Hi,

Are there any test circuits to prove a cpu is working floating about
anywhere?

I'm looking at getting hold of an 68030/68882/68040.

Thanks,
Luke.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On Tuesday, July 4, 2023 at 2:20:54 PM UTC-7, Luke A. Guest wrote:

Hi,

Are there any test circuits to prove a cpu is working floating about anywhere?

Just the "standard" free-run tests. Varies with which 680x0 you are working with. Force the appropriate control signals, and ground the data bus. Then check the address bus to see that it is incrementing properly. There are more details in texts such
as the ones by Wilcox or Clements.

Roger

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 05/07/2023 15:51, Roger Hanscom wrote:

On Tuesday, July 4, 2023 at 2:20:54 PM UTC-7, Luke A. Guest wrote:

Hi,

Are there any test circuits to prove a cpu is working floating about anywhere?

Just the "standard" free-run tests. Varies with which 680x0 you are working with. Force the appropriate control

This is basically, what I want to know. I'm after something like the Z80
LED on the address lines circuit.

I'm looking at trying to get hold of 68030 (with MMU) and 882. and later
040 (with FPU/MMU).

Can these CPU's handle being slowly clocked so that the LED's can be
seen clearly?

signals, and ground the data bus. Then check the address bus to see
that it is incrementing properly. There are more details in texts such
as the ones by Wilcox or Clements.

I don't know these books, do you have titles or can you point me to the
pdf's if available?

What would be the best way to determine whether I have the correct speed
and it's not a fake one?

I used to have an A1200 with a 68030/882@50MHz and that thing rang HOT
when Linux ran on it.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Le 05/07/2023 à 21:03, Luke A. Guest a écrit :

I used to have an A1200 with a 68030/882@50MHz and that thing rang HOT
when Linux ran on it.

Chinese blacktopped CPU and FPU ?

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 06/07/2023 11:09, Marc J. wrote:

Le 05/07/2023 à 21:03, Luke A. Guest a écrit :

I used to have an A1200 with a 68030/882@50MHz and that thing rang HOT
when Linux ran on it.

Chinese blacktopped CPU and FPU ?

No, I bought that in 1995 when I went to uni, it was direct from phase
5. It was a gold cap CPU. It only rang stupidly hot when running Linux,
because it used the MMU. can't remember if the 882 contributed to the
heat when running Linux.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

Luke A. Guest wrote:

What would be the best way to determine whether I have the correct speed
and it's not a fake one?

Check the mask of the CPU: http://www.amigawiki.de/doku.php?id=de:parts:68030_mask

Is it a XC68030 or a MC68030?

Observation in the late 1980s: the first series of 68030 labeled as
12,5MHz ran at 25MHz for hours without any problem. But they were real
hot running too fast. Means: some 25MHz CPUs and most 40MHz CPUs should
run at 50MHz but not stable running 24/7.

I used to have an A1200 with a 68030/882@50MHz and that thing rang HOT
when Linux ran on it.

In the early 1990s we used VMEbus boards from Eltec called E6. We had a
20MHz board, several 25MHz boards and some 50MHz boards. The operating
system OS-9 and our applications used the MMU and the FPU intensively.
The chips ran hot but not too hot because of the typical good cooling in
the VMEbus racks. Means: the 50MHz parts need sufficient cooling.

Regards
Ralf

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On Wednesday, 5 July 2023 at 15:51:48 UTC+1, Roger Hanscom wrote:

On Tuesday, July 4, 2023 at 2:20:54 PM UTC-7, Luke A. Guest wrote:

Hi,

Are there any test circuits to prove a cpu is working floating about anywhere?

Just the "standard" free-run tests. Varies with which 680x0 you are working with. Force the appropriate control signals, and ground the data bus. Then check the address bus to see that it is incrementing properly. There are more details in texts such

as the ones by Wilcox or Clements.

I don't know about later processors, but IIRC, I did a 'NOP' test on a 68010 by just pulling the data bus and ~DTACK low. (I kept meaning to build a circuit for my 68012s but now I've got a computer that uses one, so I don't need it.) (Actually, I think
instruction 0 is "OR.B D0, #0". (Sorry if I got the syntax wrong; it's a long time since I did any 68k'ing.))
(The top bits of the address bus will toggle at "human speeds". For a 68012 you'll have to wait a long time for A31 to toggle :-D )

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On Wednesday, July 5, 2023 at 12:03:02 PM UTC-7, Luke A. Guest wrote: ...<snip>....

Can these CPU's handle being slowly clocked so that the LED's can be seen clearly?

Do you have a logic probe? That will show you enough to confirm that the CPU is operating as expected. No LEDs needed.

I don't know these books, do you have titles or can you point me to the pdf's if available?

Willcox is ISBN 0-13-811399-8. Clements is ISBN 0-13-668120-4.

What would be the best way to determine whether I have the correct speed and it's not a fake one?

Start off slow (maybe 5 or 6 MHz) and work our way up in steps to the speed indicated on the chip. Maybe even go beyond that? I find that some Motorola/Coldfire CPUs will run 20% faster than the indicated speed.

I used to have an A1200 with a 68030/882@50MHz and that thing ran HOT when Linux ran on it.

At 50 MHz, I'm not surprised!!

Where are you located? I have an extra Clements book that I could send to you (as long as it is not too expensive to mail -- I'm retired, and Biden's inflation is killing me!).

Roger

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 06/07/2023 15:37, Roger Hanscom wrote:

On Wednesday, July 5, 2023 at 12:03:02 PM UTC-7, Luke A. Guest wrote: ...<snip>....

Can these CPU's handle being slowly clocked so that the LED's can be seen clearly?

Do you have a logic probe? That will show you enough to confirm that the CPU is operating as expected. No LEDs needed.

No. And I'm not working atm, so pretty skint.

I don't know these books, do you have titles or can you point me to the pdf's if available?

Willcox is ISBN 0-13-811399-8. Clements is ISBN 0-13-668120-4.

Found the clements book as a pdf but the other one is about £500, so
that's not happening.

What would be the best way to determine whether I have the correct speed and it's not a fake one?

Start off slow (maybe 5 or 6 MHz) and work our way up in steps to the speed indicated on the chip. Maybe even go beyond that? I find that some Motorola/Coldfire CPUs will run 20% faster than the indicated speed.

Okay. would need a bunch of crystals then.

I used to have an A1200 with a 68030/882@50MHz and that thing ran HOT when Linux ran on it.

At 50 MHz, I'm not surprised!!

They were marked as 50MHz and bought at a time you could buy an m68k CPU without worrying whether it's fake or not.

Where are you located? I have an extra Clements book that I could send to you (as long as it is not too expensive to mail -- I'm retired, and Biden's inflation is killing me!).

Am in UK.

You think inflation there is bad, try here where it's uncontrolled and
the tories have let fuel company's just take the piss completely, we
have the highest bills in the world, btw.

Thanks,
Luke.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 06/07/2023 15:41, Ralf Kiefer wrote:

Luke A. Guest wrote:

What would be the best way to determine whether I have the correct speed
and it's not a fake one?

Check the mask of the CPU: http://www.amigawiki.de/doku.php?id=de:parts:68030_mask

Yeah, have seen that. But a lot of them, even the gold top's have had
them removed and remarked.

Is it a XC68030 or a MC68030?

The ones I've been looking at the gold top PGA ones and some of the
masks are listed, e.g. "0G40W" not "00G40W" and I cannot remember what
my original 030 looked like now. And some are not lilsted, e.g. 02e23g
for an LC040 part.

Observation in the late 1980s: the first series of 68030 labeled as
12,5MHz ran at 25MHz for hours without any problem. But they were real
hot running too fast. Means: some 25MHz CPUs and most 40MHz CPUs should
run at 50MHz but not stable running 24/7.

Trust me my 030 was not a fake, it was incapable of running Linux using
the MMU/FPU within the trapdoor of an A1200, once inside a tower with
more air, it ran no problems.

I used to have an A1200 with a 68030/882@50MHz and that thing rang HOT
when Linux ran on it.

In the early 1990s we used VMEbus boards from Eltec called E6. We had a
20MHz board, several 25MHz boards and some 50MHz boards. The operating
system OS-9 and our applications used the MMU and the FPU intensively.
The chips ran hot but not too hot because of the typical good cooling in
the VMEbus racks. Means: the 50MHz parts need sufficient cooling.

See above.

If we had back in 1996 what we have now, I probably could've kept it in
the A1200 case with a fan like a laptop fan which would push the air out
of the front of the case.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 06/07/2023 15:29, marb...@yahoo.co.uk wrote:

On Wednesday, 5 July 2023 at 15:51:48 UTC+1, Roger Hanscom wrote:

On Tuesday, July 4, 2023 at 2:20:54 PM UTC-7, Luke A. Guest wrote:

Hi,

Are there any test circuits to prove a cpu is working floating about anywhere?

Just the "standard" free-run tests. Varies with which 680x0 you are working with. Force the appropriate control signals, and ground the data bus. Then check the address bus to see that it is incrementing properly. There are more details in texts such

as the ones by Wilcox or Clements.

I don't know about later processors, but IIRC, I did a 'NOP' test on a 68010 by just pulling the data bus and ~DTACK low. (I kept meaning to build a circuit for my 68012s but now I've got a computer that uses one, so I don't need it.) (Actually, I

think instruction 0 is "OR.B D0, #0". (Sorry if I got the syntax wrong; it's a long time since I did any 68k'ing.))

(The top bits of the address bus will toggle at "human speeds". For a 68012 you'll have to wait a long time for A31 to toggle :-D )

Yeah, according to the manuals, the NOP instruction isn't 0, which
would've made sense.

According to the MC68020 UM, the encoding for it is 0100111001110001.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On Thursday, 6 July 2023 at 16:18:18 UTC+1, Luke A. Guest wrote:

I don't know about later processors, but IIRC, I did a 'NOP' test on a 68010 by just pulling the data bus and ~DTACK low. (I kept meaning to build a circuit for my 68012s but now I've got a computer that uses one, so I don't need it.) (Actually, I

think instruction 0 is "OR.B D0, #0". (Sorry if I got the syntax wrong; it's a long time since I did any 68k'ing.))

(The top bits of the address bus will toggle at "human speeds". For a 68012 you'll have to wait a long time for A31 to toggle :-D )

Yeah, according to the manuals, the NOP instruction isn't 0, which
would've made sense.

According to the MC68020 UM, the encoding for it is 0100111001110001.

I don't think pulling the data bus to that would work anyway. The first thing the 68010 does is load initial values for PC and supervisor SP. It might sulk if either is odd. The OR.B will execute in the same time as a real NOP.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On Thursday, July 6, 2023 at 8:18:18 AM UTC-7, Luke A. Guest wrote: ...<snip>....

Yeah, according to the manuals, the NOP instruction isn't 0, which would've made sense.

According to the MC68020 UM, the encoding for it is 0100111001110001.

Yes, an opcode of all zeros is a variant of an OR. It might as well be a NOP. It executes without changing the processor state, so the CPU just chugs along, incrementing the address bus. No problem with the reset vector either. You'd think that the
CPU would pick up the starting PC as Zero, and just hang in an endless loop, but it doesn't. It steps through memory just fine.

You think inflation there is bad, try here where it's uncontrolled and
the tories have let fuel company's just take the piss completely, we
have the highest bills in the world, btw.

Sorry to hear that, Luke. Inflation is just another form of taxation on a populace that's stretched to the breaking point. These governments don't seem to realize that you can't get blood from a stone!

I'd really like to send you this extra copy of the Clements book that I have. It's not doing me any good to have two of them! Our postal service gives reduced rates for sending books in the U.S. Do you happen to know if that is true for the U.K. as
well? Might I get a decent postal rate in order to send it to the U.K? I think I'll check with the post office here as well.

Roger

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)

On 07/07/2023 00:25, Roger Hanscom wrote:

On Thursday, July 6, 2023 at 8:18:18 AM UTC-7, Luke A. Guest wrote: ...<snip>....

Yeah, according to the manuals, the NOP instruction isn't 0, which would've made sense.

According to the MC68020 UM, the encoding for it is 0100111001110001.

Yes, an opcode of all zeros is a variant of an OR. It might as well be a NOP. It executes without changing the processor state, so the CPU just chugs along, incrementing the address bus. No problem with the reset vector either. You'd think that the

CPU would pick up the starting PC as Zero, and just hang in an endless loop, but it doesn't. It steps through memory just fine.

You think inflation there is bad, try here where it's uncontrolled and
the tories have let fuel company's just take the piss completely, we
have the highest bills in the world, btw.

Sorry to hear that, Luke. Inflation is just another form of taxation on a populace that's stretched to the breaking point. These governments don't seem to realize that you can't get blood from a stone!

I'd really like to send you this extra copy of the Clements book that I have. It's not doing me any good to have two of them! Our postal service gives reduced rates for sending books in the U.S. Do you happen to know if that is true for the U.K. as

well? Might I get a decent postal rate in order to send it to the U.K? I think I'll check with the post office here as well.

Thanks, but I've no idea. You might be better off asking ups as they
operate here too.

--- SoupGate-Win32 v1.05
* Origin: fsxNet Usenet Gateway (21:1/5)