I'd desolder them all, then with the multimeter connected shove pin 8 of each chip against its contact while measuring for a short (you may need test clips, or multiple hands for this :D)
When you do find the short, just snip that sucker out :) Don't forget to continue testing the others, as there may be more than one faulty chip...
I've swapped out the caps today, gave it a quick test and the screen is the same, must have been on for no more than 30 seconds. IC 6 was so hot you could have fried an egg on it. Plus this chip is slightly blacker than the rest although it's not obvious to me that this is a replacement from the solder work.
Most of the IC's are 8445 STC 4116 2N. The one that ran hot is 8447 STC 4116 2N, and there's another 8447 present in IC11, although this one was not running any hotter than the others.
Might be a red herring, but thought it was worth documenting.
Balford I must admit I found desoldering pin 8 and testing a little tricky, I think at this time this method is beyond my skills, you were bang on with your advice and I know exactly what we were aiming for........but all is not lost. I do believe great progress was made today....read on.
Whilst desoldering pin 8's the thought of Zorn's lower ram replacement got ever more appealing, these 4116's are a pain in the...
I proceeded remove IC6 (Mr warm) and low and behold there was a lifted trace to pin 4, the trace was floating in the air, and I promise (it wasn't me).
So here I repaired the trace and proceeded to snip off the rest of the lower ram IC's, their legs are still present for the moment.
Plugged it in to check if the voltages were correct on the rail, first notable change - BZZZZZZZ from the coil
Multimeter gave a reading of -5, 0, 12, 5 volts at the IC legs, and a picture on screen that is a scramble of colours with a white border.
I've popped up a picture of the repaired trace, which is showing fine on a continuity test now so I'm confident in my repair.
Also here is a quick vid showing what happens on screen, what happens when i hold down the reset button, and what happens when I give it a quick press.
The next step is to solder in a socket. Then if you wish, you can try a replacement 4116 DRAM chip. If it works, jump for joy! If not you can either continue fault finding the remaining 4116 DRAM chips, or cut them out, fit sockets and install a LRR module:-)
! 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 in Somerset later in the year :)
Probably whilst you were typing I've pulled out the remaining 4116's, already taken the decision to go with replacements, since I could identify a bodged replacement of IC6 I was curious to see what was going on underneath every lower ram IC, I uncovered nothing, but at least I won't have a nagging doubt about a lifted trace elsewhere.
Sockets - Good idea, I'm on the case buying them now.
Although I have a donor board with 4116's, not very good at removing chips intact (dreams of a Hakko), so I'll probably go for the lower ram replacement module or buy a few IC's, thankfully both options are relatively inexpensive.
Now that I have the voltages, does this mean the TR4 and TR5 replacement was a success also?
Yes, that certainly means that your TR4/5 replacement was a success, judging by the heat of IC6 it almost certainly was the cause of the voltage rail short that killed these parts in the first place.
Just replace the lower RAM (looks like you've done a fine job on the repair work so far) and you'll be up and running.
OK could not wait for Sunday. Half day finish today so got straight to it.
Soldered sockets in place, checked for shorts and continuity at all stages. I'm quietly confident with my work here.
Placed a lower ram module carefully into the sockets following the guide that came with the delivery, nothing difficult there.
I'm now getting this screen, and the machine does not to react when I press the reset switch.
I'm still getting the healthy buzz from the coil, plenty of colour on screen now.
One thing I have noticed today since putting in the lower ram, after running the board for about 90-120 seconds, the Zilog Z80 is hot enough to burn me, impossible to put my finger on it for longer than half a second...hot hot hot.
I pulled out a issue 2 to compare it to and the Z80 barely got warm.(Assuming the z80 temperature is comparable between use on issue 2 or 4 boards)
Clearly somebody has had a go at this board before I arrived, and I've always wondered what this sticker was for on the Z80 :-).....hmm....maybe ....
From the video and the pics, what do you think? I'm pretty confident there's nothing wrong with the work on the ram replacement, but we can't rule anything out I suppose.
The CPU shouldn't get any more than slightly warm, so there is certainly something wrong there. Either the Z80 is somewhat dead or something is overloading one of its outputs (which may have killed it!). It does look as if it manages to signal the ULA to set the border to white, though.
It does look like a Z80 swap is the order of the day doesn't it. Thankfully I have a few untested old arcade game boards to hand that are socketed, So once I get myself a 40 pin DIL socket for this spectrum I'll try dropping a few in, whilst I'm at it I'll grab a few new z80's. At least then if I get the same results with another Z80 I guess we try to move on from there.
This board has just revealed another clue for me to check, I can see a different soldering pattern on the transistor under the heatsink, Since I suspect the replacement transistors in position TR4 and 5 were not a good match in the first instance.
Is there a way to test that this transistor TR6 is working correctly using a multimeter?
In both videos hopefully you can tell when I'm pressing the reset button.
The new Z80 is running rather cold, nothing like the heat that was coming from the one I removed.
I mentioned before that from the solder side of the board I could tell that TR6 had previously been replaced, but I did not advise on the transistor itself, today whilst the heatsink was off I tried to record this, but on this transistor there were no markings or they have worn off perhaps.
I may replace this transistor with the correct value one, just to eliminate it. From a few sites I see this as MPS2369.
There are no more obvious paths for me to take with my knowledge. I can't feel and overly warm chips, and I have tested the ULA as working in another board.
I'm wondering if I should look at IC5, but I'd certainly need help with that since I do not have the tools to program a chip, but I'm clutching now for clues.
I remembered I had the smart card from Retroleum, and that it has a rom/ram check feature, so time to use it methinks...
I decided to plug it in and it came right up on screen which was a very welcome sight
A damaged keyboard membrane prevented me from holding space to go into the test feature on switch on, but with the use of a joystick I was able to play Jetpac from a .tap file. Wayhay!!!
Here I just assumed that the Retroleum smart card had taken over the functions of IC5 and had booted the machine with it's own Rom.....I say assume(no clue)....but there's life.
Then something strange happened.
I pressed the speccy reset switch on the side and got the normal spectrum screen you see at switch on.
At this stage I thought nothing more about it and assumed this was working with the smart card and it was just toying with me.
Next, power off, remove Retroleum Smart Card, and power back up.
and I get the Spectrum switch on screen....huh????
It works ?!?!.... - so this is excellent news....but how? Before plugging in this Smart Card I was getting the garbage you see above.
I'm really baffled though. How did the mere placing of the Retroleum smart card and playing a game jumpstart the motherboard into life? It's almost like the card had a word in it's ear and reminded it what to do on it's own, go on you can do it.....
I seem to have done something by luck rather than judgement here. I'm of course happy but puzzled and the repair work must be OK throughout.
Just to recap the repair work.
TR4 and TR5 replaced.
Lower ram replaced
Z80 replaced
Hmm.
All is not good.
I moved the spectrum to another room, powered it on and we're back to the scrambled mess. Same as youtube videos above.
Minutes before I recorded this so you can see it was working at one point. This is rather annoying. https://www.youtube.com/watch?v=v7f4WZN3zjU
I've got other projects, so there is no issue buying some 4116's. I'm starting to suspect I have a dodgy lower ram module, or there's something wrong with my sockets.
8 x NOS 4116 now on order.
Could be a dry (poor) solder joint somewhere, a hair sized solder splash, a cracked PCB track/trace...
if you have a good multimeter, it may be time to get a copy of the schematic and test each and every connection for the area you have worked on. Test using a low resistance range to confirm that the track is complete between each pin of each relevant chip (test on the pins). Then using a medium resistance range, test between each pin, looking for short circuits.
If you have a 'scope, this may help cut down the above boring testing, by just going on the relevant chip pins and confirming that the wave form wiggles from 0.8V (or lower) and 2V (or higher).
! 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 in Somerset later in the year :)
Update
Quite a few posts up I documented repairing a trace on IC6 pin 4.
Viewing the board from above I have continuity on pin 4 to pin 4 of the chips to the left.
What I don't have is continuity through the socket to the trace on the board "repaired"
Damn, I guess the intermittent nature of this repair could well relate right back to the first fault/fix on the hot IC6.
Looks like the repair on the trace was a success at one point, but a flaky one at best.
So I have a lead at least, but not a trace...boom boom...
I'd hate to take the IC6 socket off, compared to when I removed the other IC's it was clear by the different chip and the traces that there was activity under IC6 before I came to the party here.
What I'm thinking now is leave the socket alone, and somehow run a jumper from the pin 4 of IC6 to somewhere on the motherboard.
I know it's rough, and I hate jumpers, weighing up the pro's and con's of removing the socket, this may be the best option. Hell, it worked for Sinclair...
! 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 in Somerset later in the year :)
That's /RAS. The same signal goes to all the other 4116 sockets in one direction. But the signal comes from the ULA pin 35. It also calls in on pins 1 and 2 of IC24.
So run a small jumper / patch wire on the track side linking the shortest path between sections where the track is open circuit.
! 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 in Somerset later in the year :)
The aim is to "make a map" of how the track/trace actually runs. This is hard if sections of it run on the top (component side) of the PCB. But not too hard if most of it runs on the bottom side. Use you meter to confirm how far it gets (tests low on the resistance range) and to confirm you have the right track :D
Then you can work out which bit is open circuit and which bit needs the jumper wire. Be aware that the physical tracks on the board DO NOT follow the order of connections as shown in the schematic.
Electrically you can use any point in the circuit to join to any other point of the same circuit. But it is best to keep jumper wires as short as possible.
! 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 in Somerset later in the year :)
THANK YOU!
As described above, I purchased some 4116 ram chips. I got them from eBay, from Russia, and they took two weeks to arrive. I made a temporary connection using IC clips and she fired up.
I'll save the replacement Lower ram module for another fix.
Got a thin piece of solid core wire and placed a jumper on the solder side of the board to the pins described a few posts up.
Reassembled and we have a working spectrum.....so thank you guys, you've been awesome and very patient. The result is a spectrum back from the dead which will be used and enjoyed.......but don't hear it from me, the Spectrum wants to say a few words.....take it away..
TR4 / TR5 replaced
Lower Ram NOS replacement
Z80 replaced
Jumper wire to fix broken trace from lower ram to ULA
Changed all capacitors
AV Mod
New keyboard membrane
Replacement rubber grommit/foot
Replacement legs/springs
Comments
Or desolder one then test, resolder and move along?
I have a donor board with 4116's, but I'm not against the modern replacement method, its interesting to me.
When you do find the short, just snip that sucker out :) Don't forget to continue testing the others, as there may be more than one faulty chip...
B
zx-diagnostics - Fixing ZX Spectrums in the 21st Century (wiki)
Sinclair FAQ Wiki
Most of the IC's are 8445 STC 4116 2N. The one that ran hot is 8447 STC 4116 2N, and there's another 8447 present in IC11, although this one was not running any hotter than the others.
Might be a red herring, but thought it was worth documenting.
Whilst desoldering pin 8's the thought of Zorn's lower ram replacement got ever more appealing, these 4116's are a pain in the...
I proceeded remove IC6 (Mr warm) and low and behold there was a lifted trace to pin 4, the trace was floating in the air, and I promise (it wasn't me).
So here I repaired the trace and proceeded to snip off the rest of the lower ram IC's, their legs are still present for the moment.
Plugged it in to check if the voltages were correct on the rail, first notable change - BZZZZZZZ from the coil
Multimeter gave a reading of -5, 0, 12, 5 volts at the IC legs, and a picture on screen that is a scramble of colours with a white border.
I've popped up a picture of the repaired trace, which is showing fine on a continuity test now so I'm confident in my repair.
Also here is a quick vid showing what happens on screen, what happens when i hold down the reset button, and what happens when I give it a quick press.
https://youtu.be/ddjk4O2xV1U
I'm thinking that to buy the lower ram replacement module is the next step, far simpler for me that individual IC's, plus I quite like the idea.
This sound sensible?
The next step is to solder in a socket. Then if you wish, you can try a replacement 4116 DRAM chip. If it works, jump for joy! If not you can either continue fault finding the remaining 4116 DRAM chips, or cut them out, fit sockets and install a LRR module :-)
Mark
Repair Guides. Spanish Hardware site.
WoS - can't download? Info here...
former Meulie Spectrum Archive but no longer available :-(
Spectranet: the TNFS directory thread
! 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 in Somerset later in the year :)
Probably whilst you were typing I've pulled out the remaining 4116's, already taken the decision to go with replacements, since I could identify a bodged replacement of IC6 I was curious to see what was going on underneath every lower ram IC, I uncovered nothing, but at least I won't have a nagging doubt about a lifted trace elsewhere.
Sockets - Good idea, I'm on the case buying them now.
Although I have a donor board with 4116's, not very good at removing chips intact (dreams of a Hakko), so I'll probably go for the lower ram replacement module or buy a few IC's, thankfully both options are relatively inexpensive.
Now that I have the voltages, does this mean the TR4 and TR5 replacement was a success also?
Just replace the lower RAM (looks like you've done a fine job on the repair work so far) and you'll be up and running.
B
zx-diagnostics - Fixing ZX Spectrums in the 21st Century (wiki)
Sinclair FAQ Wiki
Just ordered the lower ram replacement (Zorn) and sockets.
Hopefully they'll arrive this week and I'll post an update next weekend.
Soldered sockets in place, checked for shorts and continuity at all stages. I'm quietly confident with my work here.
Placed a lower ram module carefully into the sockets following the guide that came with the delivery, nothing difficult there.
I'm now getting this screen, and the machine does not to react when I press the reset switch.
https://youtube.com/watch?v=e_ou2SKF5lA&feature=youtu.be
I'm still getting the healthy buzz from the coil, plenty of colour on screen now.
One thing I have noticed today since putting in the lower ram, after running the board for about 90-120 seconds, the Zilog Z80 is hot enough to burn me, impossible to put my finger on it for longer than half a second...hot hot hot.
I pulled out a issue 2 to compare it to and the Z80 barely got warm.(Assuming the z80 temperature is comparable between use on issue 2 or 4 boards)
Clearly somebody has had a go at this board before I arrived, and I've always wondered what this sticker was for on the Z80
:-).....hmm....maybe ....
From the video and the pics, what do you think? I'm pretty confident there's nothing wrong with the work on the ram replacement, but we can't rule anything out I suppose.
Gentlemen I'm in your hands.
This board has just revealed another clue for me to check, I can see a different soldering pattern on the transistor under the heatsink, Since I suspect the replacement transistors in position TR4 and 5 were not a good match in the first instance.
Is there a way to test that this transistor TR6 is working correctly using a multimeter?
I've tested the ULA in another board, and I can confirm that is fine so good news on that score.
Fingers crossed it's just 40 points to solder and a seating of a Z80 and we're off.
Let me share with you the screens I get with the new Z80
https://www.youtube.com/watch?v=29fOT3q10Gs
and then again with the new Z80, but this time with the replacement lower ram module removed.
https://www.youtube.com/watch?v=akCrsSpoy4g
In both videos hopefully you can tell when I'm pressing the reset button.
The new Z80 is running rather cold, nothing like the heat that was coming from the one I removed.
I mentioned before that from the solder side of the board I could tell that TR6 had previously been replaced, but I did not advise on the transistor itself, today whilst the heatsink was off I tried to record this, but on this transistor there were no markings or they have worn off perhaps.
I may replace this transistor with the correct value one, just to eliminate it. From a few sites I see this as MPS2369.
There are no more obvious paths for me to take with my knowledge. I can't feel and overly warm chips, and I have tested the ULA as working in another board.
I'm wondering if I should look at IC5, but I'd certainly need help with that since I do not have the tools to program a chip, but I'm clutching now for clues.
Gents, over to you....
I remembered I had the smart card from Retroleum, and that it has a rom/ram check feature, so time to use it methinks...
I decided to plug it in and it came right up on screen which was a very welcome sight
A damaged keyboard membrane prevented me from holding space to go into the test feature on switch on, but with the use of a joystick I was able to play Jetpac from a .tap file. Wayhay!!!
Here I just assumed that the Retroleum smart card had taken over the functions of IC5 and had booted the machine with it's own Rom.....I say assume(no clue)....but there's life.
Then something strange happened.
I pressed the speccy reset switch on the side and got the normal spectrum screen you see at switch on.
At this stage I thought nothing more about it and assumed this was working with the smart card and it was just toying with me.
Next, power off, remove Retroleum Smart Card, and power back up.
and I get the Spectrum switch on screen....huh????
It works ?!?!.... - so this is excellent news....but how? Before plugging in this Smart Card I was getting the garbage you see above.
I'm really baffled though. How did the mere placing of the Retroleum smart card and playing a game jumpstart the motherboard into life? It's almost like the card had a word in it's ear and reminded it what to do on it's own, go on you can do it.....
I seem to have done something by luck rather than judgement here. I'm of course happy but puzzled and the repair work must be OK throughout.
Just to recap the repair work.
TR4 and TR5 replaced.
Lower ram replaced
Z80 replaced
Baffled but happy.
All is not good.
I moved the spectrum to another room, powered it on and we're back to the scrambled mess. Same as youtube videos above.
Minutes before I recorded this so you can see it was working at one point. This is rather annoying.
https://www.youtube.com/watch?v=v7f4WZN3zjU
I've got other projects, so there is no issue buying some 4116's. I'm starting to suspect I have a dodgy lower ram module, or there's something wrong with my sockets.
8 x NOS 4116 now on order.
Stay tuned. Comments welcomed.
if you have a good multimeter, it may be time to get a copy of the schematic and test each and every connection for the area you have worked on. Test using a low resistance range to confirm that the track is complete between each pin of each relevant chip (test on the pins). Then using a medium resistance range, test between each pin, looking for short circuits.
If you have a 'scope, this may help cut down the above boring testing, by just going on the relevant chip pins and confirming that the wave form wiggles from 0.8V (or lower) and 2V (or higher).
Mark
Repair Guides. Spanish Hardware site.
WoS - can't download? Info here...
former Meulie Spectrum Archive but no longer available :-(
Spectranet: the TNFS directory thread
! 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 in Somerset later in the year :)
Update
Quite a few posts up I documented repairing a trace on IC6 pin 4.
Viewing the board from above I have continuity on pin 4 to pin 4 of the chips to the left.
What I don't have is continuity through the socket to the trace on the board "repaired"
Damn, I guess the intermittent nature of this repair could well relate right back to the first fault/fix on the hot IC6.
Looks like the repair on the trace was a success at one point, but a flaky one at best.
So I have a lead at least, but not a trace...boom boom...
I'd hate to take the IC6 socket off, compared to when I removed the other IC's it was clear by the different chip and the traces that there was activity under IC6 before I came to the party here.
What I'm thinking now is leave the socket alone, and somehow run a jumper from the pin 4 of IC6 to somewhere on the motherboard.
I know it's rough, and I hate jumpers, weighing up the pro's and con's of removing the socket, this may be the best option. Hell, it worked for Sinclair...
Where does pin 4 of IC6 go?
We're so close here!
It's schematic time. (shudder)
See above :-)
Repair Guides. Spanish Hardware site.
WoS - can't download? Info here...
former Meulie Spectrum Archive but no longer available :-(
Spectranet: the TNFS directory thread
! 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 in Somerset later in the year :)
Is this correct.
I got the schematic here
http://www.retroisle.com/sinclair/zxspectrum/Technical/Hardware/ZXSpectrumIssue4ab-Schematics.php
http://www.retroisle.com/sinclair/zxspectrum/Technical/Hardware/ZXSpectrumIssue4ab-Schematics.gif
This is strangely enjoyable.....
So run a small jumper / patch wire on the track side linking the shortest path between sections where the track is open circuit.
Use thin sold core wire.
Mark
Repair Guides. Spanish Hardware site.
WoS - can't download? Info here...
former Meulie Spectrum Archive but no longer available :-(
Spectranet: the TNFS directory thread
! 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 in Somerset later in the year :)
Or can I run a jumper from IC6 pin 4 to Pin 1 or 2 of IC24 (I'd prefer this leaving the ULA alone)?
Note: Continuity checks out from pin 35 of IC1 to pin 1 and 2 of IC24
Then you can work out which bit is open circuit and which bit needs the jumper wire. Be aware that the physical tracks on the board DO NOT follow the order of connections as shown in the schematic.
Electrically you can use any point in the circuit to join to any other point of the same circuit. But it is best to keep jumper wires as short as possible.
Mark
Repair Guides. Spanish Hardware site.
WoS - can't download? Info here...
former Meulie Spectrum Archive but no longer available :-(
Spectranet: the TNFS directory thread
! 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 in Somerset later in the year :)
I'll see if I can get hold of some IC leg clips and jump it this way before I go getting the soldering iron out.
As described above, I purchased some 4116 ram chips. I got them from eBay, from Russia, and they took two weeks to arrive. I made a temporary connection using IC clips and she fired up.
I'll save the replacement Lower ram module for another fix.
Got a thin piece of solid core wire and placed a jumper on the solder side of the board to the pins described a few posts up.
Reassembled and we have a working spectrum.....so thank you guys, you've been awesome and very patient. The result is a spectrum back from the dead which will be used and enjoyed.......but don't hear it from me, the Spectrum wants to say a few words.....take it away..
TR4 / TR5 replaced
Lower Ram NOS replacement
Z80 replaced
Jumper wire to fix broken trace from lower ram to ULA
Changed all capacitors
AV Mod
New keyboard membrane
Replacement rubber grommit/foot
Replacement legs/springs
B
zx-diagnostics - Fixing ZX Spectrums in the 21st Century (wiki)
Sinclair FAQ Wiki