View Single Post
Old 17 May 2013, 18:14   #850
Join Date: Jun 2010
Location: out in the wild
Posts: 1,129
That's neither a complaint about the website, nor about my products. Note that 150-pin CPU slot connectors are not a product that I sell, but a necessity for making CPU cards. I have paid well over 20k EUR for the tooling, and it's up to me to decide whether I support somebody else's project or not.

One thing you have confirmed with this public complaint is that the support ticket system works: You ask, you get an answer.

Originally Posted by mugo View Post
I've already asked in the other thread, but i think Jens haven't seen it.
That's because I'm not here on a regular basis. If you want to know details about my products, you need to direct the question to the person you want an answer from. Writing it in a public forum is like "hey, I drive a Volkswagen. I have a question about it, so I print it out and hang it into a window where a Volkswagen employee may see it". You don't really expect that to happen, do you?

Originally Posted by mugo View Post
Is it possible to upgrade the 33MHz CPU on a ACA1232 by a 50MHz CPU?
You can of course solder that in, but you can't clock it at 50MHz. Max would be 40MHz with the timing that the memory controller is running. Anything above that will violate memory timing limits, and you'll have an instable system.

I don't have a memory controller with the required amount of wait states for 50MHz CPU clock for the ACA1232, and I don't even know if it's possible to fit it into the logic (haven't tried yet). Right now, the memory controller benefits from "4 states only", resulting in two state bits only. Once more wait state would require an additional macrocell, and I have my doubts that this can still be routed in the 72-macrocell CPLD. Note that the old ACA1230 used a 144-macrocell CPLD with plenty of routing resources.

Originally Posted by mugo View Post
Beside this question, am i wrong in thinking that i may benefit of the CPU change anyway, as the ceramic 50MHz version, even clocked at 40MHz would heat a lot less than the original plastic 33MHz version? Of course, i'm not worried with warranty matters...
That depends on the 50MHz version you have: Only the "C" version (mask code 1F91C) is the one with the 0,8Ám die. All CPUs with the "B" mask (I think there's three of them) have a 1,0Ám mask, so they all produce the same amount of heat.


Last edited by Schoenfeld; 17 May 2013 at 18:25.
Schoenfeld is offline  
Page generated in 0.05538 seconds with 9 queries