The Mega IIe: A Vintage Computing Adventure with James Lewis

Parker Dillmann
Welcome to circuit break from MacroFab, a weekly show about all things engineering, DIY projects, manufacturing, engine news, and Apple Mega 2e Projects. We're your hosts, electrical engineers, Parker Dillmann. And Stephen Kraig. And this is episode 420 Blaze It. Circuit breaker from Macrofab.

Stephen Kraig
Hey, circuit breakers. We have an announcement. We are running an electronic design contest on our community forums, and the theme is food devices. Go to forum.macrofab.com to find out more information about the contest and how to enter. For prizes, there's over $5,000 in cash, free prototyping services through Macrofab, and the most important thing, trophies to show that your design was one of the best entered.

Stephen Kraig
There will be a link in the show notes on where to find more information about the contest and how to enter. Our guest this week is James Lewis, AKA Bold Engineer, freelance electronics content creator. You might recognize James from the YouTube channels, add ohms and workbench Wednesdays from element 14. His interest include talking too much about capacitors, repairing vintage computers, and making unique PCBs that go into poorly designed 3 d printed enclosures.

Parker Dillmann
I don't guess that's gonna be the topic of this one.

James Lewis
It's very close.

Parker Dillmann
James has been on the show a few times now, including episode 114 where he joined Tacos about ceramic capacitors. He also joined in episode 222 to discuss testing and validating PCB assembly design, and more recently, episode 411 where we discuss reliability and component aging effects. So welcome back, James.

James Lewis
Oh, wow. I didn't realize it's this is the 4th time.

Stephen Kraig
And, the last time was just a handful of weeks ago. Yeah. That's crazy.

Parker Dillmann
I think you were one of the first people, James, on our podcast when we renamed our podcast.

Stephen Kraig
That sounds right. Yeah.

James Lewis
Yeah. I remember that because I remember listening to the, the 400 number 400, and I was thinking, okay. I have to remember there's a new name now.

Parker Dillmann
So last time we talked about component aging effects, we're kinda diving completely off track from what we normally talk about with you, James, which is, you know, components and that kind of stuff. We're gonna talk about a project that you've been working on for a long time.

James Lewis
Yeah. Yeah. Actually, depending on how you count, either 5 or 3 years.

Stephen Kraig
So what what is the project?

James Lewis
So I I call the project the Mega 2e. I took a chip out of the Apple 2gs called the Mega 2, which has all of the logic of the previous generations of Apple 2 computers on a single chip, and I built a functional computer around that chip.

Parker Dillmann
So before we dive more into that project, I I wanna know why the Apple because I didn't grow up with Apple Computers at all. I'm actually too young for at least Apple 2. Why is the Apple 2 so revered in vintage computing space?

James Lewis
You know, that's a good question, and, I think it's a different answer for different people. For me, the the Apple 2 was, I I was thinking about this, a while ago. When I was in elementary school, I went to 3 elementary schools in no, 4 elementary schools in 2 states and 3 or 4 cities. And so I saw a lot of different schools, And the one constant in every single one of those classrooms was an Apple II computer. You know, at the time, being a kid, that just seemed totally normal.

James Lewis
Right? That's, that was the computer at school, and that was just the way it was. And so for me, there's a lot of nostalgia because it was really the first computer I really learned to use. But then as an adult, I feel like a lot of times people look at vintage computers because they remember the video games. And I know I'm going to offend some Apple 2 fans, but most of the games on the Apple 2 are not very good.

James Lewis
You know, if you, like, compare it to, like, what's available on the Commodore 64, there's just, there's a big difference in especially graphics quality. And I will try to backpedal a little bit. There are good games on the Apple 2, but that's not why I love the platform. I like it because from a hardware engineer's perspective, it is a fantastic piece of engineering. Because it is an extremely open design and it has a ton of clever tricks.

James Lewis
So it's just really fun to understand how the computer works or try to understand how it works.

Stephen Kraig
I'm curious. Can you go a little bit more into when you say extremely open, what do you mean by that?

James Lewis
Yeah. So so the very first Apple 2, it came out in, like, 1977. It had about 80 chips on the entire design. Other than the roms, everything was off the shelf. There were no custom chips.

James Lewis
And then outside of the CPU, the keyboard controller, a couple of 555s, and the ROMs. Everything on that board is a 7,400 series logic chip. And so there's 2 aspects to it. 1st, if you had no documentation, you could literally figure figure out what the circuit looks like just because it's it's right there. But like computers in the early eighties, there was tons of published documentation on how it was designed, how it was built, and then things like the the ROM are completely documented.

James Lewis
Even back then when you bought, like, the first Apple 2, you got a listing of the ROM along with it. And so there were no secrets about the Apple 2, which by the way is also why there were so many clones. Because they basically sold you a copy of how to build the computer with every computer.

Parker Dillmann
So part of it was just not a lot of customs silicon inside of it.

James Lewis
Yeah. Yeah. So if you look at, like, the Apple 2 and then the Apple 2e, which by the way, the e does not stand for enhanced because there is also the enhanced Apple 2e, just to make it more confusing.

Stephen Kraig
E squared?

James Lewis
Yeah. I think e actually might have meant I think they actually meant extend it because it's fully backward compatible with everything the Apple 2 and 2 plus did, but it added some capabilities as well. But that was also the 1st computer where they took a lot of that logic and integrate it into 4 or 5 ASICs. And even then one of the ASICs is just the PAL or a gal, you know, a a gate array logic chip, that's masked. And then the others are really just implementations of the 7400 service logics.

James Lewis
And so they're just integration chips. If you compare that to, say, like a Commodore 64 at the time, Commodore never had a discrete version of their computer. There was always a handful of custom chips inside, like the PLA, for example, for the, the memory controller. And then you could argue some of the other Silicon was somewhat custom as well because, I mean, Moss did a lot of the chips for them. And so it wasn't quite as open because there wasn't like a reference for what that chip design was.

James Lewis
Whereas with the apple 2 and the apple 2e, there was essentially documentation for all of that.

Parker Dillmann
Yeah. We had, Bill Heard on who worked on Commodore.

James Lewis
Mhmm. Yeah. He's great.

Parker Dillmann
And he talked about working on the Commodore and where they had actually, they can go literally downstairs and build custom chips. Mhmm. And I think that's probably where I bet you Apple didn't have a chip fab underneath their basement, basically, to be able to do that. Garage. Yeah.

Parker Dillmann
A garage.

James Lewis
Yeah. Actually, they they basically did everything with VLSI, the company. Yeah. They didn't have that kind of capability. And I don't I mean, really none of the computer companies, I'm probably there probably was another, but I mean, that was something that was super unique to Commodore was that, yeah, they could get custom stuff done.

James Lewis
And, yeah.

Stephen Kraig
So, yeah, you're not joking. I'm I'm actually looking at the schematics right now for it, and it's I mean, these schematics are kinda brutal to visualize in your head because there's just so much stuff going around. But it's just a whole bunch of 74 100 series chips that I grabbed a random one, threw it into Google. You can still buy it from Texas Instruments right now.

James Lewis
Yeah. And by the way, so a lot of the especially, like, anything in the, like in the Apple II era, a lot of the, the Apple II and 2 plus a lot of it's LS series, which is actually harder and harder to buy today. But after the 202 plus almost everything was HC. And so there's, if, if that's not available, there's a modern equivalent. And then with the, as the kind of the computers went along later, it started to get to the really high speed stuff because they couldn't do it faster.

James Lewis
And because it was like, it was some kind of condition where they needed silicon that was faster than they could do in the custom chips. And by the way, if I can plug something real quick, I do have a project on GitHub called bitpreserve. And, it's a repository where people can contribute vintage computer schematics redrawn in KiCad. And so there's an Apple II Plus design that I actually just fixed up the other day. The 2gs is in there for obvious reasons.

James Lewis
The 2e is, was done by a friend of mine. And then there's a Commodore, Tandy, a bunch of other retro computers as well.

Stephen Kraig
Oh, that's awesome. Out of curiosity, is someone going through and validating and checking those?

Parker Dillmann
Nope.

Stephen Kraig
Okay. Well, they're they're up there. Just keep that in mind, I suppose.

James Lewis
Yeah. And what I'll say there is, like, the way I use it is if I'm trying to look up something really quick, it's a lot easier to do it with the KiCad, version of the schematic, or even if there's a PDF of it, it's just a lot easier to read. But I always have the original right next to it because it's like, okay, if I find the thing I'm looking for, I make sure on the original that they sort of match because, yeah, there's just not a good way to validate those.

Stephen Kraig
Yeah. Because these schematics I have pulled up right here, let me just call them classic schematics, where where the page where they scan them, the pages have creases where, like, going through chips or chips are off the page or or you have to kinda print it out and tie them all together.

James Lewis
When I started doing the bit preserve project, I had a ton of people ask me, well, why don't you just use, like, OpenCV and OCR, the existing scans? And I'm like, I can't read them, so I don't know how to write code to read them. Right. Right.

Parker Dillmann
Yeah. The, I remember pouring over old Atari 26178100 schematics. They're very similar. Yeah. That's actually nothing.

Parker Dillmann
Atari had their own customs to look in then.

Stephen Kraig
Mhmm.

Parker Dillmann
So they were more I guess they're kinda in between Commodore and the Apple series computers then.

Stephen Kraig
Mhmm.

James Lewis
Yeah. That's how I would that's how I view it. And so so backing way up, you know, that's to me, that's why the Apple 2 is my favorite is simply because from a hardware designer perspective, it's I can look at it and there's no secrets. It's like, here's the design. Right?

James Lewis
And I think that's just super cool. Like, that just blows me away. So on the project,

Stephen Kraig
what drove you to say, I wanna take this and run with this project and make the Mega 2 e?

James Lewis
So here's what happened. So, and this is what I say. This was like 5 or 6 years ago. I was like, you know what? I never owned an Apple II, but I remembered using them in school like I was talking about before.

James Lewis
And so I was like, I'm going to buy an Apple IIgs so that I can experience the Apple II again. And, I just want to point out that when I when I got it, I couldn't even remember how to reset it. Like, the it's like the command or, our Apple control, the power button, or reset. And I didn't remember that sequence at all. I didn't know the difference between basic and the monitor.

James Lewis
I didn't even know that AppleSoft was basic. Like, I thought that was something else, and I didn't know how to load anything off of floppy drives. I couldn't remember how to do any of that stuff. And so I was like, okay, I got to I've got to I want to take this approach of, okay, I just bought this computer. Let me go find the manuals online so that I can read about how to use it.

James Lewis
Right? And I found this one called the hardware reference guide. I was like, hey, I like the name of that. So I started reading that. And in one of the sections, it talks about this mega 2 chip and it says, the Apple 2gs contains a mega 2, which has all of the logic of an Apple 2e.

James Lewis
And all you need to add is RAM, ROM, and you would have a a fully functional apple 2e. And so I was like, alright. So I wonder if I can do that. And I want to point out because and for people that don't under don't know the 2gs architecture, this may not matter as much. But for some people that are familiar with the 2gs, so the 2gs was a 16 bit computer that was fully backward compatible with all of Apple's Apple 2 8 bit software.

James Lewis
And even their own documentation. And it's funny cause I found even earlier versions where this is a little more heavy handed. Apple tried to imply that the mega 2 is why the 2gs was backward compatible. When in reality, the 2gs uses the MEGA 2 for the Apple II video modes and as an IO controller, When software runs on the 2gs, the 2gs, its own chipset and rom is actually doing the emulation. It start it turns on and acts like an Apple 2.

James Lewis
Software has to tell it to turn into a 2gs. And so when I learned that, I got even more interested in the project because I realized at that point, even the 2gs wasn't using this MEGA 2 chip for its intended purpose. And, Apple never built a computer around the the MIGA 2. They did license it to Tiger Electronics who did make a prototype computer, but it never got out of test marketing.

Stephen Kraig
So that exists somewhere?

James Lewis
Yeah. There's there's a handful that exist. There's a person in the Apple II community that kinda shared some information with me because he has one. And, at first I was like, oh no, I did. I committed the sin of saying I've done the world's first thing.

James Lewis
And then I'm like, oh, here's this thing that is almost what I did. And I was talking to him, and he's like he's like, you gotta understand, it's all bodge wires, and, the keyboard bounces. Like, you can't even type on it. So it's not really functional.

Parker Dillmann
It was really, like, the first stab, and then then there was, like, kill the project.

Stephen Kraig
It's alpha.

Parker Dillmann
Yeah.

James Lewis
Yeah.

Parker Dillmann
Yeah.

James Lewis
Yeah. And there's there's I I have different stories on why it it was killed, but the reality

Parker Dillmann
is there's only a handful of these devices in the world. So I kind of already know the

James Lewis
answer to this, so maybe this is

Stephen Kraig
mega chip, all it takes is RAM, ROM, and maybe some glue logic, and then everything works. Right?

James Lewis
Oh my gosh. It's so funny. In my mind, I thought I was going to make a breakout board for the MEGA 2, which is a 84 pin plcc. So plastically chip carrier, I think. I can't remember what the cc, but the cool thing about that package, if, you're not familiar with it is it can either be surface mount or it can go into a socket.

James Lewis
And so it's, I kind of wish more packages were still like that, but it's kind of a cool package in that, in that regard. But it's 84.

Parker Dillmann
When you see that on a board, you're like, that circuit board's old. That's, like, has that old

James Lewis
I love

Stephen Kraig
those chips. I miss them.

James Lewis
Yeah. It you know what? And inserting those chips into the socket is one of the most satisfying things you can do, because they you get that that scratch and then this nice clunk. I I love doing it.

Stephen Kraig
When I first started in electronics though, I was very scared of those. Just how much force it takes to pull something out or push something in you, like, I am going to break this. Yeah.

James Lewis
Funny story about that. So the the mega 2 has 84 pins. One pin is VCC, 2 pins are ground, and the rest are functions. For the whole for the whole thing? For the whole thing.

James Lewis
Wow. And on one of my mega twos, I damaged one of the ground pins, and I was like, you've gotta be kidding me. Like, how lucky did I get that the only pin I damaged has a copy?

Parker Dillmann
And well, that's old. Were those chips, CMOS, or were they just NAS chips?

James Lewis
You know, I don't know. Because

Parker Dillmann
that was that because that was the old Atari stuff was all just NOS because it would they were even cheaper to just build so CMOS, you know, you have n and p type, and then so they're like, NOS was like, you only had one type.

James Lewis
Yeah. So I came across one reference that talked about, so every reference I found talked about it as CMOS, but I feel like that was also kind of a time when CMOS was still getting used for either. There was one comment about how, like, in, maybe it was in the 2e or no. No. No.

James Lewis
No. It's with the mega 2 and the way that it deals with slots. Its edge rates are a little bit slower than what the previous generation of Apple Computers had. And one of the explanations was that the chip process used is not as fast as a 74 HC series. But from my measurements, I mean, they're pretty close.

James Lewis
It's like, you know, nanoseconds difference. But it was enough that, okay, if you had a really tight timing consideration, you might be off. So my guess is it's CMOS, but I don't I don't know for sure.

Parker Dillmann
The big thing was, power consumption difference. But when back then, you know, it didn't matter. So

James Lewis
Yeah. Actually, so, you know, this whole thing. So the the PLC and then the CPU, the ROM, the EEPROM, and then a 128 k of dynamic RAM, it only draws about 250 milliamps total, which really blows me away.

Parker Dillmann
That was what I was about to ask because through one power pin, it's like, how much is it trying to pull through that one pin to power up?

James Lewis
Yeah. I I would say it's probably on the order of a 150 to 200 milliamps, and then the rest of the design draws the rest.

Parker Dillmann
That's nothing.

James Lewis
But, yeah, it's shockingly low. Here, it's so low because so I used a Raspberry Pi Pico oh, no. Sorry. Let me back up. I used RP 2040s, the Raspberry Pi microcontroller.

James Lewis
I used 2 of them because they have the PIO or the programmable IOs, which allow you to do a lot of really cool things that you can't generally do with a a microcontroller's GPIO pins. And at one point, I was like, well, why don't I use PALs or programmable array logic chips to do some of the stuff, some of the decoding that I need? And those chips draw 200 milliamps by themselves.

Parker Dillmann
I was

James Lewis
like, wait a minute. I'm going to double the current consumption by adding a single chip. It just it's like, wait. That's that's just mind boggling.

Parker Dillmann
Well, now

Stephen Kraig
that you brought up the, the the pie chips, let's, let's talk a little bit about the architecture because you because stepping back, you were just mentioning, the idea was to make this into a, kind of a breakout board, but, what did it end up becoming?

James Lewis
Yeah. Good point. Good point. So yeah. So back, I thought, okay, I'm I'm gonna make a breakout board, and I did design the PCB, and I swear I I had it fabbed, but I can't find it anywhere.

James Lewis
So maybe I I didn't actually get it built, but I was looking at it, and I was like, wait. I'm gonna have to hook up 80 jumper wires to this thing. It's like I didn't think this through. And so then I came up with a design where I put the mega, RAM, basically the what I thought were the building blocks I needed, the the mega, RAM, RAM, and, video circuit. And so I thought it was the minimum I thought I would need to make a bootable computer.

James Lewis
And after a lot of frustration, I got that to boot and then immediately crashed. And it crashed because there was no keyboard controller. And so that's the first lie is that the mega 2 needs a keyboard controller in order to work. Without something to drive the data bus when keyboard accesses happen, it bad things happen. So, yeah.

James Lewis
So that's how the architecture started to build up is it's like, okay, so now I know what the minimum blocks are to get the thing to boot, but then to make it work, I need it something to act like a keyboard. And that's where it was like, originally I was thinking to do something like with a PAL chip and decode the the keyboard stuff and blah blah blah. And then instead, I decided I'd use the Pi Pico or the RP 2040 since I could have it its IO is fast enough to watch the bus and drive the data bus when it's appropriate. I'm a

Stephen Kraig
little bit curious. What's unique about the 2040? I haven't dug too much into them. Why did you end up picking those?

James Lewis
So the number one reason was this was all during the lockdown time period where we had the component shortage. So the number one reason is they were really easy to get. So that was

Stephen Kraig
I love it.

James Lewis
It was like, okay. I can get this chip. What can I do with it?

Parker Dillmann
I wonder how many products were developed in that year time span. That is that was number one reason.

James Lewis
Yeah. Actually, it was like, I was doing a power supply design, and somebody's like, well, this is a terrible chip. Why would you ever use this? I'm like, I can get them. I just there there's nothing else available.

James Lewis
That's that's, like, that's it.

Parker Dillmann
And, honestly, they were a great price.

James Lewis
Yeah. Yeah. So the RP 2040, it's it the the thing that makes it special is that its IO pins have their own state machine and so you get. It's a 9 instruction set. You get 32 instructions and you can basically for example, you can shift in all 32 IOs in a single clock, and then you can shift those and then DMA those into RAM on the next clock.

James Lewis
Maybe it's, like, 2 clocks later without using the CPU. It's all asynchronous to whatever the, the ARM core is doing. And so like in my case, I used those GPIOs to be watching the control signals, which were clocked at about 1 megahertz. And then I had to respond and put 8 bits on the data bus within 200 nanoseconds, which is almost impossible to do with, say, like a, an interrupt on a regular microcontroller. And so that's what makes it unique because they've got these little state machines that you can program independent of the, the encore.

James Lewis
And that state machine can run those instructions at, like, a 100 megahertz or so. And so they they they can emulate you can use them to pretend like they are 74100 series logic in a lot of cases because they've got plenty of time to, to think about what they're doing.

Stephen Kraig
I did not know that. So it's almost like little really, really stripped down FPGA on every pin capability.

James Lewis
Yeah. Yeah. It's I I mean, the way I look at it is it's it's basically you're writing assembler for the GPIO pins. And so I want to be careful because there's not a lot of decode logic that you can do. Like for example, there's no if statement, and so there's some limit.

James Lewis
The PIO is really meant to help make serial bus protocols like custom protocols. So, like, you know, you've got a spy implementation that you wanna do that's kinda weird. It can handle all that for you, and then all you care about on the code side is does data show up in the in the RAM or not. Right?

Stephen Kraig
Which is exactly what you used it for, basically. Like a bus sniffer to do things.

James Lewis
Yeah. Actually, for the video, that's exactly what we did. So we went from I was gonna breadboard this thing, which was not possible. And then I built up a couple of boards that I thought was the simplified design. And then I ended up still ended up with a ton.

James Lewis
I mean, just ton of fly wires. And, you know, I mean, this is a, it's an 8 bit computer. So 16 bits for the address bus, 8 bits for the data bus, a couple of signals for the clocks. And then I had these 3 boards that were kind of connected that had to be connected together. So I had wires running around there.

James Lewis
You know, I got to this point where it was just a big mess. Now, in this case, using the PIPECO, I was able to, as a prototype, I used that to drive the data bus and I got the the the chip to actually boot. And so I got it to a prompt, and that was, like, the first major milestone of the project because that was the that was actually my first well, that was when I finally realized, okay, the MEGA 2 may actually be a fully functional chip. And so so yeah. That's when I was like, okay.

James Lewis
Now it's time to get serious and, design something more custom because the Flywire project was not fun to debug.

Stephen Kraig
Out of curiosity, what's the time frame at this point? Are you a year in? Are you a month in?

James Lewis
Yeah. So okay. So I got this idea 5, 6 years ago. So I decided and I've I've I don't remember the the dates now, but I decided, like, in say February of 21, I wanted to finally do this project. And then by July of 21, I got the chip to boot.

Stephen Kraig
Nice.

James Lewis
And so it was that's why, like, I I I must have the dates wrong because I I remember it being 6 months, but I also remember it was, like, right at the end of winter is when I said, okay. It's time to do this project.

Parker Dillmann
Did you did you submit an errata to Apple to update the documentation, not the negative 2 at that point?

James Lewis
Well, you you okay. So I I I'm so glad you brought that up, Parker, because so there is no data sheet for the mega 2. I cannot find anything. Now there was a follow on chip called Gemini that got used in a different product, But Gemini is in a different package and has a couple of different functions, and it has a feature that the mega 2 does not have. And so it's not an ex it's like some people have said, oh, that's just the mega 2.

James Lewis
It's not. It's it's clearly an iteration, but it's not the exact same chip. I do have a board design document for that chip, which at least helped identify a couple of signals that I wasn't entirely sure what their purpose was, but there is no data sheet for the, for the mega 2. I can't find anything in any of the documentation that people have from Apple. I even emailed team Woz asking to see if, there was someone there that could help me find something.

James Lewis
And the worst part was I was talking to a friend of mine. He he works at Apple, and he's like, you know, we have this internal database with all of our code names. I could just type mega 2 in and find out if there's anything in there. And I was like, okay, would you? He's like, no.

James Lewis
He's like, there's no way because he's like, I don't wanna get in trouble because you're gonna make a video about this, and then somebody's gonna be like, wait a minute. How did you get that datasheet? I was like, oh my god. So it was just like, you gotta be kidding me.

Stephen Kraig
So did you end up making your own datasheet out of this?

James Lewis
You know, that was one of my goals, and I I I never did. I do have a lot of notes about the chip, but yeah. After I found the Gemini, that was the follow on. After I found its datasheet or design document, I was less interested in doing it because a lot of the function like, the the pens are at least described in that document even if, you know, there it also describes pins that MEGA 2 doesn't have, but who cares? So

Stephen Kraig
How close are are we talking about the difference between the Gemini and the and the MEGA? Are we are we talking 90% different or 98% different?

James Lewis
It's probably the opposite. It's probably like 10 to 20% different.

Stephen Kraig
Okay.

James Lewis
There is at least 4 years between them, and so the major thing that they did with the Gemini was to make it enter so the Gemini goes into something called the Apple 2 expansion card, which plugs into the LC series Macintosh, which allows it to emulate a, or allows it to run apple 2 software. And on its, on the, that card, there is the, the Gemini, a 65 c02, a, Ram. There is no Ram. The Ram comes from the Macintosh, and then there's a disc controller, another chip that we'll talk about. And a 558 for the, the joystick.

Parker Dillmann
It's its own computer on a card at that point.

James Lewis
No. That's exactly what I

Stephen Kraig
was about to say. Like, it's it's it's what you're designing effectively or what you were.

James Lewis
Well and yeah. It is. Yeah. Actually so we'll we'll jump ahead. So I I actually and before I realized that the well, yeah, I don't think I realized the IWM was built into the expansion board, but I was like, when I was designing mine, I wanted the IWM.

James Lewis
The IWM is a chip version of the disc 2, the floppy disc controller card. I knew going into this, that I wanted that to be part of the design because like, why would I I think it would be silly to have a card. And then yeah. So then the only thing that I was kinda left wondering was why they use this other chip called slot maker. And that was that puzzled me for a really long time until I got to my second version of the project.

James Lewis
And that's when I found out that there's a keyboard signal that comes from the chip called slot maker. For some reason, the keyboard accesses have nothing to do with the slots. Slot maker doesn't decode anything. I think apple, I think the engineers ran out of pens on mega 2 and they're like, okay, well we'll just use, we'll just have slot maker generate this signal because it's connected to mega 2 over a multiplex bus. I can't figure out any reason why you have to have slot maker to use the keyboard.

James Lewis
I'm sorry. I'm I'm I jumped out a little bit there. So there's 3 control signals that you have to decode to know when to drive the data bus for keyboard access. And one of those signals comes from the slot maker. The other 2 come from MEGA 2.

James Lewis
So in addition to the RAM, ROM, CPU, and keyboard controller, you also have to have another ASIC called slot maker to make the single chip Apple 2e work.

Stephen Kraig
So okay. And I guess I'm a little bit confused. Educate me on what is Slot Maker?

James Lewis
So Slot Maker's real purpose was in the Apple 2 GS. So the Apple 2 GS was, like I said, it was a 16 bit Apple 2, but its big thing is that it had all of the major peripherals integrated. And so, like, it had a serial controller. It had, well, 2 serial controllers, one used for networking, the the disk drive controller, and a couple of other things. And so generally in an Apple IIe, you would put cards into the slots into its expansion slots for all of those functions.

James Lewis
And so the 2gs, it had those, most of those things built in, but it also had physical slots. And so slot maker allowed the computer to either use the internal slot, the internal device, or the slot for say the disc controller. You could put in your custom disc controller and disable the onboard one. And so that that was its real purpose was to decide which devices were used in the 2gs.

Parker Dillmann
This is like a multiplexer then.

James Lewis
Yeah. And it also it can it also buffered the clocks because there was a and I I I'm actually planning to look at this on a scope because it actually has to shift the clock a little bit or most of the clocks a little bit so that the timing lines up because of the way it's something to do with the way the 2 GS is designed, but it basically buffers the clocks for the slots so that they get the timing that that a Apple 2 card would expect.

Stephen Kraig
I'm I'm cringing a little bit as you're you're talking about all of this just because having to discover or having to dig through all of this information, not knowing that you need this one line to be asserted or whatever. Right. Having done things similar to that in the past, I just know that takes tons of brain power to just sit and go, why is this not working? What am I missing here?

James Lewis
Well, and actually, it's funny because so okay. Once I figured out what was going on with those key so I totally agree. Once I figured out what's going on with keyboard signals, then it's like, okay. I got the keyboard working relatively fast after that. And then it was like, okay.

James Lewis
Now I want to try since I had to have slot maker, I'm like, okay. Now I'm gonna have a slot. So I plugged in a discard, and then it was like 2, 3 weeks trying to figure out how to get the mega 2 to talk to the slot because the Apple 2e rom doesn't know how to tell mega 2 to talk to slot maker. And so I ended up finding that there's this so there's something in the the Apple 2 design called soft switches. It's basically in the c address space.

James Lewis
Pretty much every address changes something in the computer. So that's like the most obvious example is if you want to go from text to graphics mode, you hit, I think it's c 068, and then the graphics mode changes. So there was a c switch that I found in a document of all of the c switches. There's this one for slot maker on the 2gs. And I was like, I wonder if I can use that.

James Lewis
And it turned out that that c switch or soft switch, I call them c switches for some reason, would actually mega 2 would talk to slot maker and then tell slot maker to use to change the, the multiplex signal. But that took me 2 weeks to find because who who could I ask? Right? Who would have ever like, I didn't even know how to form that question because it's like, I I don't know what to do here. And so once I got the disk drive working at this stage, I built what I called rev 2, which was a backplane and the backplane had cards that plugged into it.

James Lewis
And the idea was I took each block of the apple 2 design and made them their own card so that I could basically bring up the system one block at a time. And so I got the keyboard to work and then I got the disc drive to work. And then I built a car, so I got it with an actual disc 2 card. And then I built a card based on the IWM. I got that working.

James Lewis
After that point, the rev two stuff was, I mean, I probably spent another 6 months or so, well, 6 months total. So at that point I would have spent 3 or 4 months working with the rev 2, like trying to figure out how stuff works, but like, I didn't hit any major roadblocks at that point. Everything was just like, it was just like, okay, well, I've got to find the, the address for this function. What would it be? Right?

James Lewis
But there weren't any major like, oh gosh, this this will never happen. In fact, it was actually a relatively smooth process once I was able to read from discs. By about this point now, we're about a year and a half, a year to year and a half into the project.

Stephen Kraig
I have to, I have to commend your, your persistence with with just saying this is going to happen.

James Lewis
You know, it's funny. So I I didn't mention this, here, but so this was all during, you know, the 2,000, 2,000 or 2020 1 timeframe, 2021, 2022. And so during the lockdown stuff, I started live streaming 3 days a week. And when I decided to work on this project, I basically, for 12 to 15 hours a week, I worked on this project. And I think of all of the skills I learned, the one that I found most interesting is, as I got farther into the project, I got really good at picking milestones that I could accomplish in 4 hour blocks.

James Lewis
Oh, that's cool. There were just so many times where I'm like, okay, today's stream, we're going to try to solve this. And if I didn't get it on that stream, I got it by the end of the next stream, that kind of thing. It was just like, I just somehow I and I wish this was a skill I could explain, because I think I think we could all benefit from it. But it's like Yeah.

James Lewis
I was not saying,

Stephen Kraig
I wish I organization.

James Lewis
But I just I just started to get really good about, okay, this I need this much time to get this much to work. But also that was now I think about now I say that, that was also during the time frame where I stopped running into major issues. So that's probably a bigger aspect of it. Right? It's like just like any project, I could better I I knew more about I no longer didn't know what I didn't know.

Stephen Kraig
And something else to mention, I watched your your your video that's on YouTube about this. On that backplane board where you could slot in all the different functions, you had a header for an analog discovery DLA that just sniffs all the lines. That is brilliant. That is absolutely fantastic.

James Lewis
Yeah. Yeah. So if you recall, I said in the in the first kind of the first bring up, I had these 3 boards connected together. I had an analog discovery and a digital discovery plugged in at the same time. And, I mean, there's I'll put a picture so we can put it up in the forum.

James Lewis
It was just literally a ball of wire. Right? And when I was looking at it, I'm like most of those wires are the logic analyzer. That so what can I do to get rid of those connections? And so when I laid out the backplane, I'm like, why don't I just make it so I can plug the so it was the digital discovery.

James Lewis
Why don't I just make it so I can just plug it directly in? And so it plugs in and all 16 of the address and all 8 bits of the data lines go directly in. And so then I have to then I have fly wires for the clocks or, like, control signals or whatever. But I think I said this in the video. Of everything I figured out in this project, that has been I think that was the most clever thing I did because it just it I mean, anyone that's ever used a logic analyzer, you know that it's a bad day when you have to start hooking it up.

James Lewis
Right? Because you're you can't figure out what's going on. It's a pain in the ass to plug in, and then and then you have to figure out how to trigger on the thing you need. You know, it's just it's not no one's having a good day. And this was the exact opposite experience.

James Lewis
Like, it's like I pretty much just it was, like, it took me seconds to open up waveforms and hit run and see whatever I needed to see. Yeah. That was I'm still very, very happy about that.

Stephen Kraig
Yeah. Big thumbs up on that one. I I guarantee if I ever rep project like this, I'm going to copy that.

Parker Dillmann
You you know what's really funny about that is, wow, couple weeks ago, Steven, we had, Brendan on talking about continuous integration and development. Yeah. That's in line with that kind of style is build your tools into your prototypes, so to speak.

James Lewis
Yeah. Actually and this was and I didn't make a big deal about this here, but that when I came up with this rev 2 concept, a lot of people look at the back plane and they're like, oh, so you wanna be able to plug in cards. It's like, well, not really. What I wanted was I wanted a hardware development platform for the Apple 2, or in this case, the Mega 2. And it was stuff like that.

James Lewis
It's like, yeah, why don't I build in the logic analyzer? If I had to do it again, I would also build in and I don't know how I would do this, but I didn't like how I did the power. I would have liked to been able to make better power measurements and have better control of the power rails instead of just, like, I just had, like, a header pin that I connect the logic or, the power supply to. I would like to have, like, a sense resistor and maybe some fuses or something just to get a little better idea of what's going on without having to to have a, you know, a DMM or something.

Stephen Kraig
Well, it's very clearly designed with debugging and engineering in mind because you basically made a dev kit for the Apple 2.

James Lewis
Yeah. Yeah. Love it. Yeah. I've actually and I've actually thought about like doing a design like that for the actual Apple 2e, because that's something that more people will have access to.

James Lewis
Because kind of going back to, you know, the stuff is all open, but it's like, how do you get a probe on the clock? Right? That's that's kind of a pain.

Stephen Kraig
Alright. So in terms of what's next, so we've got something bootable, and we have something that you can interact with. What happens next?

James Lewis
Yeah. So more or less, the next major thing I did was I wanted to figure out how to handle video. I didn't wanna have a composite video signal, which is what the 2gs has, And so I won't get into how the Apple 2 video works because I don't fully understand it. But what I'll say is there's a signal called serial video, and it's basically a line, a single line of luminance information for a CRT. And so it gets combined with a sync pulse, and then that's how the computer basically draws your screen image.

James Lewis
The the MIGA 2 has that serial, video signal, which you can basically frame up and use. However, the color information is it's all kind of phase encoded and trying to determine where that those edges are with, say, like, pretty much anything is just a nightmare. However, the MEGA two also has a group of signals called RGB, which there's 4 bits for RGB, and it's a 16 it's basically it's a, a lookup table for the 16 colors of the Apple 2 palette. And so what's really cool about the MEGA 2 is it always outputs in the Apple 2's highest resolution mode, but it emulates the lower resolution modes. And so the reason I bring this up is because one of the most characteristic things about the Apple 2 is that because of the way it generates video, there's these ringing effects, which are artifacts from basically NTSC and the way that it generates color.

James Lewis
And so, like, if you look at a Apple 2 screen in graphics mode, white will always have, like, a blue and red fringe on it, and it's because of the way the color gets generated. Well, the mega 2 generates its video entirely digitally, and so those fringing effects don't actually occur. And so the mega 2 actually generates them internally. It actually generates what should be fringing effects so that when you look at its output, it looks like what an Apple 2 would even though it's entirely digital.

Stephen Kraig
So wait. It's it's emulating it? That baffles my mind why they did that.

James Lewis
Well, okay. So so I I just wanna be clear, Parker, which part baffles you?

Parker Dillmann
Oh, no. No. No. That the fact that they they designed the circuitry to produce those fringe effects in their silicon

James Lewis
Yeah.

Parker Dillmann
Instead of having a crisper line or anything like that.

James Lewis
So I do wonder if there's a soft switch that I can hit to turn it off. You know, I don't know how it was implemented, so maybe it's not maybe it's like part of like the shift register when it comes out or something. I don't know. But I wonder if there's a way to turn it off. I do know there is a video mode that the 2e does not have, but the 2gs supports because the MIGA 2 does it.

James Lewis
And it's like it's called double high res monochrome, and it looks gorgeous. It's like basically an EGA level of, resolution. But frankly, I think they had to do it because so much software and so many games relied on it for their to to make the graphics look the way they wanted them to look. Without it, things would look weird. And keep in mind, it only happens when you're in color modes.

James Lewis
So like if you're in a monochrome mode, it it's perfectly clear.

Parker Dillmann
Yeah. Yeah. The I remember back when I was doing Atari video mods, and the luminosity is all digital, And the color comes out at you were talking about the phase change as chroma, which is part of the NTSC standard.

James Lewis
Yep.

Parker Dillmann
And, like a Torrey 26100 just generates that out of the chip. So you don't actually get color digitally, but you get the, the luminosity is digital. So you can Okay. So when you if you if you combine everything right, you actually get really crisp black and white, but then you kinda have to fuzz the chroma so that, it still looks clearer than what would be like on a composite image. But Yeah.

Parker Dillmann
There was no way to do, like, a HDMI output from Atari without having to do, like, a frame buffer, basically.

James Lewis
So speaking of doing a frame buffer so basically what I looked at is like, Okay, so the MEGA 2 outputs these. It basically outputs each pixel of data on this 4 bit bus. And I'm like, okay, well that gets clocked at like 14 megahertz. And so I was like, okay, well I can use another PIO to capture that data. And then basically what I did was I captured the data as it comes out, I frame it up, and then I dumped it into a, well, I guess a frame buffer that then on the, the RP 2040, that some software that I found called Pico VGA, it takes that that frame of data and then it uses another PIO to output VGA.

James Lewis
And so I thought that was going to be the most complicated part of the entire project. And I think it took like 2 weeks to to capture the data, and then figure out how to decode it and start outputting. I was actually shocked at how fast that I got that to to work. And then by this point, this was about the 20th, circuit board I had designed. And, it was the first and only one that came up with no problems.

Stephen Kraig
Nice.

Parker Dillmann
You know, hearing that you can use an RP 2040 that way is I I need to start looking to that more because back when I was doing all that other stuff was you have to use an FPGA to do the capturing of of the signals, build up a frame buffer, and then let's say, outputting VGA. I've spent probably double of 2 weeks just compiling Verilog code to get my stuff to work. Mhmm. Because you could you could change one thing, and then it's, like, 15 minutes of fitting functions for output onto the FPGA.

James Lewis
Yeah. It's always the part where it actually takes forever. Yeah. Yeah. I mean, I got lucky because I had that basically a lookup table.

James Lewis
I I basically got the pixel data from MEGA 2. If I had to look at the video signal, the lumen chroma signal, I don't know that I could have done that with an RP 2040. Maybe, But I have a feeling that would have been I think I would have needed something to decelerate that, which is what the mega 2 is actually doing. And so, yeah. So then at that point I had a, in the, in the rev two form factor, I had a fully functional apple 2.

James Lewis
It was really exciting. I was actually I thought about stopping the project right there because it's like it were it works. I still use it. I, I actually like the Rev 2 design. It's kind of a it's a beast, but it's so robust.

James Lewis
But yeah. So then I was like, alright. Well, I always wanted to make well, originally, I wanted to make a portable Apple 2, but I became less and less interested in that. And then it was just like, okay. I want to make a miniature Apple 2e.

James Lewis
And so, of course, the next step was take everything that I had built, which, I mean, altogether was 1, 2, 3, 4, 5, 6, about 8 circuit boards, and combine it all into 1, which then became what I call rev 3.

Stephen Kraig
So rev 3 is finally the single board computer that you originally sought out for.

James Lewis
Yep. Very cool. Yeah. So it has all of the it basically has everything I need to do, everything we've talked about. It has its own switching switch mode power supply.

James Lewis
I borrowed an idea from the Apple 2c where there's like a slot in the 2c, and the power supply plugs into the slot. I kinda just like that, which I'm glad I did it that way because my first power supply was terrible. It had, like, 8 volt it was 8 volts peak to peak on the 5 volt rail, or the 5 volt rail went to 8 volts peak, and then so it was 4 volts peak to peak, 3 volts peak to peak on top of the 5 volt rail. So I was really glad that I didn't have to respin a an entire 4 layer board just to change the power supply.

Stephen Kraig
Yep.

James Lewis
And then I ended up I wasn't going to put a slot, but having a slot in rev 2 was so helpful that I decided I might as well put one slot in it. And, on that one, I actually have DIP switches, so I can assign which slot it actually appears as. But yeah. So that's got it's got, like, a a modern 65 c02, 128 k RAM, the Mega 2, an IWM for the disk drive, slot maker because Because you have to. You've got to have that.

James Lewis
And then there's 2 RP 20 forties. 1 does the keyboard stuff, and it's also the it's kind of a PEMIC. It monitors the rail, and then it brings up the Apple 2 stuff once it's happy. And then there's another RP 20 40 doing the VGA output.

Parker Dillmann
So now you can Apple 2 on the plane now. Right? Is that where we're at?

James Lewis
Pretty much. It doesn't have a screen. See, originally, I was gonna build a screen into it. And instead, what I did was I modeled. I I found a, on GrabCAD, I found a Apple 2e model, which was a very good looking replica, much better than the one I drew.

James Lewis
But it wasn't a real model. It was just faces, and so I had to modify it so that it could actually be printed. And so I did that, and I was just like, I just kind of started down that path. So I started looking at this as a miniature Apple 2e, and I just kind of went right down that path and said, okay, I'm going to make this into a miniature Apple 2e.

Stephen Kraig
So now that you have this all thrown together, what's what's the the future of it? Or is it time to wash your hands and say this project is is done?

James Lewis
Yeah. So here's the here's the little secret that isn't obvious in the video. I was so burned out. So I wanted let me just mention, the the rev 3 has a ton of problems. I screwed up the ePROM, and it took me a long time to get the thing to boot, and then there's a couple of other issues that I ran into that just weren't even related to design problems.

James Lewis
But it was an absolute nightmare getting this thing up and running. And so by the time I got it the board to work, I was like, okay. Dom, I'm supposed to make a video about it. I I'm like, I'm just gonna put this down and come back to it next month. And then a year went by.

James Lewis
There's a year gap from when I was done to when I made the video.

Stephen Kraig
Nice.

Parker Dillmann
So when's RAV 4?

Stephen Kraig
You tattooed, didn't you, Parker?

James Lewis
Yeah. So so you know what I wanna do on Rev 4 is I wanna I I don't wanna solder it.

Stephen Kraig
Yeah. You're all good. It's a it sounds like it's getting crazy over there.

James Lewis
Yeah. I'm yeah. I gotta I just gotta notice that there's a severe storm coming through. Yeah. So Rev 4, I actually want to use Rev 4.

James Lewis
Alright, we're just gonna have to live with it. Yeah. Yeah. Sorry, there's a storm going on. So no joke, Parker.

James Lewis
I actually wanna use assembly service for Rev 4 because I do not want to assemble all of the off the shelf stuff on another board. That was not a fun experience. It's it's too big.

Parker Dillmann
I'll see what I can do on my end.

Stephen Kraig
Wink, wink, hint, hint. Yeah. I'll see what I

James Lewis
can do. Oh. Oh, you understood what I was getting at.

Stephen Kraig
So are are all the files available for people to look at and dig through?

James Lewis
Yeah. Absolutely. So there I've got a GitHub repo. I'll give you guys a link to put on the forum. Everything I did is up there, including the, I think you including the the case files as well.

James Lewis
But, yeah, everything is there. I I saw no reason to hide any of it.

Parker Dillmann
Before we wrap up, so what when 3 slash 5 years ago, how long did you think it would actually take you to do this?

James Lewis
Literally, seriously, honestly, frankly, I thought I was looking at a month to month and a half. But, you know, so I went to a conference for Apple 2 users, and the joke title I used there was, how I built the MEGA 2 and accidentally reverse engineered the Apple 2 computer because I didn't realize how much I was going to have to learn about the computer in order to use this chip. And in retrospect, I don't know what I was thinking. Like like, like, I really did think I was just gonna be able to hook up RAM, RAM, and, CPU and have a computer.

Parker Dillmann
I really hope, the engineer that wrote that documentation that said, oh, yeah. Just put the ram the add ram in and a power supply, and you got this thing working. I really hope that engineer is still out there and watches that video and goes, I should've I I I had that amendment that that that gotcha, but I it didn't get approved, so it didn't make the documentation.

James Lewis
They forgot to add the footnote. Yeah. I mean, yeah, at the end of the day, the mega 2 chip is an apple 2e's logic, but that is it. It's just the Apple 2e's logic. And in fact, there's a handful of things that are slightly different from the 2e.

James Lewis
So it's not even truly like somebody asked me, could you make a drop in replacement for the 2e chipset? You know, it'd be really clunky looking, but I don't think you can because like, the way it handles the keyboard is different. You'd have to add glue logic in order for the keyboard to work correctly. So it's but but for the most part, yeah, It is what they said it was, but there was a little exaggeration in on how it worked.

Parker Dillmann
The other 40 pages of documentation just was not after it. Yeah.

Stephen Kraig
So Whoops. And that's not something you can go look up. You had to stumble through all of that, which, I don't know. I I think that's absolutely fascinating.

Parker Dillmann
So, before we let you go, James, what is gonna be the next 2 month project that actually ends up being 3, 5 years?

James Lewis
I'm not even.

Parker Dillmann
We're gonna look now I'm thinking about it yet?

James Lewis
Yeah. So so The wounds have to heal. Yeah. I'm being very cautious about getting into something like this again. And to seriously answer the question, I'm looking into a lot of machine learning stuff right now for edge devices.

James Lewis
And so I'm gonna go completely the opposite direction of retro stuff and start dealing with modern things, at least for a little while after I make rev 4.

Stephen Kraig
To actually finish the project. Call it what what we say at work, done done. Done done. You've hit the first done. You did the second done.

Stephen Kraig
So for all of our listeners, there's a there's a really great YouTube video that goes into a little bit more details of this and you get to see all of James's, boards. It's on the, element 14 presents YouTube channel.

James Lewis
Yeah. I'll give you guys some links. There's actually there's 3 videos in total for the Mega 2. So there was the first video, getting to the boot, this final video, and then there was a talk I did at a conference called Kansas Fest, which is a annual conference for Apple 2 users. It was it was kind of in the in the middle.

Stephen Kraig
Awesome.

Parker Dillmann
Well, again, thank you so much, James, for joining us. For those who want to know more about James, visit baldengineered.com, which is kinda his website that combines everything that he works on. Because I actually I went there today, James, and you've got, like, 4 or 5 different it's like an umbrella website because you got, like, 4 or 5 different things going on.

James Lewis
Yeah. I try to make it look kinda like my Linktree. So it's like everything I'm doing, I'm just I just put there.

Parker Dillmann
Yeah.

James Lewis
I used to regularly write on it, but now it's just kinda like, here's everything I'm doing.

Parker Dillmann
But we'll have that link in our show notes. And, yeah. I guess that's gonna be the end of this podcast. Thank you so much, James.

James Lewis
Yeah. Thank you. Thank you, guys. It's been a blast.

Parker Dillmann
Thank you for listening to circuit break. We're your hosts, Parker Dohlman.

Stephen Kraig
And Steven Craig. Later, everyone. Take it easy.

Parker Dillmann
Thank you. Yes. You are a listener for downloading our podcast. Tell your friends and coworkers about Circuit Break, the podcast for Macrofab. If you have a cool idea, project, or topic you want us to discuss, let Steven and I and the community know.

Parker Dillmann
Our community where you can find personal projects, discussions about the podcast, and engineering topics and news is located atformdot macraff.com.