TKez wrote: ↑Tue Jul 23, 2024 12:50 am
Hey peoples!
I'm almost done modding an LG Flatron which I posted recently about here
viewtopic.php?t=60010
Pretty old and specific thread though so I'll switch to here.
With much tinkering and head scratching, I've finally got RGB working on this board.
There were a bunch of jumpers that needed to be fitted, and the thing that really stumped me, a bunch of jumpers that needed to be removed (marked as resistors with specific values on the schemtatic). LG must have really wanted switching this to Scart to be difficult!
Seems to be working ok now, except the gamma response doesn't seem right. After about %50, things curve off to black too quickly.
No amount of FBT tweaking or service menu adjustments seems to alleviate the issue which brings me to the topic of load / impedance / terminating resistors. I must admit, I've never really got my head around impedance properly.
Firstly, could mis-matched impedance cause the gamma response to be off?
Diagnostics:
With the 75ohm resistors in place at the Scart connector position on the PCB, I wasn't getting anywhere close to 0.7Vpp at the jungle chip on the RGB lines. More like 0.2Vpp and with a bit of a spike on the edges of the square waves.
(Checking the data sheet for the VCT49xy, it is indeed looking for 0.7V at the RGB inputs)
So I removed the terminating resistors completely. Now I get full 0.7Vpp clean waveforms at the jungle chip and im not seeing noise or ghosting.
Measuring between one of those lines and ground I get ~0ohms so no sneaky load resistors elsewhere.
So should I maybe try add some higher value resistors there instead of 75ohm so there is at least some load?
Or should I just be thankful it all works as it is and try deal with perfecting the gamma in software?
I ended up just putting wire jumpers at L201, L204, L206, L206, R217, R218, R219.
I tried 75ohm resistors at R203, R204, R205, R213 which grounded most of the signal so removed them completey.
My source is MisterFPGA board VGA port which I assume should be standard 75ohm, but I have no idea how to test it.
Ok this guy is in full working order. TL;DR I tweaked the terminating resistors and added missing caps at the jungle and it was happy days.
For anyone winding up here on a journey with an LG, I figure I'll make note of what had to be done, and what I learned along the way.
Firstly, the terminating resistors were pulling down the voltage too much. I also couldn't quite understand their function and why it the image looked better without them.
As I understand it now (after a long chat with ChatGPT), if the signals only have one path to follow to ground through the IC, some of the signal can reflect back on the lines and create ghosting and interence. Something like a channel of water hitting a big turbine that can't instantly start to turn. Some of the water bounces off the blades backwards and created turbulence.
By splitting the signal and creating a second path to ground, this can relieve some pressure and smooth out the flow. In the case of 75ohm resistors, it's roughly splitting the signal In half.
But now we have half signal doing work, and half going directly ground. That means to get 0.7Vpp at the chip, we now need twice as much at the source, or about 1.4V with cable disconnected.
In my case, I was only measuring ~0.84Vpp with the cable disconnected. Not enough for 75ohms. So the easiest fix in my head was change the resistors. Indeed, placing 390ohm resistors at the Scart position instead of 75ohm gave me 0.7pp at the chip.
Later I realised what the resistors marked with a star were on the schematic! They create a voltage divider with the terminating resistors and their value would be calculated to get the voltage to 0.7Vpp. In my case I should have kept the 75ohm resistors at R203, R204, R205, R213, and put 3k9 resistors in mystery positions R19,R218,R217 where I put jumpers.
A bunch of other jumpers (marked on schematic) had to be put in place to get the RGB & FB lines before I was getting the proper 0.7Vpp waveforms at the jungle, but even with them in place I still wasn't getting a picture, no matter what voltage was applied to the FBIN(78) pin. Then I saw the problem! LG had a jumper to ground at the R525 position which should be 1.2K. I removed that jumper and bam! Picture.
Now it was all working great, but there was something up with the luminance values. There was no dark greys, things seemed to clip off to black at about %33. No amount of service menu tweaking could correct it. Then I noticed the last issue. LG had also put jumpers at the cap positions at the jungle (C515, C516, C518). Once I put the correct 0.1uf caps there it was golden.
If someone could help complete my understanding here it would be great. I would have thought those caps decoupling caps, removing the DC component from the signal. But the chip expects 0.0 - 0.7V which is the same DC range as the VGA cable no?
.