Upgrading IIsi onboard RAM to 4MB
- Thread starter Zane Kaminski
- Start date
Well, the only RAM currently in the system are the 1MB onboard chips. Because one of the chips was not identical to the others, I removed them (no death chime anymore!) and replaced them with those of a battery-bombed IIsi. I cannot confirm that they are working though...
As I wrote on the other forum (see link above), I get weird voltages on the lower three bytes of the data bus during early startup until the end of the death chime. Because the bus buffers are the only ICs connected to those signals apart from CPU, ROM SIMM and the absent onboard ROM chips, it is in my opinion not unlikely that they might be the cause of the trouble apart from the possibility that there is a short circuit or other interference somewhere else on the board.
I wonder if a IIsi can boot with empty bank A, populated bank B and a separate video card...
Interesting question.
I'm going to say it's not likely because one of the first things the IIsi does at startup is to poll the sense lines on the video connector. If a monitor is detected, a portion of Bank A is buffered for use as Vampire Video's VRAM. If none is detected, the system has full use of Bank A memory and there is no performance hit.
If @Zane Kaminski was correct in his guess that the IIsi tested Bank B before it bombed, might could work? I don't see how Vampire Video could slow things down to the crawl of a bus width limited LCII however. But the problems involved stem from its IIci heritage. In that machine you want to install expanded memory in Bank B for the same reasons, no?
No IIsi should make long term use of Vampire Video as far as I'm concerned. To buy one in order to limp along with Vampire Video while saving up for video card would be the only reason to use it for any length of time back in the day. I see it as a way to get the graphics capability of a II series Mac on a budget. With the copro on the NuBus Adapter for VidCard it's quite a nice low tier II.
I'm going to say it's not likely because one of the first things the IIsi does at startup is to poll the sense lines on the video connector. If a monitor is detected, a portion of Bank A is buffered for use as Vampire Video's VRAM. If none is detected, the system has full use of Bank A memory and there is no performance hit.
If @Zane Kaminski was correct in his guess that the IIsi tested Bank B before it bombed, might could work? I don't see how Vampire Video could slow things down to the crawl of a bus width limited LCII however. But the problems involved stem from its IIci heritage. In that machine you want to install expanded memory in Bank B for the same reasons, no?
No IIsi should make long term use of Vampire Video as far as I'm concerned. To buy one in order to limp along with Vampire Video while saving up for video card would be the only reason to use it for any length of time back in the day. I see it as a way to get the graphics capability of a II series Mac on a budget. With the copro on the NuBus Adapter for VidCard it's quite a nice low tier II.
In my experience, I have rarely succeeded when throwing chips at a problem in lieu of doing all the continuity checks to verify the board wiring. I dunno if the buffers were broken in this case, since I did replace them, but I suspect they would have worked fine. I was just hoping I could replace the buffers instead of "beeping out" the problem lol.
Last edited:
I fixed the board today; the bus buffers were indeed the problem.
The problem with using a video card in the IIsi is that you are then prevented from using ethernet. So, given that, I’d rather use the onboard video, even if it’s slower.
The IIci’s setup is much better, since you have 3 slots instead of only one. That doesn’t mean that I don’t like the IIsi, though. It’s quite an interesting form factor.
- Alex
The IIci’s setup is much better, since you have 3 slots instead of only one. That doesn’t mean that I don’t like the IIsi, though. It’s quite an interesting form factor.
- Alex
I got a lot of wear out of my IIsi; it was the first Mac I ever personally owned. At one point I actually did put a NuBus expander and an 8*24*GC in it, since at the time I only used it with a dialup modem and LocalTalk. It did basic papers, Telnet and a little bit of web browsing until I got a 7200 and then a 7300.
I read somewhere that you can take the 8 chips off a 1mb RAM SIMM and use those to replace the onboard 1mb. Is this correct? I have some known-good 8-chip 1mb 80ns SIMMs, and this seems like the easiest way for me to get known-good chips to replace the onboard RAM.
Yes this will likely work as long as the rated access time is equal or lower and the pinout is the same.
Nooo!! A 30-pin SIMM is too narrow in terms of data bus width. It's got eight 1-bit chips so altogether you only get 8 bits, but the IIsi needs 32 bits of RAM data. You can grab the chips off a 72-pin SIMM and that will work.
Not true at all, especially so in the case of the Radius Color Pivot II/IIsi in combination with the Asante MacCon board for IIsi plugged into the RCPII/IIsi passthruslot. Provision is made on the two part breakout board of the MacCon. Just remove the ThickNet section from the 10bTsection and its DA15 backplane plate opening allows the breakout cable DA15 from the VidCard then bolts right up to the opening. IIRC such is illustrated in the MacCon installation docs.
Alternately, even a full size SE/30, 20 MHz compliant VidCard can be plugged into the MacCon card's PDS passthru. in that case the RJ45 for 10bT would need to be bodged out the back.
Simply drilling a hole in the VidCard's cover plate and crimping connectors onto the wire on both sides of the plate would be an elegant solution.
Better yet, a larger hole could house a WiFi antenna. Plenty of cubic available for the installing the loose inner bits if using an SD card setup or a 2.5" HDD.
Last edited:
The SCSI Rider can free up more room since that renders the hard drive bracket unnecessary in the SE/30.
Check the PCB of the IIsi, can't imagine that setup being compatible in terms of height or expansion cubic interference even if a RA socket config?
Yeah, you can put a double riser card in that accommodates an accelerator and an Ethernet/video card. It’s tight but it can be done
@Zane Kaminski what @ajacocks was lamenting would be lack of Ethernet if using a decent VidCard in Apple's standard PDS Adapter.
Accelerator/Passthru risers are great, but won't solve that particular problem. If you have a TwinSlot riser, then two PDS cards will do the job if using an SE/30 form factor VidCard in lieu of the RCPII/si.
@retr01 there's a goodly portion of cubic available under the FDD in the IIsi.
Accelerator/Passthru risers are great, but won't solve that particular problem. If you have a TwinSlot riser, then two PDS cards will do the job if using an SE/30 form factor VidCard in lieu of the RCPII/si.
@retr01 there's a goodly portion of cubic available under the FDD in the IIsi.
Sorry, I arrive after the battle @Zane Kaminski 
Maybe those are stupid questions but ...
With your discovery about "upgrading" onboard RAM, what do you think if we put faster chips instead of original ones ? Will they enhance the memory access or not ?
Does it allow access to higher video resolution to put 4MB instead of original 1MB ?
The IIsi was my first color Mac when I was young and I was frustrated by its strange memory configuration (1+4) that today I understand.
Maybe those are stupid questions but ...
With your discovery about "upgrading" onboard RAM, what do you think if we put faster chips instead of original ones ? Will they enhance the memory access or not ?
Does it allow access to higher video resolution to put 4MB instead of original 1MB ?
The IIsi was my first color Mac when I was young and I was frustrated by its strange memory configuration (1+4) that today I understand.
Interesting question about resolution! My WAG would be that the buffered main memory (Vampire Video) would be limited in ROM/memory map?
Faster RAM is never a problem, it's addressed at the system clock at the same speed as specified RAM. Under spec/marginal RAM speed is a no go as it's not responsive enough at system clock reliably.
One of my very long running projects hooking the pads up to a carrier board for addressing 8MB-32MB in Bank A. Such requires bodging four signals directly from the memory controller, IIRC. @Zane Kaminski what do you think of that possibility?
Faster RAM is never a problem, it's addressed at the system clock at the same speed as specified RAM. Under spec/marginal RAM speed is a no go as it's not responsive enough at system clock reliably.
One of my very long running projects hooking the pads up to a carrier board for addressing 8MB-32MB in Bank A. Such requires bodging four signals directly from the memory controller, IIRC. @Zane Kaminski what do you think of that possibility?