CSA Derringer 030 - tuned and broken??

[McTrinsic]

Hi all,

I recently have had the oscillator of my Derringer 030 replaced. I now have a 100 MHz oscillator instead of a 50 MHz. The 50 MHz generated the clock for a 25 MHZ bus, so now i (should) have a Derringer running at 50 MHz (i.e., the oscillator speed is halved on that board).

However, since doing so, my card does not find any RAM!

Running the D3 utility, all I get is a message saying

" ** SizeSimmSizeUp reports a bitmap of $00000
  ** SizeSimmSizeDn reports a bitmap of $00000
no DRAM found."

I already replaced the original 32MB SIMM with a 16MB SIMM from a 1230 IV, but the same message.

Any ideas?

Is my Derringer broken now??


Any help is appreciated!!

Have fun,
McTrinsic

[fragment]

Have you tried again with the original oscillator? Propably the board or the CPU doesnt like such overclocking..

[McTrinsic]

The CPU was replaced with a 68030 @ 50 MHz, so that isnt the issue (probably .

I can't simply put the original oscillator back in again, because I didnt have a socket and so had to solder it onto the board...


Any ideas anyone?


Have fun,
McTrinsic

[Ilwrath]

My guess is that your RAM latency is also determined based on the CPU oscillator.  So when you doubled your CPU oscillator, you halved the latency to your RAM, thus putting either your RAM or perhaps a built-in MMU out of spec.  

Unless there is a jumper or trace to adjust this, I'm not sure how you'd fix it without hacking more stuff on the card, though.  

[Kronos]

 :roflmao:  :roflmao:  :roflmao:

Overclocking the board by 100% and that guy even thought it might work ?

 :roflmao:  :roflmao:  :roflmao:

You might get the board to 32-35MHz, but thats allready gonna need lots of luck, nerves (to survive constant crashes) and cooling (not just the CPU).

But 50MHz ?? :crazy:  :crazy:

[McTrinsic]

Well, it basically works.

And I still think it was worth a try . The Derringer is available as a "platinum" version, with a 100 MHz oscillator giving a 50 MHz board. So why NOT try??

I will try to replace the 100 MHz with a 66 MHz, getting me to (hopefully) 33 MHz.

Or does anyone have any other idea??

Anything I might / could do?

Have fun,
McTrinsic

[McTrinsic]

What might the MMU look like, btw???


McTrinsic

[Brian]

You must remember that when overklocking a board it's not just the CPU that gets a higher clockrate... everything else on the board does too. That's why overclocked systems might get a snappier SCSI controller and why we get faster simms (and for us lucky ppl that have cards with their own boot meny set a lower speed for the ram selecting perhaps 70ns even if it's 50ns simms sitting on the board).

The 50Mhz and 100Mhz cards doesn't seem to bee the same cards and the manual doesn't specify any jumper that you could perhaps set for differnet CPU speeds so it appears that there's no simple way of speeding up these boards the "right way". The wrong way however would be to try to bypass the part where the 50Mhz is lowered to 25Mhz for just the CPU.

[Ilwrath]

The wrong way however would be to try to bypass the part where the 50Mhz is lowered to 25Mhz for just the CPU.

This would be the wrong way?  It probably would have been my first thought as to a fix...?  Of course, you may introduce problems with the processor attempting to access the bus in mid-cycle, and such.  Eh, I never was so hot at the hardware side of things.  ;-)

[Kronos]

McTrinsic wrote:
What might the MMU look like, btw???


McTrinsic

Like this:                   :-P

The MMU is intregrated in the 68030 (if it isn't a "EC"-version).

[Ilwrath]

The MMU is intregrated in the 68030 (if it isn't a "EC"-version).

Yeah...  Actually, I was thinking of 020's, not 030's.  But some 030 accelerators used the "EC" chip and then an external MMU.  It would be a fairly obvious unit on the board, though.  And would probably be identified as an MMU (and a 68xxx part number).  Most likely, it uses the integrated MMU on a full 68030.

[McTrinsic]

Well, its a full 68030, i.e. it does include a MMU.

And it is not found by a SysInfo - program.

So unless anyone comes up with a better idea, I'll go and try with a 66MHz oscillator (= 33MHz board).

Have fun,
McTrinsic