LC 475 Project - Upgrades - Mods

[ArbysTPossum]

Picture of my Q605, 52mhz with 68040RC33, 60ns VRAM and spicy o clock. I’m experimenting with the 3D printed fan duct, need to compare cpu temps with/without.

I've done something similar, and quite a bit lower budget. The normal fan blows air out through the bottom, but if you flip the fan it'll blow air into the case, but it also makes it much louder. I angled the fan backwards and made a duct out of tape. Bringing air into the case alone cools things down significantly, 040 running at 33Mhz and has a heatsink. Other components also get a bit warm, and have gotten their own heatsinks. I'm curious about how the duct works, does it make the whole thing louder and does it work well?

 

[Brooklyn]

I've done something similar, and quite a bit lower budget. The normal fan blows air out through the bottom, but if you flip the fan it'll blow air into the case, but it also makes it much louder. I angled the fan backwards and made a duct out of tape. Bringing air into the case alone cools things down significantly, 040 running at 33Mhz and has a heatsink. Other components also get a bit warm, and have gotten their own heatsinks. I'm curious about how the duct works, does it make the whole thing louder and does it work well?

I also flipped the fan around. Doesn’t seem too loud to me. I haven’t taken temps yet and I’m out of town for a bit.

Angling the fan sounds like a good idea, right now the fan blows up and hits the wall of the duct, not the best for flow.

 

[Bolle]

Or even better… move the CPU into the airflow instead of trying to funnel airflow towards the CPU:

Stole that idea from that weird Formac accelerator and applied it to a cache card.

 

[pizzigri]

Id love a cache card….

 

[Nixontheknight]

Or even better… move the CPU into the airflow instead of trying to funnel airflow towards the CPU:
View attachment 75347
Stole that idea from that weird Formac accelerator and applied it to a cache card.

I wonder if it's possible to do this with a PowerPC accelerator?

 

[Kamshaft]

Photo update!
Upgrades! BlueSCSI with wifi and 1MB of VRAM baby!

 

[Brooklyn]

Or even better… move the CPU into the airflow instead of trying to funnel airflow towards the CPU:
View attachment 75347
Stole that idea from that weird Formac accelerator and applied it to a cache card.

That’s awesome! How much performance increase do you get with the cache?

 

[Kamshaft]

I have a BlueSCSI question. I've got a bootable HD0 System.hda file, and it boots. But when I add a HD1 - Data.hda file, it won't boot anymore. Ideas?

EDIT: This mac needs to have the SCSI IDs from large number (boot drive) to lower.

 

[pizzigri]

That’s awesome! How much performance increase do you get with the cache?

It’s literally YEARS I’m pestering @Bolle asking him for performance info and a chance for one of these cache boards….

 

[Bolle]

It’s literally YEARS I’m pestering @Bolle

Hehe, I know
I just recently made that card with a proper 475 footprint. The other variants I built so far don't fit nicely into the 475.
I can't say much about the performance gain other than it's faster than without it. Any specific benchmark you're looking for?
Speedometer won't show any difference at all because the benchmarks it runs actually fit into the CPUs own cache except video score goes up a bit. Daystar PowerDemo (which was made specifically made to show off the gains by external cache) shows a 20-40% increase depending on the benchmark.

 

[pizzigri]

Well, even a comparative idea would be enough, do you think that it could compare to the perceived increase from a 33 to 40 mhz 040 cpu? also, consider that I‘m more inclined to place it inside my 650 board inside a Quadra than a 475 - well maybe both in the end

 

[trag]

Would this RAM work? https://www.ebay.ca/itm/304833489030
or

256MB (2 x 128MB) 72 Pin EDO Memory SIMM 5V 3V 60ns | eBay

Find many great new & used options and get the best deals for 256MB (2 x 128MB) 72 Pin EDO Memory SIMM 5V 3V 60ns at the best online prices at eBay! Free shipping for many products!

www.ebay.ca

Kingston made a bunch of modules for HP that work well in the Q605/LC475. They are HP part # D4893 (128 MB) or D4290A (64MB, note there's an HP 4290E DDR2 module out there) and the Kingston part number for the 128 MB is kmm53632000

 

[Kamshaft]

So, thus far I've got...

1. An overall cleaning (done)
2. Swap out the old caps with tantalum caps (done)
3. Replace the old battery (done)
4. Put an internal BlueSCSI adapter (maybe with wifi) - Received, operational, installed but not permanently fitted yet.
5. Maybe... Replace the fan with a quieter one. Noctua maybe?
6. Maybe upgrade the current 128K x 2 VRAM with 512K x2 sticks? (done)
7. Upgrade Ram to 64MB (ordered, not arrived yet)
8. Upgrade CPU to full 68040 @ 33Mhz (change the speed on the motherboard), add a heat sink (ordered, not arrived yet)

Just sharing a project update. Thanks again for the support and insight.

 

[Kamshaft]

I want to circle back to my operating system question. Some of you said to install 7.5.3 and some 7.5.5.

It's been a while, but what are the main differences in between these point releases? (if any)

Because right now I've installed 8.1 to a .hda image, runs fine. But a tad slow.

 

[ArbysTPossum]

I reserve macOS 8 and beyond for PowerPC, I never got 8 running well on a 68k Mac. I'm also interested in any differences in the point releases on a 68k Mac.

 

[Kamshaft]

Today, I got Wifi working on the BlueSCSI, other than setting up a simple FTP server on my Windows machine, browsing anything on the modern web ain't happening.. LOL

 

[Mustermann]

Is Wifi working with closed lid?

 

[Kamshaft]

Is Wifi working with closed lid?

Good question. I'll check.....

Lid on.... Nope. lol

Maybe I can put an external antenna.....

 

[Mustermann]

External antenna at a raspberry pico W seems to be possible but not easy.

 

[stynx]

Hehe, I know
I just recently made that card with a proper 475 footprint. The other variants I built so far don't fit nicely into the 475.
I can't say much about the performance gain other than it's faster than without it. Any specific benchmark you're looking for?
Speedometer won't show any difference at all because the benchmarks it runs actually fit into the CPUs own cache except video score goes up a bit. Daystar PowerDemo (which was made specifically made to show off the gains by external cache) shows a 20-40% increase depending on the benchmark.

L2 cache seems to have a big impact on the MacBench 3.0 CPU benchmark, at least for PPC Macs. The results of the FPU benchmark are not as dependent on a big L2 cache, it seems. The Speedometer CPU benchmark seems to either not favor cache at all or it runs completely in L1 cache.

A 475 cache-card would be nice...