-
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.
Mac SE/30: Scsi problems
- Thread starter nvdeynde
- Start date
I found a much more detailed schematics for the SCSI traces after digging through the Forum.
See the 3 Jpegs at top of the page below.
http://68kmla.org/forums/viewtopic.php?f=29&t=19390
Check on page 2.
For Pin 20 of the SCSI chip there’s a trace to Pin #2 ( D19 ).
Same for Pin 21 to pin #2 ( D17 ), pin 24 to pin #2 ( D13 ) and other pins of the SCSI chip
I checked the Apple Schematics but don’t find these.
Where are these D references located : its always pin #2 but what are these ?
Can someone point me into the right direction ( or post a photo with a mark where it is ) ?
These are the only unchecked points from the whole list but I doubt these would yield anything.
It was posted by phreakout but I see he hasn't logged in anymore since January so sending a PM won't help much I'm afraid..
I have the socketed version of the SE/30 board.
See the 3 Jpegs at top of the page below.
http://68kmla.org/forums/viewtopic.php?f=29&t=19390
Check on page 2.
For Pin 20 of the SCSI chip there’s a trace to Pin #2 ( D19 ).
Same for Pin 21 to pin #2 ( D17 ), pin 24 to pin #2 ( D13 ) and other pins of the SCSI chip
I checked the Apple Schematics but don’t find these.
Where are these D references located : its always pin #2 but what are these ?
Can someone point me into the right direction ( or post a photo with a mark where it is ) ?
These are the only unchecked points from the whole list but I doubt these would yield anything.
It was posted by phreakout but I see he hasn't logged in anymore since January so sending a PM won't help much I'm afraid..
I have the socketed version of the SE/30 board.
Dnn components are usually diodes/rectifiers. They only have two pins, so the pin assignment is never greater than 2. Hunt around for either little 2 pin black surface mount components (SE/30 may be too old for those to have been around yet) or little through hole components with an orange/glassy body, although they might be matte black with a silver stripe on one end. If through-hole, they'll be about the size of a 1/8 watt resistor most likely.
My SE/30s are sitting in the attic with their capacitors removed, so I do not have hands on experience with this issue (and hope I never do, yikes) but it sounds like some of the SCSI chip signals are protected by diodes, probably to avoid reverse biasing the I/O pins.
My SE/30s are sitting in the attic with their capacitors removed, so I do not have hands on experience with this issue (and hope I never do, yikes) but it sounds like some of the SCSI chip signals are protected by diodes, probably to avoid reverse biasing the I/O pins.
pbMacGeek17
Member
Gents,
I'm late to the party, but I agree with Trag and Uniserver that this is most likely a bad via or trace (I'm betting on the former). When my SE/30 went out I created a schematic tracing each of the SCSI pins back to at least one other chip. Eventually I discovered a bad via located near one of the failed capacitors. I had to make a solder bridge to restore the connection. It may be too late for this effort (and you may have already discovered a better schematic), but I'm happy to share the schematic that I created. Unfortunately I don't log in enough to see the requests in a timely fashion. If someone else that is more active will volunteer to take it on I'll forward the drawing.
I'm late to the party, but I agree with Trag and Uniserver that this is most likely a bad via or trace (I'm betting on the former). When my SE/30 went out I created a schematic tracing each of the SCSI pins back to at least one other chip. Eventually I discovered a bad via located near one of the failed capacitors. I had to make a solder bridge to restore the connection. It may be too late for this effort (and you may have already discovered a better schematic), but I'm happy to share the schematic that I created. Unfortunately I don't log in enough to see the requests in a timely fashion. If someone else that is more active will volunteer to take it on I'll forward the drawing.
The Dxx chips are at the back of the PCB under the Ram modules. Anyway, I have checked the complete SCSI chip against the 3x Jpeg files, which are very detailed, and the conclusion is: there are no broken traces or VIA's.
For that board I'm 90% sure the SCSI chip self is faulty. It noticed in FWB toolkit 2.0.6 than eventhough it recognises the hard drive, model and brand: it lists a wrong size. FWB says it's 64MB while it's a 80MB drive.
When I verify that with a good SE/30 board: it's indeed recognised as 80MB: this only adds to my case that the SCSI controller is faulty.
The other SE/30 board with dead SCSI was repaired quickly, I found 2x broken VIA's, repaired them with a wire and it was sorted.
Now I have to replace the SCSI chip: I did found one on Ebay, PLCC version, I hope this fits. But I don't have an SMD rework station to replace IC's.
Has anyone used this solution ? http://www.chipquik.com/
Have a look at the 40 seond youtube video on their website: it's very interesting how fast they can remove an SMD IC.
For that board I'm 90% sure the SCSI chip self is faulty. It noticed in FWB toolkit 2.0.6 than eventhough it recognises the hard drive, model and brand: it lists a wrong size. FWB says it's 64MB while it's a 80MB drive.
When I verify that with a good SE/30 board: it's indeed recognised as 80MB: this only adds to my case that the SCSI controller is faulty.
The other SE/30 board with dead SCSI was repaired quickly, I found 2x broken VIA's, repaired them with a wire and it was sorted.
Now I have to replace the SCSI chip: I did found one on Ebay, PLCC version, I hope this fits. But I don't have an SMD rework station to replace IC's.
Has anyone used this solution ? http://www.chipquik.com/
Have a look at the 40 seond youtube video on their website: it's very interesting how fast they can remove an SMD IC.
Don't forget that questcomp.com has the 53C80 also. I don't know what the pricing on Ebay is looking like, but it's often higher than questcomp.com. Ah, took a look and that $4.20 with free shipping deal looks pretty sweet.
The chip will be the correct size. There was only one PLCC-44 package used for the 53C80.
As far as removing it goes, I have used Chip Quik. It's been almost a decade since I used it, because I don't need it any more.
If you are deciding between buying Chip Quik and a heat gun, I'd get the heat gun.
Chip Quik is an alloy which you melt into existing solder using your soldering pencil. The allow, when mixed with conventional solder, lowers the melting point of the amalgam. This makes it possible to remove moderate sized chips using just a soldering pencil.
Optional step: Remove as much solder as possible from the pins using soldering pencil and braid.
The procedure in brief would be to coat all the pins on the chip to be removed liberally in solder flux. Then run a bead of Chip Quik on all sides of the chip. Then heat each pin/pad individually, to make sure that the Chip Quik has mixed with the solder.
Finally, go around the chip heating each bead of solder/alloy (4 beads in the case of the PLCC-44) in turn. Eventually, you'll get them hot enough that all four will be above the melting point at the same time and the chip will lift off the board. This requires patience. Do not pry the chip up, you'll lift a pad.
If you tried this without Chip Quik, your chances of ever getting the first bead to stay hot enough to release, by the time you get to the last bead, is very low to impossible.
However, if you use a heat gun and just heat the whole chip at once, then there's no need for Chip Quik. I have a Milwaukee brand heat gun with a dial adjustment on the back, which I bought at Home Depot for less than $40. However, they don't sell that model any more. The dial was perfect for this. Most heat guns only have one or two temperature settings.
When using a heat gun to desolder, it is important to mask nearby components. You can literally blow small componetns off the board. I did the first time I tried it. I use modeling clay. Others use aluminum foil. I like clay because it sticks to the board and doesn't blow away, but foil has advantages too.
I find with my heat gun set to 8 (out of 10) it takes about 2 minutes from a cold start to get the typical chip to release from the board. Maybe a little less for something like a PLCC-44, but a little more for some of the through hole chips with 32+ pins.
Once you have the chip removed, remove any remaining solder from the pads with pencil and braid. Clean the board thoroughly. Then tin just one corner pad. Set the chip in place carefully, and melt the tinned pad to settle its pin into place. Thin solder the opposite corner to secure the chip. Then proceed with soldering the rest of the chip. One can often get away with running a bead of solder across each side of the chip, and then coming back through and cleaning the excess solder off with braid. This can be a lot faster, but runs a (small) risk of hidden shorts.
The chip will be the correct size. There was only one PLCC-44 package used for the 53C80.
As far as removing it goes, I have used Chip Quik. It's been almost a decade since I used it, because I don't need it any more.
If you are deciding between buying Chip Quik and a heat gun, I'd get the heat gun.
Chip Quik is an alloy which you melt into existing solder using your soldering pencil. The allow, when mixed with conventional solder, lowers the melting point of the amalgam. This makes it possible to remove moderate sized chips using just a soldering pencil.
Optional step: Remove as much solder as possible from the pins using soldering pencil and braid.
The procedure in brief would be to coat all the pins on the chip to be removed liberally in solder flux. Then run a bead of Chip Quik on all sides of the chip. Then heat each pin/pad individually, to make sure that the Chip Quik has mixed with the solder.
Finally, go around the chip heating each bead of solder/alloy (4 beads in the case of the PLCC-44) in turn. Eventually, you'll get them hot enough that all four will be above the melting point at the same time and the chip will lift off the board. This requires patience. Do not pry the chip up, you'll lift a pad.
If you tried this without Chip Quik, your chances of ever getting the first bead to stay hot enough to release, by the time you get to the last bead, is very low to impossible.
However, if you use a heat gun and just heat the whole chip at once, then there's no need for Chip Quik. I have a Milwaukee brand heat gun with a dial adjustment on the back, which I bought at Home Depot for less than $40. However, they don't sell that model any more. The dial was perfect for this. Most heat guns only have one or two temperature settings.
When using a heat gun to desolder, it is important to mask nearby components. You can literally blow small componetns off the board. I did the first time I tried it. I use modeling clay. Others use aluminum foil. I like clay because it sticks to the board and doesn't blow away, but foil has advantages too.
I find with my heat gun set to 8 (out of 10) it takes about 2 minutes from a cold start to get the typical chip to release from the board. Maybe a little less for something like a PLCC-44, but a little more for some of the through hole chips with 32+ pins.
Once you have the chip removed, remove any remaining solder from the pads with pencil and braid. Clean the board thoroughly. Then tin just one corner pad. Set the chip in place carefully, and melt the tinned pad to settle its pin into place. Thin solder the opposite corner to secure the chip. Then proceed with soldering the rest of the chip. One can often get away with running a bead of solder across each side of the chip, and then coming back through and cleaning the excess solder off with braid. This can be a lot faster, but runs a (small) risk of hidden shorts.
I normally use a heat gun, they work great, but do practice on some junk boards first until you get the hang of doing it without frying the crap out of the board. On some particularly hairy ones I've used a combination of Chip Quik and a heat gun. That lets you use less heat and have less chance of removing nearby parts in the process or overcooking things while still being able to heat all the pins at once and reduce the risk of lifting pads. Just make sure you remove as much of the Chip Quik tainted solder as you can, or it ends up being tricky to solder the new part on. The solder stays molten for what seems like ages.
uniserver
Well-known member
once you remove the bad chip.
- determine what might be needed
- You can re-flow the pads with the iron.
- You could clean the pads with Solder Braid, (if needed) then flow new solder...
I just refux the pads, set the chip where its suppose to be, heat it up evenly till it seems all the pads are melted and the solder has re-reflowed to the legs, Then I press down slightly with my dental pick on the chip.
if i need to go over some legs with the iron i will do so.
I would like to say So far, Every-time I thought it was the IC, I was wrong, It was a trace or VIA issue
But I have seen on the web, the SCSI chip on the SE/30 is known for having issues.
- determine what might be needed
- You can re-flow the pads with the iron.
- You could clean the pads with Solder Braid, (if needed) then flow new solder...
I just refux the pads, set the chip where its suppose to be, heat it up evenly till it seems all the pads are melted and the solder has re-reflowed to the legs, Then I press down slightly with my dental pick on the chip.
if i need to go over some legs with the iron i will do so.
I would like to say So far, Every-time I thought it was the IC, I was wrong, It was a trace or VIA issue
But I have seen on the web, the SCSI chip on the SE/30 is known for having issues.
Flux is the key, use plenty of flux. If the solder is balling up, sticking to the iron, refusing to stick where you want it, or otherwise causing trouble, add a bit more flux. I usually tin a pad or two in one corner, add a bit of flux, then carefully align the chip and heat the tinned pad/pin to tack the part in place. Double check the orientation and alignment, then tack another corner. Check the alignment again to make *really* sure the chip is where you want it, then solder the rest of the pins. You can usually load up a little ball of solder on the tip of the iron and then sort of drag it across the pins. Provided there is enough flux, it will tend to flow nicely over the pad and make nice joints without bridging. If you get too much on and do get a bridge, use some desoldering braid to clean it up. The braid needs flux too, normally it comes soaked in it but it dries out over time. If you dip the braid in liquid flux it works much better.
It should go without mentioning, but use rosin based flux intended for electronics, NOT the paste or liquid flux for plumbing. That stuff contains acid and will corrode stuff like leaky capacitor electrolyte. Sometimes you can find water based flux that cleans up easier, otherwise a liberal spray of flux remover, light scrubbing with a toothbrush and a final wipedown with a paper towel soaked in a bit of flux remover will clean up the mess. If you use no-clean flux you don't really *have* to remove it, but it looks a lot more professional if you do. Take care as the flux remover will dissolve and spread the flux all over the board into a thin sticky layer if you let it, that's why I wipe it down with a paper towel.
It should go without mentioning, but use rosin based flux intended for electronics, NOT the paste or liquid flux for plumbing. That stuff contains acid and will corrode stuff like leaky capacitor electrolyte. Sometimes you can find water based flux that cleans up easier, otherwise a liberal spray of flux remover, light scrubbing with a toothbrush and a final wipedown with a paper towel soaked in a bit of flux remover will clean up the mess. If you use no-clean flux you don't really *have* to remove it, but it looks a lot more professional if you do. Take care as the flux remover will dissolve and spread the flux all over the board into a thin sticky layer if you let it, that's why I wipe it down with a paper towel.
Well it was indeed a faulty SCSI chip. I desoldered the old one with Chip Quick, put in a brand new "old stock" one I found on Ebay and the SE/30 immediately booted from it's HDD.
I left it running for 12 hours and ran hourly Norton Disk Doctor full surface tests with verify on the HDD and everything works fine.
Now FWB toolkit and Apple's HD SC Setup also recognize the HDD and were able to Initialize it, including successfull verification and mouting.
Did a clean install with OS 7.5.5 from CD-Rom: all well. Another one saved
I left it running for 12 hours and ran hourly Norton Disk Doctor full surface tests with verify on the HDD and everything works fine.
Now FWB toolkit and Apple's HD SC Setup also recognize the HDD and were able to Initialize it, including successfull verification and mouting.
Did a clean install with OS 7.5.5 from CD-Rom: all well. Another one saved
Thank you for letting us know how it turned out, and apologies for steering you away from replacing the chip. It's been a long time since anyone's problem on this board was a faulty SCSI chip, but now we can chalk one up in that category.
It's really fantastic any time I get to read about another old Mac fixed.
It's really fantastic any time I get to read about another old Mac fixed.
