• Updated 2023-07-12: Hello, Guest! Welcome back, and be sure to check out this follow-up post about our outage a week or so ago.

PowerWave weirdness | picky about video card

ppcoutlaw

Active member
I remembered something about the powerwave that I read on LEM. I had been looking for info as I own four power computing computers, powercenter, power tower pro, power tower, and I think another power center. Three g3/g4 upgrades, and one has a 210 604e in it. Anyways, the powerwave is a 604 machine, and your 225 CPU daughter card is a 604e. The powerwave may not be able to correctly run the 604e. Kinda like my 8500. It has a g3 sonnet, designed to deal with what most of the PCI powermacs require, but I dont know that it has the ability to run the 604e. I have a 180 mp w that i have thought about testing, but i need a bit more research before i try it.

 

jessenator

Well-known member
don't know that it has the ability to run the 604e
Because the PowerWave was based on the original Tsunami design? So the 604e 85/9500 used a revised design then?

and *boom*, all of a sudden the computer is recognizing all the startup key combos.
Hooked it up and tried last night—it could reset the PRAM like nothing was wrong to begin with… PCC used to brag that you didn't "have to pay extra for the keyboard" and I guess there's a reason for that!

I was talking it up in my head, with it's wired connection plus TWO ADB ports for peripherals, and the echoes of the EKII on the sides, full extended layout with F-keys and 10-key. But yeah, I can't even move, run, and strafe at the same time in DOOM with the PCC keyboard :/  I need an Apple Design Keyboard to test against, as it was more of the contemporary around PCC's rise to prominence, right?

I wonder, since I'm not really going to  be using the Mach64 elsewhere, if I should just plop both cards in there for the unlikely event that something goes awry… I still want to track down anything documented about the more modern PCI cards jacking up these 1st gen PCI designs.

 
Last edited by a moderator:

ppcoutlaw

Active member
With regards to the 604e running on the 85/9500, I can't really say. I believe there is a thread on the board that shows a crossover 8500 prototype running a 604e. It could be a revision change, as they went forward into the next model line, to test out CPU hardware. That board may or may not be modified. Apple always held onto some specs, and didn't let the market know what else their products could do. Like the 1gb limit on later emacs actually being 2gb. Maybe the same for Power Computing?

 

ppcoutlaw

Active member
I found the thread, it is an 8500 board in a desktop case. The thread is "Odd Power Macintosh 7xxx? Setup" from hyperneogeo. The CPU is an overclocked 604ev, not 604e. I should have gone looking first before blabbing. But that raises a couple of questions. 1. Can the 8500 actually run an ev, or was that board modded like the CPU? And 2. Is there anyone on 68kmla who knows any power computing secrets they are willing to spill...i mean, share?

 
Last edited by a moderator:

ppcoutlaw

Active member
Hey Jessenator, do you have a newer model of powercomputing computer that you could test the CPU in? I neglected to consider that the card itself could have an issue. A power tower pro, i think, might be useful in checking CPU functionality. Some time back i ran across a small list of CPU compatibility between the 604 powers Macs. I didnt save it at the time. I dont know if there is such a list for the clones.

 

ppcoutlaw

Active member
That is a nice one. Good speed and all. However, as with three of my Power Computing machines, there could be a G3 upgrade hiding inside. Addendum. The 8500-180 has a 604e, so I dont think I would have a problem trying out my 180mp. It shouldn't be a problem for a 604e in your machine. But I have read that the PowerComputing CPU cards can't be put in a mac, but the apple CPU cards can run in the Powercomputing machines. Maybe you can find a cheap 180/604e to try out and see if you can get some success there.

 

jessenator

Well-known member
Interestingly enough, I started to catalog each of the CPU daughter cards in hopes of IDing ebay listings that dont show or are missing the speed sticker/decal, but they didn't seem to differentiate by clock speed. This was all done on a whim, so I haven't put my cards in this list yet—these were complied from ebay.

mPDN9qr.jpg.41c68cc484517049e5ac91cd7305ada1.jpg
Also on a whim, I bought a $7 card (the one at the bottom), and it was missing its stickers, so once I get it I'll put it in.

If everymac is correct, then the PCC 604/180 is the fastest non-e 604. I don't have any other pci, pre-G3 power Macs but I'm sure you're correct in that the PCC cards could NOT be used in genuine Apple motherboards.

Eventually I want to add the Apple cards to this list as well.

 

ppcoutlaw

Active member
My 8500 started life as a 120mhz machine. I was able to get it from well known Seattle computer recycler REPC for a superb price. I fired it up and discovered it is sporting a g3 sonnet at 550mhz. It will be my appleip server as soon as I dig one of my 50gb scsis out of its hidey hole and clone the 1gb SCSI it has. It is relative identical in speed and introduction time frame to the power wave. I suspect that your machine must be equally as capable. That is why I was surprised the 225 wasn't working.

Your list will be invaluable at some point. It would be a nice reference chart for those of us who work on these typenof machines. When I have the chance, I will see if I can add any of my processors to that list...umax s900-233, 250 with coprocessor, the aforementioned power computing 210, the apple 180mp, i think that is it. Probably one or two in storage....

 
Last edited by a moderator:

jessenator

Well-known member
Grrr. I keep on discovering new information with each post... Apparently the PowerTower (not pro) 180 (not 180e) runs at a 60 MHz bus speed? even though the Tsunami design spec didn't?

The 85/9500 doesn't appear to run at 60. Ah, but the original PowerTower was Catalyst-based?? and only PCC made their Catalyst boards run at 60? My brain right now... I need some sleep.

Well, then, I guess the 604/150 from another PowerWave is the non-604e ceiling, if MacInfo.de and Everymac are to be believed.

 

MOS8_030

Well-known member
Grrr. I keep on discovering new information with each post... Apparently the PowerTower (not pro) 180 (not 180e) runs at a 60 MHz bus speed? even though the Tsunami design spec didn't?

The 85/9500 doesn't appear to run at 60. Ah, but the original PowerTower was Catalyst-based?? and only PCC made their Catalyst boards run at 60? My brain right now... I need some sleep.

Well, then, I guess the 604/150 from another PowerWave is the non-604e ceiling, if MacInfo.de and Everymac are to be believed.
Yepper, several of the Power systems ran at faster bus speeds. This was a big selling point, they were faster and cheaper than comparable Apple products.

They were totally eating into Apple's high-end sales and thus had to be terminated...

 

jessenator

Well-known member
Interesting(?) new development: with all the PowerLogix system goodies disabled and with the 604/132 CPU card Sonnet Metronome freezes the system upon opening.

MacBench specs will read it just fine, but thought that was a strange quirk. On my 4400/Starmax 5500 boards Metronome runs with whatever's in the system. Could be that that particular daughter card is wrecking havoc with Metronome, causing it to crash?  

 

trag

Well-known member
PowerWave should work fine with any 225/45 card that will fit in the slot.  

The PowerWave is essentially a (take your pick):

1) 7500/8500 with the video circuitry removed.

2)  9500 with the second PCI bus removed.

It might not (or might) work with a 210/60 or 240/60 card.   I've had the PowerTower Pro and Umax S900 up to 62MHz bus speed, provided the CPU card properly sets the CLKID pins on the CPU card.

I think I covered the 8500/8600 stuff in another thread when you were getting the PW working, but in brief:

Three revisions:

PowerMac 8500/9500, 8600/9600. 8600 Enhanced/9600 Enhanced.

The "enhanced" models are the ones with the "Kansas" motherboard and the "Mach V" CPU card which  uses the PPC604ev chip.

Apple's CPU cards 250MHz or faster were PPC604EV, Mach V, cards which only work in the two late, enhanced models.

Some third party brands sold 250 MHz PPC604e cards which will work in all the non-enhanced models, both Apple and clone.

All of the non-Mach V Apple cards, PCC cards and Umax CPU cards should work in the  PowerWave and PowerTower Pro.

Except for the PowerTower Pro and PowerWave, Apple CPU cards will not work in the Power Computing clones.

Also, PCC cards made for the PTP and the PW may not work in the other PCC clones.   The other PCC clones are based off of the Catalyst chip set and need a couple of CPU signals which aren't brought out to card pins in the Apple cards (and the PTP/PW cards?).

But, that doesn't matter, because you're testing a PowerWave.   That 225/45 card should work, if it is functional.

If  it is a Umax card, then there are three (four?) jumpers that need to be installed at the top, because Umax had this weird double-processor scheme with a ribbon cable between the CPU cards and when the CPUs were used singly, some of the cable pins needed to be shorted with jumpers.

Hopefully you're testing with no cache installed.    Those can cause issues.  On the other hand, I've read of cases where things didn't work well until a cache was installed.    I guess the extra load reduced ringing on the bus or something.

Finally, what it really sounds like is that your power supply is marginal and when the load gets above a certain point everything goes flakey.

Do you have another ATX power supply on hand you could try?

I've also seen the PCI bus on a PowerTower Pro fail because the solder joints on the PCI arbiter chip went bad.    PCI arbiter is a little square PLCC chip, with 20 or 25 pins.  I don't remember the markings at the moment.

Oh, and, having an older 132 card work and a newer 225 card not work, I have seen when there were problems with the 3.3V supply on the logic board....    Just remembered that one from about 15 years ago.

You might take a voltmeter to the 3.3V supply, although if it's something like worn out bypass caps a voltmeter might not tell the story.   In my case, one of the chips had developed an internal short that was loading 3.3V to one of the other rails, or ground, can't remember which.  It wasn't enough to shut down the PS, but it made anything that wanted 3.3V not work.    But it could just be that the old PCC power supply isn't supplying good 3.3V, in which case a repalcement ATX PS will solve that.

 
Last edited by a moderator:

jessenator

Well-known member
Thanks as always for the great reply, trag!

I think I remember you talking to the point of the Enhanced CPU cards and what the PW could handle. I took another look and the 225/45 card is definitely PCC. 

Do you have another ATX power supply on hand you could try?
I sadly borked all of the spares (admittedly not great specimens) I had when coming up with a solution to my 4400's power supply issues. You're probably right about the PSU—It's the original, and was in a smoker's environment all its life. Working this long is a feat, even for a SeaSonic perhaps. I can scrounge one up fairly easily I reckon.

I'll also check the PCI riser (if that chip is on the riser of a PW anyway). I'm a knucklehead sometimes: do you mean check the 3.3V line of the PSU or somewhere on the board? If the PSU replacement would solve it, I guess it's moot, but I didn't know if something else on the board would need checked as well.

Interesting note too about the cache. I just barely got it, and installed after testing and putting the 132 card back in for now. I'll definitely remove it when testing this mystery PCC card.

Edit: Just booted with the mystery card. Turns out I may have gotten a PW original card; 604/120. the case badge is accurate now at least :lol:  

 
Last edited by a moderator:

jessenator

Well-known member
I'll reinstall TechTool tomorrow and try it again with the PSU I just got. See how it goes.

I knew the cards got smaller as they went on, and the upgrade cards got very small indeed, but I only have seen them with the metal spacer on top to keep them secured. Didn't know they went all the way to the crossmember bar in the PW like this...

Also tried MacBench and got the missing "with cache" scores. The disk score... I tried that several times with the 225 and each time it got worse :lol:  weird.



 
Last edited by a moderator:

jessenator

Well-known member
Well, I ran TechTool Pro last night— I get the same Mathematics test failure with the 225/45 card. I also got a couple of new test fails, which I'm certain is because my SCSI drive is almost as loud as the case fan. Maybe I'll find another 604e card and see if it's just the card I have, or if it's something else.

 

trag

Well-known member
The problem could be caused by many things.   Another one that I've seen in my experience, at least on teh PPC601, was that getting heat sink grease on the pins caused an issue with the FPU.  I don't really understand how the short that might cause could affect just the FPU, but it did.   For the PPC601, the pins are around the outer edge of the chip, so that's where the heat sink grease ends up if one applies too much.   On a PPC604e, I think it would take a little work to get excess grease under the chip and amongst the BGA balls.

Probably not the problem in your case, but you know what, the heat sink grease could be worn out and the chip overheating.   As long as you're gentle, it can't hurt to pull the heat sink, clean the old grease and apply new.    And do a quick inspection to see if any has spilled over the edges of the chip.

 

jessenator

Well-known member
Not a bad idea even for due diligence. Stuff's probably nasty by now. I have some non-conductive thermal paste that I've used on a lot of things: my zip-tied PowerLogix G3 card for one, and I think even my PC Compatibility Card, which for a P100 is a bit overkill. The viper chip on the back side puts out way more heat…

I wonder if I could set up some Excel sheet to really tax the FPU and see what happens. Maybe doing something in Photoshop would be easier :lol:  

 

Franklinstein

Well-known member
Just a few things to throw in here:

The issue with the S900 was the fact that, unlike the Tsunami-based boards that had two independent Bandit PCI controllers (one provided three PCI slots plus the PCI interface to Grand Central and other onboard devices, other Bandit provided other three PCI slots), the S900 had only one Bandit PCI controller with a DEC 21152 PCI-PCI bridge chip. The S900's Bandit provided the top three PCI slots while the lower three were on the DEC bridge chip, which is why some cards won't show up properly in Apple System Profiler if they're in the lower slots. Some cards won't play nice behind the DEC bridge, either because the card or some software is flummoxed by the bridge for some reason (maybe it expects to see a hardware ID or to talk directly from Bandit to device?). This is also a problem in Tanzania clones with 5-slot PCI risers for the same reason. But that shouldn't be a concern with any PCC designs.

I'm pretty sure the last QFP-style 604 was 150MHz. They were built in PGA form up to about 180MHz, then 604e up to around 250MHz. Then the Mach V was used in a handful of cards, some with a high-speed inline cache (the 604ev Mach V supported up to a 100MHz bus speed, but no Mac chipset would run at that speed for another few years, long after the Mach V was superseded by the 750 and 7400).

I have a PCC tower of some sort (PowerWave, I think) and it had a bunch of issues with booting and stability. First it wouldn't run with the original L2 cache installed. Then it was super unstable when it did boot and it turned out the processor card was faulty. So I threw in a midrange G3 upgrade which eliminated both problems and now it works fine. I don't remember which video card it has though.

I think the bulk of the issue is, as has been covered, the Rage 128 being a pain in beige Macs. Maybe there's an update to flash to it?

 
Top