Another IIci ROM hack

[bigmessowires]

At this point the SIMM tests perfect electrically but still won't boot.

Rats. How are you doing the electronic testing-- with a multimeter, checking for shorts or opens everywhere? It sounds like you already tested pretty thoroughly for short circuits everywhere, but maybe there's an open (broken connection) that you missed? Or maybe one of the chips isn't socketed firmly, or there's some corrosion on a pin preventing it from making good contact with the socket?

I've forgotten now why you used SMD sockets instead of through hole ones. It sounds like the SMD ones are a pain in the butt!

It would be tedious, but you might try manually reading out some EPROM values by connecting appropriate high and low voltages to your SIMM connector pads with alligator clips or something. That way you'd be reading it the same way the Mac does, from the same contacts.

[Trash80toHP_Mini]

Addendum to BMOW's suggestion:

If you've got a ROM SIMM Socket, one of the larger ones would actually be better, build yourself a testing setup/prototype of the read portion of the programmer circuit in the same step. Wire wrap on a stock project/prototyping PCB would be my medium of choice for such a rig.

BTW: 15 pages & counting! Nice initial post Mr. very much EX-noob around this nook-o-da-web!

[dougg3]

I'm testing using a multimeter to be sure that every connection that is supposed to be connected is connected -- so that except for the data lines, the corresponding pins on each socket are connected together, and that there are no shorts between nearby socket pins. I also have checked that each SIMM contact is connected to the correct socket pins. So yeah, I'm looking for both opens and shorts.

It could definitely be a loose socket connection, PLCC sockets are pretty flimsy.

I used SMD sockets because they are physically smaller and because the through holes eat up a bunch of routing space on the bottom of the board. The SMD sockets aren't terrible to work with if I cut out the bottom plastic plate, so I'm going to try that.

I'm just not equipped to read values from the board yet, but I really like the idea! I don't have enough alligator clips and stuff to put a 0 for example on every address line and CS and OE. Likewise I don't have any SIMM sockets yet either

Thanks for the ideas guys! Today is a fresh new day so hopefully I can get to the bottom of it.

Edit: I just found another weird short that I hadn't found yesterday. It's not between nearby pins, but that's clearly because it's between a pin and a nearby via. I can't see it, but that HAS to be where it is. This board may be fixable after all!

[dougg3]

Success! It was that last short I just found, and the SIMM works now. That makes me happy because I know that the rev 2 board works now.

I just made another SIMM which I did more carefully, and it had one short. I fixed that and it began to work, but halfway through the crazy long startup chime, it went all nuts. It didn't work after that. I then found ANOTHER short which was NOT there before, so I'm 99% sure that the solder paste is to blame. I took the board and very thoroughly washed it with cold water, and then the short disappeared. I'm still letting it dry now so I don't know if it will work after that, but weird, huh?

I used less paste this time which helped, but I think what's happening is some extra unheated paste is getting left behind and intermittently causing problems.

Until I have a reflow oven working that can heat up ALL the paste, I will probably return to cutting out the bottoms of the sockets and using regular solder. That combination has worked for all other boards I have made with no problems. Anyway, success!

[dougg3]

And now I've been hit with another unexpected curveball.

The IIci motherboard I've been using to test my SIMMs with is no longer functional. When I hit the power button on either the back or the keyboard, nothing happens, nada. It's not the power supply -- the power supply boots my hacked IIci motherboard fine, and it's definitely putting out the 5V trickle voltage.

This is a motherboard which I have not recapped or cleaned at all, so it's possible that's all it needs, but it's funny that it just started doing it now that I'm messing with the rev. 2 SIMM. I'm not even going to touch my hacked motherboard with these SIMMs until I figure out what's going on with my now non-functional motherboard.

*sigh*

P.S. I was thinking, my SIMM programmer board should be able to test for shorts...that will be very handy. It won't be able to tell me which chip/socket it's on (except for the data pins), but that will be very helpful for testing.

[olePigeon]

Thanks for keeping us posted. I'm anxiously waiting.

Out of curiosity,did the LED eyes light up when the SIMM is inserted? Or are you testing it without the LEDs? I saw the picture you made and it looks awesome.

[dougg3]

Yep, the LEDs light up as soon as power is turned on (well, they did until my motherboard gave up the ghost). They look great!

[Arbee]

Thanks Arbee! The ASC emulation in MESS is fantastic! Awesome work! Without MESS there's no way I would have figured it all out. How did you figure out what all the bits do? Was it basically a matter of disassembling the OS/ROM and looking at how the registers are accessed?

I started with the Linux and BSD drivers, which as you know don't actually do much. Then I spent a lot of time in IDA Pro and the MESS debugger stepping through stuff and scratching my head. The Sound Manager relies on the interrupts instead of polling like the boot beep and that was "interesting" again, although there were some useful clues in Guide to the Macintosh Family Hardware, 2nd edition. Mini vMac also had a partial ASC that was in the ballpark for the wavetable mode but not quite right, and that helped to jump start that aspect.

Got the Quadra 700 to boot into Finder in MESS this weekend after a few months of hitting my head against various aspects, but there's always more work to do

[olePigeon]

Yep, the LEDs light up as soon as power is turned on (well, they did until my motherboard gave up the ghost). They look great!

I think it died from too much awesomeness.

[Trash80toHP_Mini]

Have you tested the Power Cord you were using when the IIci "Stopped Working?" That's one of the nastiest failures and one that often goes untested. It's a pain to unhook one at the plug end, so it often goes untested and the computer end is plugged right into the test PSU w/o testing. One of my IDE<->USB2/SATA adapters came with a bad power cable and I figured the whole thing was bad, when I returned it . . .
. . . they tested it and found the noob boob for me . . .

[Dennis Nedry]

It's probably the capacitors on the 'dead' IIci board. I bet they were ready to go and you've been using it more than usual. The soft power circuit is fairly simple, totally separate from any ROM-related stuff, and a schematic can be found on Gamba's site:



As you can see, even if you blew up the processor, the soft power would still kick on the power supply, so the problem is somewhere on this schematic, most likely the electrolytic capacitors. I'm actually surprised if ANY IIci still works with stock capacitors after this long. None of mine did.

[dougg3]

Got the Quadra 700 to boot into Finder in MESS this weekend after a few months of hitting my head against various aspects, but there's always more work to do

Cool! And as for how you did the ASC stuff, awesome! I started looking at the Sound Manager drivers a little bit, but the MESS code was so much simpler to follow. I think I found an interesting string resource in the System file describing the ASC driver, I might have to pull that one out if I can find it again.

I think it died from too much awesomeness.

Haha, good point! That has to be what happened

Have you tested the Power Cord you were using when the IIci "Stopped Working?" That's one of the nastiest failures and one that often goes untested. It's a pain to unhook one at the plug end, so it often goes untested and the computer end is plugged right into the test PSU w/o testing. One of my IDE<->USB2/SATA adapters came with a bad power cable and I figured the whole thing was bad, when I returned it . . .
. . . they tested it and found the noob boob for me . . .

Thanks for the suggestion jt! Unfortunately that same power cord booted the other IIci fine. Oh man that would be annoying to discover that the power cord was the problem all along

As you can see, even if you blew up the processor, the soft power would still kick on the power supply, so the problem is somewhere on this schematic, most likely the electrolytic capacitors. I'm actually surprised if ANY IIci still works with stock capacitors after this long. None of mine did.

Yeah, it was pretty surprising it was still working. I haven't checked to see if there are leaking caps but there must be. In fact I think this machine was having some kind of weird problem where soft power down would result in an immediate reboot, but that problem mysteriously disappeared. I'll take a look at it later, thank you for the info!

Do you happen to have any ideas on repairing a IIci power supply? My other machine has an Astec power supply that bit the dust, so they are both sharing this GE one. I saw the diode fix and haven't been able to try it, but the "flip a power strip off and on" trick wouldn't work for me and I'm not getting a 5V trickle voltage out of it at all.

[olePigeon]

My IIci worked fine for the past few years. I didn't even think to check for leaky caps until recently. I then recapped my IIci. Still works great. Now all I need is a Quadra 700 power supply to replace the one in my IIci. I don't think the original power supply has enough power to run my accelerator.

[dougg3]

LOL, I definitely need to re-cap and clean that "bad" logic board. I noticed the board is filthy near the soft power circuit (underneath the power supply), some from dust and some from leaky caps. I blew compressed air all over the soft power circuit and the IIci is alive again. OK for a temporary fix but I will re-cap and clean it soon.

Thanks again for the info Dennis Nedry! When you said that it would still at least start the power supply if I had blown the CPU for example, that made me feel a lot better!

[Dennis Nedry]

No problem Doug! I've never found any exact science with power supplies. If you find clues on the web, definitely try those first. Otherwise if you're really lucky you might find a big chunk out of something or a bulged out capacitor, but most of the time there won't be a visible sign. In that case, I guess I'd start by pulling out a voltmeter and trying to trace back from the trickle line until I found some juice.

I have fixed a couple of IIsi power supplies that had two smallish electrolytics that went bad with no visible signs. They shorted the trickle to ground I believe. When poking around, it helps to think about, what should this voltage be at this point, and what would make it wrong. Is there a series resistor behind this point, if so, test both sides of it. If it's good on one side and not the other, poke around the bad side for things that could be shorting it. Just basic experimentation like that, unsolder things, test if possible, solder on similar parts if you can't test, etc. Knowledge to gain and not much to lose.

You can even take a pretty good bet a lot of times and unsolder all of the capacitors one by one, testing for a short. If they don't test as shorted, maybe they're blown open, but that's not usually likely to make a power supply totally dead.

Heck, try plugging AC into the power supply just out on its own and feeding it external +5V to pin 9, ground to 5, 6, or 7, defeating the purpose of the trickle voltage. Whatever happens or doesn't happen could give you some clues. Maybe it's just the trickle, maybe it's deeper if it still doesn't wake up.

[dougg3]

Thanks for all the info Dennis Nedry! I tried your last suggestion since it didn't involve any case opening and...whoa...

So I applied +5V from my bench PSU to the /PFW pin (9) (and connected the grounds together), as you suggested.

I heard a relay click and the fan started spinning slower than usual. I measured the outputs and they were jumping wildly from where they are supposed to be (12V was jumping between 6 to 8 V, -12V was jumping between -6V to -8V, and 5V was jumping between 0 and 4V). It was drawing about 30 mA from my bench supply.

Then, I removed my bench supply and started playing around looking at the pins. I checked the +5V trickle pin (10), and the +5V trickle was there again! I didn't bother checking before I had started playing around with it though, so it's possible the trickle was already there, but I know at one point a while ago it was completely gone. So I tried telling the power supply to turn itself on by shorting /PFW and the +5V trickle, and it did the same thing it does with external power -- slow fan and too-low voltages.

Anyway, naturally after seeing that, I put the power supply into my other IIci, and suddenly, IT WORKS!

WTF is going on here? Did applying external power "fix" the +5V trickle temporarily or something? Did I accidentally make a loose solder joint good again by moving it around? Why does jump-starting the power supply like I did not allow the voltages to go all the way up? This is confusing and exciting because both my power supplies work again! Thanks for the awesome suggestions! I still have no idea what is going on here...haha

[olePigeon]

The computer gods are pleased with your work, and rewarded you with a working power supply.

[Bunsen]

Woah. Having finally taken the time to read and comprehend this thread properly, all I can say is -

What an epic adventure - and by the sound of it, this is just the beginning. dougg3, and everyone involved, take a bow.

[Bunsen]

So, to check that I'm up to speed-

I gather that the barrier to using the ROM slot as storage or general-purpose expansion is the absence of any write lines?

Also, Trash, it's worth bearing in mind that a USB *Device* port/chipset/stack and a USB *Host* port are not the same thing. The one on a computer is a host, the one on a disk is a device. There is a protocol that can do both, called USB On The Go (OTG) - but in general, an IC that acts as a device will not be able to mount (host) another device. Pardon me if I'm pointing out the bleedin' obvious here, but from your earlier posts, you seemed to be making that confusion.

Meanwhile, this tidbit caught my eye, and perhaps warrants further investigation:

How the Classic's ROM disk image is implemented

and ROM Disk driver.

Sure, still read-only, but perhaps a convenient, OS-friendly way of making use of spare ROM capacity?
*cough* SD card?

One small suggestion for the programming board- USB to RS232 or TTL cables are readily available, cheap, and infinitely useful for further hackery, so, in order to keep the custom board as simple (ergo, cheap) as possible, why not have TTL or serial headers on the board, and let the end user buy a USB converter? Some of us will no doubt have one lying about, and others (like me) have been meaning to get one anyway.

That also leaves the door open for maniacs who might want to program their ROM from a Mac serial port

[Dennis Nedry]

I'm glad to hear that your power supply is working. Have you tested the voltages on the logic board with this power supply running? You can test +5 and +12 easily in the hard drive power connector. I'm a little bit concerned that the voltages were unstable.

[techknight]

Well if your voltages are "bouncing" that means the power supply is cycling, which means there is an overload on one of the rails somewhere. Unplug the power supply from the mainboard, measure each rail point to ground in ohms. take a note of the measurements. Go to a good board, do the same thing. what you will likely find is one or more of the rails is reading lower resistance = leakage/short somewhere. have to track it down.

[dougg3]

The computer gods are pleased with your work, and rewarded you with a working power supply.

That's about the only explanation that makes any sense!

What an epic adventure - and by the sound of it, this is just the beginning. dougg3, and everyone involved, take a bow.

Thank Bunsen! I still appreciate all the info/feedback/help everyone has given me on this project, it has been a TON of fun even through the few minor roadblocks I keep running into! It's also helping me get into hardware stuff which has always been fascinating to me.

I gather that the barrier to using the ROM slot as storage or general-purpose expansion is the absence of any write lines?

Exactly -- I suppose we could tie into a write enable line somewhere (maybe?) and get read/write capabilities, but I don't know where to look for that or if it's even feasible -- maybe someone with more hardware knowledge would know?

How the Classic's ROM disk image is implemented

and ROM Disk driver.

Sure, still read-only, but perhaps a convenient, OS-friendly way of making use of spare ROM capacity?
*cough* SD card?

As long as a microcontroller/FPGA or something could translate between ROM requests and SD card accesses, but the fast timing there would probably cause problems. BMOW probably knows more about this stuff and I'm suspecting that the image of what's stored on the SD card would have to be loaded into RAM that's connected to the IIci's SIMM pins--I don't think it could read off the SD card in real time fast enough (at least that's my first guess).

One small suggestion for the programming board- USB to RS232 or TTL cables are readily available, cheap, and infinitely useful for further hackery, so, in order to keep the custom board as simple (ergo, cheap) as possible, why not have TTL or serial headers on the board, and let the end user buy a USB converter? Some of us will no doubt have one lying about, and others (like me) have been meaning to get one anyway.

That also leaves the door open for maniacs who might want to program their ROM from a Mac serial port

Good idea! It also makes the board design easier for me Since I'm going to be using a 5V microcontroller, it would be easier if I use a 5V serial port instead of 3.3V. Is that going to cause a problem? I think a lot of those types of cables are designed for 3.3V if I'm not mistaken.

I'm glad to hear that your power supply is working. Have you tested the voltages on the logic board with this power supply running? You can test +5 and +12 easily in the hard drive power connector. I'm a little bit concerned that the voltages were unstable.

Me too I just did that...thanks for the idea.

+5V reads as 5.10 V
+12V reads as 11.89 V
-12V reads as -12.18 V

Those seem ok...maybe the IIci power supply doesn't like to be started without a load connected to it?

Well if your voltages are "bouncing" that means the power supply is cycling, which means there is an overload on one of the rails somewhere. Unplug the power supply from the mainboard, measure each rail point to ground in ohms. take a note of the measurements. Go to a good board, do the same thing. what you will likely find is one or more of the rails is reading lower resistance = leakage/short somewhere. have to track it down.

Thanks for the info techknight! The weird thing is that the power supply was doing that outside of the computer with no motherboard connected to it. With a motherboard connected to it, it behaves fine. Could starting the power supply with no load cause it to choke like that and keep cycling?


SIMM Update:

The second SIMM I have been having trouble with is now functional. It had nothing to do with the solder paste -- again, it was a case of nearby SIMM contacts getting shorted to each other. Possibly this time by me scraping the contacts with my probe as I was testing continuity. Something is different about the SIMM contacts than last time -- maybe they didn't coat them with solder in the revision 1 SIMMs? I'm thinking I can solve a lot of this by sucking off the solder with my desoldering gun and maybe some wick (although the wick is expensive!) Anyway, like I said earlier, I still think I will switch back to using ordinary solder and cutting out the bottoms of the sockets just to be safe. So anyway, it's coming along and I will soon be able to start selling them once I'm comfortable with them! (With the problems I've been having I want to make sure they will be reliable before I sell to anyone).

[Trash80toHP_Mini]

How about just gluing the "bottoms" back in as spacers once you're done soldering?

Also, Trash, it's worth bearing in mind that a USB . . .

Thanks for the primer . . . I've deliberately avoided loading the USB Spec into my noggin' as yet . . .
. . . it's already bogged down with hardware hackin' mania.

p.s. We really need a USB<->Classic/NuBus Architecture Brainstorming Thread . . .
p.p.s. I really miss the maniac . . .

[olePigeon]

With the ability to make custom boot ROMs, does that make a USB NuBUS card more likely? Or is it an architectural barrier between the NuBUS design and sending USB data over it?

[Trash80toHP_Mini]

It's more a question of driver obsolescence under any NuBus Architecture OS, only PCI Architecture 'Books & Macs have the firmware/drivers available to handle USB 1 under OS9, so all 68k and, all or almost every 603e Mac or their owners are/were just SOL. Why would Apple retrofit older machines to do current tech (ancient tech by now) when they only wanted to sell new PCI Boxen?

A SCSI<->USB adapter is entirely possible, as SCSI is a general purpose interface, an open standard (which Apple botched with full documentation available as is USB. However one would need to do it for kicks and have some serious development chops to grab for the BRASS RING of NuBus Architecture Hackery.