All RAM chips destroyed??

Nagging hardware related question? Post here!
tcat
Super Gold Card
Posts: 633
Joined: Fri Jan 18, 2013 5:27 pm
Location: Prague, Czech Republic

Re: All RAM chips destroyed??

Post by tcat »

Hi Martyn,
those JUx represent wire-links/zero-ohm resistors on the Iss-5 boards...They appear immediately between the ZX8302 and the right-hand ROM socket.
QL Iss#5 Main Bord - JUx links
QL Iss#5 Main Bord - JUx links
In the picture above I only located JU1,2,6 in the red rectangle, but manual refers to JU2,3,4. Is it just me looking at the wrong place, or a manual misreference?

Thank you.
Tomas


RWAP
RWAP Master
Posts: 2834
Joined: Sun Nov 28, 2010 4:51 pm
Location: Stone, United Kingdom
Contact:

Re: All RAM chips destroyed??

Post by RWAP »

tcat wrote:
In the picture above I only located JU1,2,6 in the red rectangle, but manual refers to JU2,3,4. Is it just me looking at the wrong place, or a manual misreference?

Thank you.
Tomas
No - the picture shows JU1 to JU6 - you can only read JU1, JU5 and JU6 because the writing is not obscured by the resistors...


martyn_hill
Aurora
Posts: 909
Joined: Sat Oct 25, 2014 9:53 am

Re: All RAM chips destroyed??

Post by martyn_hill »

Hi again Tomas

As Rich said, the other positions of optional links are obscured by those three zero-ohm 'resistors' (note the single black-band == zero ohms).

It apparently proves easier to automate the component assembly process using zero-ohm resistors rather than actual bare wires...

The links in place in the picture are the appropriate configuration to match the ROMs.


User avatar
Cristian
Aurora
Posts: 960
Joined: Mon Feb 16, 2015 1:40 pm
Location: Veneto

Re: All RAM chips destroyed??

Post by Cristian »

---Repair progress update---
The MC1377P RGB to PAL/NTSC (highlighted in yellow) has been swapped/tested with another QL (iss.6). It seems to work normally. The working MC1377P from the functional QL does not affect the behaviour of the faulty QL.
Even if the MC1377P still gets warmer in the faulty QL, we can say that the problem must be elsewhere.
repair-status.jpg


User avatar
Cristian
Aurora
Posts: 960
Joined: Mon Feb 16, 2015 1:40 pm
Location: Veneto

Re: All RAM chips destroyed? - Repair progress

Post by Cristian »

--Repair progress update--
I made all the "mandatory modifications" (for Issue5 board) as per Service Manual.
There might be some little improvement in the startup sequence (or probably it's my imagination), but the substance remain the same.
The machine sometimes try to start, wit the the usual pixel pattern. No Minerva RAM error codes. Garbage slowly increase on screen, the F1/F2 menu slowly appears, then automatically red screen and cursor appear and writes (in #0) various characters (please note that the keyboard is not connected). Usually the same characters are repeated continuosly, as shown in the attached videos.

Here are the video links. Sorry for bad/blurred video quality: in reality the images and text are sharp.
https://streamable.com/ymtrxb
https://streamable.com/cfwqgf
https://streamable.com/t7q2r0

Image


User avatar
Cristian
Aurora
Posts: 960
Joined: Mon Feb 16, 2015 1:40 pm
Location: Veneto

Re: All RAM chips destroyed??

Post by Cristian »

Just tested also IC23 (ZX8302). When connected to other working machines, all seems OK. I've updated the summarizing picture in my previous post.
I still get random results at power on (mostly garbage patterns, usually vaguely regular). Sometimes QL tries slowly a real startup, as shown in the videos in my previous post.
The question is: is Minerva always able to detect all faulty RAM chips or not?
I tested / renewed many components, I don't know what else should I try. Replacing quartz oscillators? Any suggestion please?


User avatar
Pr0f
QL Wafer Drive
Posts: 1298
Joined: Thu Oct 12, 2017 9:54 am

Re: All RAM chips destroyed??

Post by Pr0f »

Chips getting warmer than usual is often a sign of "loading" somewhere on the pins - solder bridges, failed components like capacitors - some of these are not mechanically strong, and if stressed or bent can cause a short or low resistance - not what you want from a capacitor, resistors tend to fail open circuit. Some semiconductors can also exhibit odd behaviors if their input protective dioes are compromised by static discharge or higher than normal voltages at pins above VCC supply level. However, if that was the case the same component would behave the same way in another QL - so the MC1377P should be ok if it runs cooler in another QL.


User avatar
tofro
Font of All Knowledge
Posts: 2685
Joined: Sun Feb 13, 2011 10:53 pm
Location: SW Germany

Re: All RAM chips destroyed??

Post by tofro »

Watched your videos. The start-up and memory test looks severely slowed down to me.

Any slowing down hints towards a number of things the computer would be more busy with during execution of ROM code:
  • Something wrong with the 15MHz clock input of the 8302. Note this has a trimming capacitor (TC1) that could be severely de-adjusted. A scope would be handy to check the signal.
  • Delayed DTACK (memory would take longer to answer CPU requests and could slow down the CPU). DTACK is being generated by the 8301. A scope would be handy here as well to check the signal.
  • CPU busy with handling spurious interrupts from various sources. If a thorough visual inspection of the board for possible shorts to the INT line reveals nothing, a source for such errors could be the expansion port connector - If a bent pin in here - which is easy to have and very hard to see by staring into the connector... - connects to something bad it could easily create such effects.


ʎɐqǝ ɯoɹɟ ǝq oʇ ƃuᴉoƃ ʇou sᴉ pɹɐoqʎǝʞ ʇxǝu ʎɯ 'ɹɐǝp ɥO
User avatar
Cristian
Aurora
Posts: 960
Joined: Mon Feb 16, 2015 1:40 pm
Location: Veneto

Re: All RAM chips destroyed??

Post by Cristian »

First of all, thank you very much Pr0f and Tobias for your important suggestions.
Unfortunately I have no access to oscilloscopes, so for the moment I can not test the trimming capacitor. Regarding the expansion port, the pins look perfect.
tofro wrote:The start-up and memory test looks severely slowed down to me.
Yes, I confirm. The few times it occurs, the startup is extremely slow.
tofro wrote: DTACK is being generated by the 8301.
Even if it warms up somewhat, when installed in other functional machines it works apparently perfectly and remains almost cool. We can say the problem should come from elsewhere.
tofro wrote:CPU busy with handling spurious interrupts from various sources. If a thorough visual inspection of the board for possible shorts to the INT line
Could someone explain where exactly should I look for the INT line please?


User avatar
Pr0f
QL Wafer Drive
Posts: 1298
Joined: Thu Oct 12, 2017 9:54 am

Re: All RAM chips destroyed??

Post by Pr0f »

I think the pin Tobias is talking about is the EXTINTL pin - it goes from the edge connector to the ZX8302 chip, which in turn will generate an interrupt to the CPU if this line is tiggerred. QIMI mouse interface uses it, but I am guessing from the photos that this is not connected.

If you examine the edge connector through slot in the QL case on the left hand side (remove flap if it's still there), check to see all the pins are straight, and none have been bent or are touching each other - if a pin was slightly bent, it could become bent further by plugging in an expansion card, causing it to short out against neighbouring pins. It is on pin 30b - so 3rd pin in from the back of the case, top row

The next cause of action after that is to check this pin on the ZX8302 chip to 5v and GND and see if there is any short circuit there. This is Pin 2 on the chip - so second one down from the dot / notch at the top of the chip. There is a resistor between this pin and 5V line - value is 3.3K, so there should be some resistance when measuring to 5V.


Post Reply