another rampack with problems

Discussions about Sinclair ZX80 and ZX81 Hardware
bola_dor
Posts: 398
Joined: Wed Oct 02, 2019 5:32 am

another rampack with problems

Post by bola_dor »

Well This is an issue 2 16Kb rampack. it is actually my first one from about 35years ago. So I wanted to resuscitate it...
It just didn't work any more about 20years ago and was in a box until now. It gave a white screen before my intervention.
20210824_213541.jpg
Ribbon cable just did tear apart when I opened it so I replaced it with the first color clash free cable I found.
Voltages on 4116 were ok and so were the 5v for the 74LS ones.. I double checked it and the continuity across the cable in between boards.
I changed the electrolytics less the 1uF ones because I didn't have any at hand.. voltages are OK anyway.
It took me about a day to desolder all RAM chips. Two of them were bad. Did put some sockets and two new old 4116 I had.
I Tested for continuity across the bus lines . No shorts either..
And this is what I've got
GIF-210824_211850.gif
So I guess the multiplexing logics are wrong?

I have a small oscilloscope that helped me before but I don't know how exactly that circuit works. I can see lines pulsing at different rates. The only ones with a different shape are pin 2-14 on the 4116s...
Any thing to do better than desolder everyone and put new ones??
Ernesto
ZX80 USA, ZX81UK, ZX Spectrum, ZX Spectrum+, ZX Spectrum 128+ UK, ZX Spectrum +2/A, Sinclair QL, CZ1000, CZ1500, CZ2000, CZ1000Plus, CZ1500Plus, CZ Spectrum, CZ Spectrum Plus, TK83, TK85, TK90X, TK95. TS2068. And more to come :D
bola_dor
Posts: 398
Joined: Wed Oct 02, 2019 5:32 am

Re: another rampack with problems

Post by bola_dor »

Made some more tests an discovered that one "newly changed" 4116 ic was bad. And +12 and -5v were down..
I lifted up all RAM chips from the sockets. voltages were still bad. So I replaced the transistor with another ZTX 750 from a scrap rampack board I have as a donor. Voltages were restored. 12v rail was above 13v unloaded, may be normal but changed the 12v zener with a new one 1W rated.
Now I have a white screen but still no K

Used the scope to whatch all lines.. all pulsing but pin 15 /CAS that is hold High near 5v in every IC,
and pins 2/14 (D in/out) of the inner IC at the shortest row (near the transistor) is kept low..
Unfortunately I didn't find any IC layout to identify the components.
Ernesto
ZX80 USA, ZX81UK, ZX Spectrum, ZX Spectrum+, ZX Spectrum 128+ UK, ZX Spectrum +2/A, Sinclair QL, CZ1000, CZ1500, CZ2000, CZ1000Plus, CZ1500Plus, CZ Spectrum, CZ Spectrum Plus, TK83, TK85, TK90X, TK95. TS2068. And more to come :D
User avatar
1024MAK
Posts: 5101
Joined: Mon Sep 26, 2011 10:56 am
Location: Looking forward to summer in Somerset, UK...

Re: another rampack with problems

Post by 1024MAK »

For a DRAM chip to respond to an access, it has to see the correct sequence of low pulses on /RAS and /CAS.
For reading or writing, a 4116 DRAM has to first see /RAS go low. As /RAS goes low, the DRAM chip latches in the address on its address pins (A0 to A6). This address comes from the output of two of the four multiplexer chips, and should be the same as the input from the computer address bus on A0 to A6.
Then the switch signal on two of four multiplexer chips should change state, hence switching the outputs of the multiplexer chips so that now they are the same as the inputs A7 to A14 on the computer address bus.
Then, /CAS should go low, and at this point, the DRAM latches in the address on its inputs (A0 to A6). Now having the complete 14 bit address, depending on the state of the /W (or /WRITE) input, the DRAM will either store the data bit on its input (D) or output the stored data on its output pin (Q).

If /RAS is pulsing low, but /CAS is not pulsing low, then what you are seeing is either a refresh cycle, or an access to an address out of range of the address decoder. That is, either the Z80 is not trying to read or write to the RAM. Or the RAM circuitry is not recognising a valid address.

Unfortunately with no working RAM, we don’t know what the Z80 is actually doing.

Mark
ZX81 Variations
ZX81 Chip Pin-outs
ZX81 Video Transistor Buffer Amp

:!: Standby alert :!:
There are four lights!
Step up to red alert. Sir, are you absolutely sure? It does mean changing the bulb :!:
Looking forward to summer later in the year.
bola_dor
Posts: 398
Joined: Wed Oct 02, 2019 5:32 am

Re: another rampack with problems

Post by bola_dor »

1024MAK wrote: Wed Aug 25, 2021 10:34 am For a DRAM chip to respond to an access, it has to see the correct sequence of low pulses on /RAS and /CAS.
For reading or writing, a 4116 DRAM has to first see /RAS go low. As /RAS goes low, the DRAM chip latches in the address on its address pins (A0 to A6). This address comes from the output of two of the four multiplexer chips, and should be the same as the input from the computer address bus on A0 to A6.
Then the switch signal on two of four multiplexer chips should change state, hence switching the outputs of the multiplexer chips so that now they are the same as the inputs A7 to A14 on the computer address bus.
Then, /CAS should go low, and at this point, the DRAM latches in the address on its inputs (A0 to A6). Now having the complete 14 bit address, depending on the state of the /W (or /WRITE) input, the DRAM will either store the data bit on its input (D) or output the stored data on its output pin (Q).

If /RAS is pulsing low, but /CAS is not pulsing low, then what you are seeing is either a refresh cycle, or an access to an address out of range of the address decoder. That is, either the Z80 is not trying to read or write to the RAM. Or the RAM circuitry is not recognising a valid address.

Unfortunately with no working RAM, we don’t know what the Z80 is actually doing.

Mark
Thank you Mark for that detailed explanation.

Then I think that if I have a D line stuck low in one IC I have a problem either at that IC or with some connection, unless that bit is always low.

/RAS is pulsing but /CAS isn't...
Is there any possibility that the faulty rampack is preventing Z80 to try to read memory.. It should be trying to read the display file at least if still executing the rom..
/CAS comes from IC 1, Should I replace it?
Ernesto
ZX80 USA, ZX81UK, ZX Spectrum, ZX Spectrum+, ZX Spectrum 128+ UK, ZX Spectrum +2/A, Sinclair QL, CZ1000, CZ1500, CZ2000, CZ1000Plus, CZ1500Plus, CZ Spectrum, CZ Spectrum Plus, TK83, TK85, TK90X, TK95. TS2068. And more to come :D
User avatar
1024MAK
Posts: 5101
Joined: Mon Sep 26, 2011 10:56 am
Location: Looking forward to summer in Somerset, UK...

Re: another rampack with problems

Post by 1024MAK »

bola_dor wrote: Wed Aug 25, 2021 6:28 pm It should be trying to read the display file at least if still executing the rom..
Your assuming that the Z80 is actually running the ROM program. But the ROM program assumes that the RAM is working…
If the RAM is not working, there is next to no chance that the Z80 will get as far as trying to run the display.

Mark
ZX81 Variations
ZX81 Chip Pin-outs
ZX81 Video Transistor Buffer Amp

:!: Standby alert :!:
There are four lights!
Step up to red alert. Sir, are you absolutely sure? It does mean changing the bulb :!:
Looking forward to summer later in the year.
bola_dor
Posts: 398
Joined: Wed Oct 02, 2019 5:32 am

Re: another rampack with problems

Post by bola_dor »

1024MAK wrote: Wed Aug 25, 2021 6:46 pm Your assuming that the Z80 is actually running the ROM program. But the ROM program assumes that the RAM is working…
If the RAM is not working, there is next to no chance that the Z80 will get as far as trying to run the display.

Mark
😥😥😢
Ok.. I tested all lines again. Discovered two cable lines broken..
Now all control pins and Addresses are pulsing.
But got the thick bars blinking screen again..
Voltages are OK
Retested every 4116 and still OK.
Ernesto
ZX80 USA, ZX81UK, ZX Spectrum, ZX Spectrum+, ZX Spectrum 128+ UK, ZX Spectrum +2/A, Sinclair QL, CZ1000, CZ1500, CZ2000, CZ1000Plus, CZ1500Plus, CZ Spectrum, CZ Spectrum Plus, TK83, TK85, TK90X, TK95. TS2068. And more to come :D
User avatar
1024MAK
Posts: 5101
Joined: Mon Sep 26, 2011 10:56 am
Location: Looking forward to summer in Somerset, UK...

Re: another rampack with problems

Post by 1024MAK »

You may want to have a read through of this topic ;)

Mark
ZX81 Variations
ZX81 Chip Pin-outs
ZX81 Video Transistor Buffer Amp

:!: Standby alert :!:
There are four lights!
Step up to red alert. Sir, are you absolutely sure? It does mean changing the bulb :!:
Looking forward to summer later in the year.
bola_dor
Posts: 398
Joined: Wed Oct 02, 2019 5:32 am

Re: another rampack with problems

Post by bola_dor »

Finally got it working.. was hard....
Desoldered all RAM ICs and put sockets two 4116 were bad.. but then one of the new ones failed and took with him the transistor.
Cable is a nightmare.. the ribbon as the original brakes continously. Used a bunch of mono wire because it's easier to desolder and keep clean the hole to resolder. It brakes at the solder point anyway...
I desoldered most of the multiplexor ICs then discovered A1 at the 4116 side was stuck high.. then lifted all 4116 with no change so I lifted pin 12 from IC 6 (were A1 comes from) and that didn't solve it either.. Finally I suspected a short underneath the RAM sockets, A1 and +5v had continuity and them runs side by side across the 8 RAM chips. I didn't want to desolder all of them again so I smeared some flux at the sockets side and heated up pins 7 and 9 of every IC hoping it get any solder back to the pads at the other side.. It worked.. !!!!!
20210901_031208.jpg
Teaching point: small and silly things makes you fail.. that Cable is there to be hated..

Finally I wanted to clear the jailbars
20210901_015831.jpg
I tried a 1uF (105) multilayer ceramic capacitor on the 12v rail as suggested for the ZX Spectrum low RAM .. But didn't help, anything.
So I tried on 5v rail and it worked!! (again)
20210901_025131.jpg
I did solder two on top of the existing ones and a third at the back of the board ..
20210901_032028.jpg
20210901_031936.jpg
Finally (two) I want to make the WRX convertion. But that will be tomorrow....😴
Ernesto
ZX80 USA, ZX81UK, ZX Spectrum, ZX Spectrum+, ZX Spectrum 128+ UK, ZX Spectrum +2/A, Sinclair QL, CZ1000, CZ1500, CZ2000, CZ1000Plus, CZ1500Plus, CZ Spectrum, CZ Spectrum Plus, TK83, TK85, TK90X, TK95. TS2068. And more to come :D
User avatar
mrtinb
Posts: 1906
Joined: Fri Nov 06, 2015 5:44 pm
Location: Denmark
Contact:

Re: another rampack with problems

Post by mrtinb »

Great job, and great reference for future fixes.
Martin
https://zx.rtin.be
ZX81, Lambda 8300, Commodore 64, Mac G4 Cube
User avatar
1024MAK
Posts: 5101
Joined: Mon Sep 26, 2011 10:56 am
Location: Looking forward to summer in Somerset, UK...

Re: another rampack with problems

Post by 1024MAK »

bola_dor wrote: Wed Sep 01, 2021 8:26 am Finally I suspected a short underneath the RAM sockets, A1 and +5v had continuity and them runs side by side across the 8 RAM chips. I didn't want to desolder all of them again so I smeared some flux at the sockets side and heated up pins 7 and 9 of every IC hoping it get any solder back to the pads at the other side.. It worked.. !!!!!
Excellent bit of thinking there, I like 👍 8-) :ugeek:
bola_dor wrote: Wed Sep 01, 2021 8:26 am Teaching point: small and silly things makes you fail.. that Cable is there to be hated..
Yep, many others have had similar problems.
bola_dor wrote: Wed Sep 01, 2021 8:26 am Finally I wanted to clear the jailbars.
I tried a 1uF (105) multilayer ceramic capacitor on the 12v rail as suggested for the ZX Spectrum low RAM .. But didn't help, anything.
So I tried on 5v rail and it worked!! (again)
Well done on both the repair and on finding a fix for the jailbars :D 8-)

Mark
ZX81 Variations
ZX81 Chip Pin-outs
ZX81 Video Transistor Buffer Amp

:!: Standby alert :!:
There are four lights!
Step up to red alert. Sir, are you absolutely sure? It does mean changing the bulb :!:
Looking forward to summer later in the year.
Post Reply