Really well done tracking down the fault. Obviously the missing EPROM's were simple to spot but diagnosing bus contention was rather more difficult. '245 bus transceivers fail regularly (as do '04 hex inverters) so are obvious candidates. Perhaps explaining what these chips do would have been helpful with maybe a quick look at the datasheet. This would have helped explain that the outputs should be tristated when the enable line was not active but were enabled on this chip all the time regardless of the state of the enable pin. Excellent that you traced this using only a logic probe and didn't need an oscilloscope, logic comparator or logic analyser.
Right, finding this with the logic probe was a bit of a stroke of luck. Also it would have been more difficult if the rams hadn't been socketed (which I think they only were because somebody modified the pcb to have ram sockets). But of course I'm happily accepting a stroke of luck like this coming my way... ;-)
@@christophzett No problem, I enjoy your content. You have a broad understanding of these boards. Hopefully I might learn something! Just need to go back over them a couple times probably so it sticks : )
2 ปีที่แล้ว +2
Nice job with the repair! Loud and clear. I wish I could do my repair videos in english instead of french! Just to start a discussion, I would say that I am not a huge fan of the "Double the data blocks on a larger chip" technique, unless I am sure the upper address line is not left floating. And if it is, then you can use one 32kb chip to replace two 16kb and use the CS line as an upper address line 😀.
Absolutely right - it would be bad practice to let the spare address line floating. In the case of the phoenix pcb I think it is being connected to +5V permanently to the pcb traces that are there by default. I actually made sure that this is the case before applying that solution but forgot mentioning it in the vid.
You are coming across perfectly and I'm enjoying the videos! You systematically worked through something which can be very confusing i.e. contention due to a fault. That's an important topic and does crop up since many devices have active low chip enable so a poor connection or broken trace can make someone to be a permanent talker on the bus.
Excellent video again, can't wait for the next one. Regarding your 'question' at the end, your explanation is very good, and I can follow and understand your logic during faultfinding which makes it great to watch. I wonder if you could do a rom mod, and change all 8 eproms for one 27c128 eprom and tie the /ce line of this eprom to a14, and a11,a12 and a13 directly to this 27c128 eprom. you could take out ic10 (74ls138) on the cpu board and connect some pins here to route the signals to the (now) unused /ce lines of the other eproms and pick the up from there. Thinking of this, probably easier just to buy 8 eproms like you did ;-)
2 ปีที่แล้ว
Oups, almost the same question, but with 4 32kb chips 😉.
Yes I actually thought about this, too. But you are right, It would require some extra circuitry to turn the 8 chip select signals back to address lines. So too much hackery for my taste, I always like to keep the boards original. I'm also not a big fan of rom mods and high score kits and so on to be honest.
Ja, mal wieder ein Video mit reichlich u guten Erklärungen Prima!
Vielen Dank!!! Die nächsten Teile kommen bald!
We need another video, yours are excellent!!!!!!!!!!!!!!!!
*spoiler alert* First Pleiades repair is finished and so next vid coming up by the end of this week!
Really well done tracking down the fault. Obviously the missing EPROM's were simple to spot but diagnosing bus contention was rather more difficult. '245 bus transceivers fail regularly (as do '04 hex inverters) so are obvious candidates. Perhaps explaining what these chips do would have been helpful with maybe a quick look at the datasheet. This would have helped explain that the outputs should be tristated when the enable line was not active but were enabled on this chip all the time regardless of the state of the enable pin. Excellent that you traced this using only a logic probe and didn't need an oscilloscope, logic comparator or logic analyser.
Right, finding this with the logic probe was a bit of a stroke of luck. Also it would have been more difficult if the rams hadn't been socketed (which I think they only were because somebody modified the pcb to have ram sockets). But of course I'm happily accepting a stroke of luck like this coming my way... ;-)
Another great video, well done. Also I can understand your explanation of your diag and repair method clearly.
Thank you very much. Posts like yours are actually very assuring and therefore helpful and inspiring for me!
@@christophzett No problem, I enjoy your content. You have a broad understanding of these boards. Hopefully I might learn something! Just need to go back over them a couple times probably so it sticks : )
Nice job with the repair! Loud and clear. I wish I could do my repair videos in english instead of french!
Just to start a discussion, I would say that I am not a huge fan of the "Double the data blocks on a larger chip" technique, unless I am sure the upper address line is not left floating.
And if it is, then you can use one 32kb chip to replace two 16kb and use the CS line as an upper address line 😀.
Absolutely right - it would be bad practice to let the spare address line floating. In the case of the phoenix pcb I think it is being connected to +5V permanently to the pcb traces that are there by default. I actually made sure that this is the case before applying that solution but forgot mentioning it in the vid.
You are coming across perfectly and I'm enjoying the videos! You systematically worked through something which can be very confusing i.e. contention due to a fault. That's an important topic and does crop up since many devices have active low chip enable so a poor connection or broken trace can make someone to be a permanent talker on the bus.
Yes absolutely. *spoiler alert* This will also be a mayor topic on the upcoming Pleiades vid.
Another excellent repair =D
Wow thank you Chris, great to get a positive comment from the master himself! ;-)
Very instructive and very well explain. Thanks a lot for your work.
Best regards, Nicolas
Yes, love your troubleshooting and explanations!
Thank you very much!
Excellent video again, can't wait for the next one. Regarding your 'question' at the end, your explanation is very good, and I can follow and understand your logic during faultfinding which makes it great to watch.
I wonder if you could do a rom mod, and change all 8 eproms for one 27c128 eprom and tie the /ce line of this eprom to a14, and a11,a12 and a13 directly to this 27c128 eprom. you could take out ic10 (74ls138) on the cpu board and connect some pins here to route the signals to the (now) unused /ce lines of the other eproms and pick the up from there. Thinking of this, probably easier just to buy 8 eproms like you did ;-)
Oups, almost the same question, but with 4 32kb chips 😉.
Yes I actually thought about this, too. But you are right, It would require some extra circuitry to turn the 8 chip select signals back to address lines. So too much hackery for my taste, I always like to keep the boards original. I'm also not a big fan of rom mods and high score kits and so on to be honest.
Yes, also possible but again you would need an extra ic to produce the cs signal - like an or gate or so.
Hi, very instructive video thank you! Where did you order the 2732 EPROMs? Which reference did you buy?
Thank you! Got them on ebay I think from poland. I think there were some intel ones.
Nice job. Thanks
You are welcome!
Keep the video's coming!!!!!!!!!!!!!!!!!!!!
Will do! Next one due this week!
new subscriber ,lets hit 1000 !
Thank you very much!!!
We need more cow bell.
Thank you! Spread the word! ;-)
Subbed.
Thank you!