Enhanced/Improved ST motherboard - H5 Phoenix Project

[stormy]

Hi all,

For some strange reason there is little to no talk of the following project - the H5 Phoenix motherboard (Atari ST Remake Project) by Exxos. I feel that our community is quite small compared to others (like Amiga for example) And when we are lucky enough to have such projects being worked on, it is important for the community to talk about, support and spread the word! I hope that by making this post more people will be informed about the H5 Phoenix project and hopefully understand it more.

Phoenix 4160 STx Badge.jpg

What is it exactly?

The H5 is a complete re-creation of the STf motherboard (with blitter) which is physically (will fit in the original case) and electronically compatible with the original Atari STf (Not E, as the STe models don't have DIP chips for most custom ICs, where as the STf does)

Why would you want one?

Our old motherboards aren't getting any younger, problems are always plaguing people like for example: floppy subsystem failures, data-bus weaknesses, faulty traces due to poor manufacturing, inherent bugs in the design like the infamous DMA faults when using Ultrasatans etc.

Why not just re-create the original motherboard, like the Amiga guys do?

Because there are inherent design faults in Ataris original design, these design faults cause noise on the bus causing accelerator instabilities, problems with DMA causing corruption over ACSI hard disk transfers - any many more - Atari really did make the ST to 'Rock Bottom Price' and it shows.
The ST also has next-to-nothing in the form of easy expansion, unlike the Amiga, it didn't have any kind of bus expansion possibilities etc - the H5 rectifies this by adding 4x 5v bus expansion 68k headers and 2x 3.3v 68k expansion headers. This opens up the opportunity for the community to build much more amazing expansion products for our machines. This is why we need something more than just a normal motherboard copy, we need an improved product with 100% compatibility with the original.

So, if I get a H5 and populate it with the chips from my old STf motherboard, what other advantages does this give me?

* Since the system was completely hand-made and routed by Exxos (with many previous iterations, this has been a lengthy development!) The system runs CLEAN.
* You get beautiful noise-free graphics output (the H5 uses the same video DAC as the Falcon) via regular Atari monitor connection or by VGA.
* You get beautiful clean and crisp sound output.
* The floppy circuitry runs smooth and so does the DMA ACSI data transfers.
* You get 4x 5v (legacy) dedicated 68k sockets for EXPANSIONS!
* You get 2x 3.3v (modern) dedicated 68k sockets for even more EXPANSIONS!
* You can plug a Falcon 4mb RAM expansion directly into the board
* There is a SIMM slot for 4MB of RAM
* You can easily install a blitter chip
* You can install a Dreamblaster MIDI module directly on to the motherboard

What kind of expansions can I expect to use on the H5?

Since there are 4x 5v legacy 68k expansion sockets, you can fill these with any previously designed expansion that were made to run on old Atari's, like the cloudy & stormy etc, PAK, etc etc etc.

The 2x 3.3v expansions are for NEW designs based on todays accepted voltage of 3.3v, like FPGAs etc, being much more efficient not requiring voltage conversion on the addons themselves.

Here is a LINK to the custom-made H5 addons that are either already available or in the pipeline


Do you know that you can plug a Dreamblaster (Midi wavetable) board directly into the H5? So that you can listen to your MIDI music games without requiring an expensive MT-32 or Yamaha device?
Check out the following forum thread!

IMG_5842.jpeg

Where can I buy one?

From the Exxos store

Other relevant links to do with the H5:

The H5 roadmap for 2023

The H5 Forum

H5 FPGA addon development forum

Phoenix zone forum

1.jpg

2.jpg

[1st1]

4 CPU sockets to plug in expansions? Crazy as the (historic) expansions will be larger than the socket space or they are not designed to co-work with other socket expansion boards. The space should better be used to already add sensefull things like TOS 2.06 socket, IDE, autoswitch overscan, PAK, pc-speed, etc.

[stormy]

4 CPU sockets to plug in expansions? Crazy as the (historic) expansions will be larger than the socket space or they are not designed to co-work with other socket expansion boards. The space should better be used to already add sensefull things like TOS 2.06 socket, IDE, autoswitch overscan, PAK, pc-speed, etc.

The motherboard has flash-able dual-tos built in already. You can install the Trudie IDE into a socket easily, or the stormy/cloudy? It all fits. The PAK is massive anyway so of course that's gonna take up space. These days you're better off with a Terriblefire 030 card anyway. Or wait for the SEC 75Mhz booster. Wait, you're complaining that there is *too much* expansion? Well, nobody said that you *have* to use it all ... not sure what your point is.

[saulot]

I would die for a motherboard which has short ST form factor. Also project reworking STe board would be cool.
I think that point about plugging Dreamblaster into motherboard is great, but shouldn't be advertised that it will save you from getting mt-32 or similar, because mt-32 has it's own distinct sound and it is pre-midi standard device, which Dreamblaster cannot emulate (DB is something like Soundblaster/Adlib thing with opl chip).
Also it should be mentioned that shifter replacement is on the way, so there is possibility to use hdmi. Anyway it's great that we have projects like this.

[AtariOnion]

I had no idea any of this existed - thanks for summarising.

Question -

Will we always need to have an existing STF case to put the motherboard in? Any plans to actually have this available as a 'consumer ready' computer, in a refurbed (or even new) case.

Would love to buy one of these.

[Cyprian]

For some strange reason there is little to no talk of the following project - the H5 Phoenix motherboard (Atari ST Remake Project) by Exxos.

great project, I observe it on Exxos forum.

I would die for a motherboard which has short ST form factor. Also project reworking STe board would be cool.

I'm waiting for short STe motherboard

[saulot]

Short STe mainboard, I was afraid to say it loud ... But yes ..

[frank.lukas]

... the problem in my opinion is that the good original Atari Computer has to be destroyed for this. Suska is better in my eyes and offers more options, albeit more expensive ...

http://www.experiment-s.de/de/

[DarkLord]

Isn't there a lot of problems with the Suska? I visited the website, many
items are listed as "no longer in stock". The III-B has only a few left and
as you said, it's fairly expensive.

Perhaps I'm missing something but I can't recall -1- single post by anyone
using a Suska and proclaiming it's virtues.

Is anyone using one daily that could report on it? Thanks.

[stormy]

... the problem in my opinion is that the good original Atari Computer has to be destroyed for this. Suska is better in my eyes and offers more options, albeit more expensive ...

http://www.experiment-s.de/de/

You don't have to use a good/working one, you could use a faulty one. There are plenty of faulty old motherboards being sold on eBay.

[sporniket]

4 CPU sockets to plug in expansions? Crazy as the (historic) expansions will be larger than the socket space or they are not designed to co-work with other socket expansion boards. The space should better be used to already add sensefull things like TOS 2.06 socket, IDE, autoswitch overscan, PAK, pc-speed, etc.

There is a lot of expansion sockets because it is was also meant to make life easier for hardware developpers : should a port becoming faulty, a repair can be postponed by moving to another ports.

A simpler version with just one 5v expansion port and one 3v3 expansion port was planned, but you know how it is, too much projects, too little time.

[TheNameOfTheGame]

It's a very good project and kudos to Exxos for seeing the development through. I own a mainboard but just haven't had time to build it up. As with all things, time is in short supply. It would be nice to see an STE version though.

[DarkLord]

And a STacy version...

(but I'm not holding my breath on that one)

[Atari030]

... the problem in my opinion is that the good original Atari Computer has to be destroyed for this. Suska is better in my eyes and offers more options, albeit more expensive ...

http://www.experiment-s.de/de/

You don't have to destroy a board, mostly only the MMU, Shifter etc need to be pulled and they are socketed on most boards. I think the 1772 will be the only problem component. I'm building an ATX ST and not had to destroy anything. I bought the 2149's, ACIA's etc new.

The H5 is a long way down the line for me (too many projects and machines) but it is a thing of beauty. I'd love to see it in a MegaST or ATX form factor.

[stormy]

It's a very good project and kudos to Exxos for seeing the development through. I own a mainboard but just haven't had time to build it up. As with all things, time is in short supply. It would be nice to see an STE version though.

I keep seeing people ask for an STE version though, why do you think that is? Don't people understand that you'd have to manually desolder the Glue chip from an STE main board which is QFP something like 100 pins? I highly doubt you'd find many people with the skills to do that, so basically he'd have nobody who'd be able to buy the motherboard... Can you see how that's a barrier.

[Cyprian]

It's a very good project and kudos to Exxos for seeing the development through. I own a mainboard but just haven't had time to build it up. As with all things, time is in short supply. It would be nice to see an STE version though.

I keep seeing people ask for an STE version though, why do you think that is? Don't people understand that you'd have to manually desolder the Glue chip from an STE main board which is QFP something like 100 pins? I highly doubt you'd find many people with the skills to do that, so basically he'd have nobody who'd be able to buy the motherboard... Can you see how that's a barrier.

why? because the STE is more rare than the ST.
Desoldering is not needed as the STE chipsets are still on the market: https://www.best-electronics-ca.com/custom-i.htm
For me there is no difference in desoldering any Atari chips, 68k, YM or GSTMCU, I would not do that myself and I would ask for help someone more skilled than me.

Anyway I would buy a new Falcon motherboard.

[mikro]

Great summary, Stormy!

[TheNameOfTheGame]

It's a very good project and kudos to Exxos for seeing the development through. I own a mainboard but just haven't had time to build it up. As with all things, time is in short supply. It would be nice to see an STE version though.

I keep seeing people ask for an STE version though, why do you think that is? Don't people understand that you'd have to manually desolder the Glue chip from an STE main board which is QFP something like 100 pins? I highly doubt you'd find many people with the skills to do that, so basically he'd have nobody who'd be able to buy the motherboard... Can you see how that's a barrier.

Uh, sure, I can see that's a barrier, but so is the current method. Not many people can even solder DIP packages. And the updake on H4/H5 isn't really that much anyway. This is a total hobby project and we Atarians are super-small group getting smaller every year it seems. Hobbyists tend to tinker, hack and solder. Expecting a product with no "barrier" to entry, you probably should be thinking of a fully populated board with a molded case that the user can put on the desk and use right away.

As for the STE aspect, they are very rare here in the USA. I have a couple, but one is getting flakey (need to recap) and they're just getting older and older. STF(M) units are fairly easy to get although sure you have to pay a bit more than it used to be. And the STE is better imo in the basic offering with its upgraded feature set. I know the H4/H5 can be upgraded, but for games playing the STE still has advantages.

[stormy]

Yeah but you guys @TheNameOfTheGame and @Cyprian are just thinking from a * very * niche viewpoint - the viewpoint of 1. purely a consumer (not taking into account the complex and detailed work required on the business end) and 2. Also the potential for customers to buy such a complex product. Cyprian says soldering a 144 pin QFP package (MCU/Glue) is as simple as a 16 pin DIP package - I thoroughly disagree! I could not disagree more. How much emphasis can I put on that? lol

Yes, you guys do need to consider that creating a product that only 1% of the already tiny community (as TheName pointed out) is something like pie-in-the-sky.

Also you're both misunderstanding what the H5 Phoenix is, this isn't just a 'stock STf' ! Please understand, it * can be * a stock STf if you want it to be, but with the nature of future FPGA add-ons in it's many upgrade sockets you can expand it far beyond a base-line STf. You guys seem to be wanting a 'stock STe', this is a completely different philosophy.

This is why the H5 slogan is 'A platform for the future'.

If it's not for you, rather than being negative towards the project, please just be understanding and supportive.

[Badwolf]

It's a nice project and one that's a good complement to the MiniATX board. Together they cover most bases.

I had no idea any of this existed - thanks for summarising.

Question -

Will we always need to have an existing STF case to put the motherboard in? Any plans to actually have this available as a 'consumer ready' computer, in a refurbed (or even new) case.

Would love to buy one of these.

Icky (the co-developer of the Phoenix platform, of which H5 is a part) gave a nice interview to me at the Cyberlegends show at the tail end of last year. Here he displays a board, some of the addon and (whisper it) a new case under development too.

Have a look (Icky at 14m 43s):-

https://www.youtube.com/watch?v=iSuumzEbSfE&t=883s

BW

[Wayne123]

I had fun building mine, not that hard. If you need practice desoldering and soldering buy some old junk electronics to practice on. Desoldering stations are cheap now and so are decent soldering irons.

[Cyprian]

Cyprian says soldering a 144 pin QFP package (MCU/Glue) is as simple as a 16 pin DIP package - I thoroughly disagree! I could not disagree more. How much emphasis can I put on that? lol

I didn't say anything like that. Read my comment again carefully.

Also you're both misunderstanding what the H5 Phoenix is

you are not right, again. I understand it well.

rather than being negative towards the project, please just be understanding and supportive.

you are not right, again.
I am more than positive about this project. I've been following it from the beginning. I really appreciate Exxos work. And I keep my finger crossed for this project.

If it's not for you,

at that moment yes, for reasons mentioned earlier.

Anyway, please stop being so negative towards us.

[stormy]

It's nice to see that you're positive though - I just would not have guessed that otherwise. I'd rather this thread be supportive towards the H5, we're a small community and just putting something down because 'you'd rather have an STe' or 'you'd rather a Falcon' isn't productive. If you think I'm miss-interpreting you, sorry, but I just read what I see.

[TheNameOfTheGame]

It's all good my friend. If it wasn't clear before I think the H5 is a great project and accomplishment. The options and expandability are intriguing and looking forward to getting mine built .

[mikro]

If I can make a suggestion for a potential hole in the market - someone could offer a service "supply a (broken) ST, get new populated PCB in return". :) From what I could gather, many people simple lack time/skills/will to (de)solder so many components.