Thread: ACA500 tested
View Single Post
Old 02 May 2016, 16:43   #642
Join Date: Jun 2010
Location: out in the wild
Posts: 1,251
Originally Posted by kolla View Post
My mind boggles...
The explanation is easy: Practically everything on the ACA500 is connected to the same data bus: CPU, CF-cards, RAM, expansion ports. With more load on the data bus, CF-cards have a hard time driving that bus, and they may not work properly.

There is of course a technical solution called bus drivers. Chips that can be used to create subdivisions of a (data-)bus to have a known drive strength and a known time-delay.

I wanted the ACA500 to be extremely cheap. I therefore decided to treat the ACA500 as one part of the bus and the A500 host computer as the "second" part of the bus, so only one pair of data bus drivers (2x8 bits) had to be used. I already stated on that I specifically made the decision against a 4MB version because I didn't want the larger part of customers to experience worse compatibility with CF cards. I know that I may have lost a few sales and that I could have easily made 40,- EUR more per card.

For the ACA500plus, there are several subdivisions of the data bus in order to have exactly what the ACA500 doesn't have: Known load and drive strength in each section:

1) CF card#1 has it's own data bus driver
2) CF card#2 has it's own data bus driver
3) SD-Ram and the local expansion port share a data bus section
4) 68ec000 CPU and A1200 acceerator have their own data bus section
5) the A500 host computer has it's own data bus driver (similar to ACA500)
5b) clock port and the LED/display-processor are on the A500 side

Cracking up the design into so many subdivisions requires more components, more development time/testing, higher trace density - in short: It costs more money, which I did not want to charge for the ACA500.

About "personal attacking": I think only Leffmann really deviated from what I'd expect to be "discussion culture" (I'm not a native speaker, but the B-word doesn't exactly sound polite). I do know that I can't make a single product that appeals to every single customer, and I also have no problems with people who want a cheaper product with more features. I just can't serve them, and they are free to buy something else instead

However, I do have a problem with people who argue for the sake of arguing. As hopefully explained clearly enough in this posting, there is no point in arguing with the laws of physics. Open your eyes, look at the number of components on the ACA500 and explain to me why I should be technically wrong.

Schoenfeld is offline  
Page generated in 0.03909 seconds with 10 queries