Dead Performa 475

[sirlemonhead]

Hi,

I picked up a really cheap Performa 475 on ebay as the board looked OK from what I could see of the screenshots.

I'm an Apple newbie so even though the auction stated chimes of doom, I figured I might be able to get the board working correctly with little effort - not the case unfortunately.

I have no SIMM installed and no PRAM battery, but I have two VRAM sticks installed. The machine turns on and does the initial BONG noise, then the 8 notes - 4 ascending and then 3 descending which is the chimes of doom I believe? I don't get any video output. If I do the flick on/off/on again trick, I still don't get video ram. I have a new PRAM battery on the way.

The machine seems to have been kept in a damp/humid environment as the RF shield in the case was a little bit rusty, and the VRAM and RAM slots had a handful of corroded pins (somewhat greenish corrosion I think?). The included SIMM actually had most of the gold pins erroded on it's edge connector. The pins of the soldered in RAM chips were a small bit corroded and mostly cleaned up OK.

Some traces near the VRAM and Ram underneath look dark in places but I can't be sure if they're gone or not - I can't get any beep from my multimeter in continuity test when I probe either end of a dark trace using it's via holes. Though saying that, I can't get a beep either from doing similar on a perfectly good via.

PSU appears perfect. It's a Dyna Comp and the capacitors look good and each voltage rail is spot on. The capacitors on the board look perfect - the board was dusty and there was no evidence of cap leakage. Pads appear shiny. I haven't recapped yet but have caps on the way.

I found this thread - https://68kmla.org/forums/index.php?/topic/24711-finally-back-with-some-news-475/&tab=comments#comment-260992, and started checking the pins on each of the on-board RAM chips. I can successfully get continuity between pins A0-A09, RAS, CAS, WE back over to the MEMCjr chip. VCC reads 5v on each chip.

What I can't seem to find is where the data input/output pins are supposed to go - Do these also connect to the MEMCjr chip? or to the CPU? If someone could advise on that, I could start trying to find out of those are OK...

Should I wait and recap the board, or should I keep probing?

 

[BadGoldEagle]

Hi @sirlemonhead and welcome to the 68kmla and to the Mac world.

The LC475 / Quadra 605 / Performa 475 or 476 need a fresh PRAM battery in order to display video. You've got one on the way so that's good.

The chimes of death can mean a lot of things, but usually bad RAM is what's causing it. It has onboard RAM, you don't need to add extra RAM SIMMs. 

I would get a toothbrush and clean the board with soapy water and then IPA. Capacitor goo is conductive and could be shorting some components. I know I'll probably sound like a douche for saying this, but any late 80s - 90s Mac with SMD electrolytic caps with startup issues should be recapped before any further diagnosing is done. Hopefully it's either just a short or a bad chip. But even if it's just RAM, there's still a strong chance the caps will prevent it from booting. You'll still have to do it sooner or later. For those reasons, I would still proceed with the recap.

As far as rust and bad traces are concerned, pictures would be great! 

Edit: The RAM chips might be a little difficult to get hold of. Octopart reference 5 sellers but I don't know if they still stock those parts...

https://octopart.com/km44c1000bj-7-samsung-2254199

 

[Brett B.]

I would start by giving the board a very thorough wash and rinse and dry it well, and see if it comes to life.  If not, recap, paying attention to cleaning under the caps you remove.   It should boot with your current configuration of RAM and battery if you do the flick trick.

 

[sirlemonhead]

PRAM battery has been delayed - I forgot those can't go by air, I think? So I've got the caps but the PRAM battery should hopefully be with me by Friday.

I guess I was putting off recapping because I figured since the current ones are OK looking that it might be a waste of money/time to recap it. I'll do it now though so we can eliminate it as a source of potential issues.

I'll try get some pictures of the board. Maybe I could stick it on a flatbed scanner? My Dad has one of those I could use - I find photos with my phone always come out pretty dark.

The board currently has NEC chips - these specifically https://randomsonicnet.org/workshop/42(S)4400_E70.pdf

I have quite a few spare SIMMs though so I could probably salvage chips from those instead? Can the onboard slots only address a max of 4 megs?

Also, just wondering - the capacitors are probably decoupling caps, right? Could the board be tested without them present?

PS Thanks for the welcome

 

[trag]

You can probably steal the RAM chips from any eight or nine chip 4MB 72 pin SIMM.   The pinout for 1M X 4 DRAM is a JEDEC standard, IIRC.

Caps on a Macintosh logic board are typically for bypass, for sound and for power-on circuitry.  The 475 doesn't have soft power-on so probably no caps involved in power-on.    So operating without caps would probably mean no or bad sound and no bypass.   I'm not sure if I would expect the board to work or not absent bypass.   Of course, the small ceramic bypass caps are still in place and the power supply should have some late stage smoothing caps in it, so probably...

Can bigger memory be put in.   Bbraun (?) looked into the software limitations of several machines some time ago.   I don't know what, if anything he found about that particular question.   From a hardware point of view, it depends on whether there are larger capacity memory chips with compatible pinout, and then whether the memory controller implements the extra address line that would be needed.  

The 4MB on board is 1M wide, which is 20 address bits.   Addresses are multiplexed in DRAM, so that's 10 or 11 address wires.   Ten is the minimum needed, but there might be 11 if 11 X 9 addressing or 9 X 11 addressing was a thing.    Nope, datasheet shows 10 address lines with no NC pins for more.   So 10 X 10 it was.  That leaves nothing for larger memory.   It's still possible with a more complex hack, but not by simply soldering down bigger chips in the same place.

 

[sirlemonhead]

So I got the caps off.... unfortunately lost a pad for one of the caps near the audio chip. When lifting it up, it dropped and adhered itself back down at an awkward angle. No matter how careful I was, it wasn't to be unfortunately.

https://www.dropbox.com/s/5terj73dlq9gms2/p475_liftedpad.jpg

I might try to glue the pad back on, and it appears as though it was connected to the + of the capacitor beside it.. which would mean the + of the right capacitor was connected to the + of the bad pad capacitor.. does that make any sense??

Here's the black traces:

https://www.dropbox.com/s/h3vxov3fvv03184/p475_traces_1.jpg

https://www.dropbox.com/s/rixu7e60tnvuam2/p475_traces_2.jpg

This is interesting, as it shows blackened traces but these are the lines to the RAM address lines, I believe, and these are all showing continuity

https://www.dropbox.com/s/tka2xt3klixxbo9/p475_traces_3.jpg

So maybe it's not so hopeless..

Here's an example of the solder holes on the ROM socket, the topmost being dull grey

https://www.dropbox.com/s/w4r8it6vjfvr7oh/p475_rom_corrosion.jpg

 

[Brett B.]

Bummer... I have lifted a few pads too.  It's frustrating.  I have not had any luck gluing them back on - the heat from the soldering iron burns off super glue instantly.  May be better to try a piece of wire to where ever it goes.  I've also had success just blobbing up a big pile of solder that touches whatever is left of the pad/trace.  Probably not the best way to do it but it worked.

 

[LaPorta]

To expand on what Brett said, you can also scrape off more coating of the trace nearby, and solder a small piece of wire to the leg of the cap and the other end to the newly exposed trace.

 

[sirlemonhead]

Yeah I was going to just use the pad for some physical stability - would definitely be easier to just run a wire!

The board runs exactly the same without the big SMD caps. No video output with the PRAM battery installed.

Does the board do the chimes of doom without the VRAM present? I'm wondering if that could be at fault...

 

[sirlemonhead]

I've cleaned the board. It's now no worse or no better.

I have been now able to get continuity on all 4 data lines for each RAM chip to the PrimeTime chip, so it looks like those traces are OK! Unless the problem lies between CPU and PrimeTime..

 

[sirlemonhead]

It's alive!

My first ever hot air desoldering attempt was a 100% success. I removed the 10 onboard RAM chips. No pads lifted and with a socketed SIMM, now I have a grey screen with a movable cursor on it! no bongs of doom!

I've no HDD or floppy drive so I guess that's why the machine doesn't do much else?

 

[BadGoldEagle]

Congrats!

My first hot air desoldering attempt was on a Sony WM-D6C. It failed and I used the regular iron instead. But I did manage to solder the new IC back with the hot air station. I blame the shoddy chinese clone.

Did you replace the chips or did you just remove them?

 

[sirlemonhead]

Thanks

I just removed them. I don't have any spares and I have a bag of SIMMs hanging about.. I might just leave the onboard chips off.. unless there's any issue with that? 

I've just powered on the board again and it's now showing a floppy drive with the ? on it, so it's looking good.

I am still concerned about those slightly darkened traces - scrape off and reapply some new conformal coating?

 

[trag]

Ten chips?   Did you mean 8 chips?  I think the on-board RAM is only 8 chips.  Still, great that it is working now.

 

[sirlemonhead]

Yep 8, sorry

 

[trag]

Yep 8, sorry

Ah, good.  I was worried about what else you might have removed. 

 

[BadGoldEagle]

I didn't know it could run without the onboard RAM. Yeah, I wouldn't personally bother to solder new chips back in except if you want to really get the most out of your Quadra. But even so, 128mb instead of the full 132mb should still be more than satisfactory.

 

[paws]

I have 132 mb in the Quadra 650 I'm putting together. It takes ages to boot and I can't see myself ever using half of it, even doing audio.

Nice job on the troubleshooting here.

 

[sirlemonhead]

Yeah I'm running a 16meg SIMM at the moment which I figure will probably be fine!

I have a bootable OS 7 Iomega ZIP disk thats working for now so I can at least confirm the SCSI system seems to be OK

I've got a floppy drive on the way and I think I'll probably grab a SCSI2SD as buying a drive on ebay seems like it'd probably be a waste of time.