Icky - Progress

[exxos]

I need to start concentrating on the main issues I have, the keyboard / mouse repeats and the video flicker. Will scope out around the ACIAs tonight looking for ringing etc as you mention. One observation is I extended the keyboard connector using jumper leads to be able to put the keyboard away from the STF board and this increased the repeating.

Be worth scoping out both ends of the keyboard cables then. I assume you connected both gnd wires (missing pin 8 ?). Either bad power getting to the keyboard or ringing on a signal. Obviously getting worse with a longer cable so at least now we have a glue.

[Icky]

Finally fixed my flickering screen. As well as the flicker I noticed that on the flicker there was a bounce in the image.

Doing some major surgery to the SCART cable following this diagram



So the best colour settings were to use 100 ohms on the RGB for me and the TTL CSYNC is selected.

The flicker solver was the diode and capacitor - thats what stopped the bounce and white flicker.

Only thing I have left is the noise that rolls up the screen - I am thinking that this something to do with noise on the PSU as I get different noise patterns using different PSUs.

[exxos]

Only thing I have left is the noise that rolls up the screen - I am thinking that this something to do with noise on the PSU as I get different noise patterns using different PSUs.

I think the RGB buffers are picking up noise on the PSU and clock lines as mentioned before... I would assume you could probe around the drivers with your scope and may be able to affect the pattern on the screen..

[exxos]

The flicker solver was the diode and capacitor - thats what stopped the bounce and white flicker.

I really don't understand that capacitor.. Having a electrolytic on a vertical sync signal is essentially loading it.. And then connecting it to the start RGB switch line..

Shouldn't pin 8 of the RGB connector pulled up to 12V be connected to the scart RGB mode input pin 16 ?!

EDIT:
Nevermind, the pullup probably hasn't enough current to do it. Its essentially using VSYNC to charge the capacitor to keep pin 16 high.. If it didn't have the diode, the capacitor would try to get charged and discharged very fast with VSYNC and that isn't good thing to do. But still using VSYNC to charge a 47uF cap I don't really like the idea of.. it comes unbuffered from GLUE.. Will run a simulation on it....

EDIT2:

Actually 5V / 75R = 66mA... or assuming 0.7V drop across the diode 4.3V / 75R = 57mA. This excluding the initial surge of charging the capacitor from zero volts which could peek to 1.5amps. So I assume as GLUE can drive VSYNC under 75R load then it should be OK. But charging the cap from zero is acting like a short circuit on GLUE which isn't a great idea.

Incidentally, Simulation without the cap drops to like 20mA.

cct.JPG (93.63 KiB) Viewed 3820 times

Of course isn't much option in this case to do anything different. Though like I said before, I think GLUE should be isolated from the outside world and have its own buffer.. in fact I don't know FPGA output current, but that may need external buffering anyway..

[PaulJ]

I also looked at this on a scart cable I made using that schematic. It was my guess that the cap would charge and provide a crappy dc level. I choose to use the 12 volts to generate the line level instead. I never used the cable because a commercial one arrived.

[exxos]

I also looked at this on a scart cable I made using that schematic. It was my guess that the cap would charge and provide a crappy dc level. I choose to use the 12 volts to generate the line level instead. I never used the cable because a commercial one arrived.

Well this is assuming pin 16 has a 75R resistor on it.. maybe Icky can measure the scart pin on resistance on the TV direct to see ?

[Icky]

I also looked at this on a scart cable I made using that schematic. It was my guess that the cap would charge and provide a crappy dc level. I choose to use the 12 volts to generate the line level instead. I never used the cable because a commercial one arrived.

Well this is assuming pin 16 has a 75R resistor on it.. maybe Icky can measure the scart pin on resistance on the TV direct to see ?

Just measured its 75R

[exxos]

Just measured its 75R

Thanks. So will add some buffers on the next revision board. It should probably work with another 75R in series with the diode ?

[czietz]

Well this is assuming pin 16 has a 75R resistor on it.. maybe Icky can measure the scart pin on resistance on the TV direct to see ?

Just measured its 75R

This is the value to be expected from a standard-compliant (EN50049) SCART input.

This excluding the initial surge of charging the capacitor from zero volts which could peek to 1.5amps.

Parasitic inductance and resistance will in fact limit the peak current to significantly less than that.

[Icky]

Thanks. So will add some buffers on the next revision board. It should probably work with another 75R in series with the diode ?

I'll give that a try.