User avatar
intric8
Seattle, WA, USA
Website

Posted Sat Dec 11, 2021 9:48 am


For Amiga 1000 fans, the Rejuvenator by Greg Tibbs has always been a Holy Grail upgrade: almost impossible to find, and when it is found it’s old, battery eaten and still expensive.

As of today, I can happily say that no longer need be the case. It is now 100% possible to build brand new A1K Rejuvenators. Period.

As a quick refresher, the Rejuvenator is an upgrade board that gives Amiga 1000 users the following:
  • * The ability to use a physical KS ROM
    * And Internal RTC
    * Video Port, for a small handful of cards
    * ECS Denise
    * And most importantly, the ability to bring 1MB Chip, or even a staggering 2MB Chip RAM and Fat Agnus, to the original and beautiful Amiga 1000.
To do this the Rejuvenator replaced the daughterboard inside the Amiga 1000’s that mainly sold in North America as well as a tiny handful of early A1ks in Europe before it was cost reduced and the daughterboard was integrated into the motherboard.

Swapping out the NTSC daughterboard with a Rejuvenator made Amiga 1000 owners feel like they weren’t being ignored anymore whilst standing in the shadows of the younger siblings: the Amiga 500, 2000 and even the 3000 at that point.

While Commodore was literally paying Amiga 1000 owners to trade in their machines to get brand new Amiga 2000’s at a heavily reduced price, Greg Tibbs was finding a way for A1K owners to hang onto their original investment and push it further into the future the way many owners deeply believed Commodore should have offered, but didn’t.

When Greg Tibbs showed his Rejuvenator board at the AmiExpo convention in Washington DC in 1990, he told me, “Jay Miner came by on the show floor and I showed it to him and he was so happy that he hugged me!”

Jay knew. He knew this was super bad ass and something to be celebrated.

Sadly, Greg’s Rejuvenator was only on the market for about two years and ultimately was shut down after about 800 hundred boards were sold. Greg claimed to have never been paid his royalties (somewhere around $50K) but was instead given an Amiga 3000.

Thus, Greg left the Amiga world and the original Rejuvenator boards became a part of Amiga lore over the past 30 years. As many faded away into the shadows over time, the few that remained would create a total feeding frenzy for Amiga fans any time one ever stepped out into the light of day.

I happen to be the very fortunate owner of two original Rejuvenator boards. One is apparently a very early model, and the other the more common “production” model.
og-1.jpg
og-2.jpg

I decided to see if it would be possible to reverse engineer the board. I didn’t have the skills, but I knew someone in the vast Amiga fanbase must.

So, I started first by hunting down Greg Tibbs to see if it would be OK to give it a shot. And who knew? Maybe he still had some disks with schematics or something, right?

I’m not going to go into the entire story of how I got from there to here, or of how I found Greg - I have past posts and videos that go into all of that - but I did find him. And we ultimately shared conversations over the phone and via email. He not only gave me his blessing, but he honestly wanted to try and help if only as a guiding mentor to see his invention brought back to life.

Greg had an incredible memory of his work and his Amiga days. And while all of his original files had been thrown away years ago, he still remembered a great deal about all of the people, places and things he had to do to get his Rejuvenator to market.

Sadly, Greg died in January of 2020, so he never got to see his board rebuilt to completion. But he did get to see it reverse engineered most of the way, thanks mainly to an exceptional talent and man named Joe Carter, who lives in Idaho.

Joe cloned the entire board using a pre-production bare board I had sourced from one of the owners of Expert Services. I also sent Joe one of my original Rejuvenators. Joe not only cloned the board, but he also fixed a few bodged traces and other errors he discovered along the way. In fact, he assembled and tested one of his brand new Rejuvenators two years ago in early 2019! Greg did get to see that success, which made him very proud.

The only piece of the puzzle that remained were 4 special chips that contained equations we couldn’t pull out, nor was there any documentation anywhere to simply recreate them. Several attempts were made to read the chips but we simply couldn’t get it to work. We began to worry they were read-protected. And of course, Greg wasn’t able to help except with one sideways comment: that he was certain they weren’t read protected. Why we couldn’t read them, though, he had no idea.

Before completely giving up, I contacted another legendary engineer in the Amiga community based in Germany who goes by the moniker Matze, aka Matthias Heinrichs, out of Berlin.

Working with one of the administrators of A1K.org, Nelson, aka Marco Both, I took a chance and shipped one of my original Rejuvenators to Germany - the same very early model I had loaned Joe - in the hopes that Matze could unlock the secrets of the 4 chips. Keep in mind this was in the midst of the pandemic and packages being shipped to Europe were sometimes completely disappearing, or taking several months to clear various checkpoints. It was totally nerve-wracking following the shipment tracking!

But it did make it to Germany, and eventually to Matthias.

A couple of months went by and then… Matze stunned us all by declaring success! He had found a way to read the 4 chips! But how did he do it?

He didn’t use modern hardware - that’s how!
MatzeSuccess.jpg

The following quoted passages and photographs are from Matthias Heinrichs himself. All photos used in this section to show Matze's process to ultimately read the original PAL chips are owned/© Matthias Heinrichs, used with permission.


Step 1 & 2:
“I was lucky to get an ALL07A-Programer by HiLo-Systems on Ebay for 60€ and still had a Dell laptop with a real parallel-port on the docking station. I installed the software on a USB-stick using FreeDos for booting.”
02-Programmer.jpg
01-Setup.jpg

Step 3:
"The device driver is pure DOS in typical EGA-graphics - ironic that a DOS-program saved this project ;)"
03-Device Menu.jpg

Next Matthias inserted some chips into his reader to better understand how to use it.
“Of course I tried the reader using some PALs from an A3000, which could be sacrificed first.”

Once done, he had to select the appropriate chip from a HUGE device list (The AL007A can read even ancient soviet-ICs!)

Step 05 - 06:
"Now, we get into the chip-menu (notice the "checksum: 0000" field)."
05-Select Chip.jpg
06-Chip menu.jpg

Step 07:
"Hit "R for read" - the checksum changed!!!!!!"
07-Reading.jpg

Step 08:
Hit "d for display" - Whoa! A read-protected device usually shows only ones or zeros but never mixed! That means they are not read-protected! I was so happy seeing those cryptic ones and zeros.
08-Checking.jpg

Step 09:
Save that thing!
09-Saving.jpg

He repeated this steps for Rejuvenator board PAL chips U28, U29, and U30. "All ICs came from three different manufacturers. I had to [visually] match one manufacturer by its logo (It was National Semiconductor with an unknown old logo)."

Step 10:
Add some "infomagic" to the JED-files. If the pin-names are specified in the comment section above the fuses, they will be used during the equation-reconstruction-process. You see the original JED on the left and the "enriched" JED on the right. I used the Pin-names from Joes-git project.

For reconstruction the equations from the jed I used the software-kit OPAL by National Semiconductor, which I run in dosbox.

Step 11:
JED2EQU converted the read-out JEDs to equation

Step 12:
EQN2JED translated them back to the desired device. (It turns out OPAL is not y2k-safe and thinks we are in 2121!)

"That's all folks!"

-- Matthias



Once this information hit my Inbox, we all pretty much lost our damn minds! As quickly as we could, we started to try and burn the equations to our own ROMs here in the U.S.

After a few misfires, we got it working and the rest is, well, history.

Soon, some of us built our own brand new boards, updating the BOM and a few other files on Github along the way.

And as of November 21, 2021 - almost 3 years since I started this crazy project - we officially have a Github project updated to the point where others can follow in our tracks and use our research and findings to build their own boards.
new-rejuve.jpg

I’ve been asked by several folks in the past few weeks if we plan to sell Rejuvenators and for how much.

Frankly, no. We aren’t going to sell anything.

This entire enterprise was to bring the Rejuvenator back to life, but was never intended to be a way to try and profit off the Amiga community. Others can go down that road - it’s not for us. We'd considered an option to produce the boards 3 years ago, but times have changed. And these boards are not simple builds, compounded by the fact that some parts are difficult to source.

All of our time and money spent was entirely for the betterment of the Amiga community, not for lining our pockets with cash. This is a hobby about a computer we all love. That’s it. And we’ve all gotten what we ultimately wanted - brand new Rejuvenators that cost about $150 in parts alone! That’s not including the custom chips they require, of course.

So even if we did bring these to the market, we’d likely get hit with all of the FPGA crowds on one side pointing at MiSTers and stuff like that, or the software emulator folk on the other side, and just a lot of drama we really don’t really need in our lives, you know what I mean?

But for those of you out there that want to keep it “old school” and want one of these so you can use your own Kickstart ROMs and 1 or 2 MB Agnus chips, I’d be willing to bet you they’ll start to emerge in 2022 on both sides of the pond. Mark my words. Some might even be so bold as to try and put their names on the silkscreens.

Meanwhile, Joe’s not sitting still.

He’s already moving on to design a highly experimental future version that might allow for the much more common 2MB Agnuses, rather than the rare A3000 ones the original Rejuvenator uses. And when might that emerge? It might be never - but I’d never bet against Joe.

Joe, on Greg Tibbs’ passing:
He has left a mark so wide through his innovations, that he will always be remembered by Amiga 1000 aficionados as the creator of the ultimate upgrade.

A toast to Mr. Greg Tibbs, the genius who has directed a very willing portion of my life, and has taught me so much in the hobbies I love. Much respect good sir.


The Rejuvenator 2 Team:
Joe Carter
Matthias Heinrichs
Christian Stich
Matt Martin
Adrian Garay
Marco Both
and Eric Hill

We did it, Greg.

User avatar
Crispy

Posted Sat Dec 11, 2021 11:18 am

Fantastic work! In the words of Hannibal Smith, I love it when a plan comes together.

:boing: <3


User avatar
JackTheKnife

Posted Sat Dec 11, 2021 3:40 pm

Well done and very interesting epilogue video to the whole project. Looking forward what will bring 2022 <3

User avatar
obitus1990
New Orleans, LA, USA

Posted Sat Dec 11, 2021 3:46 pm

Most Excellent...and thank you again to everyone involved in the project!

User avatar
Signman

Posted Sun Dec 12, 2021 7:19 am

You really should be proud of this project and the effort and diligence it took to come to this nice conclusion.

I bought my first Amiga as a young man in February 1986 at a cost I probably shouldn't have incurred but no need to explain that here. I pushed it to the limit with memory, Digiview with camera etc. In the end the 512k chipram was such a hindrance even in Dpaint. So finally bought my brothers A500 I convinced him to buy years earlier for the extra chipram possibilities. Still I felt I bought an inferior machine even if the specs said otherwise.

Now I have a different Amiga 1000 (or just Amiga back then) and would love to do utilize this project. Great Job.

User avatar
oggie

Posted Sun Jan 02, 2022 3:05 pm

I'm about to get myself a PCB for this, and the BOM appears to missing the resistor values for r1-r14. Any chance someone knows what they are?

I pulled up the schematic, and this is what I found:
470 resistor r2,r3
10k resistor r8,r13
33 resistor r6,r7,r9,r10,r11,r12
4.7k resistor r14
3.3k resistor r4,r5
47 resistor r1

Also BOM says 47uF for C32, schematic says 22uf. Which is correct?

I also noticed the BOM is missing RTC parts:
y1
c24
c30
c21

Anyone know what to use for y1 and c24?

And finally, the silkscreen and schematic shows polarity for the decoupling caps, but standard ceramic caps don't have polarity. Are they not basic .1uf ceramic caps?

User avatar
Fantomen

Posted Sun Jan 09, 2022 4:34 am

I also have a PCB on the way, so please let me know if you find out the correct values for the components.

User avatar
intric8
Seattle, WA, USA
Website

Posted Sun Jan 09, 2022 9:56 am

Some of you guys might ask site members mattsoft, joethezombie, or obitus1990. They have already all built theirs. And if you feel the need to update the github, I'm sure that's a thing that would benefit others going forward to reduce guesswork.

User avatar
obitus1990
New Orleans, LA, USA

Posted Sun Jan 09, 2022 5:50 pm

I did not populate the RTC circuit.





Return to “Hardware”