I just tried the 2.5V solution by a voltage divider with the 560 and 180 ohm resistor:
It gave 2.25V, but the LCD TV does not switch to RGB.
The cause is the current, I found out.
I thought that offering just 1-3V was enough, but (link):
"The switching signal needs to be able to source upto 20mA into pin 16 @ 3V for switching to take place."
And we don't want that from the ZX Spectrum, as Paul describes in his documentation.
The ZX Spectrum cannot offer 20mA on 12V safely, and there is no output port offering 5V.
The best solution then is using an external supply, and I agree with Paul's solution of the LM317 to be the best.
But I do not like an external supply myself to be honest.
I can think of the alternative of a cable in between the 9V and toastrack/+2.
But I'll stick to composite.
If the LCD tv does not auto switch to RGB, can the modifcation still be made and switch to RGB via the remote control? Would the modification give a strong enough RGB signal to be stable and clear?
If the LCD tv does not auto switch to RGB, can the modifcation still be made and switch to RGB via the remote control?
Would the modification give a strong enough RGB signal to be stable and clear?
My Sony LCD does not have that option.
I think many TV's won't have it, as they use the SCART standards, so 0-0.4V for composite and 1-3V for RGB on pin 16.
IMHO the people who thought about that BLANKING signal were... Well, a bit alternative themselves.
Why demand 20mA for only a switching signal?!
One other thing I was brainstorming about:
Why not using a microcontroller with some memory to build a HDMI signal, having the microcontroller respond to any video ram change in the ZX Spectrum (so only sensing the adress and datalines basically).
So no analog circuits, no video ram reading and contention problems.
Simply a transparant innovative digital video cloning in separate ram, with HDMI output...
No more need for a ULA ;)
Ah, I see the problem... you have a Sony LCD.
My LG, Matsui and Bush LCD TV and even the bog standard no-name 14" CRT TV have the option to switch to RGB on the SCART, my LG DVD recorder has the option of Composite or RGB input on the SCART.
I use to swear loyalty at Sony products, but now I just swear at them because they don't have the features and user options of other brands - unless you download the service manuals and perform some inbuilt hacks to switch on the options that have been locked out for normal users.
Ah, I see the problem... you have a Sony LCD.
My LG, Matsui and Bush LCD TV and even the bog standard no-name 14" CRT TV have the option to switch to RGB on the SCART, my LG DVD recorder has the option of Composite or RGB input on the SCART.
I use to swear loyalty at Sony products, but now I just swear at them because they don't have the features and user options of other brands - unless you download the service manuals and perform some inbuilt hacks to switch on the options that have been locked out for normal users.
Wow, that's interesting....
So in that case it's 'enough' to just offer the correct RGB signals.
But still, there is definately no good solution for a toastrack and +2, other than (as long as people call this 'good') using 12V from separate supply or edge connector.
Yup, I can set the output of a Freeview TV box to either composite or RGB on the scart, and set the TV input to composite or RGB on the scart via the remote control.
Providing the device has an RGB output via the scart then the RGB scart input can be selected.
I have a few video recorders that do not output RGB on the scart, so I have to switch to
composite on the TV/DVD recorder so that the video recorder picture can be seen -the audio is there, just no picture. Once the TV is set to scart composite the picture is there, in not so good composite.
I just have to remember to switch RGB on/off when using the video recorder with the DVD recorder.
As to getting the Spectrum to do the same... that is a different thing. I'm still using the RF connector... nasty picture.
I usually solve this by taking a 5v feed off the internal regulator then take that through an additional regulator - a small one offering 3v at 150mA should be more than enough (they're about the same size as a transistor and readily available). In the +2 I change the links to offer 12v through pin 5 but leave the link off so pin 5 is isolated. It's then possible to connect the 3v output from the secondary regulator directly to pin 5.
I know a few people use a resistor to try and drop the voltage - the problem with that is you need to know the current draw on pin 16 at the TV end and unfortunately there's no standard specification for this so if you're going from a 12V source for example there's a risk you'll kill the TV if you don't drop enough voltage.
Hope this helps, I've done it in several +2s with good results, but as always throughly test first, particularly ensuring the regulator isn't getting too hot!
I know a few people use a resistor to try and drop the voltage - the problem with that is you need to know the current draw on pin 16 at the TV end and unfortunately there's no standard specification for this
Pin 16 is for fast switching to RGB. Therefore its input resistance is 75 Ohm. http://martin.hinner.info/vga/scart.html
That's why it is safe to simply use a resistor in series to a DC source. If you still are afraid of any overvoltage you could place a 2.7V zener diode or a green LED from this line to ground to limit the voltage. But it is really not necessary.
I've heard this a lot in the past, and I thought this was the case a while back however, I've recently learned that it's no guarantee - I know there's plenty of literature out there to suggest all control and video signals carry a 75ohm impedance but I know from a trusted friend with a great deal of experience in electronics and specifically TV circuits that this is definitely not always the case, and Pin 16 does not have an official specified impedance. The only lines guaranteed to have an input resistance of 75ohms are the video signals i.e. RGB. My friend has seen many variations between different televisions in the past years and in light of this, you really want to be bringing the voltage down to the proper level before it goes anywhere near pin 16.
However as you say there's probably no need to panic, I suspect that the vast majority of TVs will be absolutely fine but there is an (admittedly slim) chance that this won't always be the case. Since I fix computers for other people I have to guarantee that they won't damage the equipment the customer plugs them into, therefore I do ere on the side of extreme caution, but for anyone just doing this at home it's probably just fine but just be aware of the potential risks.
Now let's use 560 ohm for R2.
R3 = ( (180 + 560) * (2.5 / 12) ) / (1 - (2.5 / 12) ) = 195 ohm.
The current now is: I = 12 / (180 + 560 + 195) = 12 / 935 = 1,2 mA.
Your calculation of current does not include the impedance of the BLANKING input, which is 75 ohms. Then 75 in parallel with 195 would give about 54 ohms. This makes the total impedance 794 ohms and not 935 ohms. The current then works out at 15mA. However a bigger impact is on the input voltage to the BLANKING pin, which now works out at 0.818V.
R3 can only ever act to decrease the 75 ohm input impedance presented by the BLANKING input. For an input of 1V at the BLANKING pin, the total impedance should be 900 ohms, and hence a 900-180-75=645 ohms should be placed in series (R2 in your diagram, with R3 omitted). This yields a current draw of 13.3mA.
Paul says, SCART for 128K is not suitable for 128K+2 gray, since +2 has already BRIGHT mixed with RGB. However most cables including RetroComputerShack are marketed 128K/128K+2(gray).
Do they mix BRIGHT twice ( and overload BRIGHT signal from ULA :confused: ) ?
The grey +2 already includes circuitry to mix the BRIGHT signal but the 128 does not. Using a cable designed for the 128 will cause the BRIGHT signal to be mixed twice. This will produce an ok looking picture but the relative brightness seen between the 128 and the +2 will be different. It will probably also draw more current. Ideally each machine should use a separate cable design.
Why not using a microcontroller with some memory to build a HDMI signal, having the microcontroller respond to any video ram change in the ZX Spectrum (so only sensing the adress and datalines basically).
So no analog circuits, no video ram reading and contention problems.
Simply a transparant innovative digital video cloning in separate ram, with HDMI output...
The SPECTRA approach could be adapted for the 128K machines (I've had a prototype running to prove this) but it would require a different board. I would be interested to make such a board if there is sufficient demand. By outputting a RGB SCART signal, an external adapter box can always be used to generate HDMI or whatever might replace it in the future...
An advantage of my SPECTRA interface is that it uses the +12V from the Spectrum expansion bus to generate the necessary SCART signalling voltages and so there is no need for an external power source!
My LG, Matsui and Bush LCD TV and even the bog standard no-name 14" CRT TV have the option to switch to RGB on the SCART, my LG DVD recorder has the option of Composite or RGB input on the SCART.
Older CRTs seem to be offer this facility (my old portable Philips TV does) but I think many newer LCD/LED models don't (they expect devices to fully support the SCART specification). Some Panasonic TVs also appear to expect a PAL colour burst signal to be present on the sync line else the picture will be shifted to the left or worse can be unstable horizontally.
I know from a trusted friend with a great deal of experience in electronics and specifically TV circuits that this is definitely not always the case
I could imagine that different manufacturers support the specification to different levels of accuracy. So I think it is always better to aim for about 2V for the BLANKING signal than try be too close to the 1V or 3V limits, but I have not actually experienced such a problem myself and in theory anything in that range should be sufficient.
My friend has seen many variations between different televisions in the past years and in light of this, you really want to be bringing the voltage down to the proper level before it goes anywhere near pin 16.
I think using a voltage regulator should overcome any issues between different manufacturer's implementations.
I've tried the LM317 solution with an external 12V supply, but didn't get near 3V, and the LM317 heated up.
I used one 1N4148 diode.
I would think that a DVI / HDMI interface would be more easy to develop as it would be a digital to digital interface, so mainly software (FPGA) based (although you already did the Spectra, so you know all ins and outs).
Furthermore I think adding a SCART to HDMI converter brings that amount of extra costs that people would not be interested anymore in general.
And outputting DVI / HDMI directly would also result in a much sharper picture.
The SPECTRA approach could be adapted for the 128K machines (I've had a prototype running to prove this) but it would require a different board. I would be interested to make such a board if there is sufficient demand.
I have visions of needing several adapter boxes in series, as TV standards change...
Nevertheless I'd buy a 128k SPECTRA the 48k model is great.
Comments
It gave 2.25V, but the LCD TV does not switch to RGB.
The cause is the current, I found out.
I thought that offering just 1-3V was enough, but (link):
"The switching signal needs to be able to source upto 20mA into pin 16 @ 3V for switching to take place."
And we don't want that from the ZX Spectrum, as Paul describes in his documentation.
The ZX Spectrum cannot offer 20mA on 12V safely, and there is no output port offering 5V.
The best solution then is using an external supply, and I agree with Paul's solution of the LM317 to be the best.
But I do not like an external supply myself to be honest.
I can think of the alternative of a cable in between the 9V and toastrack/+2.
But I'll stick to composite.
Download the latest version of Bomb Munchies Ver2210 4th July 2020
My Sony LCD does not have that option.
I think many TV's won't have it, as they use the SCART standards, so 0-0.4V for composite and 1-3V for RGB on pin 16.
IMHO the people who thought about that BLANKING signal were... Well, a bit alternative themselves.
Why demand 20mA for only a switching signal?!
One other thing I was brainstorming about:
Why not using a microcontroller with some memory to build a HDMI signal, having the microcontroller respond to any video ram change in the ZX Spectrum (so only sensing the adress and datalines basically).
So no analog circuits, no video ram reading and contention problems.
Simply a transparant innovative digital video cloning in separate ram, with HDMI output...
No more need for a ULA ;)
[edit]
Or FPGA: http://www.fpga4fun.com/HDMI.html
[edit]
New topic: http://www.worldofspectrum.org/forums/showthread.php?t=47469
My LG, Matsui and Bush LCD TV and even the bog standard no-name 14" CRT TV have the option to switch to RGB on the SCART, my LG DVD recorder has the option of Composite or RGB input on the SCART.
I use to swear loyalty at Sony products, but now I just swear at them because they don't have the features and user options of other brands - unless you download the service manuals and perform some inbuilt hacks to switch on the options that have been locked out for normal users.
Download the latest version of Bomb Munchies Ver2210 4th July 2020
Wow, that's interesting....
So in that case it's 'enough' to just offer the correct RGB signals.
But still, there is definately no good solution for a toastrack and +2, other than (as long as people call this 'good') using 12V from separate supply or edge connector.
Providing the device has an RGB output via the scart then the RGB scart input can be selected.
I have a few video recorders that do not output RGB on the scart, so I have to switch to
composite on the TV/DVD recorder so that the video recorder picture can be seen -the audio is there, just no picture. Once the TV is set to scart composite the picture is there, in not so good composite.
I just have to remember to switch RGB on/off when using the video recorder with the DVD recorder.
As to getting the Spectrum to do the same... that is a different thing. I'm still using the RF connector... nasty picture.
Download the latest version of Bomb Munchies Ver2210 4th July 2020
I know a few people use a resistor to try and drop the voltage - the problem with that is you need to know the current draw on pin 16 at the TV end and unfortunately there's no standard specification for this so if you're going from a 12V source for example there's a risk you'll kill the TV if you don't drop enough voltage.
Hope this helps, I've done it in several +2s with good results, but as always throughly test first, particularly ensuring the regulator isn't getting too hot!
That's why it is safe to simply use a resistor in series to a DC source. If you still are afraid of any overvoltage you could place a 2.7V zener diode or a green LED from this line to ground to limit the voltage. But it is really not necessary.
Ingo.
However as you say there's probably no need to panic, I suspect that the vast majority of TVs will be absolutely fine but there is an (admittedly slim) chance that this won't always be the case. Since I fix computers for other people I have to guarantee that they won't damage the equipment the customer plugs them into, therefore I do ere on the side of extreme caution, but for anyone just doing this at home it's probably just fine but just be aware of the potential risks.
Your calculation of current does not include the impedance of the BLANKING input, which is 75 ohms. Then 75 in parallel with 195 would give about 54 ohms. This makes the total impedance 794 ohms and not 935 ohms. The current then works out at 15mA. However a bigger impact is on the input voltage to the BLANKING pin, which now works out at 0.818V.
R3 can only ever act to decrease the 75 ohm input impedance presented by the BLANKING input. For an input of 1V at the BLANKING pin, the total impedance should be 900 ohms, and hence a 900-180-75=645 ohms should be placed in series (R2 in your diagram, with R3 omitted). This yields a current draw of 13.3mA.
Interesting you measured 2.25V, I would have expected about 0.8V... And at 0.8V I would think it unlikely the TV would switch.
The grey +2 already includes circuitry to mix the BRIGHT signal but the 128 does not. Using a cable designed for the 128 will cause the BRIGHT signal to be mixed twice. This will produce an ok looking picture but the relative brightness seen between the 128 and the +2 will be different. It will probably also draw more current. Ideally each machine should use a separate cable design.
1V into 75 ohms means the current is 13mA. 3V into 75 ohms means the current is 40mA. For 1.5V into 75 ohms means the current is 20mA.
A very similar approach to my SPECTRA SCART RGB interface for the 48K Spectrum. ;-)
The SPECTRA approach could be adapted for the 128K machines (I've had a prototype running to prove this) but it would require a different board. I would be interested to make such a board if there is sufficient demand. By outputting a RGB SCART signal, an external adapter box can always be used to generate HDMI or whatever might replace it in the future...
An advantage of my SPECTRA interface is that it uses the +12V from the Spectrum expansion bus to generate the necessary SCART signalling voltages and so there is no need for an external power source!
Older CRTs seem to be offer this facility (my old portable Philips TV does) but I think many newer LCD/LED models don't (they expect devices to fully support the SCART specification). Some Panasonic TVs also appear to expect a PAL colour burst signal to be present on the sync line else the picture will be shifted to the left or worse can be unstable horizontally.
I could imagine that different manufacturers support the specification to different levels of accuracy. So I think it is always better to aim for about 2V for the BLANKING signal than try be too close to the 1V or 3V limits, but I have not actually experienced such a problem myself and in theory anything in that range should be sufficient.
As far as I can tell, the official SCART specification goes under the designation EN50049-1 (and the corresponding IEC designation is IEC933-1). A copy of EN54009-1 can be found here:
www.encnn.com/design-training/Peritelevision/BS_EN_50049-1 Peritelevision connector.pdf
Page 7 states the input impedance of pin 16 is 75R...
Interesting, I see that the input impedance for pin 8 (see page 6) is stated as <=10k not >10k as I have on my website and have seen stated elsewhere
I think using a voltage regulator should overcome any issues between different manufacturer's implementations.
I used one 1N4148 diode.
I would think that a DVI / HDMI interface would be more easy to develop as it would be a digital to digital interface, so mainly software (FPGA) based (although you already did the Spectra, so you know all ins and outs).
Furthermore I think adding a SCART to HDMI converter brings that amount of extra costs that people would not be interested anymore in general.
And outputting DVI / HDMI directly would also result in a much sharper picture.
Ben
I have visions of needing several adapter boxes in series, as TV standards change...
Nevertheless I'd buy a 128k SPECTRA the 48k model is great.