and/or factory rejects that were remarked, or simply not available. It's pretty difficult to buy 100-1000+ parts that were last made in the 90's and are otherwise very uncommon and get known good product. Unless it's something common like a sound chip or RAM chip or whatever, a single sourced Yamaha VDP is going to be nearly impossible to source in any kind of production quantity. Heck, I have problems getting things that went end of life 6 months ago, let alone 20-25+ years ago! If you want to see how bad the counterfeit/re-mark chip industry is, just check out ebay. Search for AY-3-8910. Any chip that is laser marked is a fake. Let's see. I see dozens of auctions for faked chips. The first results are some chips that have been painted black and a new marking stamped on. Yeah, I think GI is still making chips in 2013! (they have a 2013 date code). Seeing how GI was bought by Microchip in the early 90's, this is hilarious. And another that has been painted black and laser etched with a Microchip logo, and a 2016 date code. I think these chips were last produced some time in the early 90's, and did carry a Microchip logo, but it was always stamped and never laser etched. This is just 1 chip, now multiply that by all the sound/video/etc chips on the market. Then again, I only have had 25 years experience with the whole "making electronic hardware and selling it" racket, so maybe I don't know a lot about it. I've made and sold thousands of hi def NES adapters and all of our products at work as well and several FPGA videogame systems that have hit the market. Clearly I don't know what I am talking about!

I wasn't going to say anything about the system, but since people are talking about it here, I might as well. I think 480p only was a mistake- TVs and monitors have poor scaling and when I am testing video, there's a really obvious difference between 720p, 1080p and 480p. The latter looks blurry, like your glasses are dirty. When you do a direct A/B compare, going from 480p to even 720p seems like you've cleared the fog off a window. All the hostility against people who brought up the inconvenient truth that it was 480p only was bad for the cause. Calling such people "trolls" was disingenuous at best. It was fully reasonable to ask questions and raise concerns about it, and beating those people down for no reason other than their opinion was unprofessional and poor form. Also, showing Atari 2600 running on it, but only in VGA form and not HDMI was sort of deceptive, because there's a whole frame buffer system that has to be created before 2600 can be displayed over HDMI (unless you stick to games that output a single scanline count, and/or try to get your monitor to accept different frame rates, and don't mind it blacking out every time it changes). The system has no VGA port (though technically one could be added via the expansion port) so showing it running on a video output the board doesn't have was as Dave Jones of eevblog says.. "a little how ya doin'". Speaking of frame rate, I don't think the system outputs an exact 60.0Hz frame rate either since it's basically raw digital VGA f18a video being transcoded into HDMI, and I don't think this outputs an exact 60.0Hz frame. If you do not output 60.0 exactly, there are plenty of monitors and TVs that will choke on it and give a black screen or similar. When making a product that will be used on potentially thousands of different TV and monitor models, it must conform to the HDMI standard as closely as possible. You MUST borrow or buy an HDMI analyzer to ensure compatibility. I am not knocking their system or the creators of it at all, and I would not have said anything but it was brought up here. This is all my opinion and everyone is free to disagree with it :-)

amusingly the countdown was based on your browser's local time, so when it hit 0 for me, others still had 1 or 2 hours left!

What does the CPLD do? if you are combining the chip enables, you can simply AND all four together and then use A13/A14 on the cart edge. A 74LS08/74HC08 ought to do it.

yeah I never had the skillz or time to improve my web page. Those pics are so old they are pics of a CRT. The first pics date from 2004 when the project started, to around 2006.

I still got my day job but yeah Analogue's pretty much my full time thing at the moment. And yeah, it is a hell of a pace. haha.

Sorry 'bout that. I have been working non-stop on the mega sg since earlier this year. I am doing 12 hour days, 7 days a week. There's only one of me so unfortunately so I haven't had a chance to go back and do anything on the those yet. I still want to port the intv core to the nt mini and do some other fixes as well but it will be awhile yet.

That's not the way I operate. If I were to do TG-16, I would reimplement it from scratch.

You can press A+start to restart at the last world where you died. So if you croak in world 4, you can restart there by holding A and pressing start. This will always start you out on world N-1, though, but it is a pretty decent way to "save your progress" so to speak, vs. going back to 1-1.

That, and when the SNES port was finally released, it had a lot of bugs and glitches. Like what looks like the palette reloading during rendering which causes bright flashes for a frame between most screens. This really detracts from the immersion level and would've been an easy fix to make. There's other glitches like the screen randomly scrolling into place when you enter some screens, and the lack of smooth scrolling on the lava. (Granted that last one isn't a glitch, but it would've been trivial to add it and make things look soo much nicer.) Not to pee on anyones' parade or anything, but I kind of expect a debugged game for the asking price. There's several videos up about it on Youtube, but most of them are of the "OMG new SNES game!" variety, and seem to focus on things like the box art and manual, and other things that aren't related to the actual game itself. Once that wears off, what's the actual gameplay like? So far, no one has mentioned any of the obvious glitching or questionable play mechanics such as having to keep a button down to use the torch in the reviews at all, which is somewhat suspicious. Again, I am not trying to be negative but for the money and all the extra time it took, I was expecting a bit more.

Yeah glad people still care :-) I've been hard at work on stuff for quite a while now.

haha I wished I could tell ya but I can't. not yet, anyways.

A lot of companies made chips for Atari. They seemed to like AMI though, which is American Microsystems, not to be confused with the BIOS maker AMI. They also used IMP, Synertek, and later OKI, UMC, and others fab their chips as well. VTI was used by plenty of 3rd part game makers such as Activision for their bankswitched ROMs, the Pitfall 2 chip, and the RAM+ carts made by CBS. Several companies were making the custom bankswitched ROMs contemporary at that time, such as the F8 and F6 bankswitching schemes. Atari used VTI for the 7800 Maria. If you count Tengen, Motorola seems to be the main supplier for the ROMs and mappers and "Rabbit" lockout chip clone on their NES titles.

The issue with portability has more to do with how the hardware varies vs. the actual "core" part. You have to map the pins of your FPGA to your core's inputs and outputs. If the exact same hardware isn't present, such as RAM chips or DACs or whatever, you have to adapt the core to the new hardware so those things will play nice. It is much lower level than a software emulator running on some hardware; it's more akin to writing an emulator in assembly that directly interfaces with hardware vs. having some kind of interposing layer such as a video driver or sound driver. It's more like porting a game between the SNES, genesis, and playstation. I have written my own "framework" for all of my FPGA cores, which makes porting them to new hardware a very simple procedure in the future if it needs to be done. It has a standard set of inputs and outputs, and these connect up to some stuff that sits between the actual outside world hardware and the internal "core" part which makes it simple. This allows me to change parts and still retain the easy to build aspect of it. As for re-synthesizing the core, this is a red herring, because to port an emulator to a new system you end up having to recompile it anyways. An FPGA design is no different in this respect. The hardest part of porting is probably changing between FPGA vendors (i.e. altera/intel to xilinx). Because there tends to be vendor-specific things for PLLs and blockrams. Though I tend to use verilog only blockram instantiation vs. their "mega gizzard" (mega wizard) version, unless it's required, such as when you need true dual porting. For simple dual port or single port RAM/ROM, you can instantiate it in verilog just fine without using vendor-specific stuff.

Somehow, I don't think $500K is enough money to pay people and pay for prototyping of a full custom modern spec Intel based motherboard. Along with all the plastic molds they have to make for all the various modules (at least the tops), controller ports, the base unit itself, and all those controllers. That's probably 50-60 grand or more in mold making right there. I guess it's possible molds are already done for that stuff but I don't think they will be able to use them as-is-- at least for the base unit, so that will include money to get it retooled or redone as dictated. Then there is the whole thermal design issue, getting 35W of heat out is going to be tricky, so they will need some custom cooling solution most likely to go with it. They probably extended the deadline because they realized 500K is definitely not enough money to get all the NRE costs paid for (non-recurring engineering) and make the product too. Even if so they are still in the hole to their investors on top of it all, who will be first in line to get any money they hope to make as a profit. To recover 500K in investment money will require them selling several million dollars of product at least, before they will actually recover a penny of profit themselves.

That's a 4 phase SMPS buck regulator, commonly found on PC motherboards and video cards. That one seems to date around the athlon 64 era if I had to guess. If you check out athlon 64 mobos on google, you will see a lot of similar setups. They tend to have 3 phases but I saw a few 4 phases. It could be a year or three newer but it's from that era. That has toroidal inductors, which went out of favor for the box style ones, as inductance went down and switching frequency went up. It could also be a VRM module from a server. I did a quickie google search and found plenty of similar things but nothing exact. The other piece of PCB has a single phase buck regulator on it. I suspect if it was a mobo or video card they were careful to only take the most generic power supply components and not have anything too identifying like a CPU socket or GPU. In fact there's not really anything other than power supply in that "videogame system". All four pieces of PCB are just power supply parts. There's no chips of any kind other than power by the looks of things. I like how they used a band saw or similar to cut the PCB up to make it look broken, and the switches and stuff don't have any wires or anything attached to them.

After talking about it some more, it seems that the size of the box has grown significantly too. Their product pages claim the unit is 31*31*21cm now which is about 12.2*12.2*8.2 inches. Going by the renders, I measured the pixels across the front of it, and it was 500 pixels wide. the CDROM slot was 278 pixels wide, giving a ratio fairly close to 1.8:1. Assuming that a CDROM is 12cm, and giving 2mm of "slop" on the disc to make it easier to insert, we can calculate just how big this thing has to be. So if we assume the slot is 12.2cm, multiplying this by 1.8 gives 22cm (8.66 inches) almost exactly for the width. This means each pixel of the image is approximately .44mm. The height of the thing in the render is approximately 130 pixels, so this means it's 5.75cm high, or about 2.25 inches. From all that, I am going to say the size of the box is 22*22*5.75cm (8.7*8.7*2.25 inches) in the renders. This is a volumetric increase of 7.25 times between the render and the dimensions listed on the website (2783 cc's to 20181 cc's). edited to add: checking out the size and weight of a PS4, the Polymega seems to be awfully close to it. The 'mega is 31*31*21cm and 6 pounds The PS4 is 27.5*30.5*5.3cm and 6.1 pounds The PS4 pro is 29.5*32.7*5.5cm and 7.2 pounds The PS4 slim is 26.5*28.8*3.9cm and 4.62 pounds. I hope my math is all right. I sourced the render I used here: https://raginggazebo.com/is-the-polymega-dead/ and the 31*31*21 number is right off their product page here: https://www.polymega.com/product/polymega-base-unit/ It's under "additional information" Also listed is the weight - 6 pounds!

Yeah that's probably a reasonable guess- making something that lines up with the SOC memory that would typically be used. They've strayed pretty far from the original gameplan and are trying a hail mary pass to see if they can make a final go of it. I just think it's too little, to late. The current design is just a wishful thinking exercise IMO. I would like to see how they are going to remove 40W of heat from that little enclosure. Their previous version at E3 was overheating and was probably running on their original specs (SOC). A desktop CPU in that small box with no cooling will make a dandy coffee cup warmer I guess if nothing else.

After a bunch of WAG's and speculation, the current theory is that their hybrid emulation is as follows: That FPGA board edge footprint matches up with a standard DIMM form factor, and this board was meant to plug into a desktop motherboard's DIMM socket. The board has two ribbon cable connectors which would've connected to the element modules. It appears there's an SPD EEPROM (serial presence detect) near the right lower corner of the board, or something else similar in an 8-SO outline. The chip is a standard EEPROM, so including one wouldn't be difficult. The ultimate goal of doing this, I think, was so that they could access cartridge memory, as if it were the computer's own memory by making what amounts to a cart to PC "bridge". The FPGA would translate the CPU's memory reads and writes into cartridge reads and writes. This sounds like a good idea in theory, but I can think of a few reasons why this still wouldn't work very well. The main reason is still speed- the carts are incredibly slow in relation to the DIMM it replaced, and I think the form factor looked like DDR2 or so (I'm too lazy to check). A modern mobo wouldn't have this, but they could be using an older mobo for this, for testing and development and possibly expect to make a DDR3/4 version at some point in the future. Even if they get it to work the speed issues and more important caching would be a non-starter. If the CPU caches the cart data, that means anything with bankswitching probably wouldn't work because the underlying memory would be changing and the cache wouldn't know. If all those problems could be solved, the chipset/CPU's memory controller would need custom programming, which means hacking the BIOS and a bunch of other scary and hairy things would need to be done to make it work. Again all speculation but it's the best thing I can come up with on re-examination of the evidence presented. eta: thanks to those on discord including Drag The Waters for his theory on having to change the BIOS to make it work.