[comp.sys.amiga] CSA Comments on Amiga World Review

plouff@nac.dec.com (04/08/88)

A couple of months ago, _Amigaworld_ reviewed the CSA Turbo 020 
accelerator for the 2000 in a very screwy article.  In particular, the 
reviewer claimed that the 68020 ran slower than a vanilla 68000!  I then 
wrote a note to this conference titled "Still More 68020 Questions," 
trying to find out how the operating system treated the '020, and what 
the reviewer might have done wrong.

Well...  At the Motorola "Imaginology" seminar just completed in Boston,
Bill Reed, CSA's director of marketing, was showing off their Over 030
platform board in an A2000 running an animation demo, and in a Mac II
with the cover off.

When asked about the review, he stated, "The author was running version 
1.1, which turns the cache off."  Actually, it never turns the 68020 
cache on, but the result is the same.  He continued, "We have asked 
Amigaworld several times for a correction or retraction, but so far have 
gotten no response.  Look for our new ads which say, yes, the turbo 
board really _does_ speed up the Amiga."

Remember the AW editorial saying that their technical content was "just 
right?" :-)

When asked about problems operating with the A2090, he said something 
like "mumble, mumble... early Commodore production run has problems." 
Not knowing all the facts, I didn't pursue it. 

BTW, CSA's 68030 and 68882 platform boards are priced at $495 and $295 
respectively, without the processor chips installed.  Moto says that the 
68030 at 20 MHz is over $500 in small quantities right now.  In contrast 
small quantities of the 68020 and 68881 at 12 MHz are $152 and $129
respectively, with 16 MHz versions at $218 and $172. 

Anyone who wants to know more about 68020 accelerator or platform boards
would do well to read Motorola Ap Note AN944, "MC68020 and MC68881
Platform Board for Evaluation in a 16-bit System."  This is available 
from your local Moto sales office or by calling 1-800-521-6274.

Happy hardware hacking,

-- 
Wes Plouff
Digital Equipment Corp.		Internet: plouff%nac.dec@decwrl.dec.com
Littleton, Mass.		UUCP:     ...!decwrl!nac!plouff

plouff@nac.dec.com (Wes Plouff - NAC LAN Engr 226-7423) (04/14/88)

[Second posting... didn't see this show up on the net.]

A couple of months ago, _Amigaworld_ reviewed the CSA Turbo 020 
accelerator for the 2000 in a very screwy article.  In particular, the 
reviewer claimed that the 68020 ran slower than a vanilla 68000!  I then 
wrote a note to this conference titled "Still More 68020 Questions," 
trying to find out how the operating system treated the '020, and what 
the reviewer might have done wrong.

Well...  At the Motorola "Imaginology" seminar just completed in Boston,
a CSA marketing guy was showing off their Over 030 platform board in an
A2000 running an animation demo, and in a Mac II with the cover off. 

When asked about the review, he stated, "The author was running version 
1.1, which turns the cache off."  Actually, it never turns the 68020 
cache on, but the result is the same.  He continued, "We have asked 
Amigaworld several times for a correction or retraction, but so far have 
gotten no response.  Look for our new ads which say, yes, the turbo 
board really _does_ speed up the Amiga."

Remember the AW editorial saying that their technical content was "just 
right?" :-)

When asked about problems operating with the A2090, he said something 
like "mumble, mumble... early Commodore production run has problems." 
Not knowing all the facts, I didn't pursue it. 

BTW, CSA's 68030 and 68882 platform boards are priced at $495 and $295 
respectively, without the processor chips installed.  Moto says that the 
68030 at 20 MHz is over $500 in small quantities right now.  In contrast 
small quantities of the 68020 and 68881 at 12 MHz are $152 and $129
respectively, with 16 MHz versions at $218 and $172. 

(Anyone who wants to know more about 68020 accelerator or platform
boards should read Motorola Ap Note AN944, "MC68020 and MC68881 Platform
Board for Evaluation in a 16-bit System."  This is available from your
local Moto sales office or by calling 1-800-521-6274.) 

Happy hardware hacking,

-- 
Wes Plouff
Digital Equipment Corp.		Internet: plouff%nac.dec@decwrl.dec.com
Littleton, Mass.		UUCP:     ...!decwrl!nac!plouff

"Ashes to ashes, dust to dust
 Let the rich man live and the poor man burst"  -- The Oyster Band

daveh@cbmvax.UUCP (Dave Haynie) (04/21/88)

in article <8804132025.AA08587@decwrl.dec.com>, plouff@nac.dec.com (Wes Plouff - NAC LAN Engr 226-7423) says:
> BTW, CSA's 68030 and 68882 platform boards are priced at $495 and $295 
                       ^^^^^
> respectively, without the processor chips installed.  

Couldn't be.  No.  Never!  The 68882 is pin compatible with the 68881.
Completely.  100%.  Pop out your '881, push in your '882, and fire away.
Are you sure CSA is selling a tower board to run a 68882 in a 68881 
environment?  I can't believe it for a second.  The '030 is another story,
there's a complete pinout change there, not to mention a few other details
that have to be covered to run a 68030 in a 68020 system.

> (Anyone who wants to know more about 68020 accelerator or platform
> boards should read Motorola Ap Note AN944, "MC68020 and MC68881 Platform
> Board for Evaluation in a 16-bit System."  This is available from your
> local Moto sales office or by calling 1-800-521-6274.) 

A good start.  But Moto tends to oversimplify the interfacing needed.  If
you were to build a 68020 board directly from the Moto application notes,
you'll run into trouble, at least on an Amiga.  The notes are pretty much to
get you started, and for that keeping things simple is a good idea.

> Wes Plouff
> Digital Equipment Corp.		Internet: plouff%nac.dec@decwrl.dec.com
> Littleton, Mass.		UUCP:     ...!decwrl!nac!plouff
-- 
Dave Haynie  "The B2000 Guy"     Commodore-Amiga  "The Crew That Never Rests"
   {ihnp4|uunet|rutgers}!cbmvax!daveh      PLINK: D-DAVE H     BIX: hazy
		"I can't relax, 'cause I'm a Boinger!"

plouff@nac.dec.com (Wes Plouff) (04/25/88)

A few days ago
>  cbmvax!daveh (Dave Haynie at Commodore) said
>> I said

>> BTW, CSA's 68030 and 68882 platform boards are priced at $495 and $295 
>                       ^^^^^
>> respectively, without the processor chips installed.  
 
>Couldn't be.  No.  Never!  The 68882 is pin compatible with the 68881.
>Completely.  100%.  Pop out your '881, push in your '882, and fire away.
>Are you sure CSA is selling a tower board to run a 68882 in a 68881 
>environment?  I can't believe it for a second.  The '030 is another story,

Here's what the (undated) press release says.

"What may be even more exciting than CSA's 'Over 030' board is CSA's 
new 'FX-882'(TM) upgrade board.  This easily installed platform board 
can effectively double or triple system performance by replacing the 
MC68881 co-processor in any 680X0 based system with Motorola's new 
68882, as well as providing circuitry to increase the clock frequency up 
to 25 MHz."

CSA's value added in this case is the high-speed clock circuitry and
possibly support logic.  My memory of the actual board at the Motorola
show is quite hazy. 

>> (Anyone who wants to know more about 68020 accelerator or platform
>> boards should read Motorola Ap Note AN944, "MC68020 and MC68881 Platform
>> Board for Evaluation in a 16-bit System."
 
>A good start.  But Moto tends to oversimplify the interfacing needed.  If
>you were to build a 68020 board directly from the Moto application notes,
>you'll run into trouble, at least on an Amiga.  The notes are pretty much to
>get you started, and for that keeping things simple is a good idea.

Absolutely.  I just wanted to inject a few facts into the floating
discussion, and show that it's not a 10-minute project, and not rocket
science either, but is straightforward engineering. 

Last night a friend brought over the May issue of _Amiga Sentry_, in
which you write about your A2620 board.  Nice job on the board!  Nice
description!  Your article touches nicely on the range of design
problems in adapting a fast 32-bit processor to the Amiga. 

Certainly the ap note doesn't cover the design of no-wait-state systems, 
which the Amiga/68000 system is.  It certainly doesn't cover the 
subtleties of DTACK timing for chip RAM.  Your article says more about 
putting a 14 MHz processor in a 7 MHz system than anything on the 
subject I have seen in public print, especially regarding cycle-by-cycle 
timing differences between the '000 and '020.

But the ap note remains the best (and only) introduction to 68020 
platform boards I have seen.

--  
Wes Plouff			Phone:	  (617) 486-7423
Digital Equipment Corp.		Internet: plouff%nac.dec@decwrl.dec.com
Littleton, Mass.		UUCP:     ...!decwrl!nac!plouff