David Cook
Well-known member
I recently purchased a IIsi from eBay without photos of the inside. Risky, I know. However, the rear expansion slot had a video connector, so I figured there was at least a video card and PDS adapter inside.
I was pleasantly surprised by an unusual PDS adapter with a cache SIMM. PSI 030 BUS ADAPTOR REV B P/N E-0K-0-51C 1990/1991. The adapter has the FPU installed. (There was also a video card, but that isn't the focus of this post.)


There's a PAL on the adapter. I assume this is for the FPU.

I don't like the look of that ceramic capacitor.

Nice bend on electrolytic C1! When I recapped C2, I made the mistake of installing a taller capacitor. It barely fits -- rubbing against the power connector for the hard drive. C2 really needs to be small like the original designer installed.

The cache module is labeled PSI CACHE-It!-IIsi REV B. P/N EA-64K-20-rb 1990/1991. Based on the 64K part number (and tests presented later), this is a 64KB cache.

A cache simm is a pretty cool concept for 1990. It would be cool to swap in a larger cache, but I assume this is a proprietary pinout.


I could only find one mention on the Internet. https://archive.org/details/TNM_030_Bus_Adapter_Card_for_Macintosh_IIsi_-_PSI_20170827_0383

Notice they say the cache is rated for 25 MHz. Conspiracy theorists can point to this as further evidence that the IIsi was intended to be 25 MHz. I suppose this means that this will work in an 'overclocked' IIsi, although sadly the FPU installed in my adapter is only marked for 20 MHz.
Are you jealous of my really cool unique cache card? Well, read on...
I ran Norton System Info speed tests on System 7.1 on the IIsi. I compared the DayStar FastCache (with the control panel software and without) vs the PSI vs the stock IIsi. The PSI cache did not compare favorably.

Notice that BlockMove is particularly bad (below stock) on the PSI cache, but not on the DayStar cache. I wonder why?
Video performance is generally better with a CPU cache. Still DayStar constantly beats PSI.

Just to be sure Norton isn't wrong, I ran my cache tester program. Results were consistent with Norton.

In conclusion, the IIsi receives a moderate performance boost from a CPU cache. DayStar is better than PSI.
- David
I was pleasantly surprised by an unusual PDS adapter with a cache SIMM. PSI 030 BUS ADAPTOR REV B P/N E-0K-0-51C 1990/1991. The adapter has the FPU installed. (There was also a video card, but that isn't the focus of this post.)


There's a PAL on the adapter. I assume this is for the FPU.

I don't like the look of that ceramic capacitor.

Nice bend on electrolytic C1! When I recapped C2, I made the mistake of installing a taller capacitor. It barely fits -- rubbing against the power connector for the hard drive. C2 really needs to be small like the original designer installed.

The cache module is labeled PSI CACHE-It!-IIsi REV B. P/N EA-64K-20-rb 1990/1991. Based on the 64K part number (and tests presented later), this is a 64KB cache.

A cache simm is a pretty cool concept for 1990. It would be cool to swap in a larger cache, but I assume this is a proprietary pinout.


I could only find one mention on the Internet. https://archive.org/details/TNM_030_Bus_Adapter_Card_for_Macintosh_IIsi_-_PSI_20170827_0383

Notice they say the cache is rated for 25 MHz. Conspiracy theorists can point to this as further evidence that the IIsi was intended to be 25 MHz. I suppose this means that this will work in an 'overclocked' IIsi, although sadly the FPU installed in my adapter is only marked for 20 MHz.
Are you jealous of my really cool unique cache card? Well, read on...
I ran Norton System Info speed tests on System 7.1 on the IIsi. I compared the DayStar FastCache (with the control panel software and without) vs the PSI vs the stock IIsi. The PSI cache did not compare favorably.

Notice that BlockMove is particularly bad (below stock) on the PSI cache, but not on the DayStar cache. I wonder why?
Video performance is generally better with a CPU cache. Still DayStar constantly beats PSI.

Just to be sure Norton isn't wrong, I ran my cache tester program. Results were consistent with Norton.

In conclusion, the IIsi receives a moderate performance boost from a CPU cache. DayStar is better than PSI.
- David