The Sonically Pure PagesOriginally Web posted Wednesday, 27 October 2004.
Content last modified
Friday, 25 January 2008
.
External links last verified Wednesday, 26 January 2005.
Notwithstanding the excellent information i found when searching xlr8yourmac.com and other WWW pages, nowhere could i find an explanation of what the dip switch settings on my MAXpowr G3 300 MHz CPU/200 MHz cache Internal Processor Upgrade Card did, nor recommended settings beyond some very skimpy information in the otherwise excellent Newer Technology manual.
Unless/until better information comes along, this page attempts to fill the knowledge gap by presenting the results of some tests i did. If anyone reading this page, especially Newer Technology engineers and/or technicians, cares to contact me with additional information, i will gladly add it to this page.
My motivation for spending the time diddling with the dip switch settings was to improve stability and eliminate some very obvious and distracting video artifacts. The Newer Tech MaxPower Pro 250MHz PPC750 card Review page at xlr8yourmac.com discusses video artifacts not being a problem at color depths other than Thousands. While i have seen this in other circumstances, this was not the case with my 8500/MAXpowr G3 300: problems were at least as bad, if not worse, at Millions of colors, and were visible even with 256 colors.
The same review and other sources mention that the problem is restricted to built-in Macintosh video, and is totally eliminated when an add-on PCI video card is used. This is true, and is due to the separate clocking of the PCI bus. The onboard video is clocked at the processor bus speed, which is typically altered (sped up) by accelerator cards. The artifacts are not restricted to cards from Newer Technology by any means! Not wanting to put more money into this mid-1990s Macintosh, nice as it is, and using the video glitching as a “warning canary” for actual and potential stability problems, i chose to attempt to make the stock onboard video glitch-free.
Also mentioned on the same review page is lowering processor bus speed to eliminate the artifacts. While the MAXpower Control control panel offers all sorts of nifty adjustments, none of them alter processor bus speed. Not coincidentally, no combination of control panel settings i could find solved the video artifacts problem.
Testing of the various dip switch settings was only performed on one configuration:
Previous testing revealed that the control panel settings had no effect on CPU clock speed, nor bus speed… just MAXpowr backside cache speed. Testing consisted of looking for visual artifacts around startup icons during startup, and repeatedly moving a single Finder window on the desktop. The latter test was especially good at revealing any video artifacts in a very short period of time.
No effort was made to be consistent about resetting NVRAM/PRAM. I found no difference in results whether those items were reset or not, as long as the MAXpowr card remained in its socket. The machine was cold booted, tested for glitches, shut down, then the dip switch settings changed and the process repeated. There were occasional crashes into MacsBug due to some aspect of the Control Strip during startup. No attempt was made to reproduce these crashes nor troubleshoot them. If there was to be a crash, it happened when the “incorrect date” dialog was displaying. Testing was done with an external power supply in place of the normal PRAM battery, hence the date being reset to its default. Casual inspection hints at the crash being due to times where the control strip preferences file was last modified at a time and date in the future, relative to current (incorrect) system time.
| Speeds are in MHz, as displayed by NewerTech Gauge PRO 1.1 | |||||
|---|---|---|---|---|---|
| DIP Switch Settings | CPU Speed | Cache Speed | Processor Bus Speed | Memory Performance Speed | Comments |
| All Off | 302.1 | 201.4 | 46.5 | 52.5 | |
| 1 On, Rest Off | 302.1 | 201.4 | 46.5 | 52.5 | No discernable change in operation vs. all Off |
| 2 On, Rest Off | 300.0 | 200.0 | 50.0 | 50.1 | The “nice round numbers” setting. Highest bus speed found. |
| 3 On, Rest Off | 300.7 | 200.4 | 43.0 | 50.1 | No video glitching! Slowest, safest bus speed found. |
| 4 On, Rest Off | 302.1 | 201.4 | 46.5 | 52.1 | No discernable change in operation vs. all Off |
| 1&4 On, 2&3 Off | 302.1 | 201.4 | 46.5 | 52.7 | Other than the control strip crash, no discernable change in operation vs. all Off |
| 2&3 On, 1&4 Off | 302.1 | 201.4 | 46.5 | 52.7 | Other than the control strip crash, no discernable change in operation vs. all Off. I expected another bus speed. |
| 1&2 On, 3&4 Off | 300.0 | 200.0 | 50.0 | 50.2 | Not discernably different than having just #2 On |
| 1&2 Off, 3&4 On | 300.7 | 200.4 | 43.0 | 50.1 | No video glitching |
| 1&3 On, 2&4 Off | 300.7 | 200.4 | 43.0 | 50.1 | Video glitches appeared with this setting, even with the slowest processor bus speed |
| 2&4 On, 1&3 Off | No startup: Flashing ? | ||||
| 1,2,3 On, 4 Off | 302.1 | 201.4 | 46.5 | 52.6 | No discernable change in operation vs. all Off |
| 1,2,4 On, 3 Off | 300.0 | 200.0 | 50.0 | 50.2 | Not discernably different than having just #2 On |
| 1,3,4 On, 2 Off | 300.7 | 200.4 | 43.0 | 50.2 | No video glitching |
| 2,3,4 On, 1 Off | 302.1 | 201.4 | 46.5 | 52.7 | No discernable change in operation vs. all Off |
| All On | No startup: Flashing ? | ||||
This Siber-Sonically Pure Page is 
, 
, and 
Transitional compliant.