exxos's Atari, Amiga & retro forum

christophe.griard wrote: ↑Thu Apr 30, 2020 7:15 pm I made the measurement again on pin 22 and 59 and I confirm the +5V. I made during these two new measurements a picture during power on on these two pins.

It is strange When CPU comes out of RESET, the pins should be low.. but also strange that reset itself does actually change the state of the pins.

So I can only assume either the CPU is faulty, or some other chips on the bus is causing the error like Videl.

How good is your soldering ? Because at this point I would carefully lift the reset pin on videl ( I think it was the top right pin so easy as on corner of the chip ) and solder a very thin wire to that pin and connect to 5V.. This will disable Videl... Then you can check A0,A1 again to see if it goes to 0V like other address pins..

I look that the 2 pins with a magnifier and I do not see anything wrong/different from other CPU pins.

I looked again at the signal on these two pins and nothing different from before and still +5V. In addition, if I push the reset button, the +5V on these two pins is not affected by the reset and stays at +5V.

I also checked like you the Voltage (still without ROM and without RAM) on all pins of expansion port that correspond to addresses (from pin 26 to pin 48) and only pin 47 and 48 are at +5V (like seen on the CPU pins 22 and 59... So I did not make a mistake when I measured pin 22 and pin 59 directly on the CPU).

Yes the reset pin on the VIDEL is the pin n°1 at the upper right corner of the VIDEL.

How good is my soldering skills? I know how to / can perform "classical" solders like capacitors on a motherboard but I never did such solders on such small pins.
I have some flux to do such solders, I saw some videos on Youtube people explaining how to proceed for example to desolder/solder such a chip like the VIDEL but it needs practice and I never did it... So I am sure I am not able to do it properly and as you told me some days ago, it would be better to go to the company that is used to such SMT chip soldering.

Having said that, I think I can desolder the VIDEL reset pin and connect it to + 5V. You have to be very careful and it will stress me out because I am someone who takes very good care of what I have in general and it will hurt me to make such a treatment to my Falcon... but it is to make him in better health so...). The issue is that maybe I won't be able to solder it again to the motherboard in case the VIDEL chip seems not to be involved in the failure...

What would be your advice to desolder the VIDEL pin : do it with my soldering station with heat :
- better do it with a soldering iron (I have also a less powerful soldering iron but equipped with a manual pump; next buy will be a desoldering station with a gun and an electrical air pump)...
- or desolder it with hot air (I have both on my soldering station)?

Is there another way to diagnose if VIDEL is faulty (for example by scanning output pins of the VIDEL or whatever)?

Heat from soldering iron, flux, and gently lift pin with a small needle is how I do it.. If your not confident to do this.. then I suggest you don't try it.

My idea is to hold the videl in reset mode to see if the CPU then behaves.. there is no other way, other than to remove the videl completely and change it and hope that is the bad chip.. but there really is not many more options left now.

Thank you very much Exxos for all your advice and answers.

I think that desolder it gently without doing any other damage is something I can do. Resolder it would be another story...

I fully understand the purpose of this VIDEL deactivation: I will therefore try this "mod" to try to further diagnose my baby.

Many thanks again for your precious help ...

I'll probably try to do this tomorrow. I will obviously keep you informed of the result.

Have a nice end of the day Exxos.

See you tomorrow.

Hello Exxos,

I succeeded in desoldering properly and gently the pin 1 of VIDEL (see below picture).

I am still without ROM and without RAM.

In that situation with the pin 1 of VIDEL connected to nothing, I checked the voltage for addresses on expansion port and checked that every address pin is at 0V unless pin 47 and 48 still at +5V.

Then I checked than +5V was still present on pin 24 of expansion port.

Then I connected a wire between pin 24 of expansion port to pin1 of VIDEL... And same result : pin 47 and pin 48 are still at +5V.

I did not solder a wire on pin 1 (I put on wire on the pin 1 and I kept it pressed hard) but I am sure of the electrical connection between pin 1 and pin 24.But unfortunately, same result...

Well done!

So something else on the bus is causing those errors... Problem is that with lots of ICs on the bus, it be difficult to work out which one. It could well be the CPU is simply faulty . the CPU should drive those pins low and it isn't. You already checked resistance to 5v so its not shorted to 5v. You also say solder joints are OK.. But if possible check with scope directly those faulty address lines and make sure CPU is really still at 5v. A broken solder joint from CPU to pads could cause same error as CPU drives address pins all low, so if there is bad connection to pad, it will remain at 5v. So you need to check directly on CPU pin... Or maybe safer to check resistance from CPU pin to expansion port with power off of course.

I will have a think if there is any more tests you could do, but likely you will just have to start changing chips. Likely looking you need to try changing the CPU first. But if your going to get someone else to do that, you might consider changing combel and video at the same time to be sure.

Hello Exxos,

Thanks for the "Well done" and thanks for your feeback.

Yes I measured yesterday the resistance to 5V for each pin (directly on CPU pin) with a 4,6 kOhms for Ax addresses and 10 kOhms for DX addresses.

Concerning the +5V on A1 and A2, I first measured them on CPU pins directly...

Then, when you checked on your Falcon that pin 47 and 48 were at 0V, then I checked the signals on the expansion port with the same results (0V on all addresses pin unless pin 47 and 48). I tought I was stupid to check these signals directly on the CPU as they can be much more easily checked on the expansion port...
The +5V was measured on pin 22 and pin 59 directly on the CPU.
I will rechecked that to be more than 100% sure.

I will nevetheless check also the resistance between each pin of the Ax and Dx CPU pins and the corresponding pins on the expansion port.

Concerning the chip replacement, I should receive in some days/weeks a new VIDEL and a new COMBEL but not a new CPU. I did not find these chips on your webstore: I imagine I can order them only by Best Electronics or is there any other source of spare parts?

I will checked the things I mentioned above this evening and I will keep uou uptodate.

Thanks again for your help Exxos.

christophe.griard wrote: ↑Fri May 01, 2020 7:30 pm Concerning the chip replacement, I should receive in some days/weeks a new VIDEL and a new COMBEL but not a new CPU. I did not find these chips on your webstore: I imagine I can order them only by Best Electronics or is there any other source of spare parts?

I don't sell any Falcon main chips, there is nowhere to get them from other than BEST electronics unfortunately... I would suggest buying the CPU from them anyway, rather than some other site.. So many fake CPUs about is better just simply not to risk soldering in a faulty one..

I'm nowhere as skilled as Exxos and others but before you get into the nasty job of replacing ICs, I would still open up the schematics and went from the beginning - if you are confident that BERR, HALT and friends are OK, then yes, check connections for data and address bus, control signals, cold joints, ... it's very easy (but dull & painful) work but could show you some basic issues - like some GAL (or even capacitor!) is not outputting anything or at some place there is broken connectivity.

I have had numerous Falcons repaired (even ones in a really bad shape), we even fixed a Falcon which had a totally destroyed NVRAM holes and PCB paths around it (a lot of wiring across the whole PCB to various chips...), we didn't have to replace any of the main chips even once.

Btw, a nice diagnostics method is also fire up the Falcon, have the PCB accessible and ... knock on each chip. Sounds super stupid, does miracles.

Just my two cents.

I just checked the continuity between the Address 1 to Address 23 on the expansion port and the corresponding pins on the CPU (putting my probes directly at the outlet of plastic core of the CPU and everything is ok (resistance less than 1 Ohm) : so the solders are ok between CPU and expansion port.

As there is no A0 corresponding on the expnsion port, only A0 was not tested.

Thanks Mikro for your information: I did try to apply some pressure on each of CPU, COMBEL and VIDEL when Falcon was on with no result... Unfortunately...

And yes the CPU is not halted and there is no bus error (seen on both the expansion port and on the CPU pins). What is not normal is the fact that without ROM installed, addresses A1 and A2 are High instead of Low (all other addresses are low)...

I have already check some resistors on the motherboard (both on upper side and rear side) with no issue but I have not tested all resistors => so it is something I will do. For the capacitors it is much more tricky for me as you have to desolder them to test them so...

If we would have found th root cause of the failure I would raise the question to replace all capacitors of the motherboard : at the moment, the motherboard is very clean (upper and rear side) with no leakage and as we do not find the root cause I won't change the capacitors at this stage unless a bad capacitor could lead to make my Falcon fail...

As we tested with the VIDEL deactivated it can be the CPU or the COMBEL that are faulty...

Maybe I should try to get in contact with Mr Czuba to see if he could find something else after explaining him all the checks we made...