New Amiga motherboards from Jens

[dJOS]

Originally Posted by Daedalus

I just picked that number from the air. Yes, parts would be cheap but development not. Think of the cost of the existing custom USB solutions - you'd need similar levels of development with a similar raw materials cost so it's not too far a stretch to think that adding USB to the board would add a cost not far off that of the stand-alone card.

As for the cost of the Raspberry Pi, there are a couple of reasons for this: First is the volume of Raspberry Pis produced which is already around 50,000 times higher than the number of boards Jens will sell. That's a different galaxy when it comes to economies of scale, buying parts etc. Second, the RPi is a not-for-profit venture so no profit is included in the price. This also means the RPi Foundation can avoid taxes and so on which reduces the cost even further. And third, it's using a SoC design that already includes pretty much all the interfacing electronics that are needed "for free", which is certainly not the case for the Amiga motherboard.

Don't get me wrong, I'd love to see USB etc. on this new board too, but it doesn't make sense to include all the things that anyone wants when it will increase the price for everyone. I'd much rather pay €350 for a board and modules that does what I want it to do than €450 for an all-in-one board that does a lot of stuff that I don't want.

And that's why I suggested integrating a cheap soc as a co pro to run the extra hw. In theory you could use the soc to offload all the tcp IP, usb, auto config management etc.

Anyway, unfortunately I'm not a developer so I've no idea how hard this would be.

[pandy71]

Originally Posted by Daedalus

I just picked that number from the air. Yes, parts would be cheap but development not. Think of the cost of the existing custom USB solutions - you'd need similar levels of development with a similar raw materials cost so it's not too far a stretch to think that adding USB to the board would add a cost not far off that of the stand-alone card.

Don't get me wrong but HW is not problem - XMOS 4 - 6 core (tiles) CPU cost in single quantities around 4 - 5$ , PCB, some buffer logic i would estimate for 5$ additionally with assembly.
Based on XMOS you can emulate MC68000 bus in real time (so no need to use CPLD/FPGA) and as such any I/O connected to it. So cost are realistic - rest is code - and i agree programmer (especially good one) work is costly (but i assume also that people doing things for Amiga are more like hobbyist - this however can be different from Jens perspective).
I think 50E (and not less even for small board) is simple price of such things - you spent more on cinema ticket for two or lunch in restaurant (i mean real restaurant not McDonald or KFC) so... so why bother?
Every hobby is expensive (except collecting bier canes maybe ).
Jens doing this as main business so i don't blame him - fact that he is one of the last dedicated to small/classic computer make his work even more impressive.

Originally Posted by Daedalus

As for the cost of the Raspberry Pi, there are a couple of reasons for this: First is the volume of Raspberry Pis produced which is already around 50,000 times higher than the number of boards Jens will sell. That's a different galaxy when it comes to economies of scale, buying parts etc. Second, the RPi is a not-for-profit venture so no profit is included in the price. This also means the RPi Foundation can avoid taxes and so on which reduces the cost even further. And third, it's using a SoC design that already includes pretty much all the interfacing electronics that are needed "for free", which is certainly not the case for the Amiga motherboard.

Don't get me wrong, I'd love to see USB etc. on this new board too, but it doesn't make sense to include all the things that anyone wants when it will increase the price for everyone. I'd much rather pay €350 for a board and modules that does what I want it to do than €450 for an all-in-one board that does a lot of stuff that I don't want.

Well and this is my point or Jens focus on people willing to pay for something worth maybe 10$ for example 350$ or on hobbyists for 25$.
Not sure what Jens planning but on one side he may prefer to sell 50 boards for 350$ instead 350 board for 25$ - less work, better money...
What You will do on his side?

To conclude - i believe such board should not happen but if it happen this new board should offer improvements (for example CPLD/FPGA UMA bridge to overcome CHIP<>CPU limitations [and looks like this was addressed by Jens - so UMA and AGA cycles are no longer limiation] - not some strange DSP connectors etc).
I have more than few of old Amiga units - some of them intended to be source for parts (how many working A500 you can have/may need ? 2? 3?).

Originally Posted by dJOS

And that's why I suggested integrating a cheap soc as a co pro to run the extra hw. In theory you could use the soc to offload all the tcp IP, usb, auto config management etc.

Anyway, unfortunately I'm not a developer so I've no idea how hard this would be.

Well i have this idea from a long time and i expressed this multiple times - you can even imagine to feed video data in digital way from Denise to MIPI port in RPi and as such overlay Amiga video over normal RPi video so in theory RPi can be fancy flicker fixer with 3D, GPGPU etc - yes, board with such capabilities will be more costly than RPi itself but with RPi you should be able to fit inside 100$(E).
You can even go further and try to emulate MC68000 on RPi so no real CPU need to be used (some FPGA based glue logic required to interface classic Amiga chips with RPi is necessary for more complex task but it can be part UMA bridge and MIPI port).

[BigFan]

Originally Posted by NovaCoder

There's no point in adding new hardware or improving the AGA chipset because no bugger is going to code for it and Jens will struggle to get the OS updated to support it.

USB, LAN and the IndyMrk2 are different because they don't require any new hardware, software or drivers....they are all proven Classic hardware add-ons.

Sir, yes sir!

That's the way the cookie crumbles ....

[Amiga1992]

Originally Posted by Mrs Beanbag

dear Jens i would also like moon on a stick, i can has moon on a stick pls

well done.

We see in this thread the full scale of why the Amiga community can't have nice things, and especially, how twisted and different each person's idea of what Amiga should be today is.

[pandy71]

Originally Posted by Akira

well done.

We see in this thread the full scale of why the Amiga community can't have nice things, and especially, how twisted and different each person's idea of what Amiga should be today is.

Same for Linux community (or even more) .

[Schoenfeld]

Originally Posted by AEV

Hey Jens, for the external floppy would you consider adding a pin header or leave a place on the board where people could solder there own on (not hard but voiding warranty) to plug in a ribbon cable connection for an external floppy drive.

I'll probably go for the DS0/DS1 option on the internal connector. If people will use a drive at all, it'll be somewhere internal. I highly doubt that the DB23 connector will be used by anyone (and if so, only for part-time use), and it's way too expensive "just for the sake of having it" (expensive by means of financials as well as board real estate).

Originally Posted by IanP

Keyboard options are bound to be limited to re-using an old A500 or A1200 keyboard given the target housings unless somebody hacks together or sources an alternative new keyboard solution.

Having a PS2 translator on a mainboard that uses the original keyboard-cases sounds odd to me. If I'd target some modern form factor that uses external keyboards, then yes, a PS2 keyboard would make sense. At this point, however, you'd be creating something truly Frankenstein.

Originally Posted by IanP

I'd like to see a real time clock included as standard on the motherboard.

If people have an RTC on the accelerator that they choose (for example a Blizzard), they will have paid for two in that case. That's the part I want to avoid. These OKI/Ricoh-compatible chips aren't exactly easy to find, and putting them on a board as standard means you make everyone pay for it - even those who don't need it (because they want a gaming machine) or who already have one. You'd be surprised about the low percentage of RTC modules sold vs. the amount of A1200 accelerators: Only few people consider an RTC worth spending money for. Sure, everyone takes one if it's free, but at the end of the day, every chip must be paid for by someone.

Originally Posted by dJOS

If you can build and sell the base model raspberry pi with usb and HDMI etc for $25 then surely Jens could include them on the Amiga reloaded without exploding the cost? Heck he could possibly include a small arm soc to handle the io, tcp etc etc and not burden the CPU with these tasks.

You still need drivers on the Amiga-side, and that's the most expensive part. I read that a lot: People consider my products expensive, because they count the $$ worth in parts on the board, and compare to the sales price. Hardly anyone sees that I have monthly bills to pay while I develop this stuff, and the same is true for my employees. So RPi as IO for the Amiga? Rather not, as it would create a 50k-EUR piece of software that talks to a 30-EUR piece of hardware. How many illegal copies do you think would exist within a blink of an eye? Any who do you think will end up paying such an expensive development?

Originally Posted by kolla

How does one do early diagnostic (screen colour codes) with b/w composite signal? Or will this board have a different type of diagnostic than real (or should I say "other") Amiga systems?

I will most likely go for a few LEDs as early diagnostic (where "early" means before the monitor can show characters).

Originally Posted by pandy71

Real cost of fast HW (maybe even with MC68k DMA-like coded on X-MOS for example) USB is around 5 - 10E - but i understand there is cost of software development (driver + app + testing) and profit - however surely not 50E...

Here's another example of a Mr. know-it-all who considers the price of the components as the main factor for the price of a product. This may be true if you produce a 40-ft container of some PC product, but not a small-scale production run (where "small scale" is anything below 50.000 pcs in the first run). If you can make a USB host controller with all the required software for 50 EUR - why don't you just do it? I'd be more than happy to quit making RapidRoad and buy your product instead. Oh wait, you'd serve all those customers... OK, well, just come out with your product.

Originally Posted by sardine

he said the aca500 had 2mb ram because of a "technical reason", and after all the complaints low and behold,, wait we now can do aca500 8mb ram

The ACA500 was mainly a 5V-design where only the CPLD was a 5V-tolerant 3.3V-part. The ACA500 plus is over 60% more expensive than it's predecessor (130,- Vs. 80,- EUR), and uses mostly 3.3V parts. This factors in that I'm losing those customers who want a very low-cost product: I won't be able to make an 80-EUR accelerator with the specs of the ACA500plus.

Originally Posted by sardine

also note this new reloaded amiga will fully support ACA12xx cards, it might support some blizzards but not his own apollo cards thats not 100% compatible when 99% of 1200 out there just work out of the box.

Note that the "Apollo" designs are not designed by me, but bought by me, so I could shelve the world's worst Amiga products that are responsible for too many Amigans leaving the hobby due to unstable computers. So no more discussion about Apollo, please: Use them if you want to, but don't bother me.

Originally Posted by Mrs Beanbag

exactly... although i kind of see the logic of not including one if it's not needed, as i see it there are two options:

1. leave a space for an optional 020 to be soldered to the board, on request, or,
2. produce a very basic 020 card

The "on request" part implies that there's a space on the board (=PCB real estate that everyone would be forced to pay for) plus support circuitry that makes the CPU work. How is that any cheaper than making the cut in the most natural location: The CPU port? I'm pretty open about the cost of the materials for the lowest-cost CPU card: The ACA1221 is pretty much sold "at cost" in it's lowest config: just under 80,- EUR incl. 19% VAT. If you remove the CPU, you save maybe a dollar, and if you remove the connector, you might save another few dollars. Give or take some extra stuff around it, you still need to calculate 70,- EUR for a CPU+RAM system that comes without CPU. This would be sitting there idle - totally useless to a good part of the customers who choose another CPU. I just hate wasting resources like that.

The audio discussion was a little lengthy, but definitely worth looking into. Thanks for that.

Jens

(next Q&A next monday!)

[Amiga1992]

Man, Jens even takes time to respond to all our ramblings. We should be very glad that he is still here and still working on this product instead of fleeing because we're all a bunch of cunts.

Thanks Jens, for all your hard work.

PS: please put Svideo on the motherboard as standard

[Jope]

Originally Posted by Schoenfeld

I'll probably go for the DS0/DS1 option on the internal connector.

Cool! I'm waiting with interest to see what features end up being on there in the final product. :-)

[pandy71]

Originally Posted by Schoenfeld

Here's another example of a Mr. know-it-all who considers the price of the components as the main factor for the price of a product. This may be true if you produce a 40-ft container of some PC product, but not a small-scale production run (where "small scale" is anything below 50.000 pcs in the first run). If you can make a USB host controller with all the required software for 50 EUR - why don't you just do it? I'd be more than happy to quit making RapidRoad and buy your product instead. Oh wait, you'd serve all those customers... OK, well, just come out with your product.

Well , Jens, truly flattered but... - i pointed only why for me this can be a problem - don't take this too serious... especially that for many this is OK.
I know prices, You know prices - rest is marketing and business target - some people will appreciate 25E product others prefer to pay 50E for same.
I would be glad to develop something but i know also realities of making business and simply i prefer rather to do open hardware and open software instead struggle with all complaints...
And don't get me wrong - i respect what are You doing from years even if not being Your customer so thumbs up and good luck with business! (honestly no sarcasm).

[jbenam]

Originally Posted by Akira

Man, Jens even takes time to respond to all our ramblings. We should be very glad that he is still here and still working on this product instead of fleeing because we're all a bunch of cunts.

Thanks Jens, for all your hard work.

This x1000. Thanks again for all your hard work Jens

Now I'm waiting for someone to suggest adding PowerPC to the Amiga Reloaded...

[BigFan]

Originally Posted by jbenam

Now I'm waiting for someone to suggest adding PowerPC to the Amiga Reloaded...

Don't miss the show on this thread http://eab.abime.net/showthread.php?t=79029

[Retrofan]

Originally Posted by BigFan

Don't miss the show on this thread http://eab.abime.net/showthread.php?t=79029

Yes, I see. If you had answered Jens when he started and he had opened a thread asking for ideas as he was going to make an accelerator probably it would never have happened.

[Mrs Beanbag]

Originally Posted by Schoenfeld

The "on request" part implies that there's a space on the board (=PCB real estate that everyone would be forced to pay for) plus support circuitry that makes the CPU work. How is that any cheaper than making the cut in the most natural location: The CPU port?

Thanks Jens, i understand your decision on this part and agree with it.

However i thought all the support circuitry for a 14MHz 68020 would already be contained in the chipset, or would be required anyway for accelerators to connect to the bus. But. i am not the expert here.

[TjLaZer]

First off thanks to Jens for producing some amazing hardware! I am a fan.

I'll be honest though I'm not happy about the b&w composite and no external floppy port. Semi unhappy about the removal of two audio RCA jacks but at least they are there. I feel If the original case is used then all the ports should be available.

But again this will be one kick arse AGA Amiga 500 replacement motherboard with CF storage and fast AGA graphics and CPU upgradability. Very excited to get one!

[vulture]

@Jens

dunno if it has been answered already, I've read all 12 pages so far, but not thoroughly:

Do you think blizzard ppc cards + bvisions will be supported? And how about mediator boards?

Thx!

[eXeler0]

Ok I have zero intentions of lecturing Jens of manufacturing costs, however I will use some simple logic to question some of the "facts" in his latest (long) post.
Some ppl point at the cost of pure hardware while Jens sees the final number after software dev has been paid for. That sure makes sense from his perspective...Fine... however...

Capable AMD GPUs were available to PPC Amiga ppl (nemo board, sam etc) long before there were any drivers that actually utilized the the GPU in a meaningful way. Still ppl went out and bought them (It took a good while but now they are selling drivers for it that actually make more use of all that power.).
My only point here is that the *final* price of a certain feature Jens is giving is not *actually* known. If he released hardware features without the software (because they were cheap to implement) who knows what hobbyists would do with it after a while. And if someone wants get a few bucks for it they could sell the drivers for each such feature.
That being said... I am not proposing this as an actual strategy.. just pointing out that the calculations were based on some specific assumptions and sometimes things work out a bit differently.

(Unless Jens has plans to produce new / updated reloaded boards every so often (doubtful) then this is the hardware a lot of Amigans will be "stuck" with for the foreseeable future. )

[ptyerman]

It does seem you would be better off getting a replacement A1200 board or having one recapped, from the feature list been proposed on this new board. It's definitely not something I will be going for. Shame really, but it just reads like a wasted opportunity.

[dJOS]

Originally Posted by ptyerman

It does seem you would be better off getting a replacement A1200 board or having one recapped, from the feature list been proposed on this new board. It's definitely not something I will be going for. Shame really, but it just reads like a wasted opportunity.

My thoughts exactly.

[AEV]

Thanks for answering my ext drive question Jens, totally understand your reasoning.

I think people rather than complaining about what this board will initially lack should look to the future.

I'm sure this is just Jens first step toward creating a more advanced FPGA based Amiga. Once he's tested his motherboard out he can replace the AGA chipset with FPGA based equivalents. In the future if we're lucky we might even be able to upgrade this board, the chipset will be socketed after all.

[Arnie]

Here we go round the mulberry bush.

"Nah you don't want a 020, stick a 030 in there."

"030! Are you mad? 040's what you need"

"040! Do us a favour, what's wrong with you? 060 or nothing matey"

"But, 020 is O.K for me, I can't afford anything else"

And around the bush we go again.