CT60e is coming :)

[alexh]

Thanks Willy. Money to be sent ASAP.

[Estrayk]

BTW: Interested to see how quick the exxos stock is depleted......

I've always wondered why he (Exxos) sells two types of 060. 004/006 and 002
Really there´re any difference?

There are. Later revisions generally produce less heat, and can work stable when overclocked.

ah, I didnt know. thx Adamk


BTW, payment sent ! waiting impatiently.

thanks for all Willy.

[alexh]

BTW: Interested to see how quick the exxos stock is depleted......

I've always wondered why he (Exxos) sells two types of 060. 004/006 and 002
Really there´re any difference?

There are. Later revisions generally produce less heat, and can work stable when overclocked.

Not true. The ones he sells are all Rev6 MC68060RC50 with the same mask set E41J and so use the same technology and produce the same heat when overclocked.

Some have 002 stamped on them and others have 004/006.

He says they are "a later production run" but doesn't provide the date code which is stamped on the chips.

They are the same chip, from the same fab, using the same process, using the same mask geometry. And without the date code they could have been made at the same time.

[AdamK]

I've always wondered why he (Exxos) sells two types of 060. 004/006 and 002
Really there´re any difference?

There are. Later revisions generally produce less heat, and can work stable when overclocked.

Not true. The ones he sells are all Rev6 MC68060RC50 with the same mask set E41J and so use the same technology and produce the same heat when overclocked.

Some have 002 stamped on them and others have 004/006.

He says they are "a later production run" but doesn't provide the date code which is stamped on the chips.

They are the same chip, from the same fab, using the same process, using the same mask geometry. And without the date code they could have been made at the same time.

It is possible that they have the same mask, but still have diffrent performance/thermal characteristic, bacause of imperfect production technology. It is a known rumor that some producers tested chips from the same production runs and marked them differently, according to stability with diffrent speeds. This may be the case here.

[1st1]

Would (turn on "dream mode" now) something like the Vampire be compatible it?

Yes, why not. Falcon's expansion bus where the ct60e docks in is very similar to Amiga 1200's trapdoor expansion connector. Both are based on the 680x0 bus protocol, both are 24 Bit adress bus, but the one is 16 and the other is 32 Bit data bus. But that's only small issue. And A1200 Vamp is announced, they still have trouble getting trapdoor connector. Currently Vampire is based on 68000 socket and bus protocol (but misses a few signals ST needs), so they are already very close to expand the platform. I think this year we already will see ST booting with Vamp.

[troed]

68060 rev6 have 3 sub-revisions : -002, -004 and -006
Some -002 and -004 cannot reach 100 Mhz, don't know why exactly...
The best is the -006, I have two of them and they are working fine at 105 Mhz...

http://www.amiga.org/forums/archive/ind ... 64303.html

I'm not saying it's true, I'm saying that some believe it to be at least.

(Personally I would expect Freescale chips to have been done on a newer node than the Motorola ones - and the lore is that -002/004 are Motorola and -006 is Freescale)

[alexh]

I would expect Freescale chips to have been done on a newer node than the Motorola ones

They have the same mask version number. There is no PCN that I could find.

The only PCN after the announcement of the MC status of the 1E41J mask (by Motorola) was from 2013 which announced the MC68060RC50 would be EOL (no longer sold) from 21/11/2014

https://www.nxp.com/docs/en/product-cha ... N15939.htm

[rudis]

Great news. Money send.

Now i need a Falcon..

[alexh]

It is a known rumor that some producers tested chips from the same production runs and marked them differently, according to stability with different speeds. This may be the case here.

But they are marked with the same speed (50MHz). And the numbers on the side of the chip (the ones where the 002/004 and 006) do not correspond to any of the different parametics under which Motorola/Freescale sold 68060. I would only be speculating as to what they might mean.

It is possible with the date code to determine if all 006 parts are later parts. Perhaps over time a fab becomes more experienced with a geometry node and produces slightly higher grade parts?

[Sturm]

Any news of the Supervidel ?

[troed]

It is possible with the date code to determine if all 006 parts are later parts. Perhaps over time a fab becomes more experienced with a geometry node and produces slightly higher grade parts?

Freescale marked 68060s are later in time than Motorola marked (for natural reasons). Lore is that 002/004 are Motorola branded, 006 is Freescale.

I have two 68060, one Motorola -004 and one Freescale -006.

Regarding improvement over time, yes, that's reasonable to assume.

But I'm pretty sure that technology was improved during 6 years. Usually all 060 rev6 produced by Motorola are 2000. And 6 years in semiconductor business is really a lot to improve as ytou know.

http://www.amibay.com/showthread.php?27 ... guaranteed


edit: I could compare overclockability with my 060s with my CT60e, but I'm waiting for a better heatsink before I want to push it. Am I correct in that the CT60 temperature .acc and .cpx are wrong for CT60e? I get 55 degrees at 75MHz with one of them and 75 degrees with the other. But I'm not sure those values have any real meaning.

[joska]

Same here, temperature readings does not make sense. It starts immediately in the high 60's, and climbs another 10-20 degrees before it stabilizes. However, the CPU is barely warm to touch, even at 95MHz.

[mikro]

I guess you know this: http://powerphenix.com/ct60/english/fpu.htm ?

[alexh]

I guess you know this: http://powerphenix.com/ct60/english/fpu.htm ?

Yes

http://www.amiga.org/forums/showpost.ph ... stcount=15

[troed]

Same here, temperature readings does not make sense. It starts immediately in the high 60's, and climbs another 10-20 degrees before it stabilizes. However, the CPU is barely warm to touch, even at 95MHz.

Hmm. I downloaded the CT60e Utility Demo that willy has provided that reads out the actual data. It agrees with Didier's .acc and the .cpx.

If those values are to be trusted then it seems my Freescale -006 rev 6 68060 runs about 20 degrees _hotter_ than the Motorola -004 rev 6 68060, all other things equal (same heatsink, same re-applied cooling paste, exact same CT60e, clock speed [75MHz], environment).

The -004 starts out at about 45 and ends up at 55-57 or so after a while. The -006 starts out at 66 and ends up at 76 after few minutes.

I'm not really too fond of unplugging/replugging 204 pin PGA chips but I will try to make one other switch back and have an opinion on how warm the heatsink "feels" as well. I'd say when the -006 is at 76 the heatsink feels "warm" (but not _hot_).

[Xevious]

Board is in, gathering the rest of the components.......

Got a bonus, two items for the price of '1':
* ct60e board
* and a "skippers pipe"

So many thanks to willy

[troed]

I continued my investigation into accurate temperature readings in another thread: http://atari-forum.com/viewtopic.php?f= ... 50#p338498

In short, if you measure the resistance between two pins on your 68060 before you plug it into the CT60e, you will be able to get a more accurate feel for the actual CPU temperature while running with the temp .acc/.cpx.

If you don't, well, it can be (at least) 30 degrees C wrong

/Troed

[jd]

Where is everybody ordering the CT60e from?

[alexh]

Everyone who pre-ordered got a private email from Willy with the price and how to order. Send him a PM

[PeterS]

Would (turn on "dream mode" now) something like the Vampire be compatible it?

Yes, why not. Falcon's expansion bus where the ct60e docks in is very similar to Amiga 1200's trapdoor expansion connector. Both are based on the 680x0 bus protocol, both are 24 Bit adress bus, but the one is 16 and the other is 32 Bit data bus. But that's only small issue. And A1200 Vamp is announced, they still have trouble getting trapdoor connector. Currently Vampire is based on 68000 socket and bus protocol (but misses a few signals ST needs), so they are already very close to expand the platform. I think this year we already will see ST booting with Vamp.

Back in dream mode...

Would it make sense to create a Vampire based board that replaces the 040 or 060 allowing it to be used in CT60's, Hades and Milan ? Basically just a minimal cpu replacement.

[alexh]

The vampire doesn't have a 680x0 compatible MMU and the author says it never will have which means it will only (at the moment) be (mildly) compatible with an LC060 which I don't believe is falcon compatible. I believe you need a 680x0 MMU for a Falcon?

[mikro]

LC060 which I don't believe is falcon compatible. I believe you need a 680x0 MMU for a Falcon?

If my memory serves well, it should be. The translation registers work even without the MMU. The biggest difference is that TOS is not copied to RAM but stays in ROM in such case.

However I do remember reading in one of the release notes that support for this non-MMU setup had been dropped. Could have been the new CTPCI releases.

[1st1]

The last Word in MMU Support on vamp is not spoken yet. They just have other priority for the Moment. Some day it will get relevant for them as their target for one day is to support any software running on 68k, and that includes also Linux68k. Linux without MMU is what....?

[PeterS]

I don't know anything about the Atari fpga's but do any of them have mmu support ?

[alexh]

Depends what you mean by MMU. If you mean 680x0 MMU then no.

The Atari ST chipset has a part called the MMU but it isn't what you might consider as a real MMU.