Upgrade kit for the A3640 (2 wait states removed)

[Cosmos Amiga]

Remove two wait states for a general speedup. Works well with a 68040 and with the 040-060 adapter from Mozart & Matze

- Five new sockets
- Three new Gal 22V10 (10ns)
- Two new Gal 16V8 (7ns)


21 euros including Priority shipping in gift on my Paypal

 


What is it ?
It's a freeware firmware upgrade by SpeedGeek for A3640 owners seeking a
performance boost. The new state machine logic eliminates 2 extra wait states
from every A3640 bus cycle so your 68040 can access the A3000/A4000
motherboard resources at full speed. This logic has been tested at clock
speeds from 25-50 MHz*. It works with either a 1x or a 1/2x clocked
motherboard (other mods are required for a 1/2x clocked motherboard).

How much faster will my A3640 be ?
It's hard to quantify but you can expect Fast memory and ROM speeds to
improve by 2-3 MB/sec. Chip memory, Zorro Bus and Custom chips will
probably have < 1 MB/sec increase. Your mileage will vary with clock
speed, ROM and RAMSEY speed settings, etc.

 
*The new U207 and U208 were tested by apollo630
since they were created to avoid cut and jumper mods on the originals. If you
don't want to use them then cut the trace on U207 pin 9 and solder jumper it
to pin 10. Also, cut the trace on pin 3 U208 and jumper it to GND or a pull up
resistor if you have the inverted version (from the old EAB archive).

(U207 is the IC with 1 white dot)
(U205 is the IC with 2 white dots)
(U204 is the IC with 5 white dots)

(U203 is the IC with 3 white dots)
(U208 is the IC with 4 white dots)

[kinmami]

A3640 must be 3.1 or 3.2 version?

[Cosmos Amiga]

Should works on any revision...

Please, ask SpeedGeek to be sure !

[slayer]

Too bad this wasn't available a year back when I had 4 of my A3640s redone and upgraded to 3.2.

Nice to see these improvements coming out!

[SpeedGeek]

A3640 must be 3.1 or 3.2 version?

Version and other questions already answered on EAB:

http://eab.abime.net/showthread.php?p=778105

Too bad this wasn't available a year back when I had 4 of my A3640s redone and upgraded to 3.2.

Nice to see these improvements coming out!

WTF? This A3640 upgrade was first released on EAB 3.5 years ago! It was later moved to Amibay because some idiots on EAB posted negative comments about it. Since then, many people have done this upgrade and have been very pleased with the results.

[Oldsmobile_Mike]

WTF? This A3640 upgrade was first released on EAB 3.5 years ago!

*A LOT* of info posted on EAB and Amibay never makes it over to amiga.org.  Or if it does, it dies in a single thread that receives zero comments and so drops off the homepage quickly, before most people even have a chance to see it.  It's kind of disappointing and frustrating, really, especially since the threads that *DO* stay at the top of the homepage for months, are usually full of trolling/FUD/other garbage.  

[Motormouth]

Is anyone willing to offer a soldering service for this speedgeek upgrade, I am definitely interest in it but time and albeit my surface soldering skills are not great.

[biggun]

This is great.
I would have loved to have this on my 4000/040 some years ago.

[guest11527]

This is great.

Is it? Does anyone know why the wait states were there in first place? Surely not because some engineers said, "Ok, folks, let's make the product a bit slower so it sells better". Surely, we have nowadays faster RAM and faster logic, but can this be applied on this card? Likely not. Thus, I would probably give that a second thought and wonder what exactly the wait states are good for.

[biggun]

Is it? Does anyone know why the wait states were there in first place? Surely not because some engineers said, "Ok, folks, let's make the product a bit slower so it sells better". Surely, we have nowadays faster RAM and faster logic, but can this be applied on this card? Likely not. Thus, I would probably give that a second thought and wonder what exactly the wait states are good for.

ok good point.

My assumption is that this was tested now and is known to work reliable.
Someone should verify this of course.

I naiv assumed is that one of the reasons for the slowness of the 3640 might have just been "time-to-market".
E.g as soon as it worked it was sold - without giving the developers time for tuning.

[psxphill]

I naiv assumed is that one of the reasons for the slowness of the 3640 might have just been "time-to-market".
E.g as soon as it worked it was sold - without giving the developers time for tuning.

As it's commodore, you can assume they sold it before it worked. An insight into what was going on at commodore at the time http://www.wonkity.com/~wblock/a4000hard/defibust.html

Whether those wait states are ever required is something that anyone doing this upgrade will have to consider, with the hack it might only work in a limited combination of a3640/buster/motherboard/phase of moon etc. If it works for you and you don't have an issue modifying classic hardware then I don't see the problem. If you have problems in the future then it will always be the first thing that you will doubt.
 
 Dave was human, he made mistakes. They might have been added under an assumption they were required, or they might not even have been intended. You could ask him, he might remember.

[ChaosLord]

Is it? Does anyone know why the wait states were there in first place? Surely not because some engineers said, "Ok, folks, let's make the product a bit slower so it sells better".

The Engineers did not run Commodore.  Mehdi Ali did.  And Mehdi Ali said stupid crap like that all the time.

I assume that Dave simply did not have 1 or more of the following: [the time, the energy, the skill or the permission] to improve the design and eliminate the wait states.

Most (possibly all) other accelerator cards do not have a bunch of unneeded extra wait states thrown in and they all work just fine.

 Surely, we have nowadays faster RAM and faster logic, but can this be applied on this card? Likely not. Thus, I would probably give that a second thought and wonder what exactly the wait states are good for.

If someone removes the wait states and can then prove that it causes some sort of problem then I would love to hear about it.

As for me, if this had been available in the years 1992-1997 then I would have bought it instantly as I was rocking an A4000/040 back in those days and I always hated that my hardware was performing at a level less than perfection.

[SpeedGeek]

I guess there is no Amiga forum where you can avoid negative comments from the ignorant, incompetent or resident site trolls... :rolleyes:

First of all, Commodore engineering didn't add the extra wait states to the A3640. They were already incorporated into the "040 Bus Adapter Application" designed by Motorola. This application was intended to serve as a design example for 040 designer's. It was never optimized nor intended to become the "Default" design for 040 cards (Unfortunately, that's exactly what happened).

Since, Commodore engineering was ordered to produce the cheapest possible 040 card, it's very easy to understand why they used the Motorola design with as few modifications as possible.

For you German guys, Mozart was actually one of the first A1K.org members to do this upgrade:

http://www.a1k.org/forum/showthread.php?t=37954

[psxphill]

Since, Commodore engineering was ordered to produce the cheapest possible 040 card, it's very easy to understand why they used the Motorola design with as few modifications as possible.

Section 7 from http://bitsavers.trailing-edge.com/pdf/motorola/68000/MC68040_Designers_Handbook_1990.pdf ?
 
 According to Dave the "cheapest possible 040 card" came out of engineering when it was going to be used in  the a3200/a3400.
 
 http://www.bigbookofamigahardware.com/bboah/product.aspx?id=1943
 
 "So, anyway, my buddy Greg Berlin was in charge of the A3200 and A3400, the only thing I did was fix his expansion bus when it didn't work. He was as annoyed about these models as everyone else, and so, kind of as a skunkworks project, he had Scott design "the cheapest '040 board known to mankind" as he put it, the idea being that, even if these systems went out and didn't impress anyone, we at least had this add-on to turn your A3200 into more or less an appropriate 1992-ish computer."
 
 Things didn't improve greatly when A3200/A3400 was killed to be replaced with the A4000. I think engineering were just grateful to be allowed to release a computer with AA.

[Oldsmobile_Mike]

I guess there is no Amiga forum where you can avoid negative comments from the ignorant, incompetent or resident site trolls... :rolleyes:

https://www.youtube.com/watch?v=coOxrm4CU0U