Apollo 1240 25MHz - random/negative RAM values

[apnovy]

Good evening.
First post - it's a pleasure to read and learn from you. Thank you.


Most of the time: after power ON black screen and LED falshing or just stays ON, randomly able to load WB3.0 (from floppy) but chip RAM and fast RAM showing random values, sometimes negative, it's changing every 3-10sec.


Hardware was 20 years off. A1200 1D.4; Apollo 1240 25MHz - original state, no mods (4MB or 16MB which was working 20 years ago)

I have checked all basic things (cleaning, visual inspection Apollo, removed battery - no sign of leakage, new ATX PSU, reseated A1200 ROMs, reseated EPROM, new Xtal 50MHz, measured main power +12V, -11.6V, 4.91V without Apollo, +12V -11.6V, 4.84V with Apollo, timing fix done: E123C and E125C removed, Apollo: new 47uF 16V, reseated MACH131 and 210).

Apollo with RAM - random values in WB:
https://mega.nz/file/msNhTSjL#au3oJ4..._cEy869RIlMaOw

https://mega.nz/file/SlFjUCyT#3SEFUu...CYf6ffnnNe9ZfM

https://mega.nz/file/HgNTxI7Z#TWFgXt...9che027zrxUShg

RAM removed:
https://mega.nz/file/CoFHySyD#4VgK8Y...3z1LO-FJYMMPZc

https://mega.nz/file/XxE3CARK#kHrrEo..._McE63Cx5_xRpI

Without Apollo - A1200 works without any issues.
Without RAM in Apollo - A1200 works without any issues (but the pain is unbearable).

What could be the cause of this Apollo strange behavior?


Cheers!

[daxb]

If the problem only happens when RAM is added to your Apollo card then I would try other RAM. Or at least run RAM test tools.

[pandy71]

If possible try to use alternate power supply - it may be issue not necessarily with too low voltage but too high current ripple (some DMM's may be incapable to correctly measure momentary voltage drops / sags). It would be good to verify power quality with some oscilloscope. Clean, stable power lines are key to correctly working system.

[apnovy]

Thanks for replies.

daxb: I'm not able to run RAM test on Amiga. 95% of the time all I have is "black screen".. I think I was "lucky" to see random RAM values. I'll try to find some other old hardware to test RAM (even they worked with this Apollo 20y ago it's good idea to verify )

pandy71: I believe my PSU is OK but I'll test it (and try other PSU bought in 2021).
Unfortunately I don't have expensive tools to make all the tests I'd like

BTW Made some basic measurements while "black screen". One of 74F257AD has 4,84V on pin 1,2,3,5,6,10,11 all the time. Other three, closer to OKI chip, have 3,85V on pin 1 "s", 0V on pin 9,10,11,12,13,14,15. All four have: pin 16 "vcc" 4,84V -ok, pin 8 "GND" 0V -ok, pin 15 "n OE" 0V -ok.
Time to read 74F257AD specs..

What is typical Apollo behavior when one of 74F257AD or MACH210 is damaged?
Would it be possible for Apollo to run (without RAM) and damaged 74F257AD at all?

[daxb]

Does the early startup screen (hold both mouse buttons on startup) work? If so, can you start without startup-sequence and is the system stable while showing the CLI window? IF yes you can run RAM test tools.

[apnovy]

Thanks.
Not possible to enter early startup screen.

Following pandy71 suggestion to use second PSU - still black screen.

A1200 with Apollo, RAM in Apollo, powered from 2nd PSU (decent ATX from 2021, Amiga in "black screen mode").

1) without additional wire from PSU to FDD pins on mainboard:
mainboard socket: 12.02V, -12.03V, 4.81V.
mainboard FDD pins: 12.00V, 4.76V.
edge expansion port: 12.00V, -11.99V, 4.76V

2) WITH additional wire from PSU to FDD pins on mainboard:
mainboard socket: 12.07V, -11.99V, 4.93V.
mainboard FDD pins: 12.06V, 4.90V.
edge expansion port: 12.06V, -11.99V, 4.90V

There are no signs of cap leakage at all, no smell, clean shiny capacitor soldering points but most of caps were never replaced. Few days ago I replaced two 470uF 16V because they were a little bit swollen on top.

I need to make more tests..
Maybe mainboard needs full recap. I know it wouldn't hurt but I'd like to make more tests, learn and read.. I still think it's the Apollo issue.

If someone is sure I need to do full recap please write "Don't think, RECAP!"

About my previous questions -> <-

[apnovy]

pandy71 - thank you, even if it's still far from "solved". I'm learning..
For example I measured
C236 has 0.65V on "+" and 0.04V on "-".
C239 has 4.78V on "+" and 0.00V on "-",
C235 has 2.45V on "+" and 2.40V on "-",
C459 has 1.98V on "+" and 0.00V on "-".
C214 has 0.28V on "+" and 2.44V on "-".
C409 has -0.19V on "+" and 0.65V on "-" (cap relaced a few days ago)

Other thing I noticed without Apollo and FDD (only RGB and keyboard) from PSU power ON to Amiga welcome screen takes 16 sec
With Apollo was like 2-3 sec.

Cheers!

[apnovy]

Firstly thank you all for sharing knowledge!
Sorry for reviving old threads.
Some spare time - at last.

Problem - wrong RAM detection. With any RAM in Apollo - most of the time black screen, Power LED blinking in loop, Couldn't get to boot-up screen or early startup screen.
Without RAM Apollo 1240/25MHz is correctly detected, WB is working.
https://www.flickr.com/photos/199714...posted-public/

Solution.
I've found two resistors 33ohm with cold solder (before desoldering: 147ohm and 112ohm). After desoldering both 33ohm. I resoldered all 33ohm resistors and Apollo 1240/25 is working with RAM for 2 days now.
https://www.flickr.com/photos/199714...posted-public/

I've also repaced 47uF 16V capacitor.
https://www.flickr.com/photos/199714...posted-public/

I also made TimingFix and standard recap for Amiga 1200 (1D4)