Dragon's Lair Fans - Arcade Lifestyle
General Chat => Technical Area => Topic started by: Laszo on September 16, 2015, 12:04:59 AM
-
New topic, I am fed up with my non working Centi pcb so I wanna get to the bottum of this:
Problem: two pcb's with sound and random number generation missing. All mushrooms are lined up to the left. Any suggestions ... Pokey and big blue are ok.
-
Sofar I replaced the ic sockets for both processor and pokey. Sound and random number where back. :) But not for long. Now it is working sometimes but then it is faulty again..... :evil:
-
Welkome to the wonderful world of PCB repair :)
-
those randomly appearing errors really suck... it sounds like a connection problem... try to resolder the vias from the lines going from the cpu to the pokey... maybe this helps...
-
I'll try a basic one: did you replace the edge connector already ?
Was extremely loose on my Centi.
-
The edge connector could be a problem later on but for now it does not even work in my test setup.
The funny thing is everytime I have it fixed it is faulty again the next day.
Two days ago i reseated the roms and centi worked again. Yesterday I restarted the pcb and it was faulty again. This time reseating did not work but after 5 minutes some soft cracking noises were heard and after restarting the pcb worked.
Keep you posted.
-
Maybe it's not your PCB but your power supply that is acting weird?
To rule it out you could try to feed your PCB with a "modern switcher" so the +5 -5 and +12 are 100% reliable and see if the same problems occurs.
-
Reseating is not a real solution to the problem.
Replace the sockets and clean the legs with a pencil eraser or something similar.
I've been using my new "microsope" to check both pins from ROMS and the pins from the sockets and it's actually pretty shocking what you see and already a small miracle that it often _does_ (kina) work.
Sadly it's very hard/impossible to make pictures with through the microscope else it would be fun to show some here.
-
Sadly it's very hard/impossible to make pictures with through the microscope else it would be fun to show some here.
I did a test with my Phone straight to the micro microscope a while ago, not too bad if you ask me..
(SD card contacts)
(http://www.opdenkelder.com/pics/micro.jpg)
-
Oh yes that's a better idea....I tried my regular camera but the lens is way too bulky...
-
Maybe it's not your PCB but your power supply that is acting weird?
To rule it out you could try to feed your PCB with a "modern switcher" so the +5 -5 and +12 are 100% reliable and see if the same problems occurs.
i'm afraid so, it is allready on the test bench. With a switcher ....
All pins on roms, pokey, 6502, ect. show 'normal' output waves. The only signal that is looking strange is the pokey signal that goes in at pin 30 and the sound that comes out at pin 37.... No idea how the should look because even when working it looks weird 😩
-
Still stuck, who knows how the random number generator works?
At pin 30 and 37 i get a patern on the scope as long as the lm324 is not in it socket. With an lm324 in place pin 5,6,7 of the lm are all high at 5,17 volt.
-
LM324's are notorious for going bad.....for some strange reason especially on Centi boards.
The one on my board was kaputt too. I figured this out with the "wet finger" method ;)
Seriously: I had no sound and was touching around the LM324 (which is an amplifier (OP-AMP) chip) and suddenly I started hearing some faint sound. Replaced it and sound back.
From what you write it's pretty clear the LM324 is toast and shows a full short (you could measure on the LM) between the inputs.
If that is true and the POKEY survives getting +5V on the sound output (pin 37) it would say the POKEY is a pretty rugged design, but I can sure imagine it's not going to like it :)
You say it's socketed, so maybe at one point in time it was put in backwards and failed like that because of this.
I have a quite a few of them if you need one or two, PM me your address if you don't have'm.
-
This is what I found about POKEY's random number generation:
Pokey uses a LFSR (https://nl.wikipedia.org/wiki/Linear_feedback_shift_register) to generate random numbers.
They are repeating and predictable (good emulation reproduces it perfectly). So, not really random and even ignoring the repeating the distribution isn't necessarily random either.
Pokey random numbers repeat every 131,000 or so cycles - if 9-bit Poly noise is enabled then it's every 512 cycles.
So, it's not truelly random, but good enough for games and it was also excellent to HAVE a harcware random generator in the Atari 8 bit computers, most competitors needed to do it in software which cost valuable processing time.
Another good thing about it is that it can be emulated perfectly f.i. in MAME and Atari 8 bit emulators :)
-
Little update I AM STUCK :-\ I tested the lm324 is seems to be good, replaced it anyway and ... Did not have 324 but some 224 and 124 and still no luck.
Used a scoop and everthing seems fine except output on pin 37. So how can it be that both pokeys failed in two seperate boards and both did not just die but where failing with ups and dows, half working working inbetween ect?
Put in some pics...
-
More pics, what is what I will post later.
-
Little update, took both pokey chips to level42 for testing in his atari 800. Both pokey chips turned out faulty..... So Centi is killing my pokey's ... 😩😩😩. But why? Any suggestions, maybe to much AC noice?
Anyone any ideas?
Well I got myself some new pokey's but at 10 E a piece I don't want to feed them to Centi without knowing what is wrong ... The ghost in the machine must be exocisted. :evil: :evil: :evil:
-
Last update, in the end I replaced the pokey with a nos one and it all worked out. Don't know why the old pokeys died, but the new ones have been working great for the past 6 months.
...... The Ghost left the machine.?????