As you can see from Gunnars statement in "Feature Creep" CLICK HERE the people in the team are not just playing around. There are quite different tasks, most of them are kind of software and as a constant reader of the forum you see that they make quite good progress. But most of the things are done in emulation because real hardware is missing. And this is the difference to the mentioned open source projects. The open source projects exactly know how their system will behave because they are coding on real existing hardware. So they can tell in advance exactly how their final product will be. Unfortunately we don't. We know that the new NatAmi hardware will work in some way because it is based on the "old" prototype. I think it was not mentioned before - the "old" C-One prototype was about the speed of an A3000 achieved by technology slightly better than an A3000. With the new board design we are skipping two generations of FPGAs and are also skipping two generations of DRAM. This is not just a slight update this is quite a "technology warp" for the Amiga system.

In other words some basic hardware concepts are changing and need to be implemented:
- new FPGA generation
- new memory generation and therefore new register arbitration design
- new video out circuit
- new audio out circuit
- additional SynchronZorro bus

Because of that the next board which is the first real Natami board will be an "natami evaluation board". This board will not necessarily be the then following developer board. It contains all necessary hardware and peripherals to put the natami in operation on the new technology but in contrast to the current "high end" proposal it will have a smaller FPGA and 16bit RAM instead of 32 or 64. The duty of this board will be to provide a full hardware environment on which the following developer boards can be based on.

The next major step for the hardware will be to provide such a board and have OCS/ECS/AGA running with an MC68060 processor. Without any of the new extensions. This will be the base of the developer board on which new feature integration steps will take place.

There will not be a random toying around with strange ideas - there will be an implementation and integration with following exact documentation in a feature by feature style.

It will not be "the big problem" implementing "some" new features. But it will need some care how to implement them in an Amiga like style and contributing to it rather than trashing the Amiga philosophy and implementation. We explicitly do NOT want to create addons in PC GFX card style and patch them together; every two months a completely new design and new interface. This is not the way. We will spend quite some effort in taking care about all interdependencies of the whole Amiga system concept and design.

And now to the part why this article is in the "news" section. As you already know there is a 68060 CPU board in schedule. My expectation was to have it working by now - but unfortunately it is not. At least not the one shown in the picture of the Hardware section. This phrasing suggests a backup plan... from which I will tell later.

I experienced some further problems with the CPU board in addition to getting familiar with the RoHS soldering process. Accidentally I burned the first board while soldering with the wrong temperature curve. OK, that's bad luck I thought but no problem in doing it right next time. So the next day I soldered two boards (with the right temperature this time) and they looked really well. No shorts, no misplacements. I assembled two boards at once because I was completely sure there won't be further problems.

Because there is no baseboard in which I can plug in the board I need to solder two thin wires for power supply onto it. With checking those wires on the first board for polarity I found out that there is a shortcut between ground and 3.3V power. So I carefully checked the complete board but found no defect. So I got suspicious and had a look on other board. The other board and also all other unassembled PCBs had this same shortcut. So it seemed to be a systematical error and I doublechecked the layout data. But found no error. One day later I wrote an email to the manufacturer and asked them if they could see a connection between the nets in the production data. Only one hour later an email came back saying "sorry, there was an error in the CAM preprocessing, new boards are being manufactured which will be shipped in a few days - excuse the inconvenience".

So I am now waiting for the next set of error-free PCBs to arrive. This will take at least until next week.

Thank you for sharing your plan and explaining details to us! Good luck with your next board set.

mfG,
Bartek

You mention new generations of DRAM and FPGA.  Which kind of DRAM and FGPA is planned to be used in the new update?

Thanks again for the work.

question:  Did you ever read about or use any of the Sage Computer's from the Mid 1980's?  The company later was known as Stride Micro... an interesting design.  The 050 and latency goals you have set forth remind me of the original Sage concept... very interesting.

Thanks again.

Cheers

No.

Thomas does only handsoldered the very first 1 or 2 test boards.

That Thomas even handsolders BGA chips is exceptional - as many people are not able to do this.

If the first tests are done successful then all the boards will be automatically be soldered.

Cheers

the problem with leds it that you need to replace it with one of the same BIN value.

Kuddo's to Thomas for handsoldering not one but more BGA packages on those prints.

If he does, I believe we can expect NatAmi in 11+ months. If he doesn't (done by a robot or a company) I can assume that we gain 1 month? Let's say 9-10+ months?

The Ball Grid Array is one of if not the hardest thing to hand solder.

BTW it gets placed by a robot and cooked in a oven if it's SMD/SMT production.
mains with ease in one day you could make a natami or 200. if not more.
This depense on the production facility. ;)

sorry to be offtopic as usual

I don't see how you can base a timeline on Thomas post.

What Thomas said is that he is working on a development board.
Thomas said nothing about the timeline.
The board Thomas is developing right now could be available next month or at any other time.
I would not start guessing timlines from an outside view.

I would regard any Natami board is Natami.
I would assume that the first batch will be "full" Natami/AMIGA compatible as the next verions, and later version also.
I think for interested users/developers also the very first Natami model might make good sense.

Cheers

I didnt say "it will be ready at: " I said "based on estimation it won't be ready until: ". And if you read carefully my post and Thomas post, anybody can really make an estimation. You dont have to be a rocket scientist. You need time to get the cpu cards, to prepare them, to design the pre-natami, to be manifactured, to test on it, to design the devboard, to be manifactured, to test it, etc! It is not going to happen in 6 months! I would be more than happy if you can proove me wrong. You know that. I am waiting for natami like a child waits for xmas :-)

The truth is that Thomas is working on the mainboard layout already since very long time. Thomas started first designing the mainbaord, only because we wanted to test a charge of 68060 chips in parallel  Thomas quickly designed the CPU card in parallel to the mainbaord.
I say quickly here as the mainbaord has more parts on it and is therefore more complex.

The thruth is also that Thomas has designed two different CPU card versions and that he has working CPU cards. And Thomas is already hording the parts the manufacture the mainboard to prevent delays.
 
So the fact is that Thomas does not start with the mainboard only after the CPU card but rather otherwise around. And I would not dare to guess a timeline out of Thomas post.
It could be that Thomas will be ready with the mainboard desing next week. But it could also be that he will be ready in some month and even if he is ready it could be that something odd happens (some parts are not available anymore etc) and that this will delay everything again.
 
Also you seem to wait for the 2nd or 3rd version of the board.
I see no reason for this. If the 1st version is working then it will be as good Natami too.

I personally would want to get the 1st model. :-)

Keep going on sharing your experience with us.
I soldered some DCE flicker fixer on an A4000T using two sockets to rehigh the FF module and it was a mess.
I'm not an electrical engineer but sometimes it is necessary to understand the basics of computer electronics.
I was please to read you Thomas.