| Electronics > Repair |
| What more i can do? |
| << < (108/112) > >> |
| Swake:
You too, and also for everyone here, have a perfect 2025 ! |
| m k:
--- Quote from: 222Lab_Test222 on December 26, 2024, 11:07:42 pm --- --- Quote from: m k on December 26, 2024, 12:16:59 pm --- Why did you scope only half of the data? --- End quote --- Half the data mean? --- End quote --- D0 to D7 only once. So it was like #382? First bigger interference of #477 means the start. So first D0 is up, so 0d0d. 76543210 00001101 But the end part is very different than #382. So one more time, with three sets you can start voting. --- Quote from: 222Lab_Test222 on December 26, 2024, 11:16:17 pm --- --- Quote from: m k on December 26, 2024, 06:06:59 pm --- If you like you can check that A4 of all faulty boards. If shape is the same you have functioning connection between CPU and program ROM. If that single high is not there you have a similar fault that the original test subject has. --- End quote --- If I do not get similar signals then that hints CPU is faulty? Or is there any way to know CPU is ok or not. --- End quote --- You can't prove negative. CPU is also so complex that you can have a faulty section even when the thing seems to be ok. One example is memory test, back in the day it was doomed to be inaccurate and only accurate thing being an operating system. And even then it was still like which end or something in between. Internal CPU errors are rare, like one day adding doesn't go anymore. Usual problem area is external world interfacing part. New processors with embedded extras are different. So in practice you must diagnose enough faulty pins by judging them as not right, so comparing them with known good state. But your CPU is regular and it is socketed, so "just" find an old regular computer and test them. Digging the chip out is not necessary easy, check few videos. In this case you can also use measurements, since that earlier wrong data was 0d0d and other ROM content is 0d, you can make an educated guess. You swap ROMs and do the same measurement again. But since you don't know what you don't know there are no short cuts, partial measurement doesn't tell much, you must do it all again. Don't hassle either, be sure that chips are correctly inserted and legs are straight. With old and new measurement you can do a comparison. If the data is still 0d0d you can guess that it is not from a ROM at all. If it is e9e9 you can guess that it is from the swapped ROM, and this can affect CPU operation in measurable way. Since e9 is a jump you must have addresses measured also, so that you can check how that jump is executed. If that jumped address then correlates to next data bytes you can say that the CPU is actually seeing what you have measured. If jump address is something else you can do the next educated guess. |
| asis:
next... |
| 222Lab_Test222:
--- Quote from: asis on December 27, 2024, 02:00:48 pm ---next... --- End quote --- Yup these are good |
| 222Lab_Test222:
--- Quote from: m k on December 27, 2024, 11:41:02 am --- --- Quote from: 222Lab_Test222 on December 26, 2024, 11:07:42 pm --- --- Quote from: m k on December 26, 2024, 12:16:59 pm --- Why did you scope only half of the data? --- End quote --- Half the data mean? --- End quote --- D0 to D7 only once. So it was like #382? First bigger interference of #477 means the start. So first D0 is up, so 0d0d. 76543210 00001101 But the end part is very different than #382. So one more time, with three sets you can start voting. --- Quote from: 222Lab_Test222 on December 26, 2024, 11:16:17 pm --- --- Quote from: m k on December 26, 2024, 06:06:59 pm --- If you like you can check that A4 of all faulty boards. If shape is the same you have functioning connection between CPU and program ROM. If that single high is not there you have a similar fault that the original test subject has. --- End quote --- If I do not get similar signals then that hints CPU is faulty? Or is there any way to know CPU is ok or not. --- End quote --- You can't prove negative. CPU is also so complex that you can have a faulty section even when the thing seems to be ok. One example is memory test, back in the day it was doomed to be inaccurate and only accurate thing being an operating system. And even then it was still like which end or something in between. Internal CPU errors are rare, like one day adding doesn't go anymore. Usual problem area is external world interfacing part. New processors with embedded extras are different. So in practice you must diagnose enough faulty pins by judging them as not right, so comparing them with known good state. But your CPU is regular and it is socketed, so "just" find an old regular computer and test them. Digging the chip out is not necessary easy, check few videos. In this case you can also use measurements, since that earlier wrong data was 0d0d and other ROM content is 0d, you can make an educated guess. You swap ROMs and do the same measurement again. But since you don't know what you don't know there are no short cuts, partial measurement doesn't tell much, you must do it all again. Don't hassle either, be sure that chips are correctly inserted and legs are straight. With old and new measurement you can do a comparison. If the data is still 0d0d you can guess that it is not from a ROM at all. If it is e9e9 you can guess that it is from the swapped ROM, and this can affect CPU operation in measurable way. Since e9 is a jump you must have addresses measured also, so that you can check how that jump is executed. If that jumped address then correlates to next data bytes you can say that the CPU is actually seeing what you have measured. If jump address is something else you can do the next educated guess. --- End quote --- Seems more complicated than I expected. |
| Navigation |
| Message Index |
| Next page |
| Previous page |