Re: Manufacturing New 650x - Need help

DerTrueForce · Post by **DerTrueForce** » Fri Dec 16, 2016 7:38 am

EDIT: BigEd has alerted me that the head post has been deleted. He's also recovered it, and asked me to edit it in, so that things continue to make sense. Therefore, the original head post was the following:

Quote:

Hello, my name is Kevin and I'm currently researching stuff for founding my own microchip manufacturing plant.
I also want to prodce some of the 650x chips, namely the 6502 itself, aswell as the 6509 and the 6510... as well as RAM and I/O chips...

But sadly I can't find any of the schematics I'd need for that... (Either I'm just to stupid or to drunk - propably the later ^^)

Anyhow... If anyone could supply me with some schematics, or with contacts of people who are know to have such, please write me an email at: kevin.giebel[at]hotmail[DOT]de

If I would get the schematics in 2017, I could start production in 2025... I know, this is late...

BUT chip-manufacturing equipment is expensive...

Thanks in advance...

Greetings,

Kevin

BTW The company is going to be called "MOS Technology Belgium K.G."

OK, this is interesting. I just said that I'd like to do something like what you describe(off in another thread), and then you come out with this!

The name may not be legal to use, but don't quote me on that...
Also, I have a sneaking suspicion that this may be a troll post or scam, but I'm going to keep an open mind and see how it goes.

As to your question, I don't actually know how you'd get hold of such schematics. I would probably try to track down the original designers. They might be able to help, and even if they can't, if they endorse the project, that will almost certainly be a great marketing tool.

Some questions:
Are you planning to:
produce reproduction NMOS parts only, or do you plan to make CMOS workalike versions as well?
make higher-speed versions of these parts? As in, faster than the 14MHz maximum of the WDC parts?
make a user-programmable microcontroller?
make new/updated 65-series chips?
(Probably the most important one): do you plan to have a (probably online) shop, and sell to the public, or just to distributors? Because I think it'd really suck if you didn't sell to the public.

DerTrueForce · Post by **DerTrueForce** » Fri Dec 16, 2016 8:20 am

OK, That's fair enough. I was asking mainly out of curiosity; the CMOS versions was the main thing I was interested in. The other one was selling direct to the public.Which you still haven't answered, BTW.

ttlworks · Post by **ttlworks** » Fri Dec 16, 2016 8:53 am

Hi, Giebels2609, and welcome to the forum.

If you are posting in a foreign language for the first time,
it might be that the "filters" inside your head are not adjusted for this,
what makes you sound like a "troll" to the native speakers,
and you probably are not aware of this effect.
Don't worry, that's only temporary.

Please post a short resume and a list of your skills into the "introduce yourself"
section of the forum so we would be better able to help you.

Also, it would be nice if you post a few pictures of your previous projects
here in the "hardware section", rusty Lada 1200 cars and Elbonian submarines
would count if they would be 6502 powered.

Costs for buying chip manufacturing equipment are high,
costs for running chip manufacturing equipment are high, too.
There probably won't be a high demand for your chips on the market,
making the price per chip a little bit astronomical.

Would suggest that you try to aim for the SID.
Most of the inner workings of the SID have been reverse_engineered,
and true audiophiles probably won't mind paying >2000€ for a clone
of an obscure sound chip from the 80s.

BTW: from the Google Earth pictures the village you live in looks nice.

Only 415 inhabitants or such...

GARTHWILSON · Post by **GARTHWILSON** » Fri Dec 16, 2016 9:12 am

There's an Austrian company that plans to introduce a new model of professional tape recorder next year, 2017, because of the persistent demand for analog audio equipment: https://horchhouse.com/project-r2r You can still buy professional recording tape (the last thing in the news page at the link was posted 7 hours ago), floppy discs (even 8"!), discrete logic, and there are couple of companies still making quality slide rules.

Setting up wafer fab is extremely expensive by comparison, but if you have the means and it's a hobby like someone paying millions of dollars for a yacht plus $15,000 a month just for the slip, why not? You won't hear me telling you to forget it!

I worked for a transistor manufacturer in the mid-1980's. I did not work in wafer fab, but had to go to that building frequently, and occasionally into the wafer fab cleanroom itself. Besides the fab, there was the area where they did autoprobe which is where individual dice on each wafer were probed to identify bad wafers or bad sections of a wafer before scribing them and breaking them into individual dice and mounting them; then there were stations where the dice were mounted into packages, the stations where the bond wires were added, a machine to spot-test the strength of the wire bonds, X-ray machines to look for voids under the dice (since these were power transistors and had to dissipate a lot of heat), capping, testing for leaks (gross leak and fine leak), production test of final product, and in some cases, burn-in testing. Quality control was involved at every step. I have no idea how the masks were designed and made for the wafer fab. (I was in applications, so my concern was for what the finished product could do.) There was also the ion-implantation machine which I suppose cost millions of dollars by itself. In the applications engineering lab, there was about $100,000 of test equipment per engineer.

I think what will happen though is that you learn a lot in the process, and end up improving the old parts rather than making them just the way they were, and you'll find a cheaper way to do it than buying your own wafer fab. Even WDC doesn't do their own wafer fab. But then, I don't make my own PC boards either, even though I've designed dozens over the years and brought them to market for my employer. Now with custom boards becoming quite affordable for the hobbyist, I will definitely be getting more and more boards made for my own uses. Heck—by 2025, maybe you'll be able to send your files to a fab house and get finished ICs in the mail in a few weeks!

ttlworks · Post by **ttlworks** » Fri Dec 16, 2016 9:34 am

Giebels2609 wrote:

BTW How the hell do you know where I live

Google knows.

DerTrueForce · Post by **DerTrueForce** » Fri Dec 16, 2016 9:52 am

ttlworks wrote:

Hi, Giebels2609, and welcome to the forum.

If you are posting in a foreign language for the first time,
it might be that the "filters" inside your head are not adjusted for this,
what makes you sound like a "troll" to the native speakers,
and you probably are not aware of this effect.
Don't worry, that's only temporary.

Umm... The thing that made it seem like a troll post/scam was the subject matter itself. It seems too good to be true to me. I am sorry that I was not specific. The answers you gave to the questions made me less doubtful, as they seemed more realistic than what I imagine a scammer or troll would say.

Giebels2609 wrote:

Oh, sorry, I forgot that. Of course I'll make the CMOS available to the public... Anything else would be dogsh*t (in my oppinion)

I agree that the NMOS parts would probably be worse. AFAIK, they have a less-than-stellar reputation around here.
The thing they would be good for, though, is replacement parts for things like C-64s, VIC-20s, Apple Is and ][s; that sort of thing. CMOS equivalents might throw those older designs out of whack, or people might want exact replacements.
I would suggest this: Produce and sell the NMOS, at least initially, for those purposes.
You'll probably have to get the CMOS versions designed. AFAIK, making a CMOS equivalent requires a CMOS design based on the original. If and when you do, get them tested in the old machines, and see if they work properly. If they work, you could then discontinue the NMOS parts, knowing that the CMOS ones will do just as well, if not better. If they don't, then you may have to keep selling the NMOS, despite it being not the best.
This is only my suggestion, mind. You don't have to use it.

And your English is not as bad as some I've seen on this forum, so I wouldn't worry too much.
The RAM and ROM would probably be the bulk of your business, but it's not such a bad thing. The profits on the memory might be enough to balance the price of the 65-series hardware.

Another thing that comes to mind is: you could talk to WDC. They are fabless, but I'm under the impression that they might have chip designs you could license.

DerTrueForce · Post by **DerTrueForce** » Fri Dec 16, 2016 10:14 am

They sell a number of things, including:
the 65C02, which is a 6502, in CMOS, with several improvements,
the 65816, a 16-bit processor. Sort of a successor to the 'C02,
a couple of microcontrollers, based on the aforementioned processor cores,
the 65C51 ACIA,
the 65C22 VIA,
the 65C21 PIA.(these three are CMOS versions of the NMOS chips of the same names.)
This is possibly incomplete and inaccurate list, but it's what I can think of off the top of my head.
AFAIK, all their hardware is rated for up to 14 MHz. The old NMOS ones were rated for 1, 2, and 4 MHz, I think.

KC9UDX · Post by **KC9UDX** » Fri Dec 16, 2016 10:47 am

Giebels2609 wrote:

DerTrueForce wrote:

They sell a number of things, including:
the 65C02, which is a 6502, in CMOS, with several improvements,
the 65816, a 16-bit processor. Sort of a successor to the 'C02,
a couple of microcontrollers, based on the aforementioned processor cores,
the 65C51 ACIA,
the 65C22 VIA,
the 65C21 PIA.(these three are CMOS versions of the NMOS chips of the same names.)
This is possibly incomplete and inaccurate list, but it's what I can think of off the top of my head.
AFAIK, all their hardware is rated for up to 14 MHz. The old NMOS ones were rated for 1, 2, and 4 MHz, I think.

Yes, I know that with 65c02... but what I mean is it is faster than the original

It's also slower than the original. I don't think speed is an issue. No one as far as I know minds putting a CMOS chip in place of an NMOS one. The trouble is in the way it works. If you could make a CMOS 6509 and 6510, you'd sell just as many (but likely more) than if you made them NMOS. The same may be true about original 6502. If you made a CMOS 6502 with the identical (buggy) operation of the original NMOS 6502, you'd sell as many as if you made them actually NMOS, and again, you'd probably sell more.

You seem to be extremely ambitious. If you are going to consider some of these things, you may as well also consider some like the 65EC02, in my opinion.

Tor · Post by **Tor** » Fri Dec 16, 2016 12:12 pm

As for you location, I didn't try, but if you're using your handle elsewhere then just putting it into google and follow all the links, and combining the info for a couple more searches, will probably find your hometown in a minute or so. Even if somebody else use the same handle there are enough clues to filter them out - you already mentioned Belgium, for example.
I know it won't be difficult to find my exact location, down to the house address. Even if I never wrote it anywhere. But I had to do that kind of search over on another forum where I am a moderator, and a problematic user (harassing others) tried to circumvent getting banned by creating multiple accounts. A little search found all the details, problem solved.

Tor · Post by **Tor** » Fri Dec 16, 2016 1:41 pm

Giebels2609 wrote:

Huh

Er, what? Not sure what that's supposed to mean.

Tor · Post by **Tor** » Fri Dec 16, 2016 2:44 pm

Oh, indeed. It's very very difficult to hide on the net these days.

Arlet · Post by **Arlet** » Fri Dec 16, 2016 3:11 pm

I'm pretty sure there's not enough demand for 6502 devices to warrant a new manufacturing plant.

cbmeeks · Post by **cbmeeks** » Fri Dec 16, 2016 4:57 pm

Arlet wrote:

I'm pretty sure there's not enough demand for 6502 devices to warrant a new manufacturing plant.

Agreed.

No one obsesses over vintage computers more than me and I wished I had replacement parts for every vintage computer there is.

I think a better solution are small FPGA boards like the F18a did for the TMS9918. A small board with level shifters, etc. could be made as drop-in replacements for much of the chips out there.

Sure, it's not 100%. But it's certainly better than just running an EMU on your computer and a heck of a lot cheaper than starting up your own fab plant.

That guy who made the F18a sells quite a few of them I imagine. And at nearly $100 each, probably nets him a little pocket money.

GARTHWILSON · Post by **GARTHWILSON** » Fri Dec 16, 2016 8:03 pm

KC9UDX wrote:

No one as far as I know minds putting a CMOS chip in place of an NMOS one.

Some will want the so-called illegal op codes of the NMOS. The CMOS instruction set is better, but when people only had NMOS, they helped themselves to op codes that got new functions when the CMOS came along.

Quote:

If you are going to consider some of these things, you may as well also consider some like the 65EC02, in my opinion.

We have the data sheet for the 65CE02 on this site at http://6502.org/documents/datasheets/mo ... 02_mpu.pdf . That was Commodore's creation but they didn't make very many. It's a lot more powerful than the 65C02.

Tor · Post by **Tor** » Fri Dec 16, 2016 8:07 pm

But this sounds terribly expensive. I doubt even Bill Gates would fancy paying out of his own pockets for a semiconductor fab. It's something that huge corporations do, that was true even back in the day of micro, not nano, scale processes.