I inserted a 220 uH coil in the +12VA line going to the video chip IC14. (board issue 4a - lifted one side of R62 and put the coil between R62 and board). That, and replacing all electrolytics that showed an ESR value over 0.5 ohms, removed most of the video noise.
Ok, back to topic.
I wanted to give this a try but I think it is R42 and not R62.
If it is R42, and according to schematics is a 1Kohm resistor (am I reading right?), would it make a difference the dc resistance of the coil?
I have not looked at the schematic, but I can say that it will not work with a 1k ohm resistor in series with the 12V rail...
Typical resistors for supply filters range from 1 ohm to 120 ohms.
! 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 mislead myself searching in the vicinity of IC14 when I should be looking around the voltage converter circuit.
Yes, there is a R62 indeed and its value is 15 ohms. ;)
So I guess 15 ohms + 6 or 7 ohms would be significant.
Back to original topic: if you look closer at screenshots with the worst shadows around characters, you may notice that those shadows occur all around those characters.
Read: light shadow on a scanline above, when ZX Spectrum hasn't even output the black pixels below it! Or light shadow before a character, when (timing-wise) what happens is that voltage levels in composite video signal drop (=darker), not a little jump up before going down. Some "overshoot" may occur, but that would show on trailing edge (=to the right of characters, or in the left-most black portions). Likewise below characters, that is: a whole scanline later than when characters were sent by the ZX Spectrum!
Which leaves as only logical conclusion that these are artefacts caused by the image processing in LCD screen etc (a direct effect of the different way that LCD screens render an image as compared with a CRT). And thus fix should be searched in sharpness / brightness / contrast settings.
Vertical banding (like shown in 2nd picture in this thread) is often caused by power supply fluctuations. But here you should be looking at the large(r) electrolytic caps, especially those close to voltage regulators (or -converters) and in video encoder / ULA supply lines. A suitable coil in series may do wonders here. But first replace those old electrolytic caps if that wasn't done yet.
Horizontal banding may be caused by too small value of series capacitor in composite video output. 47uF may work but is too small really. 100 uF is better, 220 uF may be better still. Not much point going far beyond that though.
Trying to improve picture quality you can do the following:
Take a ferrite core. Wind two wires in parallel as often as possible (using thin flexible wire or isolated thin copper wire). Then you have two starts and two ends. solder a Cinch connector on both of the two starts. Look for that end belonging to the inner pin of the Cinch connector (npt that ground pin). solder it to the composite output pin of the Spectrum. now take the other open end (that belonging to the ground pin of the Cinch cable). now switch on the Spectrum and try to connect the mentioned (ground) end of the wire to different ground pins on the Spectrum as ULA-ground pin, gorund pin of the 1889 and so on while observing picture quality.At the end solder this wire to the ground pin that gives best picture quality.
Back to original topic: if you look closer at screenshots with the worst shadows around characters, you may notice that those shadows occur all around those characters.
Read: light shadow on a scanline above, when ZX Spectrum hasn't even output the black pixels below it! Or light shadow before a character, when (timing-wise) what happens is that voltage levels in composite video signal drop (=darker), not a little jump up before going down. Some "overshoot" may occur, but that would show on trailing edge (=to the right of characters, or in the left-most black portions). Likewise below characters, that is: a whole scanline later than when characters were sent by the ZX Spectrum!
Which leaves as only logical conclusion that these are artefacts caused by the image processing in LCD screen etc (a direct effect of the different way that LCD screens render an image as compared with a CRT). And thus fix should be searched in sharpness / brightness / contrast settings.
The white outline around the dark or black pixels is due to the picture processing in the LCD TV. See if you can find a menu entry / setting to switch off sharpness and any picture improvement options.
Vertical banding (like shown in 2nd picture in this thread) is often caused by power supply fluctuations. But here you should be looking at the large(r) electrolytic caps, especially those close to voltage regulators (or -converters) and in video encoder / ULA supply lines. A suitable coil in series may do wonders here. But first replace those old electrolytic caps if that wasn't done yet.
I recommend that ALL the electrolytic capacitors are replaced. For the best picture quality, use low ESR capacitors for the inverter/converter power supply section (see list below).
Note on some boards, the power supply section may have had Sinclair recommended modifications done. Also the ID of C78/C80 changed between issues/boards. For boards that have been modified, this 22uF capacitor will not have a silk screen ID.
The working voltages show are the minimum. 25V or 35V types are okay as well.
Horizontal banding may be caused by too small value of series capacitor in composite video output. 47uF may work but is too small really. 100 uF is better, 220 uF may be better still. Not much point going far beyond that though.
Yes, use a 100uF low ESR with a 100nF ceramic wired in parallel for best bandwidth. But this will not solve the banding if you are already using a 100uF type.
Note: not all modifications tried on all board issues (I have mainly tried stuff on an issue 2).
! 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 :)
Horizontal banding may be caused by too small value of series capacitor in composite video output. 47uF may work but is too small really. 100 uF is better, 220 uF may be better still. Not much point going far beyond that though.
Would this also apply to s-video output?
If not, what would be the way to reduce horizontal banding in s-video?
Horizontal banding may be caused by too small value of series capacitor in composite video output. 47uF may work but is too small really. 100 uF is better, 220 uF may be better still. Not much point going far beyond that though.
Would this also apply to s-video output?
If not, what would be the way to reduce horizontal banding in s-video?
Yes, the luminance signal has the same or slightly bigger bandwidth as a composite signal. The chrominance (colour) channel does not need such a high value, but no harm is done by fitting a 100uF capacitor here.
! 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 inserted a 220 uH coil in the +12VA line going to the video chip IC14. (board issue 4a - lifted one side of R62 and put the coil between R62 and board). That, and replacing all electrolytics that showed an ESR value over 0.5 ohms, removed most of the video noise.
Following Tuukka's advice I soldered the coil in series with R62, which was reduced so that total resistivity is about 15 ohm.
! 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 :)
Ok, I have noticed two benefits on picture quality:
- If I crank up brightness I no longer see wavy lines zig-zagging across the screen. Very nice, picture feels clean and solid.
- I can increase the color setting of monitor with very little bleeding.
If it wasn't for the jail bars and I would say it's almost on par with +2 CVBS output! =D>
However, I previously did the s-video mod on this spectrum according to rga24 specifications, so composite video mod will surely suffer from dot-crawl.
Horizontal banding may be caused by too small value of series capacitor in composite video output. 47uF may work but is too small really. 100 uF is better, 220 uF may be better still. Not much point going far beyond that though.
Would this also apply to s-video output?
If not, what would be the way to reduce horizontal banding in s-video?
Yes, the luminance signal has the same or slightly bigger bandwidth as a composite signal. The chrominance (colour) channel does not need such a high value, but no harm is done by fitting a 100uF capacitor here.
Mark
I tried fitting a 220uF capacitor on the luminace output and the monitor lost picture.
Can't understand why, but perhaps something about the way the s-video mod works on the spectrum is out of spec?
I don't know the exact reason (as I don't know the details of the circuit that you are using), but too large a value may cause problems. Whatever drives the output has to charge and discharge the capacitor. If it cannot do this, there will be signal loss...
! 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 have just made a picture comparison between my 48K and 48K+ and started to wonder if there are other factors influencing picture quality.
I noticed that my 48K has a clearer background (less jail bars, no 'hot air column effect', almost no interference from keyboard pressing). Could this be because I have the upper and lower ram modules installed (in my case zorn's) instead of the true IC's?
I realised that my 48K+ has a mettalic plate holding the membrane against the back of the keyboard. Could this cause interference in some way? Should it not be earthed?
I noticed that my 48K has a clearer background (less jail bars, no 'hot air column effect', almost no interference from keyboard pressing). Could this be because I have the upper and lower ram modules installed (in my case zorn's) instead of the true IC's?
These use modern SRAM chips, not old DRAM chips. That could be the reason (they work completely differently internally). The SRAM only needs a +5V supply. The old 4116 DRAM runs off a +12V supply (the same supply that the video circuits run off). Also keep in mind that different issue Spectrum boards are all slightly different. And even the same issue board may use different components. All this has a greater or lesser effect.
I realised that my 48K+ has a mettalic plate holding the membrane against the back of the keyboard. Could this cause interference in some way? Should it not be earthed?
Nothing in any ZX Spectrum is earthed! The PSU does not have an earth wire... Or did you mean connected to the 0V / DC ground?
Even if you did earth this metal plate, it is unlikely that it would make any difference to the Spectrum's video output. It may help reduce interference emitted from the Spectrum slightly though.
Connecting it to the 0V / DC ground would not make any difference either.
The biggest thing that causes interference on the video picture, is the logic, memory and CPU chips that the Spectrum is made from. They are electrically very noisy, and some of this noise is bound to find a way into the video circuitry.
! 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 don't know the exact reason (as I don't know the details of the circuit that you are using), but too large a value may cause problems. Whatever drives the output has to charge and discharge the capacitor. If it cannot do this, there will be signal loss...
Mark
Now that you mention it, I recall Richard A. saying there was no need for such a capacitor.
Perhaps it's wiser to listen to the author of the mod and just leave things as they are.
I realised that my 48K+ has a mettalic plate holding the membrane against the back of the keyboard. Could this cause interference in some way? Should it not be earthed?
Nothing in any ZX Spectrum is earthed! The PSU does not have an earth wire... Or did you mean connected to the 0V / DC ground?
Even if you did earth this metal plate, it is unlikely that it would make any difference to the Spectrum's video output. It may help reduce interference emitted from the Spectrum slightly though.
Connecting it to the 0V / DC ground would not make any difference either.
Mark
Sorry, I meant connected to DC ground.
I just thought that might have some influence because the 48K is plastic all round the membrane.
! 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 biggest thing that causes interference on the video picture, is the logic, memory and CPU chips that the Spectrum is made from. They are electrically very noisy, and some of this noise is bound to find a way into the video circuitry.
Mark
What you said has left me thinking about the potential benefit of more modern/alternative manufacturer replacements.
For instance, any idea if there is a substitute for the Z80 that could be electrically less noisy than Zilog's original?
Edit: Just found a post where Balford refers to a CMOS cpu (drawing less current). Could that one be of interest?
I had good results while reducing the (relatively) large output gotten via a large capacitor (220uF) by means of a 'volume control' created with a 220 Ohm variable resistor. I got the strong idea that this controlling had effect on the level or the shape of the sync pulses, as I could get a more quiet picture. Of course the final brightness was adjusted on the monitor.
Of course this is all related to the specs of the input of the used monitor(s). Two, in my case.
Best thing I found for reducing video interference was replacing C5 to C8 (the 12v decoupling caps for the lower RAM) with 1nf ceramics. On mobile at the moment, but if you search my post history you'll find it :)
Best thing I found for reducing video interference was replacing C5 to C8 (the 12v decoupling caps for the lower RAM) with 1nf ceramics. On mobile at the moment, but if you search my post history you'll find it :)
! 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
B
zx-diagnostics - Fixing ZX Spectrums in the 21st Century (wiki)
Sinclair FAQ Wiki
Ok, back to topic.
I wanted to give this a try but I think it is R42 and not R62.
If it is R42, and according to schematics is a 1Kohm resistor (am I reading right?), would it make a difference the dc resistance of the coil?
Typical resistors for supply filters range from 1 ohm to 120 ohms.
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 was looking in the wrong place! :-B
I mislead myself searching in the vicinity of IC14 when I should be looking around the voltage converter circuit.
Yes, there is a R62 indeed and its value is 15 ohms. ;)
So I guess 15 ohms + 6 or 7 ohms would be significant.
Read: light shadow on a scanline above, when ZX Spectrum hasn't even output the black pixels below it! Or light shadow before a character, when (timing-wise) what happens is that voltage levels in composite video signal drop (=darker), not a little jump up before going down. Some "overshoot" may occur, but that would show on trailing edge (=to the right of characters, or in the left-most black portions). Likewise below characters, that is: a whole scanline later than when characters were sent by the ZX Spectrum!
Which leaves as only logical conclusion that these are artefacts caused by the image processing in LCD screen etc (a direct effect of the different way that LCD screens render an image as compared with a CRT). And thus fix should be searched in sharpness / brightness / contrast settings.
Vertical banding (like shown in 2nd picture in this thread) is often caused by power supply fluctuations. But here you should be looking at the large(r) electrolytic caps, especially those close to voltage regulators (or -converters) and in video encoder / ULA supply lines. A suitable coil in series may do wonders here. But first replace those old electrolytic caps if that wasn't done yet.
Horizontal banding may be caused by too small value of series capacitor in composite video output. 47uF may work but is too small really. 100 uF is better, 220 uF may be better still. Not much point going far beyond that though.
Take a ferrite core. Wind two wires in parallel as often as possible (using thin flexible wire or isolated thin copper wire). Then you have two starts and two ends. solder a Cinch connector on both of the two starts. Look for that end belonging to the inner pin of the Cinch connector (npt that ground pin). solder it to the composite output pin of the Spectrum. now take the other open end (that belonging to the ground pin of the Cinch cable). now switch on the Spectrum and try to connect the mentioned (ground) end of the wire to different ground pins on the Spectrum as ULA-ground pin, gorund pin of the 1889 and so on while observing picture quality.At the end solder this wire to the ground pin that gives best picture quality.
Greets Ingo.
Note on some boards, the power supply section may have had Sinclair recommended modifications done. Also the ID of C78/C80 changed between issues/boards. For boards that have been modified, this 22uF capacitor will not have a silk screen ID.
The working voltages show are the minimum. 25V or 35V types are okay as well.
Yes, use a 100uF low ESR with a 100nF ceramic wired in parallel for best bandwidth. But this will not solve the banding if you are already using a 100uF type.
Note: not all modifications tried on all board issues (I have mainly tried stuff on an issue 2).
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 :)
Would this also apply to s-video output?
If not, what would be the way to reduce horizontal banding in s-video?
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 :)
Following Tuukka's advice I soldered the coil in series with R62, which was reduced so that total resistivity is about 15 ohm.
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 :)
Ok, I have noticed two benefits on picture quality:
- If I crank up brightness I no longer see wavy lines zig-zagging across the screen. Very nice, picture feels clean and solid.
- I can increase the color setting of monitor with very little bleeding.
If it wasn't for the jail bars and I would say it's almost on par with +2 CVBS output! =D>
However, I previously did the s-video mod on this spectrum according to rga24 specifications, so composite video mod will surely suffer from dot-crawl.
Yes, I consider it worth the trouble.
I tried fitting a 220uF capacitor on the luminace output and the monitor lost picture.
Can't understand why, but perhaps something about the way the s-video mod works on the spectrum is out of spec?
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 my 48K has a clearer background (less jail bars, no 'hot air column effect', almost no interference from keyboard pressing). Could this be because I have the upper and lower ram modules installed (in my case zorn's) instead of the true IC's?
I realised that my 48K+ has a mettalic plate holding the membrane against the back of the keyboard. Could this cause interference in some way? Should it not be earthed?
These use modern SRAM chips, not old DRAM chips. That could be the reason (they work completely differently internally). The SRAM only needs a +5V supply. The old 4116 DRAM runs off a +12V supply (the same supply that the video circuits run off). Also keep in mind that different issue Spectrum boards are all slightly different. And even the same issue board may use different components. All this has a greater or lesser effect.
Nothing in any ZX Spectrum is earthed! The PSU does not have an earth wire... Or did you mean connected to the 0V / DC ground?
Even if you did earth this metal plate, it is unlikely that it would make any difference to the Spectrum's video output. It may help reduce interference emitted from the Spectrum slightly though.
Connecting it to the 0V / DC ground would not make any difference either.
The biggest thing that causes interference on the video picture, is the logic, memory and CPU chips that the Spectrum is made from. They are electrically very noisy, and some of this noise is bound to find a way into the video circuitry.
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 :)
Now that you mention it, I recall Richard A. saying there was no need for such a capacitor.
Perhaps it's wiser to listen to the author of the mod and just leave things as they are.
Sorry, I meant connected to DC ground.
I just thought that might have some influence because the 48K is plastic all round the membrane.
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 :)
What you said has left me thinking about the potential benefit of more modern/alternative manufacturer replacements.
For instance, any idea if there is a substitute for the Z80 that could be electrically less noisy than Zilog's original?
Edit: Just found a post where Balford refers to a CMOS cpu (drawing less current). Could that one be of interest?
My zodiac sign is to blame!
Of course this is all related to the specs of the input of the used monitor(s). Two, in my case.
I'm convinced that one doesn't pose a problem.
However, on my 48k+ the further I move the keyboard/membrane/metallic plate away from the motherboard the less interference is shown on screen.
B
zx-diagnostics - Fixing ZX Spectrums in the 21st Century (wiki)
Sinclair FAQ Wiki
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 :)