Difference between 48k and 48k+?

[PaulJ]

Is there a difference between the motherboards of the 48k rubber model and the 48k plus to handle the plus specific keys (edit, delete, quotes etc) or is it the membrane?

I have a fully working plus apart from the specific keys so wondering if the donator has swapped the boards and just added the reset switch.

[PeterJ]

The motherboards are the same [mention]PaulJ[/mention] so the keyboard should work fine.

https://www.dataserve-retro.co.uk/conte ... k/d65.html

The main boards in the Spectrum+ are the same as those in the rubber key version from Issue 3B onwards, so all of the issues affecting the earlier spectrum are still present in the Spectrum+.

I remember Sinclair did upgrade kits too, for the rubber key version. Thinking back that was a very good deal!

For information, if you read the upgrade instructions below there is talk of a 22K resistor (which I assume is linked to the reset switch). The Spectrum+ also looks like it had a larger heatsink.




https://drive.google.com/file/d/1DtHh1_ ... sp=sharing

[Ast A. Moore]

None whatsoever. The keyboard connectors are identical all the way through to the Spectrum 128K. It is the membrane construction that handles the additional keys by combining them via an extra layer and a modified conductive-paint track layout.

[PaulJ]

mm thanks...
Will have to dig a little deeper then

[Ast A. Moore]

there is talk of a 22K resistor (which I assume is linked to the reset switch).

It actually is meant to reduce the value of a pull-up resistor in one of the keyboard lines. Apparently, the resistance of the additional tracks of the new membrane was a little too high in some cases, so Sinclair provided this just-in-case solution. I personally don’t remember ever seeing a factory board that had a resistor of a value that was anything other that 10K in that position.

[PeterJ]

Thanks [mention]Ast A. Moore[/mention]. I love reading through these old documents!

So could this be linked to Pauls issue by any chance?

[Ast A. Moore]

I love reading through these old documents!

They’re a great read, aren’t they?

So could this be linked to Pauls issue by any chance?

Quite possible, although first and foremost I’d suspect mechanical damage in the membrane itself.

[1024MAK]

Yes and no...

Plus machines as produced tended to use issue 4 boards and above. But throughout all the 16k/48k/+ machines and issue 1 to 6a boards, the keyboard connector wiring was the same. Hence a user or service centre could take any issue board and transfer it to a plus keyboard and case.

If the board has been used in a plus case, there is likely to be the remains of solder on the legs of capacitor C27, as this is the most common place where the reset switch wiring connected to the board.

Yes, Sinclair included a resistor with the upgrade. The apparent issue was that with some upgraded machines, the user could not get the “STOP” keyword. This is the issue that the extra resistor was intended to fix.

What is the problem you are having, which keys are not working?

Mark

[PaulJ]

What is the problem you are having, which keys are not working?

Mark

The newly added single keys specific to the Plus, so things like EDIT, DELETE, quotes, full stop.
They are (or were.. see below) working fine when using SHIFT 0 for delete, but not the single DELETE key.

I have 2 new membranes and tried both in there, and tried others as well.
A few hours ago I stripped it down, shuffled the key may and membrane around.. re-assembled and now it worse
It seems very fussy.. after about 15 strip/assembles.. some keys worked, others didn't.. and they changed each time!
I had to take a breather or the whole thing would have gone out of the window

[Joefish]

I had a similar problem after replacing the membrane with a new one. The original Spectrum+ upgrade we did didn't need the extra resistor; all the keys still worked. But with the new membrane in I had to find a copy of the old Sinclair upgrade instructions and fit it to get all the keys to work.

The extra functional keys are just wired in the membrane so they trigger two other keys at once. This meant some parts of the circuit had longer tracks than before and messed with the delicate balance of power levels going on in the switch grid. Hence sometimes the need for an extra resistor to adjust things.

[Sparky]

Are the membranes you can buy different for the 48 and the plus?

[PeterJ]

£1.25 on eBay. [mention]PaulJ[/mention] with free postage.

https://www.ebay.co.uk/itm/293825852316

[PeterJ]

Are the membranes you can buy different for the 48 and the plus?

Yes. The + has extra keys. The Spectrum + and 128 have the same membrane. The rubber key Spectrum one is different.

https://www.dataserve-retro.co.uk/conte ... /p967.html

[PaulJ]

£1.25 on eBay. @PaulJ with free postage.

https://www.ebay.co.uk/itm/293825852316

They are 32k.. the instructions say 22k... ?

[PeterJ]

They are 32k.. the instructions say 22k... ?

Sorry [mention]PaulJ[/mention],

That's my rubbish eyesight. I hate getting older!

https://www.ebay.co.uk/itm/1-4W-Metal-F ... SwQFZbgbd6

You may need to pick the value from the drop-down list.

[PaulJ]

Ordered
Fingers crossed.

[Guesser]

The Spectrum + membrane design is terrible and it's a wonder they ever worked properly at all in my opinion.
I've found them extremely fussy when fitting, and had one new membrane replacement with an apparent minor manufacturing error, but really they're just horrible fiddly things.

If all the regular keys work but the extra keys don't, it sounds similar to the issue I had, of the connection between the two sets of membrane layers being unreliable/missing. In my case I had the lines for the extra shifting working and so on but not the main letters (I think it may depend which of the two tails is making a bad connection).

Here's a crude and entirely not-to-scale diagram showing roughly the how the Spectrum + membrane is constructed, and a couple of cross sections showing what I believe was the issue with mine and how I solved it.
The drawing is done mostly from memory in mspaint, and membranes from different sources are all slightly different so don't take it as being exactly what your membrane should look like, it's "for illustrative purposes only"



There are 4 conductive layers, on three sheets of plastic. They are connected together by face-to-face contact of traces on the different layers' "tails" by the keyboard "clamp" blocks, against short stubs of rubber from the bubble dome mat. One of these layers forms a sort of "bridge" to connect the traces on the main keyboard ribbon to the traces on the opposite side of the centre sheet.

It seems to me that if the tails are not aligned correctly in the clamps (or the inner tail has been left excessively long) then no amount of tightening of the clamp screws will make any difference (and they're easy to crack/strip, so don't over-tighten them, they only need to be snug!!).
With the clamp only squeezing across all three pieces of plastic, the connection between the "bridge" and the main tail is basically left down to luck.
I was finding the membrane would work while testing it with the case open, but as soon as I closed the case it would stop registering keys again! I believe it was the particular curve in the ribbons with the keyboard removed that was keeping the layers pressed together, and when I moved the keyboard back to its proper place, that tension reduced and the connections went open circuit again.

This is a long way of saying, if one or other layer of keys is flaky, or the keyboard randomly doesn't work at all, check the clamping of the layer interconnections because it may be that all that's holding the outer layers in contact is a bit of sellotape!

[PaulJ]

The Spectrum + membrane design is terrible and it's a wonder they ever worked properly at all in my opinion.

Wow.. well.. that's fixed one problem.. and stopped me throwing out of the window.

Initial build & test worked apart from those pesky Plus specific keys = clamps in place.

All other builds had little or no keys working at all = clamps off for testing!

Clamps back on now and back to original issue of Plus specific keys.. hopefully a resister may fix that.
At least I can use the machine for gaming now.

What sort of weird fanlged method is that of securing a contact!! tut tut..

Thanks

[SpeccyKev]

I just put a new membrane in a Spectrum+

First thing that happened was none of the specific Spectrum+ keys worked. I think I had tightened the screws on the back of the steel plate too much. I loosened them up and put it back together and it all worked fine, apart from the dedicated cursors on the bottom row. (strangely the 5,6,7.8 keys work as cursors...)

I think it is something to do with the clamps not making the different layers connect properly. I'm going to have a fiddle and see what I can do.

[Hedge1970]

I like this description about half way down he explains how important the tension of the screws are in getting a fully working keyboard

https://www.dataserve-retro.co.uk/conte ... k/d32.html

Good luck

I just put a new membrane in a Spectrum+

First thing that happened was none of the specific Spectrum+ keys worked. I think I had tightened the screws on the back of the steel plate too much. I loosened them up and put it back together and it all worked fine, apart from the dedicated cursors on the bottom row. (strangely the 5,6,7.8 keys work as cursors...)

I think it is something to do with the clamps not making the different layers connect properly. I'm going to have a fiddle and see what I can do.

[SpeccyKev]

Yep, it’s crazy what passed for technology 30 years ago. We used to have to lean a pile of books on the AV plug to get a clear picture from it.

[spider]

Seem to recall years ago a friends toasty rack 128 intermittently would turn the 'cursor' keys into the regular 5/6/7/8 , seemed to vary a bit if you pressed at an angle though too. I can sort of see how if the numeric line makes contact without the shift line its going to do that, same membrane as the 48K+. I don't remember the Amstrad based units doing it, do realise its a different membrane!

I did read a while back (can't recall the link now sorry) on a membrane guide that they had seen small pieces of tape 'factory fitted' so it seems Sinclair may have experienced a few issues perhaps even when they were new. Seemed to also suggest as has been mentioned already the tail clamps and the backplate screws do not need winding down to "withstand an earthquake", over tightening them is going to cause concerns.

[Ast A. Moore]

over tightening them is going to cause concerns.

That is true for most (if not all) screws. Most people don’t understand the physical principle of the screw, and over tighten them all the time.

[Guesser]

I don't remember the Amstrad based units doing it, do realise its a different membrane!

I'm not sure how the +2 membrane works, I know the keyboard moulding carried over to the black +2 and +3 has integral clamp positions. On the +3 and black +2 Amstrad did away with building the "logic" for the shifted keys into the membrane, and made the matrix bigger and moved the decoding into the gate array.

I did read a while back (can't recall the link now sorry) on a membrane guide that they had seen small pieces of tape 'factory fitted' so it seems Sinclair may have experienced a few issues perhaps even when they were new.

As far as I'm concerned it's just a bad design, prone to misalignment, so it wouldn't surprise me if some were wonky from the factory.

[1024MAK]

As far as I'm concerned it's just a bad design, prone to misalignment, so it wouldn't surprise me if some were wonky from the factory.

As discovered repeatedly, this ‘thing’ of trying to make a keyboard where two different pairs (sets) of contacts are used and HAVE to make and break simultaneously was always a bad idea. Further, with the plus membrane, two different physical membrane layers have to be pressed together to make electrical contact. This is part of the function of the keyboard clamps. Any misalignment and the keyboard does not work correctly.

All this is due to the keyboard matrix being only designed for 40 keys (5 x 8 matrix). At the time that the ZX Spectrum was designed, 40 keys was considered fine. After all, it had been okay for the ZX80 and the ZX81.

Hence the five keyboard input lines were included in the ULA. A 40 pin chip that was already very tight on available pins for all the functions it was doing. So even though the ULA input port has seven available bits (the eighth bit is the cassette input or Ear bit), due to the lack of ULA pins and no foresight that a larger keyboard may be needed, that was that.

In retrospect, IMHO, it would have been better to use a separate commonly available TTL LS logic tri-state buffer chip for the keyboard input lines. The ULA would still had the control circuits for the input port, so a single ULA pin could have controlled this extra chip. This extra chip would then have seven inputs, thus expanding the keyboard matrix to 7 x 8 and hence allowing a 56 key keyboard. If only this had been done with the original design.

Could it have been done with the introduction of the ZX Spectrum+? Yes, but, such a design change would have involved a new PCB design, a redesign of the ULA, obviously a new keyboard and changes to the ROM code. This would have also likely meant a different matrix layout of the original 40 keys, and hence incompatibility with the rubber key model. All extra expensive for Sinclair and the keyboard and ROM code would be make it incompatible with the earlier rubber key model.

The only other alternatives would have been to use extra circuitry and have the extra keys mapped to a different input port. But you still have the problem of needing changes to the ROM code, and a new PCB design. Or put in additional logic in a new custom chip and get this to deal with the extra keys while retaining the existing ULA with the 5 keyboard input lines. But a new custom chip costs money and you still need a new PCB design.

Mark