I should clarify that the card does work in a IIgs for some things, but the current design doesn't quite work for IIgs video as it can't tell the difference between 16-bit and 24-bit addresses. This isn't a problem for IIe/II+ software running on the GS, as that software is unlikely to access anything outside the 128KB window the Mega II chip can see. I'm working on a new design that more closely simulates the behavior of the Video Overlay Card with regard to the M2SEL and M2B0 signals. There will probably be some tradeoffs which will make it more of a fork than a version bump. The screenshots I took this weekend were from the Total Replay splash screens of a few games, and GSOS starting up. I would not classify GSOS as really usable yet with the card due to all the memory accesses outside the Mega II window corrupting the card's screen contents.
David, I'm curious if the current design has any unused IO pins on the Pi Pico? (Not only for the IIgs problem, but more-so for people wanting to add extra peripherals. I see /M2B0 is only available on the IIgs in slot 3, but I can't even find info on which pin. What about M2SEL? I found this article on those two signals which might help people understand their function: (Page 9-10) mirrors.apple2.org.za/Apple%20II%20Documentation%20Project/Computers/Apple%20II/Apple%20IIGS/Documentation/Apple%20IIGS%20Technical%20Notes%2065-79.pdf
M2SEL and M2B0 are available on every slot in the ROM 03 GS, but on the ROM 00 and ROM 01 systems M2B0 is only on slot 3. Apple IIgs Tech Note 68 has some of the info, but the Apple IIgs Hardware Gude has the actual pin numbers on the slots. Pin 35 is M2B0, and Pin 39 is M2SEL. On the IIe and II+ these pins had different functions. M2B0 and Q3 (Pin 37) are needed for proper video decoding on the GS, and M2SEL is needed for cards that do their own address decoding outside the select lines for their slot. (An example of a card which does this is the CFFA3000, or the Prometheus VersaCard). Before working on this project I had not done any hardware design for the IIgs, and I don't think Mark Aikens had either before designing the original AppleII-VGA project. When I spec'ed out a chip to combine the slot select signals (what the ATF16V8 "PicoPal" does), I had not accounted for the extra signals needed on the GS.
Ah great info, thanks. Is that PicoPal ATF16V8 still a production IC? It was my understanding all PAL chips were OEL at this point. (Maybe only the DIP versions that I usually use.)
ATF16V8 is still in production, and is Microchip (formerly Atmel)'s replacement for the PAL & GALs of the same series, it does both Registered and Combinatorial logic and is pin & fuse map compatible with the GAL16V8. All the logic is written in CUPL, which is a very simple language for logic equations, but the compiler is very crash prone and almost never gives you useful error messages when it finds something it doesn't like in your code.
@@GrizzLeeAdams heh ok that's cool. I've used WinCUPL before for GALs and yes, what a hard program to use. That is cool it's still made! And I guess can it be programmed on a MiniPro with an adapter?
It’s so refreshing to see a project where the person is not trying to make money from vintage computing hobbyists as is unfortunately so often the case. Also, an open source project can grow beyond its current capabilities or morph into something else entirely. Thank you David for your generosity!
I am trying to make a little bit from the project, but almost all of the money made is getting funneled right back into making more boards and scaling up the production. A solder stencil holder and pick & place are next on my shopping list. Right now I'm assembling all the boards by hand and doing extensive inspection and testing to ensure every board shipped is to the highest standards.
@@GrizzLeeAdams There are people out there who need plug-and-play, you cater to them and you finance your hobby. There’s nothing wrong with that. I was aiming at those projects that are commercialized, then disappear off the face of the planet when the manufacturer has had enough. It’s very nice that you documented and made the sources public. I just wish it was the rule and not the exception in the community.
Indeed. There have been many retrocomputing projects that have disappeared when their creators did. One of the Atari retro community passed away suddenly at 53 not too long ago, taking several of his products with him and more recently, one of the SS-30/SS-50 bus community members had an unexpected illness that stopped production of several projects. (He still isn't back in production but has started selling blank boards at least.) What I've taken from these examples is if I were to rely on those products for income, I would include those in my estate planning--either to release those publicly upon my death or to have the IP bundled up and ready for my next of kin to sell to someone else who could either continue producing them or release them outright.
@@wesley00042 same with Amiga 'community'where a bunch of assholes buys IP of as much to profit of it. Or people that have custom IC internal schematics but don't release them.
Not trying to be argumentative or flame you, but what's wrong with making money from the vintage computing hobby? It's not as if this is some sacred calling in which all projects and development efforts must be for non-profit only for the betterment of humankind. We're not talking about saving lives or providing clean water to impoverished parts of the world, it's just a niche hobby where some enthusiasts are fond of old computers. If there's some quality of life improvement somebody put the time and effort to develop, I don't see anything wrong with them selling that for profit. After all, folks can always choose not to buy, since these are not essential for the enjoyment of the hobby.
26:47 for anyone with more modern pc wanting to see this fractal on terminal, here's the python version: for y in range(-12, 12): for x in range(-39, 39): ca = x * 0.0458 cb = y * 0.08333 a = ca b = cb for i in range(0, 15): t = a*a - b*b + ca b = 2 * a * b + cb a = t if a*a + b*b > 4: if i > 9: i = i + 7 print(chr(48+i), end='') break else: print(' ', end='') print('')
Loved seeing the Pi Pico used like this. The basic fractal program got my attention so I figured I'd add a couple more data points :D. I timed the program running micro PyBasic on a Pi Pico and on a Teensy 4.1 micro controller. To appreciate the times you should know that micro PyBasic is an unusual implementation of Basic designed as a teaching tool. It's an interpreter written in Python, (an interpreter itself) ported to run on a microcontroller version of Python. The fractal program took 107 seconds to run on the Teensy 4.1 and it took 1071 seconds to run on a Pi Pico.
@@retiredwizard Snap that's fast! And what's crazy is that the second processor core isn't utilized, at least from what I've heard. Difficulties getting MMBasic to scale across multiple cores or something.
I have to chuckle every time you use 'Drol' as the demo game, because in Afrikaans (a language that I speak and also gets spoken in South Africa), 'Drol' translates to 'Turd' 🙂
I have much stuff from the mid 70's and early 80's from when I was actively doing microprocessor development, and instead of dumping it in the trash, I'd like to send it to you, Adrian.
I bought a scanline doubler - Indivision Mk3 - for my Amiga 1200. It sports an HDMI-output and gives that pixel-perfect output. Makes all my old Amiga games look better than ever.
FYI, In AppleSoft BASIC, the "LET" keyword is completely optional. Also, AppleSoft BASIC only pays attention to the first two characters of identifiers. So, a variable named "BLACK" could also be referred to as "BLUE".
I've been trying on and off to "port" a BASIC Battleship game from Applesoft to GW-BASIC. Is there any discrepancies between the two versions of BASIC I should know about that'll affect coding?
@@Chaos89P Not that I'm aware of. The most common problem when translating from one version of BASIC to another is reserved variables. For example, on the Commodore 64 (and all other Commodore 8-bit computers), TI and TI$ referred to the internal "jiffy clock". Other versions of BASIC did not have an equivalent. That won't be a problem here, though, since neither Applesoft nor GW-BASIC have any reserved variable names.
@Chaos89P Anything involving graphics may need translating, simply because of the graphics being different on Apple ][ series machines and those using GW-Basic. But you can take advantage of the differences. Back around 1979 a friend and I converted a "Star Trek" type space war game from BASIC on an HP-1000 minicomputer to my Apple ][. We went from no graphics and no sound to adding a bit of both. It took a while to type it all in from a hardcopy listing printed from the HP, then translating any functions without equivalents, before adding and improving it. Good luck with it, and have fun!
I like using LET because it lines up the left side of the code. The code looks sloppy without it, with all the assignment statements starting one character before every other line.
@Bill Chatfield I think the one best reason to not use LET was on low RAM systems, where every byte saved helped. But I may be mistaken, it's been quite a while since I programmed in Applesoft BASIC! It's tokenized, I think.
Great video. I've been falling this project from afar for a few months. Thought it was really cool, and I was glad to see that it used a Pico, but I didn't realize David was able to pack so many features into such a small, unassuming card. Absolutely brilliant! While I still love my VidHD card for HDMI output, it's great to see a new open-source, budget-friendly option for getting a decent video signal out of an Apple II. I still think the Apple II series is among the most "future-proof" line of PC's ever made. So much potential for expansion, as this project clearly demonstrates.
Wow, I just watched the new video and I have to say I'm impressed! I love that a VGA output and Z80 for CPM were added to an open-source board. Additionally, the possibility to add more functionality with simple firmware updates is really intriguing. Great video!!
Mainly commenting to appease the almighty algorithm! Thanks for sharing this awesome piece of tech with us, Adrian! I hope everyone is doing well and having a great day!!
Thanks! I did the PCB layout originally with the II+ in mind, but it should be easy enough to use a short VGA extension in a IIe as long as all the slots aren't filled. The future revisions will probably have a notch in the PCB at the back and a separate VGA breakout on a cable so you can mount the VGA connector to the case without running into other slots.
This is incredibly awesome and hopefully it can set the stage for future similar projects enhancing EVERYTHING by snooping and responding as a device to the system bus! Even simulating RAM chips which some systems may be running out of in the near future. SIMPLY OUTSTANDING!
Oh fun, benchmarking basic performance! Couldn't resist typing it in on a few of my machines: Philips NMS-8250 MSX2, Z80 @ 3.58MHz: 851s; Philips NMS-8250 MSX2, Z80 @ 7.14MHz: 409s; Panasonic FS-A1ST MSX turboR, R800 @ 7.16MHz: 145s. My conclusion is that MSX-BASIC is apparently very slow floating point wise compared to the others (at least the Amstrad, with a similar CPU, less accurate?), the R800 is doing not so bad but it's a heavily optimized Z800 variant that does most instructions in a single clock cycle. Thanks for the great videos!
The Raspbreey Pico is an excellent basis for an 8 bit retro computer peripheral system. I've used it myself for a virtual "Winchester" interface for the ACT Apricot computers. Only really enough pins for an 8 bit data bus though with off-board address decoding using a GAL. Oh, in case you want some more BASIC fractal benchmarks: Sinclair QL (7.5MHz 68008) stock: 260 seconds. Sinclair QL (7.5.MHz 68008) with Lightning fast-math SuperBASIC extensions: 191 seconds. Sinclair QL with Miracle Systems Super Gold Card expansion (24MHz 68020) with Lightning fast-math SuperBASIC extensions: 19 seconds.
Genuinely awesome - I'm going to have to make some to play with. Many thanks to all concerned. PS: It would be great if flipping between colour and monochrome could also be handled with a key combination.
Always exciting to see the Apple II line continuing to be developed and loved so many years later. I ran a BBS on mine, so I had a clock card so it could track time, an extended 80 column card with its own RAM (mine had 512K but it had space for up to 1Mb that I used as a RAMdrive) and a SCSI card with an external 10Mb hard drive. I never got as far as installing a zip chip to speed up the processor - but it was an enticing idea. I remember taking my SCSI card and drive over to a friend who had a Laser clone that ran 3.6mhz and marvelling at how fast it could load a gif image from the hard drive rather than writing line by line. The IIgs existed in those days but I would never have envisioned that it would be possible to output graphics at VGA level. Exciting project!
Incredible! The possibilities are absolutely endless with this! Every card that ever existed could theoretically be emulated this way not to mention future cards that haven’t even been thought of yet. Amazing work!
Great card and very clever idea! The same concept would work on any retrocomputer which has full system bus access through some port and the video memory is mapped into the overall address space. In addition, it could be a great system diagnostics and debugging tool too (essentially spying on system bus activity and visualizing it).
I’m so excited for projects like this, especially anything that shows promise for getting IIgs RGB graphics on a more modern display since solutions can be hard to find, finicky or very expensive.
I have some data for the spreadsheet; although more "for fun" than a 1:1 comparison, since the devices in question only have limited displays, I had to "print" it on a 24 character printer and scaled the fractal to 23*15 characters - call it "print mode resolution" :) Epson HX-20: 230 seconds Sharp PC-1403H: about 550 seconds (I missed the last line but it was sloooow) TI-74 Basicalc: 260 seconds
@@adriansdigitalbasement Thanks! I can't attach pictures in comments, but I have a post with the resulting paper strips. The entries might need an asterisk due to the reduced resolution though 😁
Another card to try to get working on my Franklin Ace 2200! Already have a few PICOs laying around and a few of the other components. Looking forward to building it!
It is quite impressive interfacing modern hardware with old to do this, certainly something I could never wrap my head around doing, I tend to just plug stuff together and hope it doesn't go bang!!! :P
That's really cool, I'm impressed how much the conversion improved the color burst mode. I always forget the Apple II has it's own method for Artifact Colors. I think it was a good solution for bitd between text and color images even if it had some speed bumps like needing a switching monitor, or a monochrome screen and a color SDTV. I keep hoping DOS-CGA will have a Composite revival to try and undo the past internet damage from not just the ridicule of it's base colors but also the fetishization of those same colors too in pixel art from everybody not using Artifact Color Mode. 😐
How about adding wifi and some version of uIP and from there you could even remote control an Apple II/etc from a PC, add files, backup the entire system, snapshot it, inspect memory while running, etc! The possibilities are virtually ENDLESS, thank you Adrian from showcasing this project!
I have an Apple II plus frankenstein build with a Unitron clone motherboard that has really bad, noisy composite video, this seems like the perfect workaround!
Yeah! The Rev 4 Apple ][ I used it on has terrible, nearly unusable Composite video (due to it being so non standard) so this just gave that machine pixel perfect output.
David should get in touch with the My Life in Gaming folks. This is exactly the kind of thing they like to cover, though they focus more on game consoles than computers.
I was very confused why you were showing an image of a listing in Sheets...but then saw that you made this into a science project! Thanks for sharing the sheet with us!
This is really cool already, but the fact that the feature list can be expanded on in the future is a very exciting prospect. I already have a composite to vga box that I use with my IIE when I want to use color, but I think it'd be neat to get one of these cards as well. Especially because of future possibilities.
Amber for sure. I have a small amber CRT that I love using with my Apple //e. Would be nice to have a larger amber display and get that one off my workbench for a while.
Ayyy, I love seeing the RP2040/Pi Pico used in things like this, its a very cheap MCU compared to the price AVRs can get upto or even SBCs such as the pi4, Zero or alternatives can get upto.
Having implemented several PC ISA sound cards/chips on the Pico (namely Gravis Ultrasound, Adlib, CMS, and Tandy 3-Voice), I want to get my hands on one of these boards and implement emulation of the Mockingboard!
Nice! I picked up a MS CP/M card for my //e recently and it's been fun playing with CP/M again. But it, like most Apple Z80 cards, is slower than standard CP/M. I'd heard of the Applicard, but as you said, they are pretty rare and I didn't know much about it... Also, have you seen they are making (have made?) a Fujinet for the Apple computers? (As well as Commodore and Coleco Adam). Be interesting to see/hear what you think of Fujinet at some time. (I do have one for my Atari 400 (with 48k RAM).)
The FujiNET is already available for Apples and it too includes PCPI CP/M functionality. It's turns the Apple into a terminal to connect to the CP/M running on the Fuji's processor which is very C/PM-esque indeed.
12:00 It's just ClearType. Look how much more readable the text is when we use RGB subpixels individually to increase the resolution. I'm joking of couse, but could it have been a thing on composite if they really put more thought into it?
The slimmer blue vga connectors like what looks like you have, will just barely slide past the port nearest the PSU, by the way. I was pleasantly surprised after Amazon false advertised having an extender in stock.
Great product! However, the seller has already "retired" the card? Any ideas for where to purchase such a card? (not at all confident about my precision soldering skills!)
Adafruit is now selling a version of their RP2040 feather board that outputs DVI over an HDMI connector. It could be interfaced to the Apple II and do the same thing.
This is so cool, I thought of a couple more thing that may be possible to do on it, #1 IBM PC emulator, ram would probably be limiting, and Mocking board support.
Looks like a cool alternative to the VidHD card. I had a VidHD and it worked great until the microcontroller shorted out and emitted smoke. Nowadays you can’t buy the VidHD any more due to part shortages.
Apple 2 indie developers continually do this for unknown reasons. Hell, even some cards from back in the 80s/90s did this (example: audio cards on the IIGS outputting to a stereo 3.5mm jack... which did not face the backplane but instead internally). Other cards had no backplane connector but instead internally has a ribbon cable that went to a screwed-on backplane connector (think: SCSI and parallel cards). The whole Apple 2 expansion card ecosystem was bizarre in this regard and I've never seen anyone in all my years explain why they do this.
Mainly because the II/II+ has just slots for cables, and they are the most common targets for upgrades. As others have said, on the IIe and IIgs there is an array of D-Subs in 9, 19, 25 pin variants, and it would be an absolute pain to insert a card cleanly that had to match up to a cutout, and would lock down what slot / machine it could go in. The compromise most of us have gone with is internal facing connectors (which was common on the II/II+ expansions) so you can use short extensions to reach an available cutout on the back of the IIe or IIgs. I may start stocking 1.5ft VGA cables and gender changers as an addon option in my store. They could be wedged into a slot on the II/II+, or mounted in one of the DE9 holes on the IIe or IIgs.
Single pixel graphics could be increased by running a convolution algorithm to detect the single lines and only thicken the single lines in the RPi modified video memory map.
You may automate the truck scope from the computer. I have used a quick screenshot, and developed a full characterization test for analog filters. Unlocked DS1054Z, in my case.
Some notes on the C128 performance in 80 column mode. First of all basic 7 is slower than most traditional MS basic variations on 8 bit machines because of supporting more than 64k and all the bankswitching which results from that. But also, the 'screen editor' code for supporting 80 column mode is really really inefficient.
Adrian, I am curious on how does this implement the color fringing filter for the VGA signal. I can see that the VGA signal is properly generating color which is also artifact based (so we don't see the alternating lines as we do in monochrome mode) but it seems to be efficiently discarding undesired noise on single pixel white bars such as used in text.
Pedantry: The Apple II doesn't have a colorburst, the patterned hi-res monochrome is all it can output, but due to limitations of NTSC that is picked up as color information.
The Apple II outputs a colorburst signal when in color mode. It does not output a chrominance subcarrier. The colorburst signal occurs at the beginning of each scan line, and it is used by the TV set to synchronize the decoding of the chrominance signal. As you indicate, the patterns in the monochrome signal are decoded as color information. In text mode, the Apple II does not output the colorburst signal, which indicates to the TV set that there is no chrominance subcarrier to decode, so it decodes the signal as monochrome only.
I'd been thinking about using the RP2040 (the Pi Pico CPU) for doing a bus interface board - it's a trivial thing to get going and doesn't require hugely complicated board designs. Good to know I was on the right track...
One correction, if itvis monitoring bus activity and one ram chip output is stuck, then it bay drag the bus low/high enough to garble the frame buffer data.
Intersting...my IIgs has the "manila adapter with Nec 1550m (blurry but 50hz modes) or actual with Samsung 510m (sharp but no 50hz and GS-OS still not perfect". I need another solution.
Hi Adrian, I've just built one of these from scratch and was wondering where to find the configuration tool as I have some issues with mine and want to determine if it's a config issue or if I need to do some repairs. Love the channels BTW :)
My eyes perked up on the Basic09 score. I assume it had the benefit of being compiled in to pcode instead of having to run through a native basic interpreter like all of the other machines.
So if this card works by monitoring bus activity, could it also be used as a diagnostic card with the right software? If it expects X routine on the bus and sees Y instead, report that? I'd imagine it wouldn't be as comprehensive as a diag rom, but still could be useful to see if there is some bus activity on an otherwise dead machine.
Few comments. No one I know of "ever" used LET to assign a variable in apple basic. It basically does nothing except take up token space. As for boolean. They are possible, just tricky. As an example you could do something like this: A = 0 + ABS(A < 1) Applebasic converts this sort of thing to either 0 or -1, as though it was part of an if/then. Things like AND, OR, etc. however work different based on which "version" of basic is used. The older ones where bitwise, while later ones returned 0 or -1, supposedly. In any case, it is possible, just... quirky and probably not well documented.
What about the comb filter, which was a feature of most high quality composite monitors of the time? For example, the mouth of the character in drol looks like teeth on your composite monitor, but looks kinda like a scarf on the VGA output. I also like to test this with Alice in Wonderland from Windham Classics.
The RP2040, with two ARM cores and more than 256KB of internal RAM, could emulate the entire Apple II not only the video controller. But anyway, I still think an FPGA is better suited than a microcontroller for this application. And in this case too, even an small FPGA would be enough to recreate the complete Apple II. In fact I just did that some weeks ago and was able to sinthesize an Apple II clone with VGA graphics, PS2 keyboard, and floppy disks emulated from an SPI Flash memory (read only). It took about 1400 logic cells out of 7680 in an Lattice FPGA.
I've had more fun with the rp2040 based microcontrollers like the pico and just building boards using the rp2040 than with pi's or other sbc's (aside from the new one I've been using with 16gb of RAM, getting closer and closer to true desktop power in a relatively tiny sbc). They're fun little easy to use microcontrollers. I've made an 8bit computer emulated using a few to handle different aspects of the computer like the custom chips back in the day (software now versus logic, but in spirit there idea is similar). I planned on making a game hand held with this approach for fun but other things came up.
Apple II/II+ have notches for cables to snake out the back of the case, and on the IIe and IIgs there is an array of D-SUB connectors in the 9, 19, 25 pin sizes. On the IIe and IIgs you end up using internal extension or header cables to get the connector to the opening you want it in. Sometimes it can be a pain to fit connectors in a fully loaded system as they bump into the cards.
i've seen a few small form factor pc cards use a vga connector on a riser cable, a separate revision of this card could work with that and give people the option to fix the dsub connector or have it dangle outside of the case.
@@amirpourghoureiyan1637 The IIgs specific version I'm working on does just this. That board however will lack the WiFi chip since the IIgs has 12-bit RGB and I'm using the RP2040 bare chip on that design to gain access to all the RP2040's I/O pins.
I should clarify that the card does work in a IIgs for some things, but the current design doesn't quite work for IIgs video as it can't tell the difference between 16-bit and 24-bit addresses. This isn't a problem for IIe/II+ software running on the GS, as that software is unlikely to access anything outside the 128KB window the Mega II chip can see. I'm working on a new design that more closely simulates the behavior of the Video Overlay Card with regard to the M2SEL and M2B0 signals. There will probably be some tradeoffs which will make it more of a fork than a version bump. The screenshots I took this weekend were from the Total Replay splash screens of a few games, and GSOS starting up. I would not classify GSOS as really usable yet with the card due to all the memory accesses outside the Mega II window corrupting the card's screen contents.
David, I'm curious if the current design has any unused IO pins on the Pi Pico? (Not only for the IIgs problem, but more-so for people wanting to add extra peripherals. I see /M2B0 is only available on the IIgs in slot 3, but I can't even find info on which pin. What about M2SEL? I found this article on those two signals which might help people understand their function: (Page 9-10) mirrors.apple2.org.za/Apple%20II%20Documentation%20Project/Computers/Apple%20II/Apple%20IIGS/Documentation/Apple%20IIGS%20Technical%20Notes%2065-79.pdf
M2SEL and M2B0 are available on every slot in the ROM 03 GS, but on the ROM 00 and ROM 01 systems M2B0 is only on slot 3. Apple IIgs Tech Note 68 has some of the info, but the Apple IIgs Hardware Gude has the actual pin numbers on the slots. Pin 35 is M2B0, and Pin 39 is M2SEL. On the IIe and II+ these pins had different functions. M2B0 and Q3 (Pin 37) are needed for proper video decoding on the GS, and M2SEL is needed for cards that do their own address decoding outside the select lines for their slot. (An example of a card which does this is the CFFA3000, or the Prometheus VersaCard). Before working on this project I had not done any hardware design for the IIgs, and I don't think Mark Aikens had either before designing the original AppleII-VGA project. When I spec'ed out a chip to combine the slot select signals (what the ATF16V8 "PicoPal" does), I had not accounted for the extra signals needed on the GS.
Ah great info, thanks. Is that PicoPal ATF16V8 still a production IC? It was my understanding all PAL chips were OEL at this point. (Maybe only the DIP versions that I usually use.)
ATF16V8 is still in production, and is Microchip (formerly Atmel)'s replacement for the PAL & GALs of the same series, it does both Registered and Combinatorial logic and is pin & fuse map compatible with the GAL16V8. All the logic is written in CUPL, which is a very simple language for logic equations, but the compiler is very crash prone and almost never gives you useful error messages when it finds something it doesn't like in your code.
@@GrizzLeeAdams heh ok that's cool. I've used WinCUPL before for GALs and yes, what a hard program to use. That is cool it's still made! And I guess can it be programmed on a MiniPro with an adapter?
This is why the retro scene still is so interesting; modern solutions for old 'opportunities'.
It’s so refreshing to see a project where the person is not trying to make money from vintage computing hobbyists as is unfortunately so often the case. Also, an open source project can grow beyond its current capabilities or morph into something else entirely.
Thank you David for your generosity!
I am trying to make a little bit from the project, but almost all of the money made is getting funneled right back into making more boards and scaling up the production. A solder stencil holder and pick & place are next on my shopping list. Right now I'm assembling all the boards by hand and doing extensive inspection and testing to ensure every board shipped is to the highest standards.
@@GrizzLeeAdams There are people out there who need plug-and-play, you cater to them and you finance your hobby. There’s nothing wrong with that.
I was aiming at those projects that are commercialized, then disappear off the face of the planet when the manufacturer has had enough. It’s very nice that you documented and made the sources public. I just wish it was the rule and not the exception in the community.
Indeed. There have been many retrocomputing projects that have disappeared when their creators did. One of the Atari retro community passed away suddenly at 53 not too long ago, taking several of his products with him and more recently, one of the SS-30/SS-50 bus community members had an unexpected illness that stopped production of several projects. (He still isn't back in production but has started selling blank boards at least.) What I've taken from these examples is if I were to rely on those products for income, I would include those in my estate planning--either to release those publicly upon my death or to have the IP bundled up and ready for my next of kin to sell to someone else who could either continue producing them or release them outright.
@@wesley00042 same with Amiga 'community'where a bunch of assholes buys IP of as much to profit of it. Or people that have custom IC internal schematics but don't release them.
Not trying to be argumentative or flame you, but what's wrong with making money from the vintage computing hobby? It's not as if this is some sacred calling in which all projects and development efforts must be for non-profit only for the betterment of humankind. We're not talking about saving lives or providing clean water to impoverished parts of the world, it's just a niche hobby where some enthusiasts are fond of old computers. If there's some quality of life improvement somebody put the time and effort to develop, I don't see anything wrong with them selling that for profit. After all, folks can always choose not to buy, since these are not essential for the enjoyment of the hobby.
26:47 for anyone with more modern pc wanting to see this fractal on terminal, here's the python version:
for y in range(-12, 12):
for x in range(-39, 39):
ca = x * 0.0458
cb = y * 0.08333
a = ca
b = cb
for i in range(0, 15):
t = a*a - b*b + ca
b = 2 * a * b + cb
a = t
if a*a + b*b > 4:
if i > 9:
i = i + 7
print(chr(48+i), end='')
break
else:
print(' ', end='')
print('')
Loved seeing the Pi Pico used like this. The basic fractal program got my attention so I figured I'd add a couple more data points :D. I timed the program running micro PyBasic on a Pi Pico and on a Teensy 4.1 micro controller. To appreciate the times you should know that micro PyBasic is an unusual implementation of Basic designed as a teaching tool. It's an interpreter written in Python, (an interpreter itself) ported to run on a microcontroller version of Python. The fractal program took 107 seconds to run on the Teensy 4.1 and it took 1071 seconds to run on a Pi Pico.
There's also the Picomite which runs MMBasic, written in ANSI C and it supports VGA output. Makes me wonder how much faster it would be.
@@jimtekkit A lot :D. The program ran in less than 4 seconds in MMBasic on the Pi Pico.
@@retiredwizard Snap that's fast! And what's crazy is that the second processor core isn't utilized, at least from what I've heard. Difficulties getting MMBasic to scale across multiple cores or something.
@@jimtekkit How would you envision the second core being utilized?
I have to chuckle every time you use 'Drol' as the demo game, because in Afrikaans (a language that I speak and also gets spoken in South Africa), 'Drol' translates to 'Turd' 🙂
I have much stuff from the mid 70's and early 80's from when I was actively doing microprocessor development, and instead of dumping it in the trash, I'd like to send it to you, Adrian.
I need a mailing address!!!!
David did a great job, respect. It did all the work of VidHD that cost a bit more, with more firmware option up to imagination.
I bought a scanline doubler - Indivision Mk3 - for my Amiga 1200. It sports an HDMI-output and gives that pixel-perfect output. Makes all my old Amiga games look better than ever.
The C64 mode might be silly on an apple II machine, but damn it made me smile instantly when you switched over to it.
FYI, In AppleSoft BASIC, the "LET" keyword is completely optional. Also, AppleSoft BASIC only pays attention to the first two characters of identifiers. So, a variable named "BLACK" could also be referred to as "BLUE".
I've been trying on and off to "port" a BASIC Battleship game from Applesoft to GW-BASIC. Is there any discrepancies between the two versions of BASIC I should know about that'll affect coding?
@@Chaos89P Not that I'm aware of. The most common problem when translating from one version of BASIC to another is reserved variables. For example, on the Commodore 64 (and all other Commodore 8-bit computers), TI and TI$ referred to the internal "jiffy clock". Other versions of BASIC did not have an equivalent. That won't be a problem here, though, since neither Applesoft nor GW-BASIC have any reserved variable names.
@Chaos89P Anything involving graphics may need translating, simply because of the graphics being different on Apple ][ series machines and those using GW-Basic. But you can take advantage of the differences.
Back around 1979 a friend and I converted a "Star Trek" type space war game from BASIC on an HP-1000 minicomputer to my Apple ][. We went from no graphics and no sound to adding a bit of both. It took a while to type it all in from a hardcopy listing printed from the HP, then translating any functions without equivalents, before adding and improving it.
Good luck with it, and have fun!
I like using LET because it lines up the left side of the code. The code looks sloppy without it, with all the assignment statements starting one character before every other line.
@Bill Chatfield I think the one best reason to not use LET was on low RAM systems, where every byte saved helped. But I may be mistaken, it's been quite a while since I programmed in Applesoft BASIC! It's tokenized, I think.
Great video. I've been falling this project from afar for a few months. Thought it was really cool, and I was glad to see that it used a Pico, but I didn't realize David was able to pack so many features into such a small, unassuming card. Absolutely brilliant!
While I still love my VidHD card for HDMI output, it's great to see a new open-source, budget-friendly option for getting a decent video signal out of an Apple II. I still think the Apple II series is among the most "future-proof" line of PC's ever made. So much potential for expansion, as this project clearly demonstrates.
I agree! Apples make great 8 bit workstations! I have had one on my workbench since the beginning.
Wow, I just watched the new video and I have to say I'm impressed! I love that a VGA output and Z80 for CPM were added to an open-source board. Additionally, the possibility to add more functionality with simple firmware updates is really intriguing. Great video!!
Mainly commenting to appease the almighty algorithm! Thanks for sharing this awesome piece of tech with us, Adrian! I hope everyone is doing well and having a great day!!
What a brilliant project. Thanks for sharing this and helping with it, Adrian!
Early in the product cycle but definatly a revolution... Great form factor.
Thanks! I did the PCB layout originally with the II+ in mind, but it should be easy enough to use a short VGA extension in a IIe as long as all the slots aren't filled. The future revisions will probably have a notch in the PCB at the back and a separate VGA breakout on a cable so you can mount the VGA connector to the case without running into other slots.
This is incredibly awesome and hopefully it can set the stage for future similar projects enhancing EVERYTHING by snooping and responding as a device to the system bus! Even simulating RAM chips which some systems may be running out of in the near future. SIMPLY OUTSTANDING!
I used to have a IIe. This is a cool product. I really like the VGA. The expansion slots are paying off again.
Oh fun, benchmarking basic performance! Couldn't resist typing it in on a few of my machines: Philips NMS-8250 MSX2, Z80 @ 3.58MHz: 851s; Philips NMS-8250 MSX2, Z80 @ 7.14MHz: 409s; Panasonic FS-A1ST MSX turboR, R800 @ 7.16MHz: 145s. My conclusion is that MSX-BASIC is apparently very slow floating point wise compared to the others (at least the Amstrad, with a similar CPU, less accurate?), the R800 is doing not so bad but it's a heavily optimized Z800 variant that does most instructions in a single clock cycle. Thanks for the great videos!
Thanks, they have been added.
The Raspbreey Pico is an excellent basis for an 8 bit retro computer peripheral system. I've used it myself for a virtual "Winchester" interface for the ACT Apricot computers. Only really enough pins for an 8 bit data bus though with off-board address decoding using a GAL.
Oh, in case you want some more BASIC fractal benchmarks:
Sinclair QL (7.5MHz 68008) stock: 260 seconds.
Sinclair QL (7.5.MHz 68008) with Lightning fast-math SuperBASIC extensions: 191 seconds.
Sinclair QL with Miracle Systems Super Gold Card expansion (24MHz 68020) with Lightning fast-math SuperBASIC extensions: 19 seconds.
Thanks I added it!
Genuinely awesome - I'm going to have to make some to play with.
Many thanks to all concerned.
PS: It would be great if flipping between colour and monochrome could also be handled with a key combination.
Pretty awesome to see the community still pouring some love into these old hunks and making new hardware giving it new life.
Always exciting to see the Apple II line continuing to be developed and loved so many years later. I ran a BBS on mine, so I had a clock card so it could track time, an extended 80 column card with its own RAM (mine had 512K but it had space for up to 1Mb that I used as a RAMdrive) and a SCSI card with an external 10Mb hard drive. I never got as far as installing a zip chip to speed up the processor - but it was an enticing idea. I remember taking my SCSI card and drive over to a friend who had a Laser clone that ran 3.6mhz and marvelling at how fast it could load a gif image from the hard drive rather than writing line by line. The IIgs existed in those days but I would never have envisioned that it would be possible to output graphics at VGA level. Exciting project!
Incredible! The possibilities are absolutely endless with this! Every card that ever existed could theoretically be emulated this way not to mention future cards that haven’t even been thought of yet. Amazing work!
Fascinating - I had no idea what this did. Or really about cassette bias signal. So I learned a lot today. Thank you!
Great card and very clever idea! The same concept would work on any retrocomputer which has full system bus access through some port and the video memory is mapped into the overall address space. In addition, it could be a great system diagnostics and debugging tool too (essentially spying on system bus activity and visualizing it).
Pico seems truly a gift for so many retro projects
I’m so excited for projects like this, especially anything that shows promise for getting IIgs RGB graphics on a more modern display since solutions can be hard to find, finicky or very expensive.
really amazing indeed! I am very impressed! Thank you for showing it to us!
Open Source is the best! That is a great little card. Thanks to you both.
Hate it, totally anachronistic hardware. (It it had a real Z80 on it, it would have been another thing.)
I just gotta say, thats a kickin wallpaper my dude.
I have some data for the spreadsheet; although more "for fun" than a 1:1 comparison, since the devices in question only have limited displays, I had to "print" it on a 24 character printer and scaled the fractal to 23*15 characters - call it "print mode resolution" :)
Epson HX-20: 230 seconds
Sharp PC-1403H: about 550 seconds (I missed the last line but it was sloooow)
TI-74 Basicalc: 260 seconds
Awesome, thanks! I've added the results the the spreadsheet. Great channel too by the way!!
@@adriansdigitalbasement Thanks! I can't attach pictures in comments, but I have a post with the resulting paper strips. The entries might need an asterisk due to the reduced resolution though 😁
Another card to try to get working on my Franklin Ace 2200! Already have a few PICOs laying around and a few of the other components. Looking forward to building it!
It is quite impressive interfacing modern hardware with old to do this, certainly something I could never wrap my head around doing, I tend to just plug stuff together and hope it doesn't go bang!!! :P
That's really cool, I'm impressed how much the conversion improved the color burst mode. I always forget the Apple II has it's own method for Artifact Colors.
I think it was a good solution for bitd between text and color images even if it had some speed bumps like needing a switching monitor, or a monochrome screen and a color SDTV.
I keep hoping DOS-CGA will have a Composite revival to try and undo the past internet damage from not just the ridicule of it's base colors but also the fetishization of those same colors too in pixel art from everybody not using Artifact Color Mode. 😐
Open Source hardware and software. The last bastion of human freedom in tech.
For this you would need to switch to a RISC-V cpu, then yes.
This definitely deserves a revisit once more features and such are added, sounds very cool!
How about adding wifi and some version of uIP and from there you could even remote control an Apple II/etc from a PC, add files, backup the entire system, snapshot it, inspect memory while running, etc! The possibilities are virtually ENDLESS, thank you Adrian from showcasing this project!
I have an Apple II plus frankenstein build with a Unitron clone motherboard that has really bad, noisy composite video, this seems like the perfect workaround!
Yeah! The Rev 4 Apple ][ I used it on has terrible, nearly unusable Composite video (due to it being so non standard) so this just gave that machine pixel perfect output.
If you check out Mark Aiken's AppleII-VGA project, this is exactly what he went through and why it all started.
David should get in touch with the My Life in Gaming folks. This is exactly the kind of thing they like to cover, though they focus more on game consoles than computers.
I was very confused why you were showing an image of a listing in Sheets...but then saw that you made this into a science project! Thanks for sharing the sheet with us!
This is really cool already, but the fact that the feature list can be expanded on in the future is a very exciting prospect. I already have a composite to vga box that I use with my IIE when I want to use color, but I think it'd be neat to get one of these cards as well. Especially because of future possibilities.
Green monochrome and amber monochrome modes would be nice options to also have in the menu.
Amber for sure. I have a small amber CRT that I love using with my Apple //e. Would be nice to have a larger amber display and get that one off my workbench for a while.
Ayyy, I love seeing the RP2040/Pi Pico used in things like this, its a very cheap MCU compared to the price AVRs can get upto or even SBCs such as the pi4, Zero or alternatives can get upto.
This project looks awesome - thanks for covering it and giving it exposure.
Having implemented several PC ISA sound cards/chips on the Pico (namely Gravis Ultrasound, Adlib, CMS, and Tandy 3-Voice), I want to get my hands on one of these boards and implement emulation of the Mockingboard!
Nice! I picked up a MS CP/M card for my //e recently and it's been fun playing with CP/M again.
But it, like most Apple Z80 cards, is slower than standard CP/M. I'd heard of the Applicard, but as you said, they are pretty rare and I didn't know much about it...
Also, have you seen they are making (have made?) a Fujinet for the Apple computers? (As well as Commodore and Coleco Adam). Be interesting to see/hear what you think of Fujinet at some time. (I do have one for my Atari 400 (with 48k RAM).)
The FujiNET is already available for Apples and it too includes PCPI CP/M functionality. It's turns the Apple into a terminal to connect to the CP/M running on the Fuji's processor which is very C/PM-esque indeed.
Excellent piece of hardware. Kudos to David.
That Agon Light is super fast. Will we see a review soon on this channel?
22:42 : Finally the Apple ][ has a good looking display.
12:00 It's just ClearType. Look how much more readable the text is when we use RGB subpixels individually to increase the resolution. I'm joking of couse, but could it have been a thing on composite if they really put more thought into it?
That spreadsheet with the results? That was impressive to see the agon light pretty much in a class of its own compared to the other 8-bit computers!
Yeah, I'd have loved Adrian to cover the Agon, Commander X16, and the random PC at number one!
I found the MicroM8 Apple emulator sure brings back a lot of memories with the Apple II
What an awesome use of pi Pico.
Great work! Congratulations!
wow this is as amazing project for vintage enthusiasts like me. Thanks for sharing.
wow thats outstanding color
24:44 what's up with the blurry, shimmering text on the left side?
The slimmer blue vga connectors like what looks like you have, will just barely slide past the port nearest the PSU, by the way. I was pleasantly surprised after Amazon false advertised having an extender in stock.
Great product! However, the seller has already "retired" the card? Any ideas for where to purchase such a card? (not at all confident about my precision soldering skills!)
I was wondering that myself! I would love to have something like this, but I'm not entirely sure how to go about building it myself.
Very cool video Adrain. It would be a cool test to run on a Apple //c+ at 4MHz.
Thanks for the content.
What EEPROM programmer are you using and which software for programming EEPROM are you using?
Open source hardware is fantastic.
Excellent and definitely on my wish list. Thanks for sharing
At 5:33 , the motherboard would look much better in 1440p or 4k resolution. Your viewers would appreciate high-res.
Adafruit is now selling a version of their RP2040 feather board that outputs DVI over an HDMI connector. It could be interfaced to the Apple II and do the same thing.
This is so cool, I thought of a couple more thing that may be possible to do on it, #1 IBM PC emulator, ram would probably be limiting, and Mocking board support.
Looks like a cool alternative to the VidHD card. I had a VidHD and it worked great until the microcontroller shorted out and emitted smoke. Nowadays you can’t buy the VidHD any more due to part shortages.
Very fascinating. I wonder why he couldn't layout the PCB with the VGA connector on the other side of the board.
Apple 2 indie developers continually do this for unknown reasons. Hell, even some cards from back in the 80s/90s did this (example: audio cards on the IIGS outputting to a stereo 3.5mm jack... which did not face the backplane but instead internally). Other cards had no backplane connector but instead internally has a ribbon cable that went to a screwed-on backplane connector (think: SCSI and parallel cards). The whole Apple 2 expansion card ecosystem was bizarre in this regard and I've never seen anyone in all my years explain why they do this.
Couple of reasons. Lack of space being the main one. Plus, the back plane of the II+, IIe and IIGS are all different.
Mainly because the II/II+ has just slots for cables, and they are the most common targets for upgrades. As others have said, on the IIe and IIgs there is an array of D-Subs in 9, 19, 25 pin variants, and it would be an absolute pain to insert a card cleanly that had to match up to a cutout, and would lock down what slot / machine it could go in. The compromise most of us have gone with is internal facing connectors (which was common on the II/II+ expansions) so you can use short extensions to reach an available cutout on the back of the IIe or IIgs. I may start stocking 1.5ft VGA cables and gender changers as an addon option in my store. They could be wedged into a slot on the II/II+, or mounted in one of the DE9 holes on the IIe or IIgs.
Single pixel graphics could be increased by running a convolution algorithm to detect the single lines and only thicken the single lines in the RPi modified video memory map.
You may automate the truck scope from the computer. I have used a quick screenshot, and developed a full characterization test for analog filters. Unlocked DS1054Z, in my case.
WOW, great board. Good video, thank you
Some notes on the C128 performance in 80 column mode.
First of all basic 7 is slower than most traditional MS basic variations on 8 bit machines because of supporting more than 64k and all the bankswitching which results from that. But also, the 'screen editor' code for supporting 80 column mode is really really inefficient.
Adrian, I am curious on how does this implement the color fringing filter for the VGA signal. I can see that the VGA signal is properly generating color which is also artifact based (so we don't see the alternating lines as we do in monochrome mode) but it seems to be efficiently discarding undesired noise on single pixel white bars such as used in text.
Pedantry: The Apple II doesn't have a colorburst, the patterned hi-res monochrome is all it can output, but due to limitations of NTSC that is picked up as color information.
The Apple II outputs a colorburst signal when in color mode. It does not output a chrominance subcarrier. The colorburst signal occurs at the beginning of each scan line, and it is used by the TV set to synchronize the decoding of the chrominance signal. As you indicate, the patterns in the monochrome signal are decoded as color information. In text mode, the Apple II does not output the colorburst signal, which indicates to the TV set that there is no chrominance subcarrier to decode, so it decodes the signal as monochrome only.
@@Mueller3D Ah yes, crucial difference.
@@Mueller3D Yes, this is correct!
I'd been thinking about using the RP2040 (the Pi Pico CPU) for doing a bus interface board - it's a trivial thing to get going and doesn't require hugely complicated board designs. Good to know I was on the right track...
Very cool project!
Absolutely genius!
One correction, if itvis monitoring bus activity and one ram chip output is stuck, then it bay drag the bus low/high enough to garble the frame buffer data.
Will it work in a Laser 128? Only 7 minutes into the video, pardon me if you tried it in the video. Heh.
Just got my first apple II. What do I do next? I have No software😢
Wow, this is just awesome!
ARCHON ! Goosebumps !
we need a replacement Commodore 5717 mouse controller those are fairly simple gates: modes: Joystick or Proportional
Intersting...my IIgs has the "manila adapter with Nec 1550m (blurry but 50hz modes)
or actual with Samsung 510m (sharp but no 50hz and GS-OS still not perfect".
I need another solution.
No need (nor desire for) CP/M, but having VGA output is cool! 🎺🎶
Apple II Forever.
Hi Adrian, I've just built one of these from scratch and was wondering where to find the configuration tool as I have some issues with mine and want to determine if it's a config issue or if I need to do some repairs. Love the channels BTW :)
That is the power of FOSS, have you sent your basic program for inclusion in the project?
I wonder if a game could be made to use the z80 and 6502 at the same time for incredible graphics.
Sort of mind blowing, makes me think about if you could use a microcontroller to create an 8-bit sound card or VGA card for an IBM PC compatible?
My eyes perked up on the Basic09 score. I assume it had the benefit of being compiled in to pcode instead of having to run through a native basic interpreter like all of the other machines.
Yeah there is no other explanation for that insanely fast speed at just 1mhz!
So if this card works by monitoring bus activity, could it also be used as a diagnostic card with the right software? If it expects X routine on the bus and sees Y instead, report that? I'd imagine it wouldn't be as comprehensive as a diag rom, but still could be useful to see if there is some bus activity on an otherwise dead machine.
Few comments. No one I know of "ever" used LET to assign a variable in apple basic. It basically does nothing except take up token space. As for boolean. They are possible, just tricky. As an example you could do something like this:
A = 0 + ABS(A < 1)
Applebasic converts this sort of thing to either 0 or -1, as though it was part of an if/then. Things like AND, OR, etc. however work different based on which "version" of basic is used. The older ones where bitwise, while later ones returned 0 or -1, supposedly.
In any case, it is possible, just... quirky and probably not well documented.
What about the comb filter, which was a feature of most high quality composite monitors of the time? For example, the mouth of the character in drol looks like teeth on your composite monitor, but looks kinda like a scarf on the VGA output. I also like to test this with Alice in Wonderland from Windham Classics.
would love to see this on a IIc, perhaps using the cpu slot as a pass-thru
The RP2040, with two ARM cores and more than 256KB of internal RAM, could emulate the entire Apple II not only the video controller.
But anyway, I still think an FPGA is better suited than a microcontroller for this application. And in this case too, even an small FPGA would be enough to recreate the complete Apple II. In fact I just did that some weeks ago and was able to sinthesize an Apple II clone with VGA graphics, PS2 keyboard, and floppy disks emulated from an SPI Flash memory (read only). It took about 1400 logic cells out of 7680 in an Lattice FPGA.
I've had more fun with the rp2040 based microcontrollers like the pico and just building boards using the rp2040 than with pi's or other sbc's (aside from the new one I've been using with 16gb of RAM, getting closer and closer to true desktop power in a relatively tiny sbc). They're fun little easy to use microcontrollers. I've made an 8bit computer emulated using a few to handle different aspects of the computer like the custom chips back in the day (software now versus logic, but in spirit there idea is similar). I planned on making a game hand held with this approach for fun but other things came up.
Is there a reason why the VGA connector is oriented internally? Are Apple II expansion cards inaccessible through the backplate?
That's right. Apple II doesn't have slot aligned holes on the back.
Apple II/II+ have notches for cables to snake out the back of the case, and on the IIe and IIgs there is an array of D-SUB connectors in the 9, 19, 25 pin sizes. On the IIe and IIgs you end up using internal extension or header cables to get the connector to the opening you want it in. Sometimes it can be a pain to fit connectors in a fully loaded system as they bump into the cards.
i've seen a few small form factor pc cards use a vga connector on a riser cable, a separate revision of this card could work with that and give people the option to fix the dsub connector or have it dangle outside of the case.
@@amirpourghoureiyan1637 The IIgs specific version I'm working on does just this. That board however will lack the WiFi chip since the IIgs has 12-bit RGB and I'm using the RP2040 bare chip on that design to gain access to all the RP2040's I/O pins.
@@GrizzLeeAdams awesome, this project is gonna be a lifesaver with how expensive official apple monitors have become in the last few years!
i have a FastChip with a 65816 running at 16Mhz on Apple //e that should run that program fast... love this project!!