OK, so it seems fixed (knock-knock). To sum my adventure up + to possibly help others in a future I try to compile some hints about NVRAM and Falcon repair in general.
1. Never try to mess with Falcon hardware, especially meaning soldering work, if you are not very
experienced. Having skill like "I can solder something here and there" is simply not
2. Even if you are experienced iron man :), keep in mind Falcon has multi-layer PCB and its contacts can be destroyed not only by scratches but also by much heat from soldering iron! (my case)
Ok, so even although you were careful (I was not) with soldering iron, something stopped to work after NVRAM replace (better said, after NVRAM -> socket+NVRAM replace).
3. Download NVRAM (Dallas) datasheet (attached) and Falcon schematics (from http://www.janthomas.org.uk/files/falcon.zip
). Identify and check basic things on NVRAM -- +5V on pins 1/24, GND on pin 12.
4. Check further things -- 3 signals from Combel (see post from enito or check Falcon scheme for RTCxx signals) to NVRAM, data bus (d0-d7) connection (against Videl, Combel, CPU, ... whatever you like) to NVRAM, RxW signal (anywhere on Falcon MB). Be careful with data bus on ACIAs, as they have d0-d7 but they are upper 8bits, not lower !!! So don't use them for connection validation.
5. Now, tricky thing. TOS (EPROM) is connected on NVRAM data bus, not general Falcon one!!! So you must check d0-d7 from NVRAm to TOS! Rodolphe was not sure but it's possible SDMA chip is also connected to NVRAM and not general data bus.
6. Ok, now you should have some broken traces found. How to repair: if it's easy soldering (like signal from NVRAM), solder it on front or back side of MB, as appropriate, there's no problem with that. Problems begin with destinations -- from NVRAM. For example, soldering on Combel, TOS chip (with very small pins)... it's worth checking if there isn't some hole near the pin connected to it, i.e. you wont solder on pin itself but on the hole, you can use back side of MB then, too. Very useful.
7. You've done all soldering and still not working. Very important thing: check everything again
. It's not empty words, I really mean it, best to let it check to someone else than who did the soldering / first checking. Everything from good connections between damaged traces to right number of pin (!).
8. Checking: it's more or less safe to use "beeper" to check, at least for data and control signals, I didn't destroy my Falcon with that :) But for checking on CPU, FPU, GALs, ... better use resistance measuring, i.e. if you see unlimited resistance => no trace, ~0 Ohms resistance => working trace.
9. Ok, you're sure everything is connected and Falcon still shows nothing, time to check if reset signal is delivered to CPU and if some most important chips receive their clocks. In Falcon schematics archive is very useful diagram by Rodolphe showing clocks in Falcon with connections and values. Very useful to check. Also AS signal on CPU is useful to check, if you get 1-2 UPs and then nothing, it's bad (access fault reading TOS ROM).
10. You receive all clocks, see reset signal when measuring and still not working? It's possible some pins / contacts on chips seems to be ok, when you push them, you receive values but as soon as you get your hands out, chip is disconnected. I recommend to let Falcon running and try to push / knock on MB / chips and to see if something doesn't change. Sounds silly? It shouldn't, it's true!
So, this is what I learned here. Used stuff: multi-meter device, Falcon schematics, Dallas data sheet, Rodolphe's clock diagrams, oscilloscope. Short summary of my problems / causes:
- I suck at soldering :)
- trace D5 from NVRAM to data bus was broken
- trace from D5-D7 from NVRAM to TOS EPROM were broken
- RTCDS from Combel to NVRAM was broken (found after 3rd check in totally hopeless mood!)
- RTCDS from Combel to NVRAM was soldered on Pin18 instead of 17 ;-)
- XRESET from U11 to CPU wasn't delivered (measured by oscilloscope), we didn't solve this mystery, it got fixed after next step, probably some side effect
- Pin12 (GND) on NVRAM was badly soldered (good contact only after push / knocking on MB)
You see, quite horror. NEVER try to solder if you can't do it!!! NEVER !!!
I want say public and huge thank you to -XI-/Satantronic
for his excellent soldering skills, Jookie
for his useful and pragmatic point of view on our work ("hey, here's the bug, ok, next will follow, don't worry" :) and amazing ability to deduce things from foreign schemes for foreign hardware (he didn't do any work on any Falcon before, not even change of hard disk :) and of course Rodolphe Czuba
for his useful tips what to check (we would never find out TOS EPROM is behind NVRAM for example).
You do not have the required permissions to view the files attached to this post.