16K RAM Pack Schematic / Chip Layout

[TMAOne]

Ian's blog & progress report.

So Pack #2 has a short. Even I should be able to find & fix that,...

Searched with the continuity test on my meter, comparing beeps with another reference board when I wasn't sure. Couldn't find the source, but pins 8 and 9 are clearly shorted.

Only one other place to look. Under the damn RAM chip sockets I put in there close to 30 years ago. I assume it was me that did it. I don't have anyone else who can take the blame, much as I'd appreciate having someone else take the bullet for me.

Started unsoldering and pulling sockets. What a mess! There were many, many circuit traces missing under the sockets.

Some history: My electronics is all self-taught. So much so that in 1982 when I wanted dearly to assemble this newfangled computer kit I was excited to bring home, I ran out and bought a soldering GUN. Don't laugh--I didn't know any better!

That ZX81 still works, knock wood. But the original RAM pack died at some point.

So some time during my Dark Ages I must have attacked this old RAM pack with a pick axe and hand grenades. I don't know what I was thinking. Did I not NOTICE that half the runners came up with the chips when I pried them out? Aparently not.

Anyway, I found the short. It was an almost invisible solder blob bridging the runners of pins 8 and 9, as I expected thanks to you guys.

I also mapped out a total of 17 jumpers that will be necessary when and if I still try to rescue this pack. I think little wires into the chip sockets along with the chip legs would work okay.

So where do I stand? Well, pin 9 is now a perfect +4.99V. Pin 8 is +10.44V, not great but probably nominal.

Pin 1 is still at -0.7V, which seems to be the noise level on my meter.

So I replaced or swapped everything below the horizontal ground line as seen in the diagram, i.e. the 5.1V Zener (swap with a donor board), 1uF cap x2 (swap), 1N4148 x2 (got plenty of those) and replaced the resistor. Still no -5V.

I think I have to wait for my transistors and other reliable parts. The transistors could be a common point of failure affecting both of my still errant packs.

So this project is officially on the shelf for some weeks.


My wife will be happy if I come to bed at night for a while instead of hunching over the electronics bench. She's been good about it really, occasionally throwing scraps of food down the stairs with a picture or two of the children.

I don't know why I'm being so dogged about this. I have one working pack, #3, which only looks bad on the Zeddy I'm testing with because of it's bad ULA. So I need another Timex from somebody's basement, not another pack.

Anyway, cheers,
Ian

[sirmorris]

She's been good about it really, occasionally throwing scraps of food down the stairs with a picture or two of the children

[1024MAK]

My electronics is all self-taught. So much so that in 1982 when I wanted dearly to assemble this newfangled computer kit I was excited to bring home, I ran out and bought a soldering GUN. Don't laugh--I didn't know any better!

Don't worry. A lot of people are self-taught. And even those people who have been educated in the art of soldering have made plenty of mistakes along the way. I've see other people make a right mess of a job, and I've done it
The most fun one that I was responsible for, was when I connected up a 12V dc telephone unit at work to a 50V dc supply (we had both 12V dc and 50V dc type units and both types used the same type of dc supply lead & DIN socket). Needless to say, the logic chips did not enjoy the extra voltage, and two blew their tops off! . They all had to be replaced . Luckly me, there was a spare, so I just got the stores system to deliver some new chips, and no one was any the wiser

I don't know why I'm being so dogged about this. I have one working pack, #3, which only looks bad on the Zeddy I'm testing with because of it's bad ULA. So I need another Timex from somebody's basement, not another pack.

Because it's fun! It's even more fun when you hit a snag, bang your head against the nearest wall in frustration. Go to bed. Wake up in the middle of the night with an idea. Then when you get a chance, go with your "new" idea, and find that you have solved the problem...

Keep it up, how else are we to enjoy your adventures

Mark

[1024MAK]

She's been good about it really, occasionally throwing scraps of food down the stairs with a picture or two of the children

ROTFL

[TMAOne]

You're right Mark; it is fun, and that's exactly why I'm doing it. No other reason. For close to three decades I've been annoyed by the fact that my RAM pack had died, and I didn't know how to go about fixing it. Now thanks to the internet and you helpful guys I'm looking to make my first "digital repair". My soldering skills are pretty good now. De-soldering too. I can pull chips out with only a minor amount of brute force and usually stand a 99% chance of saving the board and the chip. Let's see, 8 chips x 16 pins x 2 boards,... 256 pulls with the de-soldering bulb--a little daunting but I can do it while there's a football game on my little shop black & white. Uhm, that would be CFL, (Canadian Football League), which is not "real" football but rather a game similar to the American game. Ours is better. The players aren't, but the game is.

both types used the same type of dc supply lead & DIN socket

I can see where that would be an easy thing to do. My aforementioned wife plugged the supply from a cordless phone into my bedroom Sangean ATS-909 shortwave receiver, and did not understand why I was upset. "What's the difference? It fit." I had some explaining, but fortunately no repairs, to do to set that right.

Anyway, back to my pet software project which uses my 32K modified ZX81. I'll pick up this thread when my parts come from China and I can fix one or both packs.
Ian

[XorA]

I can pull chips out with only a minor amount of brute force and usually stand a 99% chance of saving the board and the chip. Let's see, 8 chips x 16 pins x 2 boards,... 256 pulls with the de-soldering bulb

If the chips are bin jobs its much easier and safer to snip all the pins with side cutters and use tweezers to pull out the pins individually then clean the holes.

[TMAOne]

Well it's been a while since I've heated up the soldering iron on these errant RAM packs of mine. I got back to it for an evening, so it's time for an update.

To review, I have 1 good pack, and 3 packs in various states of post meatball surgery. Only one of those is in a state even to be powered up. It has no -5V. (Of the other two, one is realistically hopeless, and the other has been stripped of parts and has no ribbon bridge. I may, if buoyed by any success, try to hybrid those two and see what I get, but that's a stretch for some day in the future.) I've been concentrating on the one with no -5V, which also has had at least one of the 4116's fried. If I can fix the voltage, I would undertake to remove, socket, and replace the RAMs with replacements I have available.

So what did I do? Why, I bought another pack on eBay of course. Bad news though, it's defective too. (I got a complete refund.) So what I have to work with now is a "virgin" bum pack. Guess what? It has no -5V. Sounds familiar, that.

Okay, new plan for my limited diagnostic skills. Take apart my one and only precious working RAM pack, and measure voltages at all the points in the "power section". Then do the same with the newest defective pack, and note where the -5V seems to be going astray. They fortunately have the same layout. (Not all my packs do.) The following photo shows the points at which I took measurements.

Working Voltages_2.jpg

(878.78 KiB) Downloaded 2292 times

Here are the measurements I got:

Point___Working Unit_______Bum Unit
_1________-11.620__________-10.770
_2___________.005____________.004
_3_________-3.350___________-1.940 !
_4__________9.420___________9.640
_5_________-4.930____________.050 !
_6___________.004____________.004
_7__________9.280___________9.330
_8___________.000____________.000
_9__________9.390___________9.610
10___________.000____________.000
11___________.000____________.000
12__________9.260___________9.330
13_________12.110__________11.020
14__________8.910___________9.200
15__________9.410___________9.500
16_________12.110__________11.060
17__________8.900___________9.260
18__________0.450___________0.260
19___________.004____________.004
20___________.004____________.004
21__________9.400___________9.510

"!" denotes the ones I thought were significantly different.

Here's where my lack of training, knowledge, or intelligence breaks down,... I'm not good at relating the physical parts to the schematic, and I'm quite lost at understanding how the circuit is supposed to work. It looked to me like the 5.1V zener might be the bad guy, so I replaced it with a new one. (I have new parts to work with now.) No dice, measurements unchanged.

Here is a copy of the schematic and a photo of the component side of the bad unit, for reference.

TimexSinclair1016_Schematics.png

(219.07 KiB) Downloaded 2292 times

HPIM1349_2.jpg

(935.66 KiB) Downloaded 2292 times

Does anyone know what I should try to replace next?
Ian

[RetroTechie]

Here's where my lack of training, knowledge, or intelligence breaks down,... I'm not good at relating the physical parts to the schematic, and I'm quite lost at understanding how the circuit is supposed to work.

Hmm... edge connector signals are named, you can find in ZX81 schematic / docs where on the edge connector each signal is, lines represent connections, and IC pin numbering is standardized. Power supply points are easy to locate too. From there it's simply a matter of following circuit board traces (and perhaps take notes / make drawing of component locations). Timeconsuming, yes. Difficult, no.

Personally I wouldn't bother powering up a RAM pack while power circuitry is in unknown or non-working state. Because it's easy to kill the RAM chips that way, and they are much more difficult to find & replace.

Instead, why not isolate power circuit from logic + RAM chips, and hook up to an external supply? A PC supply is probably not suitable for this purpose, if you don't have a +12 / -5V supply around (take +5V from ZX81), just put something together quick & dirty using an old transformer, bridge rectifier, a few capacitors, and a 7812 / 7905.

If there's any electrolytic capacitors in the logic + RAM part you isolated, and it's still original: replace unseen (with new bought ones, not ripped from other RAM packs or equipment!). Then test on ZX81 to see if RAM chips still work. If not, problem is with a RAM chip (likely), or logic IC (unlikely but possible). Unsolder & swap with known working ones until RAM pack works again. Test the ripped RAM chips in a socketed RAM pack (or a ZX Spectrum!) to find out which ones were bad & which ones are still okay.

Then solder some resistor(s) as dummy load onto that RAM pack's power circuity. Test that, and if not okay: why not replace all electrolytic capacitors + semiconductors (in the power circuit) in one go? It's not many components, and all cheap parts & easily available parts. For electrolytic capacitors, that's likely the right thing to do anyway. Diodes cost pennies, and for the transistors (oh wait, just one!) it should be easy to find a replacement type.

Then when RAM is working and the power circuitry works normal again: remove those dummy load resistors, re-connect power circuit to RAM + logic part, done.

One a side note: why go through all that trouble when internal RAM expansion is so easy? (and avoids RAM pack wobble). Just save those housings + boards as collectors item or for some mass storage, replacement SRAM project or something.

[1024MAK]

Here's where my lack of training, knowledge, or intelligence breaks down,... I'm not good at relating the physical parts to the schematic, and I'm quite lost at understanding how the circuit is supposed to work. It looked to me like the 5.1V zener might be the bad guy, so I replaced it with a new one. (I have new parts to work with now.) No dice, measurements unchanged.

<snip>

Does anyone know what I should try to replace next?

I agree with the comments about isolating the power supply circuit from the DRAM chips.

But first, search the internet, or use a library if you have one to learn about some basic electronics. You do not need to read and understand all the maths and theory, just learn about the different types of component and get an idea what they look like, the symbols and figures used on schematic diagrams, and their basic functions.

A dummy load is simply a resistor that consumes power. So using Ohms law, to consume 5mA of current at 5V:

5V / 5mA = 5V / 0.005A = 1000 ohms = 1kΩ.

Note k means kilo, m means milli and μ means micro. Resistance is measured in ohms (Ω), voltage is measured in Volts (V) and current is measured in Amps (A). Power is measured in Watts (W). These are all SI units.

1mA is 1/1000 of an Amp, so 1mA = 0.001A
1μA is 1/1000 of 1mA, so 1μA = 0.000001A

Mostek MK4116 DRAM (MAK).pdf

(25.34 KiB) Downloaded 246 times

So using the typical figures, the +12V line (VDD) needs to be able to supply 8 x 27mA = 216mA. So using Ohms law again, 12 / 216mA = 55.55Ω. However, this value is not available, so use the nearest standard value, which is 56Ω. This resistor will dissipate 2.5W so it needs to be a 3W type.
The +5V (VCC) line, we are not interested in, because this comes from the 5V regulator inside the ZX81.
The -5V (VBB) line needs to be able to supply 8 x 50μA = 400μA. So 5V / 400μA = 12.5kΩ. Again use the nearest standard value, which is 12kΩ.

TimexSinclair1016_Schematics Power ccts.png (37.32 KiB) Viewed 5550 times

I suggest you desolder and remove the following components:
The two 22μF electrolytic capacitors,
the 100μF electrolytic capacitor,
and the two 1μF electrolytic capacitors.
Then desolder and remove the ZTX750 transistor.
Using a meter, you can now test a lot of the remaining components using the resistance range for the resistors and the diode test range for the diodes (remember to test each diode both ways round).

If no additional faulty components are found, fit brand new replacement electrolytic capacitors and a new ZTX750 transistor or a new ZTX751 transistor (these are still sold by electronics companies).

Then test the power supply section

Mark

[Paul]

Hi Mark.
Thank you very much about this detailed instruction. Up to now I had not bothered repairing my faulty 16K RAMpacks. But with this help even I might be able to get it done. I think I will give a try within the next weeks.
Specially what to desolder in order to do easy measurement on the remaining parts is very good.
Kind regards
Paul.