Next on to my ZX81. Bit off topic, but what the hell. This suffered from the classic missing back porch problem associated with the 2C184 ULA.
To remedy this, I purchased a which adds the back porch, and outputs composite video.
The installation instructions tell you to install the device beside the heatsink, and to solder the flying leads to the underside of the board. This would mean a dangling connector, so I gutted the modulator and added it there:
Once the modulator lid is back on, the modification's not noticeable (apart from the insulated wire I used):
[/SIZE][/SIZE]
Thanks for posting this. My ZX81 was suffering from the same missing back porch.
My ZX81 is fitted in a rather ugly looking Crofton ZX81-Adaptakit case. This as a video out and some electronic on the keyboard to boost the composite signal. Unfortunately this only made the picture look worse. I removed these parts and replaced them with the ZX8-CCB, the picture is now perfect.
Okay, newest arrival chez Balford is an Issue 4A rubber key Spectrum.
I deliberately went for something that looked tatty and broken on eBay, as I'm looking for a case to house my Harlequin board once Ingo sends them out and I get around to assembling.
Once it turned up though, it immediately became apparent that the case and faceplate was in much better nick than I expected (and cleaned up rather well), but also upon opening I found a yellow PCB. So the decision was made, this was a keeper :)
Usual lower RAM defective stripes along with a very noisy display greeted me upon power on. Voltages checked out, apart from a slightly low -5v (measured -4.65v).
I'd read other posts that suggested changing C44/C45 to cure display noise, also dodgy caps elsewhere couldn't be helping the -5v/12v generation.
A swift recapping later and the noise was gone, and voltages were closer to spec.
Plugging in my diags board revealed that IC9 had bought the farm, so replacing just this part gave me a successful boot to the copyright message with a full 48K of working memory.
The membrane, which was working initially, decided to give up the ghost just as I was putting the machine back together. Not to worry though, replacement on order and this machine will be 100% again.
very sexy SRBP PCB - That's only the third I've ever seen, although apparently they are not at all rare.
Here's no.4
I was surprised when I opened the case as I've never seen a yellow one.
Unfortunately this one doesn't seems to work very well. I doesn't tune in, I can just make out the Sinclair research bit, but it crashes after a minute or 2, and buzz from the speaker.
I can see now it has faulty capacitors, 3 are visibly bad.
One other thing I noticed while cleaning is that the faceplate is secured by metal tabs, not double sided tape.
I had thought this was only the case on early machines, but either way I'm not complaining, it'll make membrane replacement much easier :)
B
Out of 9 loose spectrums I had in a draw that I got with ebay bundles, the first 2 I pulled out had these tabs. An issue 4A and 6A. I always thought it was a later thing.
An issue 3B Spectrum bought from AmiBay It was sold as defective for to be used as parts. When I declared interest to buy it, I had only one thought: I'd try to repair it.
Here is the Spectrum straight out of the box. Some of you may see an RCA cable sticking out of the back, suggesting a composite mod.
I of course connected it and there was no picture as I had been told, so I opened it up and saw the nature of the composite mod. The core of the cable had been soldered into the wire going into the RF modulator. The shield of the cable had a solder blob suggesting that it should be soldered into the RF modulator casing but it was hanging loose.
This was going to be an easy fix or so I thought. I unsoldered the RCA cable and proceeded to do a proper composite mod.
Here is the Spectrum opened and with the composite mod done. My joy was short lived as I still couldn't get any kind of picture. I measured voltages from the lower memory and got pretty much zilch on all the three rails. I first turned my attention to the 7805 regulator. It had experienced a violent death as there was a piece missing from it as you can see below.
This only occurred to me at a later point but when you look at the PSU that came with the Spectrum, it starts to look pretty obvious, what has caused it's demise.
Well, you'll get proper polarity for every other fraction of a second. :roll:
It was time to begin acquiring some parts...I went to my local electronics store and bought a new 7805 regulator plus 8 16-pin DIL sockets. I also bought 8 new 4116 chips from here as I suspected I'd have to change them anyway.
When I got home I swapped the 7805 and tried powering up.
Well, at least I got an image, of sorts. So +5V rail was fine but still no life on -5V and +12V rails. I needed to change both TR4 and TR5. As you may know TR4 is really picky about the type of transistor used limiting you practically to ZTX650 or ZTX651, which aren't easy to come by. Luckily I was able to find some 651s cheap from Bowood Electronics, so I placed an order. TR5 contains a ZTX213, which can be substituted with BC557. That I could find from Helsinki. Luckily, my sister lives there so I asked her to pick it up and send to me.
In the meantime, I snipped the old 4116s off and soldered the sockets in their place. Here's a few pictures of the procedure:
Now it was just waiting for the rest of the parts to arrive...
A few days later, the memory chips and the BC557 arrived. I had desoldered both TR4 and TR5 already, so it was simply a matter of soldering the new transistor in place.
Also the ZTX651s arrived so I soldered that in too:
I quickly checked that I was getting voltage to all three rails. The +12V rail was bit low but apparently within tolerance so I proceeded to power the Spectrum up.
Well, it's different but still not quite what were after. At this point I noticed that one of the traces coming from the CPU had lifted and cut.
Well, that obviously needed some fixing. I also began to wonder about the low value on the +12V rail. I had tried to look for the correct orientation from the BC557 datasheets and compared it to the ZTX213. Based on those I had concluded that the BC557 would come the same way as the ZTX213. A Spectrum fixing video on YouTube proved me wrong. The collector and emitter are the other way round. D'oh!
I unsoldered the TR5 and turned it around and also fixed the missing CPU rail with a piece of wire.
The moment of truth. I now had proper value on the +12V line. Crossing my fingers I re-seated the 4116 chips, hoping I had not burned them already.
First, I was greeted with this:
I managed to get it booting properly with reboot but I was still getting this striped screen every now and then, especially after a long period of being turned off.
I reworked all the new RAM socket joints and changed the upper memory RAM chips and I am now seeming to be constantly getting the copyright text as shown below on every boot.
It still needs a new membrane, which I've ordered and it's also missing a few case screws and the screw that fastens the motherboard into the case bottom. A new case bottom wouldn't hurt either as the current one is slightly warped from heat. Nevertheless, it's a nice feeling to revive one piece of microcomputing history.
Thanks to TheCorfiot over at AmiBay and Zetr0! I could not have done it without your help and support. :)
Latest one from me is a bit extreme - it's probably not possible to get a Spectrum more dead unless you torch the PCB :) This isn't finished yet but thought I'd post progress as I go.
This came to me as part of an eBay lot a while ago, and was duly dispatched to my spares box. I almost threw it out, but got to thinking that since I have the spare parts and also a spare case to rebuild it, why not do just that!
Here was the starting point:
As well as missing all of the IC's, all the electrolytic caps had been snipped, and it was missing both BA157 diodes, the 5v1 Zener and a couple of 1N4148's, mostly in the power section. The edge connector had also been poorly reflowed, leaving it uneven.
After cleaning up the board as best I could (including the edge connector) First step was to replace the coil, power socket, electrolytic caps and the missing diodes. Once this was done, I performed Zorn's resistance checks and once everything checked out, powered up.
A quick verification of voltages showed everything was in order, and swiftly on to the next step :)
Once the ULA and CPU were in place, a quick power on was attempted, however there was no video output and the CPU appeared not to be running (no clock signal from CPUCLK on the ULA).
CPUCLK turned out to be shorted to ground, this took a while to find, but a solder splash was discovered in the region of C52. This removed, the CPU clock looked more sane, but still no video.
Probing the ULA outputs showed that /Y (luminance - the B&W signal) was again shorted to ground. Comparison with another 4A board on the solder side revealed a bridge between through holes that looked factory till it was removed.
Shorts fixed, we get the following display:
All good, this is exactly what I'd expect from a Spectrum missing all it's RAM and with no ROM installed (the pattern is the ULA picking up floating values from the data bus since there's no RAM). B&W image since the LM1889N PAL colour encoder chip isn't installed yet.
So, got a quick few minutes earlier. Installed the LM1889N and powered on, got colour so the video section's now all good.
Checked the address/data lines for shorts to ground/5v, all good there too.
So only one thing for it, solder in lower RAM and ROM sockets, and go for broke:
Next jobs are:
- Fit EAR/MIC/Keyboard molex sockets
- Install upper RAM sockets and IC's
- Test thoroughly
I have a suspicion that the edge connector's still not clean enough (I removed a lot of excess solder from it, but not all as I ran out of braid). Once that's done to my satisfaction, I'll attach a few interfaces to it and see how I get on :)
! 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 :)
Wow! That is some job! Wish I had this level skill now. I used to but I think it's all gone now.:( I have a 4b board here to repair but I just feel not up to it.
@Protek - yep, it's an intact RF modulator, but composite modded using the single wire solution, so I could easily restore it to RF out if needs be.
@Firefox11 - I see you're in Ireland too - I'd be happy to have a look at the board if you'd like to get it working but don't have the time or motivation.
So, tonight's progress! Keyboard molexes in and keyboard functionality tested and working. EAR/MIC sockets also installed but I've not tested saving or loading yet.
Most importantly, we have 48K of working RAM:
My diagnostics board doesn't want to play well with this board for the moment, I suspect the edge connector isn't making good enough contact, I'll need to clean this up further once I pick up some more desoldering braid.
However my DivMMC worked fine once carefully positioned, so I was able to run the tape version of my diagnostics, which assumes lower RAM is good and just tests the upper 32K.
To continue, I'm awaiting a keyboard membrane, once this arrives I'll put together the exterior (I have one of RWAP's excellent remanufactured faceplates ready also).
Once that's done I'll also perform a proper memory soak test, test tape operations and generally give it a good workout :)
Not much to report on tonight, apart from diagnostics board FAIL (the edge connector on the 4S was actually fine - it was the socket on the diags board which wasn't making proper contact. Duly replaced.)
This done, I ran through some ULA, sound and most importantly memory soak testing which ran for over an hour, and which the machine passed with flying colours :)
Membrane arrived today, so coupled with a day off work and some spare time, I got to finishing this.
I had a spare 48K case with rubber keymat, coupled with a new membrane and one of RWAP's excellent faceplates and we have the following:
I knew the serial number of the board (the same eBay lot that contained the PCB also had what I would guess was it's 48K+ case, I passed this onto a forum member for their Harlequin since).
Serial number reinstated (not exactly true to the original typeface, I may redo this depending on how anal I'm feeling):
I must admit, I've really enjoyed reading this thread, it's very interesting to see Speccy's being brought back to life.
I dug my box of computers out of the garage and was pleasantly surprised to find I had two ZX Spectrum's rather than one. Inspired as I am, I'll refurbish the ropey one and get it looking and working good as new!
A question though - the brick power supply for the spectrum, is the dc output (that's fed to the spectrum) just rectified and smoothed or is it regulated? Just thinking that it'd be worth recapping any electrolytics in the PSU as well as the spectrum itself.
The standard Spectrum power brick is an unregulated supply that puts out about 13v off load and 11v or so on load. Part of the reason that the heatsink gets so hot is that the regulator has to dissipate this extra voltage as heat when producing the 5v regulated supply.
I myself don't use the Speccy supplies but a 9v regulated PSU rated for 2 amps. This way, the regulator don't have to work as hard :)
Definitely replace all the electrolytics in the computer, you can do the PSU ones too but it isn't as critical.
I find that the unloaded output voltage on some can get to around 14 to 15V.
If the PSU has been used a lot, it is still worthwhile renewing the capacitor.
Alternately if you have a oscilloscope, you can see the output waveform under load and make a judgement call...
! 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 :)
For my 'preservation' Spectrum's, I'll just recap, but I have a thought to get a spectrum+ for using and then do a bunch of 'life extending' mods. I'm thinking a dc/dc converter to replace the 7805, heat sink the ULA and see if there is any low-power equivalents of the other components on the MB - I'm sure other will have done similar.
If you're into modifying, then the single best thing you can do from a reliability and power consumption perspective is install a Lower Ram Replacement (LRR) module from zx.zigg.net.
This'll remove the single most unreliable aspect of the Spectrum (4116 lower RAM) and result in a power saving of over 200mA.
I've done this and other mods, they're floating round in other posts but I can sum up here if you're interested.
If you're into modifying, then the single best thing you can do from a reliability and power consumption perspective is install a Lower Ram Replacement (LRR) module from zx.zigg.net.
This'll remove the single most unreliable aspect of the Spectrum (4116 lower RAM) and result in a power saving of over 200mA.
I've done this and other mods, they're floating round in other posts but I can sum up here if you're interested.
B
Thanks for the info, I'll have a look at that.
If you do have a list of mods, I'd really appreciate it, it's also be good to have them all in one space.
I've started having a look at a replacement for the 7805 in the speccy, and think that this might be a contender, being a direct drop in.
I've started having a look at a replacement for the 7805 in the speccy, and think that this might be a contender, being a direct drop in.
A number of us on here have fitted this into a Spectrum, and it's the second most worthwhile mod after the LRR board :) It also means you can get rid of the heatsink, which makes for much easier edge connector cleaning (a pencil eraser does the job nicely, FYI).
The same guy who does the LRR board also does a URR version to replace the upper RAM. If you've done all the above and are hungry for more, you can replace the Z80 with a modern CMOS variant, and replace the 74LS logic on the board with 74HCT variants.
I've done all of the above to an issue 4A board and took measurements along the way:
Standard Spec 48K (Issue 4A) 620mA
- with switch mode regulator -220mA (400mA)
- with Lower RAM module -105mA (295mA)
- with Upper RAM module -35mA (260mA)
- with CMOS CPU -55mA (205mA)
- with 74LS replaced with HCT -15mA (190mA)
As you can see, it's possible to get down to 1/3rd of the original power consumption, which has implications if you ever want to run a setup off battery power etc.
This is a great thread. Very informative and helpful. I have never done any real restoration but its great just to see what others can do with a bit of time and skill and interest.
I liked the cleaned up ZX spectrum Plus - and the idea of socketing the [RAM?] chips I presume.
Not that it is very useful in a Spectrum, but the CMOS Z80 CPU can run from stopped to a bit beyond the marked maximum frequency.
On some development/experimentation/hobbyist boards people run them at silly slow speeds in order to see how the Z80 communicates via the address, control and data buses by using a logic probe or simple LED circuits.
! 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 :)
Comments
Thanks for posting this. My ZX81 was suffering from the same missing back porch.
My ZX81 is fitted in a rather ugly looking Crofton ZX81-Adaptakit case. This as a video out and some electronic on the keyboard to boost the composite signal. Unfortunately this only made the picture look worse. I removed these parts and replaced them with the ZX8-CCB, the picture is now perfect.
John
zx-diagnostics - Fixing ZX Spectrums in the 21st Century (wiki)
Sinclair FAQ Wiki
I deliberately went for something that looked tatty and broken on eBay, as I'm looking for a case to house my Harlequin board once Ingo sends them out and I get around to assembling.
Once it turned up though, it immediately became apparent that the case and faceplate was in much better nick than I expected (and cleaned up rather well), but also upon opening I found a yellow PCB. So the decision was made, this was a keeper :)
Usual lower RAM defective stripes along with a very noisy display greeted me upon power on. Voltages checked out, apart from a slightly low -5v (measured -4.65v).
I'd read other posts that suggested changing C44/C45 to cure display noise, also dodgy caps elsewhere couldn't be helping the -5v/12v generation.
A swift recapping later and the noise was gone, and voltages were closer to spec.
Plugging in my diags board revealed that IC9 had bought the farm, so replacing just this part gave me a successful boot to the copyright message with a full 48K of working memory.
The membrane, which was working initially, decided to give up the ghost just as I was putting the machine back together. Not to worry though, replacement on order and this machine will be 100% again.
Here's a pic for those who are interested:
B
zx-diagnostics - Fixing ZX Spectrums in the 21st Century (wiki)
Sinclair FAQ Wiki
I had thought this was only the case on early machines, but either way I'm not complaining, it'll make membrane replacement much easier :)
B
zx-diagnostics - Fixing ZX Spectrums in the 21st Century (wiki)
Sinclair FAQ Wiki
Here's no.4
I was surprised when I opened the case as I've never seen a yellow one.
Unfortunately this one doesn't seems to work very well. I doesn't tune in, I can just make out the Sinclair research bit, but it crashes after a minute or 2, and buzz from the speaker.
I can see now it has faulty capacitors, 3 are visibly bad.
Out of 9 loose spectrums I had in a draw that I got with ebay bundles, the first 2 I pulled out had these tabs. An issue 4A and 6A. I always thought it was a later thing.
zx-diagnostics - Fixing ZX Spectrums in the 21st Century (wiki)
Sinclair FAQ Wiki
An issue 3B Spectrum bought from AmiBay It was sold as defective for to be used as parts. When I declared interest to buy it, I had only one thought: I'd try to repair it.
Here is the Spectrum straight out of the box. Some of you may see an RCA cable sticking out of the back, suggesting a composite mod.
I of course connected it and there was no picture as I had been told, so I opened it up and saw the nature of the composite mod. The core of the cable had been soldered into the wire going into the RF modulator. The shield of the cable had a solder blob suggesting that it should be soldered into the RF modulator casing but it was hanging loose.
This was going to be an easy fix or so I thought. I unsoldered the RCA cable and proceeded to do a proper composite mod.
Here is the Spectrum opened and with the composite mod done. My joy was short lived as I still couldn't get any kind of picture. I measured voltages from the lower memory and got pretty much zilch on all the three rails. I first turned my attention to the 7805 regulator. It had experienced a violent death as there was a piece missing from it as you can see below.
This only occurred to me at a later point but when you look at the PSU that came with the Spectrum, it starts to look pretty obvious, what has caused it's demise.
Well, you'll get proper polarity for every other fraction of a second. :roll:
It was time to begin acquiring some parts...I went to my local electronics store and bought a new 7805 regulator plus 8 16-pin DIL sockets. I also bought 8 new 4116 chips from here as I suspected I'd have to change them anyway.
When I got home I swapped the 7805 and tried powering up.
Well, at least I got an image, of sorts. So +5V rail was fine but still no life on -5V and +12V rails. I needed to change both TR4 and TR5. As you may know TR4 is really picky about the type of transistor used limiting you practically to ZTX650 or ZTX651, which aren't easy to come by. Luckily I was able to find some 651s cheap from Bowood Electronics, so I placed an order. TR5 contains a ZTX213, which can be substituted with BC557. That I could find from Helsinki. Luckily, my sister lives there so I asked her to pick it up and send to me.
In the meantime, I snipped the old 4116s off and soldered the sockets in their place. Here's a few pictures of the procedure:
Now it was just waiting for the rest of the parts to arrive...
A few days later, the memory chips and the BC557 arrived. I had desoldered both TR4 and TR5 already, so it was simply a matter of soldering the new transistor in place.
Also the ZTX651s arrived so I soldered that in too:
I quickly checked that I was getting voltage to all three rails. The +12V rail was bit low but apparently within tolerance so I proceeded to power the Spectrum up.
Well, it's different but still not quite what were after. At this point I noticed that one of the traces coming from the CPU had lifted and cut.
Well, that obviously needed some fixing. I also began to wonder about the low value on the +12V rail. I had tried to look for the correct orientation from the BC557 datasheets and compared it to the ZTX213. Based on those I had concluded that the BC557 would come the same way as the ZTX213. A Spectrum fixing video on YouTube proved me wrong. The collector and emitter are the other way round. D'oh!
I unsoldered the TR5 and turned it around and also fixed the missing CPU rail with a piece of wire.
The moment of truth. I now had proper value on the +12V line. Crossing my fingers I re-seated the 4116 chips, hoping I had not burned them already.
First, I was greeted with this:
I managed to get it booting properly with reboot but I was still getting this striped screen every now and then, especially after a long period of being turned off.
I reworked all the new RAM socket joints and changed the upper memory RAM chips and I am now seeming to be constantly getting the copyright text as shown below on every boot.
It still needs a new membrane, which I've ordered and it's also missing a few case screws and the screw that fastens the motherboard into the case bottom. A new case bottom wouldn't hurt either as the current one is slightly warped from heat. Nevertheless, it's a nice feeling to revive one piece of microcomputing history.
Thanks to TheCorfiot over at AmiBay and Zetr0! I could not have done it without your help and support. :)
B
zx-diagnostics - Fixing ZX Spectrums in the 21st Century (wiki)
Sinclair FAQ Wiki
This came to me as part of an eBay lot a while ago, and was duly dispatched to my spares box. I almost threw it out, but got to thinking that since I have the spare parts and also a spare case to rebuild it, why not do just that!
Here was the starting point:
As well as missing all of the IC's, all the electrolytic caps had been snipped, and it was missing both BA157 diodes, the 5v1 Zener and a couple of 1N4148's, mostly in the power section. The edge connector had also been poorly reflowed, leaving it uneven.
After cleaning up the board as best I could (including the edge connector) First step was to replace the coil, power socket, electrolytic caps and the missing diodes. Once this was done, I performed Zorn's resistance checks and once everything checked out, powered up.
A quick verification of voltages showed everything was in order, and swiftly on to the next step :)
Once the ULA and CPU were in place, a quick power on was attempted, however there was no video output and the CPU appeared not to be running (no clock signal from CPUCLK on the ULA).
CPUCLK turned out to be shorted to ground, this took a while to find, but a solder splash was discovered in the region of C52. This removed, the CPU clock looked more sane, but still no video.
Probing the ULA outputs showed that /Y (luminance - the B&W signal) was again shorted to ground. Comparison with another 4A board on the solder side revealed a bridge between through holes that looked factory till it was removed.
Shorts fixed, we get the following display:
All good, this is exactly what I'd expect from a Spectrum missing all it's RAM and with no ROM installed (the pattern is the ULA picking up floating values from the data bus since there's no RAM). B&W image since the LM1889N PAL colour encoder chip isn't installed yet.
More soon :)
B
zx-diagnostics - Fixing ZX Spectrums in the 21st Century (wiki)
Sinclair FAQ Wiki
Checked the address/data lines for shorts to ground/5v, all good there too.
So only one thing for it, solder in lower RAM and ROM sockets, and go for broke:
Next jobs are:
- Fit EAR/MIC/Keyboard molex sockets
- Install upper RAM sockets and IC's
- Test thoroughly
I have a suspicion that the edge connector's still not clean enough (I removed a lot of excess solder from it, but not all as I ran out of braid). Once that's done to my satisfaction, I'll attach a few interfaces to it and see how I get on :)
B
zx-diagnostics - Fixing ZX Spectrums in the 21st Century (wiki)
Sinclair FAQ Wiki
Only, that's more of a rebuild than a repair :o
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 noticed that you have installed an RF modulator. Is it used as such or is there some kind of composite mod?
@Firefox11 - I see you're in Ireland too - I'd be happy to have a look at the board if you'd like to get it working but don't have the time or motivation.
So, tonight's progress! Keyboard molexes in and keyboard functionality tested and working. EAR/MIC sockets also installed but I've not tested saving or loading yet.
Most importantly, we have 48K of working RAM:
My diagnostics board doesn't want to play well with this board for the moment, I suspect the edge connector isn't making good enough contact, I'll need to clean this up further once I pick up some more desoldering braid.
However my DivMMC worked fine once carefully positioned, so I was able to run the tape version of my diagnostics, which assumes lower RAM is good and just tests the upper 32K.
To continue, I'm awaiting a keyboard membrane, once this arrives I'll put together the exterior (I have one of RWAP's excellent remanufactured faceplates ready also).
Once that's done I'll also perform a proper memory soak test, test tape operations and generally give it a good workout :)
B
zx-diagnostics - Fixing ZX Spectrums in the 21st Century (wiki)
Sinclair FAQ Wiki
Very impressed Mr B :)
Not much to report on tonight, apart from diagnostics board FAIL (the edge connector on the 4S was actually fine - it was the socket on the diags board which wasn't making proper contact. Duly replaced.)
This done, I ran through some ULA, sound and most importantly memory soak testing which ran for over an hour, and which the machine passed with flying colours :)
Here's a quick video of things:
B
zx-diagnostics - Fixing ZX Spectrums in the 21st Century (wiki)
Sinclair FAQ Wiki
I had a spare 48K case with rubber keymat, coupled with a new membrane and one of RWAP's excellent faceplates and we have the following:
I knew the serial number of the board (the same eBay lot that contained the PCB also had what I would guess was it's 48K+ case, I passed this onto a forum member for their Harlequin since).
Serial number reinstated (not exactly true to the original typeface, I may redo this depending on how anal I'm feeling):
And obligatory 'in use' shot:
Stick a fork in her, she's done :)
B
zx-diagnostics - Fixing ZX Spectrums in the 21st Century (wiki)
Sinclair FAQ Wiki
I dug my box of computers out of the garage and was pleasantly surprised to find I had two ZX Spectrum's rather than one. Inspired as I am, I'll refurbish the ropey one and get it looking and working good as new!
A question though - the brick power supply for the spectrum, is the dc output (that's fed to the spectrum) just rectified and smoothed or is it regulated? Just thinking that it'd be worth recapping any electrolytics in the PSU as well as the spectrum itself.
The standard Spectrum power brick is an unregulated supply that puts out about 13v off load and 11v or so on load. Part of the reason that the heatsink gets so hot is that the regulator has to dissipate this extra voltage as heat when producing the 5v regulated supply.
I myself don't use the Speccy supplies but a 9v regulated PSU rated for 2 amps. This way, the regulator don't have to work as hard :)
Definitely replace all the electrolytics in the computer, you can do the PSU ones too but it isn't as critical.
B
zx-diagnostics - Fixing ZX Spectrums in the 21st Century (wiki)
Sinclair FAQ Wiki
If the PSU has been used a lot, it is still worthwhile renewing the capacitor.
Alternately if you have a oscilloscope, you can see the output waveform under load and make a judgement call...
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 :)
For my 'preservation' Spectrum's, I'll just recap, but I have a thought to get a spectrum+ for using and then do a bunch of 'life extending' mods. I'm thinking a dc/dc converter to replace the 7805, heat sink the ULA and see if there is any low-power equivalents of the other components on the MB - I'm sure other will have done similar.
This'll remove the single most unreliable aspect of the Spectrum (4116 lower RAM) and result in a power saving of over 200mA.
I've done this and other mods, they're floating round in other posts but I can sum up here if you're interested.
B
zx-diagnostics - Fixing ZX Spectrums in the 21st Century (wiki)
Sinclair FAQ Wiki
Thanks for the info, I'll have a look at that.
If you do have a list of mods, I'd really appreciate it, it's also be good to have them all in one space.
I've started having a look at a replacement for the 7805 in the speccy, and think that this might be a contender, being a direct drop in.
Will
A number of us on here have fitted this into a Spectrum, and it's the second most worthwhile mod after the LRR board :) It also means you can get rid of the heatsink, which makes for much easier edge connector cleaning (a pencil eraser does the job nicely, FYI).
The same guy who does the LRR board also does a URR version to replace the upper RAM. If you've done all the above and are hungry for more, you can replace the Z80 with a modern CMOS variant, and replace the 74LS logic on the board with 74HCT variants.
I've done all of the above to an issue 4A board and took measurements along the way:
As you can see, it's possible to get down to 1/3rd of the original power consumption, which has implications if you ever want to run a setup off battery power etc.
Hope this helps!
B
zx-diagnostics - Fixing ZX Spectrums in the 21st Century (wiki)
Sinclair FAQ Wiki
I liked the cleaned up ZX spectrum Plus - and the idea of socketing the [RAM?] chips I presume.
I take it that the Z80 available from farnell will be OK, just run underclocked (link)
B
zx-diagnostics - Fixing ZX Spectrums in the 21st Century (wiki)
Sinclair FAQ Wiki
On some development/experimentation/hobbyist boards people run them at silly slow speeds in order to see how the Z80 communicates via the address, control and data buses by using a logic probe or simple LED circuits.
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 :)