Amiga.org
Amiga computer related discussion => Amiga Hardware Issues and discussion => Topic started by: danbeaver on March 02, 2013, 01:34:31 AM
-
To Clarify: The black stripe of Diode 800 points TOWARD the motherboard.
A failure of 5 volts on pin 25 of the SCSI port (failure = lack) when powered up, or pin 25 "goes to ground" when powered down means diode D800 is going to ruin your day. I have two A3000D's and the silk screening is opposite on both (one points in and one points out) and neither diode was correct, with both pointing out.
__________________
x x x x x x x x x x x x x
\ o x x x x x x x x x x x /
__________________
o = Pin 25 looking end-on of the female SCSI socket
-
Yep, it's certainly been documented over the years.
http://amiga.serveftp.net/A3000_HardwareGuide/scsi-termpower.html
Another common problem is that people plug a Centronics printer cable into the external SCSI port, which shorts the +5V termination supply to ground, which causes the diode to go short circuit and the excess current causes the track on the bottom side of the board to fuse open circuit.
-
Yes, I found everything BUT which way the diode goes.
Yep, it's certainly been documented over the years.
http://amiga.serveftp.net/A3000_HardwareGuide/scsi-termpower.htm
Another common problem is that people plug a Centronics printer cable into the external SCSI port, which shorts the +5V termination supply to ground, which causes the diode to go short circuit and the excess current causes the track on the bottom side of the board to fuse open circuit.
404 File Not Found
The file /A3000_HardwareGuide/scsi-termpower.htm couldn't be found.
-
Yes, I found everything BUT which way the diode goes.
404 File Not Found
The file /A3000_HardwareGuide/scsi-termpower.htm couldn't be found.
Should be http://amiga.serveftp.net/A3000_HardwareGuide/scsi-termpower.html
-
The diode protects the machine from excess voltage/inbound current from the SCSI port, so it faces the port with its cathode/stripe.
While you're at it you might as well add a 1A picofuse to prevent smoking the trace in case of short circuit (been there, done that - no, not a printer cable but a Parnet cable).
-
Fixed many a A3000 with this diode shorted,backwards or burned traces over the years
lack of term power is why many people think their scsi is flakey or problematic. Its not always a case of a old WD chip revision. it really should of been fused. I think my original 3000t manual shows a fuse in the schematic but it was probably removed for cost reduction knowing commodore.
-
Mech,
I knew you would know, but after a very bad guess(or two) I just pulled the schematics, checked it with a multimeter and found the the black stripe "points toward the port" but that pin is going toward the motherboard.
Funny that it worked the way it had blown (+5 powered, but ground when power off). Also had to resolve whether the new GALS I burned were flakey.
Oh, I found blacked traces underneath, but they tested good.
Would one place the fuse in line (between the diode and port) as makes the most sense to a chemistry major, or some other place?
-
Yes, put the fuse in series, the order doesn't matter. :)
-
Again as a chem type person, I would want the fuse to blow before the diode; granted with my soldering skills, I would end up replacing the fuse and diode at the same time. But anything that works to protect the MB traces is better than redoing them.
@Mech (Mike), Do you work on (others) stuff like this? I try to refurb my stuff, but there are sticky areas (SMD for example) that didn't exist when I was a kid (soldering my O27 train stuff). I learned my DC current from that hobby. I remember you reprogramming NEC USB (EP)ROMs so I know you have a "skill set."
-
The fuse's main purpose is to protect the board; the diode is easy to replace and in the same cent price range anyway. Once the tripping current is exceeded and it blows, the electrical circuit is interrupted and there's no more current to do any harm.
Alternatively, you could put a resistor in series with the diode, limiting the current to a non-destructive level. However, the termpower design is poor enough as it is (delivering only 4.3V, Vterm should be more in the range of 5.25V), so that would require a bit of redesign... Ever done that, mechy? ;)
-
Thanks for pointing out the truncated URL above, fixed now. I'll make a note in that section which way the diode should be, it's clearly not obvious to everyone.
The design would have been far better with a self resettable fuse or something, but was either removed to save money or was simply an oversight. If you wanted to, it would be tidy enough to lift one end of the diode and solder an axial fuse in series with it. A series resistor won't be ideal as you won't be able to supply the required current and maintain the output voltage, i.e. you'll have issues with voltage drop across it.
4.3V at the terminator won't be an issue for active terminators which regulate this to 3.3V anyway. Even passive terminators which use a simple 220/330 Ohm resistor divider to derive 3V on tri-stated lines will only suffer a 14% error (2.58V as opposed to 3V) which is still well within spec. Actually I just noticed in this Wikipedia article (http://en.wikipedia.org/wiki/Parallel_SCSI#Termination) that the termination power is generally at 4.3V anyway. Seems Commodore's design wasn't all bad.