View Single Post
Old 31 May 2016, 18:27   #421
Join Date: Jun 2010
Location: out in the wild
Posts: 1,245
Originally Posted by traal View Post
So instead of preconfiguring it with the fewest options, preconfigure it with the most options?

Why not make it more like ordering a computer online: let everyone configure it however they want at the time of purchase? It would save buyers from needing to pay more money again later to finally get what they want. I think people are resentful when they have to pay twice like that.
I haven't made a decision yet. From what I found so far, all E13G CPUs work fine @42MHz (even the 16MHz-rated ones!), so I might as well enable that "always" and treat it like will treat that frequency of the ACA500plus: Have it always available, but don't guarantee it.

For the ACA1221EC, this would mean you pay for a 28MHz accelerator, and if it works at 42: Good for you.

For the ACA500plus, you'll pay for a 14MHz accelerator. If it works at 21, 28 or 42: Good for you.

If the higher frequency doesn't work, it's just bad luck. However, I do have to make a difference for those accelerators that use a CPU with a mask that is known to "not like" overclocking. I'll have to sell these cheaper.

Originally Posted by traal View Post
Have you considered the 68EC030? I think it has the full 32-bit address bus so it won't be limited to 16 MB.
The 68EC030 has the full address space, that's correct. It lacks the MMU, though, so Linux-68k folks would not be happy.

When I did my very first A1200 accelerator, the ACA1230, I had lots of EC030-25 and developed the MapROM feature, so lack of MMU doesn't hurt that much. I do have a few EC030's left in stock, but the majority of CPUs I have includes the MMU. You don't save much money on EC030 versus full-030. It's mainly a matter of availability, not of price. I found a nice stock of brand new 68030 (partly Motorola, partly Freescale), which is why I have taken the time to improve the ACA1233 one more time.

Schoenfeld is offline  
AdSense AdSense  
Page generated in 0.03905 seconds with 10 queries