Will there be an ACA1240 or ACA1260?

[donpalmera]

One company quoted $19.95.
Most slightly higher (although not as silly as US pricing).

Long time lerker .. first time poster..

I have a few questions; You are raving about these 68060FE133 chips being the best thing since sliced bread but have you actually tried any? It's very easy to say that someone should make a board based on those chips.. but it begs the question; If it is so easy why aren't *you* doing it instead of bad mouthing Jens for not doing it?
What's left of the "amiga community" seem to come up thing "I want a turbo card with n gigs of DDR, PCI blah blah" on a daily basis but there seems to actually only be three or four people actually still making serious efforts on producing new hardware (Jens, MikeJ etc).

Back to technical details.. Ok, so you can get these 68060FE133 parts from Chinese parts brokers. They don't exist according to Freescale but some of the documentation on some of the late 68k parts is a bit patchy.. i.e. what are all of the different 68SEC000 part numbers about? I suspect it's ROHS compliance or something but none of the datasheets or Product Change Notifications seem to go into any detail about what the different versions of the 68SEC000 are. The details on the 3.3v static 68040V parts is equally thin on the ground. So lets give these 68060FE133 parts the benefit of the doubt. It's possible that Motorola/Freescale produced them for one customer or something and there just happens to be a bunch of them around.

I'm not sure if you have any experience with Chinese parts brokers.. but there is a reason why Digikey etc have $200+ for certain 68k parts and a parts broker will quote you ~$20 for the same part. The reason is the Chinese broker has no idea if the parts are real or work and they don't care either. I did an experiment with 68SEC000 parts; I bought ~$200 worth of sample parts from about 5 or 6 brokers. The prices ranged from about $6 each to as low as $2 for 68SEC000FU10 (Motorola marked, 10Mhz). I have 5 - 10 parts from each broker.. To a certain extent you can use the mask revision,date code and assembly stamp on the underside of the chip to verify the parts. There are 3 masks of the 68SEC000 that I know about and via the PCNs on Freescale site I can work out if the date code is weird for that mask and from some other PCNs I can guess where a chip of that date code should have been packaged (Hong Kong or Malaysia). Out of all of the parts I think only one order *seems* like real parts. A lot of them are obviously suspect by the weird texture of the top of the chip (i.e. it feels like it has been sanded to remove the real markings) or the markings on the chip are just wrong. I have a theory that Chinese brokers will re-label chips to order (i.e. they have a pile of chips that are equivalent or at least the same form factor and they re-label them to suite whatever the customer has ordered). To test this I ordered some Freescale 68SEC000AE16 parts.. These should be Freescale marked (the part number is a freescale one, not a Motorola one) and should be a .5mm QFP not a .8mm QFP like the ones I had order up to that point.. I intentionally didn't put what package that the part should be in the Request For Quote. Out of all of the 20 or so quotes I got I had only one that stated the parts they had were LQFP as they should have. I ordered some from 2 or 3 of the brokers... guess what turned up a few weeks later? .8mm pitch QFPs with AE part numbers (If you haven't followed what I'm saying so far.. they shouldn't exist).

This is for hobby stuff so I don't really mind as long as the chips work. So I made up a small test jig with a BeagleBone that resets the chip and clocks it a few times to check if the stobes etc start doing the right things. I tested one chip from each batch (I have to solder them to a carrier, so I can't test them all) and they all seem to pass this basic "looks like a 68000 starting up" test. So .. what do you reckon is inside of these things? Without decapping them I will never know. I don't have the stuff to do it or I would have done it already.

Back to the 68060FE133 and Jens or whoever making a turbo card with them. The only place you can get them is Chinese brokers which are shady to say the least (See above). If you have followed MikeJ's FPGAArcade posts you will realise how difficult it is to get existing-according-to-Freescale 68060's that haven't been relabelled to a newer mask revision.
So any of these you buy should be treated as suspect from the get go. If you are shipping a product to end users you can't go shipping those parts without doing a lot of testing. They could be working chips with the part number changed to make them look faster.. they could be reject parts that have been saved and labelled, they could be empty plastic lumps with legs sticking out of the side. Who knows? If you could trust Chinese brokers I would have a pile of the 68040V's on my desk (They quoted $20 for those when the Digikey price is ~$200) and I would be designing a machine to use them in.. a 3.3v 68040 + FPGA sounds good to me. Hell, I can get some of those BGA 68060 parts from China for ~$20 .. maybe I should get some of those too? Maybe I would get really lucky and have real parts that worked 100% but it's more likely that they will be utter crap and the broker will disappear when I contact them about it.

The only person that seems to have actually tried those chips seems to be the Netami guy.. but I can't find anything that says if they actually worked or not, what the revision register said etc. Does anyone have a link? At least with the 68060 there is some data you can verify the revision against burnt into the chip..

The only cheap way I would recommend sourcing 68k parts is via surplus sellers in the US or similar. I have sourced some 68SEC000 parts for better than Digikey prices and recently some bog standard 68040FE parts for a tenth of the Digikey price this way. You can't be 100% sure they aren't fakes but it's a lot less likely... But for people shipping a product this is really still too risky IMHO and the quantities you can get hold of are in the tens of units (Maybe enough considering the market?). So you either have spend ages sourcing and testing the hell out of cheap parts or pay the crazy price that decent distributors want. Sounds like an awful lot of hassle for a product that might ship 50 units and for customers that harp on about how they could do better all the time..

[donpalmera]

What he said, basically.
I keep seeing people posting about how "x should do y" all the time. Building an accelerator isn't just a matter of soldering a processor onto a board, shoving on a few RAM chips, and packaging it in a nice glossy box. It takes a LOT of time and effort to design new hardware like accelerators, jamming a processor into a machine that was never designed for it. The more different the CPU, the more work (and therefore cost) it is - and the 68060 is very different to the 68020/30.

This is what prompted me to post to be honest. I'm sure there are a few really skilled software devs and hardware guys still around.. but they are definitely in the minority... you could probably count them on your hands alone. Aside from them there seems to be two types of people. Type 1: "I want 68060/Soft core in FPGA/PPC etc!! I must have it. NOW!" Type 2: "68060/Soft core in FPGA/PPC is so easy, I'm not sure why no one has done it yet. I wrote an enhanced 68060 core while I was on the toilet one morning.". The Type 2 people get the Type 1 people all excited by spouting drivel. I don't have the time or desire to mull over "the good old days" or "what if..." hence I have never posted before. I read amiga.org to keep up to date with interesting stuff like the FPGAArcade. Some of the people here seem to be on crack or have no grasp on reality. It's a shame people like Piru seem to have left. They might have been a bit sour but they added some much needed reality.

I doubt there are many people left "in the scene" that could put together a 68000 or 68020 system from scratch and those are the easy ones. The 68040 and 68060 are monsters in comparison. Yes.. it's nice to have a bit of a day dream when you see some cheap 68060s on Alibaba or whatever but unless you have the skills of Jens or MikeJ you'd be better off keeping it to yourself.

[donpalmera]

@ the 68060FE133 haters

Various ppl have bought them.  They used them.  Nobody has ever reported them as being fake or defective.  So why the hate?

I have yet to find one report of them working.. links please. The only thing I could find that went into any detail was the thread on MorphOS zone or whatever it was called.. and no one there could provide/find an example of one of these things working either. You're also assuming that all of these chips that are floating around the Chinese brokers are the same thing.

If everything is so crystal clear you'll have some evidence to back up what you're saying.. If you are really sure about this I will order a tray of them. If they turn out to be relabelled parts or just outright fakes you'll give me my costs back for the parts and making up a test jig right? :roflmao:

[donpalmera]

The Natami guys tested them, they're exactly what they seem to be, someone just needs to make a board to test them with a regular Amiga. They seem to be a Chinese knockoff, i.e. a replica.

So basically, even the people that have supposedly "tested" them have no idea what they are. If they do work I doubt they are "replicas". There is no profit in reverse engineering the 68060 for the amount of units you could sell. I'm not sure how insane you have to be to imagine anyone in China spending years reverse engineering the 68060 to produce "replicas" of something they might sell a few tens of units of.. but it's pretty high up on the scale.

It is much more likely that one guy in Shenzhen got some other QFP 68060s (rejected parts, recycled parts from telecoms hardware..) and relabelled them and told all the brokers about them... and this is why if you google that part number you get 3 posts from Amiga forums about how great these chips are from people with no evidence than "my friend's mate's sister's dad's mate's next door neighbor's dog's friend apparently used them and said they are ok". Who are the Natami "guys" anyhow? There seemed to be loads of team members when only one person was doing any work and that guy decided to go it alone.. shame not one of all of those team members actually posted anything showing those chips working..

[donpalmera]

If they didn't work then Thomas would have told me.  It would be a really big deal.  That is my view.

Great.. a view.. I'm looking for evidence here. If they do work where are all of the posts showing them off? If you google this part number there are only about 3 results that aren't Chinese brokers and those results are all threads like this.. not a single result about one of these things working. Strange for something so amazing eh?

The way I remember it is that Thomas plugged one into his 060 card and it worked and it did not have MMU or FPU.  

You could try datamining the Natami forum.
.

There is no evidence there either... more he said she said. No one has posted "They work, they clock up to xyz, the mask revision according the the register in the chip is...". For something people are so sure about there is so little evidence it's laughable.

Supposedly some Atari guys are using them.  You could search Atari forums if u want.

Of if they are really $20.00 why not just buy one and see what happens?

And we're back to "my mate Dave down the pub said they work, so they must work". There's nothing about these chips in the Atari forums either.
Having actually ordered parts like these from brokers I can tell you that the unit cost might be $20 but it will cost a lot more than that to actually have something in your hands.
You need to order 5 parts at least and tell the broker you want n hundred in the future to get those prices. You will then pay $30 or so to have them Fedex them... and then you'll pay a fee for daring to pay with something traceable like Paypal. So we're talking about at least $100 at the end of the day... then I would need to put together a little board to test them. I can get a 4 layer 10cmx10cm board made up for about $100 shipped.
So we're now talking $200 to test some chips that are "100% working according to Dave!!1!1!!!". Yeah, I think I'll pass on that. My barter still stands but the way you just side stepped it with more "someone down the road out of their head on smack said they worked" makes me think you don't have the confidence in these parts to put your money where your mouth is..

[donpalmera]

Thanks Don,
Its nice having some backup on this.
After all, as I've said repeatedly, Freescale says there is no such thing as an FE133.

I would love to be proved wrong. It would save me so much time to have a 3.3v 68040 or 68060 vs having to wire up a mass of level convertors to a 5v 68040..
The only way anyone will know what they are is to buy some, test them (and publish better results than "some guy somewhere on the internet said they worked, I just can't actually give you a link to it..") and take some fuming nitric acid to the chip package to see what is inside.
Until someone does that all this harping on about Jens being a bad guy for not making a turbo card out of these parts is basically just rude. I'm not sure why Jens bothers doing anything for this "community" to be totally honest.

[donpalmera]

Its a capacitor and it looks like it might have leaked.

It's a clock crystal with it's can soldered down to the board (very normal) with some flux residue left over that has probably collected some dirt over the years...

[donpalmera]

and think that any manufacturer could make a few $ in the process.  I can't speak for the others here though.

- If it was easy Jens or someone else would have done it.. it's not easy and wouldn't turn a profit. So no one has done it. Sourcing the parts, putting together a schematic, laying out a board with a ton of layers, creating all of the glue logic and software etc is not a weekend job in the shed at the bottom of the garden.

- The Apolo design is a dead end. This is according to Jens, someone who has built similar products .. knows what he's talking about. A "reliable source".

- 68060s that aren't very very suspect are very very expensive if you can get them at all.

The best you could hope for is that a tray of 68060s finds it's way into Jens hands from a reliable surplus dealer (i.e. not Alibaba) and he decides to have a go for ****s and giggles. I suspect there are enough 060 cards already floating around in the market for the people that really want them though.

[donpalmera]

True, I regret bringing that up.
And the fact that LC or EC parts are still available isn't a reason to call for their use either.
After all, a $220 part without an FPU or MMU isn't that useful.

The problem with using LC or EC parts would be that basically all Amiga software is frozen in time and when that freeze happened it was assumed that all machines with an 060 had a working FPU. So LC and EC parts are basically useless for the Amiga..
If you want a super fast Amiga on the cheap then your best bet is to wait for the FPGAReplay and hope that their 68K core kicks ass... or lower your expectations with regards to this dead technology and just have some fun with it instead.

[donpalmera]

Sorry man, don't agree with you there.  I sold a ton of 68LC040 A4000's in my day and they were plenty adequate for most people then.  Still awesome today for what they are running.  Frozen in time, just like any other retro platform.  Dead?  No.

Well, I did say "68060".
The 68k (000,010,012,020..) series is officially dead. Freescale will be taking the last orders for parts next year. I have a feeling they stopped fabbing any of the real 68k series years ago either way. Dead technology. The Amiga technology is very much dead if you ignore the FPGAReplay.

Of course the "amiga community" or whatever isn't dead but there aren't many people left with any of the skills required to make any progress on the hardware side of things. There are two type of people here as I've said before. The people that want stuff and the people that can't work out way no one has done it before because "it's sooo easy".. there is no third group of people that actually produce anything.

[donpalmera]

How many layers does an A1200 060 accelerator have?

I've never made one so I can't say. Restricting yourself to things you have first hand experience of is something that's sorely missing around here..
A board with a 68SEC000, some level convertors and a CPLD for glue logic required 4 layers to route and it was barely possible then.. 4 layers is basically the most complex board you can do as a hobbyist. Unless you happen to do this stuff for a living..

And after all of this.. why would you use the QFP parts and not use the BGA parts that are actually documented as existing and can be bought for similar prices from Chinese brokers*.

I remember correctly Jens is the only person in the world right now that has the edge connectors for the A1200 trap door. It might be a good idea to listen to what he says. :elvis:


* Likely to be fakes.

[donpalmera]

I remember correctly Jens is the only person in the world right now that has the schematics and the rights to produce the Apollo 1260 accelerator.

I wonder why he bought the rights for the 060 accelerator?

Because he had the money to do so and the rights were for sale. He might have considered using them. Who knows. I don't tell you what to do with your stuff do I?
Again, if you are so sure about everything get some cash and offer to buy them from him. I'm sure if you have a big enough pile of dosh he'll be happy to get rid of them.

I wonder why he keeps cranking out 020 accelerators but not 060 ones?

Because in reality opposed to magic-fairy-pony dream land producing working products is more complicated than wishing them into existence.
I see trays of real 68020s on Yahoo Auctions every few weeks. There seem to be some stock piles of the lower end chips around in the US,Europe and Japan.. so maybe it's an availability of parts thing? Maybe there is more market for basic cards that will allow 90% of the Amiga software out there to run just fine? Maybe most of the people in the market for a turbo card right now only want to run games via WHDLoad so there is no point wasting time on anything else?

The only things we can say for sure are: Making an 68060 card isn't a walk in the park. The 68040 and 68060 are considerably more complex than the 020/030. Jens has the designs of a "working" 68060 which he said more or less that he thinks is a piece of crap. You can agrue all you like that Jens is wrong.. but I know who's opinion I trust over people that have zero experience with this stuff. Jens has produced a bunch of turbo cards and other hardware .. sum total of hardware produced by people with big ideas about using shady chips here.. safe to say zero right?

I wonder why he keeps cranking out 030 accelerators but not 060 ones?

Because it's working for him? Maybe he's actually managing to make a bit of money out of them? I'm not sure how you guys have the cheek to say basically "someone else should risk their time and money doing something for me even though I have no intention of actually listening to anything they say".

[donpalmera]

Can u name a single piece of software that makes such an assumption?

I have never met a program that did such a thing.

Any application that tries to make use of the FPU or MMU..

Tons of ppl used LC and EC parts and I never heard of them not working.

And we're back to "my mate Dave said his dad had a special 68060 specially imported by James Bond" again. Did you not get the memo about a lot (most?) LC and EC 060 parts having working MMU and FPU?

[donpalmera]

Unless he finds a good cheap source of guaranteed genuine, tested 68060 chips; and I don't think that's going to happen any time soon.

Wait.. random Chinese sellers on Alibaba aren't reliable? Ones that can offer products at less than 10% of the price of trusted distributors like Digikey etc? Offering products that have never officially existed? I am shocked.
On a serious note.. stop making so much sense.

[donpalmera]

There are posts and pics and videos proving that they do work. NatAmi's softcore CPU was never completed so all tests were done with 060 (regular and FE) CPU daughterboards.

Please link them then..

Many of us followed the NatAmi forum daily, so I can pretty much say that we know more on this matter then you do. There is no need for such an offensive attitude here.

Sigh, you followed some forums.. great. These chips very well may have "worked". But as I said.. I have 0.8mm pitch AE package 68SEC000s that don't exist that apparently work too.. would you ship a product based on these parts?

The fact that none of you even brought up the fact that Motorola mentions CQFP parts in their documentation when they apparently don't exist surprises me to be honest. I would have thought you would have mentioned it right away..