On 2005-06-06 18:00, lensman wrote:
thats great - now got my 48k working too - thanks for that - now im off to see if I can get a ZX81 and the Cambridge computing joyskick working
Excellent...
Good luck on the ZX81, but remember they're evil incarnate when it comes to tuning them into a modern TV. If it is broken, though, the membrane is pretty much the same as a Speccy's, but harder to replace....
Your problem is the keyboard membrane has decayed with age. If you open up the machine you'll find the two membrane legs which should plug into the sockets on the board may have split - if you're lucky they'll have split at the bottom where they plug into the main board. If you're unlucky, they'll have split at the top where the legs sprout from the keyboard itself.
If you are lucky, then you can probably fix it with a pair of scissors - just carefully cut the membrane legs just above where the split is, and then cut the plastic insulator part of the leg so that the metal is exposed again when it's plugged into the board.
If you're unlucky, you'll need a new membrane - you can get them on eBay or from places like Sintech - they're a pretty easy fix.
Hope that helps some - take a look inside and if you're not sure what you're seeing feel free to ask!
Hello again, Sir!
Thanks very much for the advice. Just opened up the machine and the two legs seem to be intact. Just noticed, though, that the black box where the right-hand leg plugs into the base is leaning slightly, towards the top of the machine. Might this be significant? :)
On 2005-06-06 19:27, Jason Arnopp wrote:
Thanks very much for the advice. Just opened up the machine and the two legs seem to be intact. Just noticed, though, that the black box where the right-hand leg plugs into the base is leaning slightly, towards the top of the machine. Might this be significant? :)
Perfectly normal, I'm afraid, they were built like that. If there's no cuts in the membrane legs - and they can be quite hard to spot: look for thin white lines traversing the leg, check right up where the membrane joins the keyboard where the leg is folded... if there's too much creasing there that may well stop the contact.
If you have a multimeter handy you can check the membrane contacts half way up the leg are connecting to the pins on the board...
On 2005-06-06 19:27, Jason Arnopp wrote:
Thanks very much for the advice. Just opened up the machine and the two legs seem to be intact. Just noticed, though, that the black box where the right-hand leg plugs into the base is leaning slightly, towards the top of the machine. Might this be significant? :)
Perfectly normal, I'm afraid, they were built like that. If there's no cuts in the membrane legs - and they can be quite hard to spot: look for thin white lines traversing the leg, check right up where the membrane joins the keyboard where the leg is folded... if there's too much creasing there that may well stop the contact.
If you have a multimeter handy you can check the membrane contacts half way up the leg are connecting to the pins on the board...
Thanks again for the advice.
Ahhh, darn, I've no idea what's going on. Still couldn't see anything on the legs, so hopefully put it back together and came up 'snake eyes'. :) Still no response from the keyboard at all. Perhaps I should just go back on eBay - or buy a Sinclair joystick so I can actually play games on my +2... :)
Are you the 'Jason Arnopp' the rock/metal journalist who wrote fopr kerang in the late 80's/ear;y 90's and eho's greatest claim to fame was 'outing' Michael Barrymore?
Yes, I'm the Jason Arnopp who wrote for K! and did end up in the Barrymore story, but I made sure I didn't 'out' him. :) It was just a silly story about him being drunk in the same hotel bar as I was sitting in! You've got a great memory, mind...
Getting back on topic for a moment - I successfully 'repaired' another Interface 1 over the weekend. Knackered ULA unfortunately, but I repaired the rest of it (someone had hacked it up and removed chips etc.) so if anyone wants one of them fixing too... ( as long as the ULA is good, anyway!).
_________________
--
Ian Gledhill www.retroreview.com
[email="IanG@+nospam.+amiganet.org"]IanG@+nospam.+amiganet.org[/email] (remove text between plus'es!)
[ This Message was edited by: Spirantho on 2005-06-13 12:08 ]
My original Speccy died a long time ago after I lent it to a friend. From memory, when turned on it comes up with a miriad of coloured blocks and flashing attributes. What could be the cause and can you fix it?
Could be anything, and yes I can. :) Probably the lower RAM has lost its -5V and/or +12V, but it could be the ULA or something.
PM me if you want me to have a look at it for you!
_________________
Ian Gledhill www.retroreview.com
[email="IanG@+nospam.+amiganet.org"]IanG@+nospam.+amiganet.org[/email] (remove text between plus'es!)
[ This Message was edited by: Spirantho on 2005-06-15 10:27 ]
If you have the keyboard working, you can type:
PRINT PEEK 23732+PEEK 23733 * 256
If the answer "65535" appears, that's a 48K Speccy. If the answer is "32768" it's a 16K one. If it's anything else, it's almost certainly a 48K with some dodgy RAM at the top!
I picked up a +48k at a boot sale a few weeks ago but I couldn't get it to load any games. I just tried your test on it and it think's it's a 16k speccy! (it's not though, as it's an issue 6a motherboard with 2 RAM banks.)
So, are you still offering this service? And is it repairable?
I have just got my old speccy down from the loft and some of the keypresses are slow/dead. Its an expanded 16k spectrum with a foxx memory upgrade, also in 1984/85 I had to replace the keyboard and opted for a SAGA Emperor keyboard. No other repairs have ever been done to this machine since new.
The screen is brown not white, so I don`t know why that is.
I also would like to replace the SAGA keyboard with an original rubber keyboard.
Can the keyboards(entire top half of speccy), still be obtained? And if so how much?
p.s.
I also have a battered grey +2 with a homemade ear socket to attach a cassette deck(ala Crash TechTips). Is it possible to give it a Sinclair 128k toaster rom?
On 2005-09-03 09:10, FrankT wrote:
I have just got my old speccy down from the loft and some of the keypresses are slow/dead. Its an expanded 16k spectrum with a foxx memory upgrade, also in 1984/85 I had to replace the keyboard and opted for a SAGA Emperor keyboard. No other repairs have ever been done to this machine since new.
The slow keypresses will be down to your keyboard... even non Sinclair keyboards get old, but they can usually be fixed with a bit of cleaning.
The screen is brown not white, so I don`t know why that is.
If it's an Issue 2 Speccy, that's not surprising... but can be fixed as there's little presets inside which you can turn with a screwdriver to get it good again. If necessary you can change the ULA too to get a slightly better piccy, though you need to change some other resistors too (not hard).
I also would like to replace the SAGA keyboard with an original rubber keyboard.
Can the keyboards(entire top half of speccy), still be obtained? And if so how much?
Cheapest way to do that is to buy another Speccy. :)
Seriously, a new membrane will set you back about ?10, and I think I may be able to find a lid off an old Speccy if you really need one.... but most of them are still attached to Spectrums!
p.s.
I also have a battered grey +2 with a homemade ear socket to attach a cassette deck(ala Crash TechTips). Is it possible to give it a Sinclair 128k toaster rom?
You have a +2, not a +2A, so there's really very little point. The ROM is identical apart from the copyright message anyway, I think.... I suppose it could be done with an EPROM and a bit of wire.... I can do it if you really want it...!
On 2005-06-06 11:04, Spirantho wrote:
And also, in case it's of any help, at: http://www.retroreview.com/iang/TestSpec2.zip
there is a Zip file which contains a ROM (and the assembler listing, zasm-compatible) to test Speccies!
Once I'd replaced _all_ the lower 16K RAM chips in my dead Speccy, I realised it had to be something else, so I wrote a program that writes to screen RAM every bit pattern, and then writes it to other RAM ($6000-$7800) and copies it to screen RAM, hence testing write and read abilities. It also displays all the $FE port inputs so you can test keyboard/tape.
The clever bit though is that it doesn't _use_ any RAM - all variables are held in registers, so that means no subroutines or anything. Nasty. It also means that it will work with any Spectrum as long as the Z80 is good - great debugging chips!
Once I'd done this I could see that one of my RAM chips was getting interfered with (!) by one of the upper RAM chips! A quick replacement later and yay! It works!
Maybe that ROM could be of use to others, hence why I've posted it here.
Just received my 27e512 eproms, very nice and easy to erase. OK, trying to fix my grey +2 now: I have suspected ram fault (coloured flashing blocks). I first burnt your test rom. Spectrum initialized and screen looked just like in emulator. How should I read the screen?
If there is any hint I burnt next ordinary 48k rom and initialized +2 with it. It started nicely with (C) text. When I made PRINT PEEK 23732+PEEK 23733*256 answer was 32767 and not 65535! So there must be something wrong with ram.
What would be the best way to find the faulty ram/rams? Is it possible to modify the TestSpec2.rom so that it can check whole 128k ? And what is the meaning of the picture TestSpec2.rom displays?
Timppa
[ This Message was edited by: Timppa on 2005-09-09 22:32 ]
I could make it check the 128K memory banks but it'd be just as easy to write something in BASIC, to be honest... I think the machine should still boot up. I could be wrong, though... I may look into it.
As for the display, I've updated the archive (www.retroreview.com/iang/TestSpec2.zip for those with short memories!) with an annotated picture of it running. You should get an identical display. Anything else and you've got dead RAM - apart from the little black box which actually will usually have a white stripe on a real Spectrum. Looks like Spin doesn't quite emulate it 100%. :)
The important thing is that the columns on the left go through every permutation 0-255, and are identical to the columns at the far right. If you have a black or white line going all the way down all the columns, that line represents a dead RAM chip. Ditto with interference. It should all look precisely the same as the jpg in the archive.
Can you take a close-up screenshot of your display?
This means that you can now burn this onto an 8K EPROM, and switch this EPROM with a Multiface EPROM - no soldering required!
Now, when you start up your very-ill Speccy, and press the red button, up comes a screen from which you can see exactly which RAM chips are knackered!
Still only works on the lower 16K at the moment, but as usually you can use BASIC to work out the upper RAM....
On 2005-09-12 10:42, Spirantho wrote:
I could make it check the 128K memory banks but it'd be just as easy to write something in BASIC, to be honest... I think the machine should still boot up. I could be wrong, though... I may look into it.
As I am able to boot now with original 48k rom, i tried to poke and peek some addresses. Addresses below 32768 are fine (I get same peek what I poke), but addresses above get strange answers. +2 manual says that bank 2 starts from 32768. Should I assume it is (at least) broken. How are the physical ICs (IC17-IC32) mapped to banks (0-7). Are the in same order IC17,IC18 = bank 0, etc..?
As for the display, I've updated the archive (www.retroreview.com/iang/TestSpec2.zip for those with short memories!) with an annotated picture of it running. You should get an identical display. Anything else and you've got dead RAM - apart from the little black box which actually will usually have a white stripe on a real Spectrum. Looks like Spin doesn't quite emulate it 100%. :)
Yes, the picture was the same. From that I assumed at least some rams work, and then tried with 48k.rom.
Now I'm ready to start changing ICs. I perhaps first try to change ICs 21 and 22 if they make the bank 2.
Cheers
Timppa
[ This Message was edited by: Timppa on 2005-09-13 23:38 ]
Hmm... I've never tried to fix a +2 (or anything other than a 16K/48K to be honest) so I can't say without the schematics... are they online anywhere? I can find the +2A, but not the 128K or +2 scematics, so I can't say what the chips are likely to be....
Right, I've looked at my Speccy 128 (and the 128 I'm currently trying to repair - Hi Soren :) ).
The RAM chips are the ones all the way along the bottom, and I suspect the lower numbers are the bottom 64K, so the higher numbers are the higher 64K, so it's the higher ones that you need to look at.
I've just uploaded http://www.retroreview.com/iang/TestSpec4.zip
which does a bank-switch into bank 3 and tests that RAM. You'll notice there's a new graphic column - that's data that been printed via $C000 bank 3. If this doesn't look identical to the other columns, the pixel column(s) which is/are different are your culprits.
I've not tested this ROM on a real Spectrum, yet, only my emulator (as it's my emulator I can give it 'dead' RAM and everything and it seems to work) so I'd be interested to know what happens if/when anybody tries this....
Comments
Excellent...
Good luck on the ZX81, but remember they're evil incarnate when it comes to tuning them into a modern TV. If it is broken, though, the membrane is pretty much the same as a Speccy's, but harder to replace....
Hello again, Sir!
Thanks very much for the advice. Just opened up the machine and the two legs seem to be intact. Just noticed, though, that the black box where the right-hand leg plugs into the base is leaning slightly, towards the top of the machine. Might this be significant? :)
Best,
Jason
wonder if thatll happen to us lot if we get alzimers
Perfectly normal, I'm afraid, they were built like that. If there's no cuts in the membrane legs - and they can be quite hard to spot: look for thin white lines traversing the leg, check right up where the membrane joins the keyboard where the leg is folded... if there's too much creasing there that may well stop the contact.
If you have a multimeter handy you can check the membrane contacts half way up the leg are connecting to the pins on the board...
Ahhh, darn, I've no idea what's going on. Still couldn't see anything on the legs, so hopefully put it back together and came up 'snake eyes'. :) Still no response from the keyboard at all. Perhaps I should just go back on eBay - or buy a Sinclair joystick so I can actually play games on my +2... :)
Are you the 'Jason Arnopp' the rock/metal journalist who wrote fopr kerang in the late 80's/ear;y 90's and eho's greatest claim to fame was 'outing' Michael Barrymore?
I've just read my previous post for the first time. Damn!! must proofread more carefully before I hit the 'submit' button...
It's ok Karingal ... sometimes we get excited when we spot celebs and forget to proofread so no worries there mate.
Skarpo
:)
My wife will tell you that!!
he read barrymore and poofread :oops:
_________________
--
Ian Gledhill
www.retroreview.com
[email="IanG@+nospam.+amiganet.org"]IanG@+nospam.+amiganet.org[/email] (remove text between plus'es!)
[ This Message was edited by: Spirantho on 2005-06-13 12:08 ]
Cheers
PM me if you want me to have a look at it for you!
_________________
Ian Gledhill
www.retroreview.com
[email="IanG@+nospam.+amiganet.org"]IanG@+nospam.+amiganet.org[/email] (remove text between plus'es!)
[ This Message was edited by: Spirantho on 2005-06-15 10:27 ]
I picked up a +48k at a boot sale a few weeks ago but I couldn't get it to load any games. I just tried your test on it and it think's it's a 16k speccy! (it's not though, as it's an issue 6a motherboard with 2 RAM banks.)
So, are you still offering this service? And is it repairable?
Yes and yes. :)
Should be quite easy to fix if you have the right tools... and if you don't, then that's when I can help. :)
PM me if you want me to have a look at it and I'll give you my address etc.
PM me if you want me to have a look at it and I'll give you my address etc.
[/quote]
I bet your postman likes you .... not
Coils - easy to re-wind yourself if you know how.
Unfortunately, the ZTX650 isn't the hard one to get - it's the ZTX213 and to a lesser extent the ZTX313s that are the trouble...
As for the coil, I don't know how. :) They're fairly intricate I think... but if you can repair them I can send you some. :)
The screen is brown not white, so I don`t know why that is.
I also would like to replace the SAGA keyboard with an original rubber keyboard.
Can the keyboards(entire top half of speccy), still be obtained? And if so how much?
p.s.
I also have a battered grey +2 with a homemade ear socket to attach a cassette deck(ala Crash TechTips). Is it possible to give it a Sinclair 128k toaster rom?
Thanks.
My ZX Art Music Page
Carlos Michelis Theme
The slow keypresses will be down to your keyboard... even non Sinclair keyboards get old, but they can usually be fixed with a bit of cleaning.
If it's an Issue 2 Speccy, that's not surprising... but can be fixed as there's little presets inside which you can turn with a screwdriver to get it good again. If necessary you can change the ULA too to get a slightly better piccy, though you need to change some other resistors too (not hard).
Cheapest way to do that is to buy another Speccy. :)
Seriously, a new membrane will set you back about ?10, and I think I may be able to find a lid off an old Speccy if you really need one.... but most of them are still attached to Spectrums!
You have a +2, not a +2A, so there's really very little point. The ROM is identical apart from the copyright message anyway, I think.... I suppose it could be done with an EPROM and a bit of wire.... I can do it if you really want it...!
Let me know if I can be of assistance further. :)
Just received my 27e512 eproms, very nice and easy to erase. OK, trying to fix my grey +2 now: I have suspected ram fault (coloured flashing blocks). I first burnt your test rom. Spectrum initialized and screen looked just like in emulator. How should I read the screen?
If there is any hint I burnt next ordinary 48k rom and initialized +2 with it. It started nicely with (C) text. When I made PRINT PEEK 23732+PEEK 23733*256 answer was 32767 and not 65535! So there must be something wrong with ram.
What would be the best way to find the faulty ram/rams? Is it possible to modify the TestSpec2.rom so that it can check whole 128k ? And what is the meaning of the picture TestSpec2.rom displays?
Timppa
[ This Message was edited by: Timppa on 2005-09-09 22:32 ]
As for the display, I've updated the archive (www.retroreview.com/iang/TestSpec2.zip for those with short memories!) with an annotated picture of it running. You should get an identical display. Anything else and you've got dead RAM - apart from the little black box which actually will usually have a white stripe on a real Spectrum. Looks like Spin doesn't quite emulate it 100%. :)
The important thing is that the columns on the left go through every permutation 0-255, and are identical to the columns at the far right. If you have a black or white line going all the way down all the columns, that line represents a dead RAM chip. Ditto with interference. It should all look precisely the same as the jpg in the archive.
Can you take a close-up screenshot of your display?
http://www.retroreview.com/iang/TestSpec3.zip
is now uploaded. The only difference is that we're now Multiface compatible! (Thanks Sokurah for the idea!)
This means that you can now burn this onto an 8K EPROM, and switch this EPROM with a Multiface EPROM - no soldering required!
Now, when you start up your very-ill Speccy, and press the red button, up comes a screen from which you can see exactly which RAM chips are knackered!
Still only works on the lower 16K at the moment, but as usually you can use BASIC to work out the upper RAM....
As I am able to boot now with original 48k rom, i tried to poke and peek some addresses. Addresses below 32768 are fine (I get same peek what I poke), but addresses above get strange answers. +2 manual says that bank 2 starts from 32768. Should I assume it is (at least) broken. How are the physical ICs (IC17-IC32) mapped to banks (0-7). Are the in same order IC17,IC18 = bank 0, etc..?
Yes, the picture was the same. From that I assumed at least some rams work, and then tried with 48k.rom.
Now I'm ready to start changing ICs. I perhaps first try to change ICs 21 and 22 if they make the bank 2.
Cheers
Timppa
[ This Message was edited by: Timppa on 2005-09-13 23:38 ]
The RAM chips are the ones all the way along the bottom, and I suspect the lower numbers are the bottom 64K, so the higher numbers are the higher 64K, so it's the higher ones that you need to look at.
I've just uploaded
http://www.retroreview.com/iang/TestSpec4.zip
which does a bank-switch into bank 3 and tests that RAM. You'll notice there's a new graphic column - that's data that been printed via $C000 bank 3. If this doesn't look identical to the other columns, the pixel column(s) which is/are different are your culprits.
I've not tested this ROM on a real Spectrum, yet, only my emulator (as it's my emulator I can give it 'dead' RAM and everything and it seems to work) so I'd be interested to know what happens if/when anybody tries this....