Sorry I don't follow you. Should I solder the cut pin 6 to GND or just forget about this mod and restore the RGB amp?Originally Posted by Calpis
EDIT: GND to pin 6 doesn't help.
Fudoh on shmups forum have tried this on two model 1's and it didn't help on either of them :/
Last edited by Bearking; 04-12-2011 at 07:23 AM.
Konsolkongen
Saw this site, thought it might have a degree of interest on the RGB issue?
http://www.retrocorner.net/index.php...e_modification
You lifted/cut pin 6 right? Then you don't have to do anything because the oscillator causing the interference stops.
Here's what could be going on:
Pin 6 is the input to an inverter/NOT gate, pin 5 is the inverter's output which is right next to the blue input. Normally pin 5 and 6 go to a crystal which makes a feedback loop which causes the inverter to toggle at 3.58/4.43 MHz. The inverter is digital so it outputs a square wave which has high frequency components @ 10, 18, 21 MHz etc at a higher voltage than the small video signal. To save money the MD uses a single crystal oscillator that controls all timing in the console, it divides the 53ish MHz down to 3.58/4.43 and feeds the '1145's pin 6 (input) with that. Even though there is no crystal, pin 5 (output) still toggles, but I don't think it has a load to absorb any energy so it becomes a very short range transmitter.
Another possibility is that the two signal wires (PCB traces) run in parallel (as is definitely the case with the blue and composite output traces). Two conductors have a very small capacitance across them, but it's enough to have two signals couple if they are close enough and the frequency is high enough.
Another possibility (the last I think) is that there is significant switching noise from the oscillator. When some types of logic switch from one state to another they draw significant current for just an instant. All chips are wired to the power rails in parallel and same with components inside a chip. All wires have some inductance to them and when the chips switch levels and current goes through the wire, the inductance causes the component's local power rail to bounce up and down (it's local because it's separated from the main one by resistance, capacitance and inductance based on the wire's high frequency properties, length and composition). This may affect one subcircuit, one chip, or multiple chips, but basically circuits on that local rail may fail (or couple the rail jitter as interference) if they use the rails as a voltage reference which many circuits do. For this reason it's common for all digital chips to have "bypass" or "decoupling" capacitors very close to every chip's power rail. They act as a local energy storage and low pass filter to keep the their local voltage stable and noise from other chips out. I haven't checked but I don't think the MD has a bypass cap near the RGB encoder, though it probably should.
those who can't make, mod
Hell of a post. Wish I had the wherewithal to understand it all.Does this mean that lifting pin 5 on the CXA could fix Bearking's continuing interference issues?
No because pin 5 is the out of the in which is pin 6. If nothing is going in nothing is coming out...
...I think.
I'm think I'm starting to understand the issue, and also beginning to wish I had knowledge like Calpis does!
Seems weird that it works for some of us and not for others...cable issue as shown in Bacteria's post?
Last edited by Twimfy; 04-12-2011 at 06:34 PM. Reason: Removed arrogance/added padding.
Yes exactly, if the oscillator is not running it's held in a single state = no AC = no interference
those who can't make, mod
If that's the case then cutting the trace leading to pin 6 as far back as possible might work? But if this is the case the jailbars should have been improved some just by cutting the pin, right? It looks exactly the same as before.
I'm back to using the RGB amp btw :)
Konsolkongen
Cutting pin 6's trace wouldn't matter, it's too far from pin 4. If pin 5 (the output) actually went somewhere which I assume it doesn't, then that could be the trace possibly in parallel with pin 4.
those who can't make, mod
Makes sense :)
Konsolkongen
Sorry, I worded it wrong. You could trace the signal back on the motherboard and sever its connection someplace. I'm going by pictures found on the net as well, but according to this image, 6th pin on the CXA leads to C28 and R23 (highlighted on the picture). Removing the connection from there is simpler and undo-able. According to schematics found on google, C28 is what you should disconnect, R23 is some blob of resistors connected to many other things.
I've highlighted the chip and pin in the picture (this is not my motherboard by the way...which it was, the region jumpers are in a much nicer place!)
As far as I can see the traces for that pin run under the chip so clipping it is probably the best option unless there is something on the underside of the board you can do, or you can follow the trace back far enough to it's origin but as you can see there's a lot going on in that corner.
On the plus side the legs are pretty hefty and could be fixed by even the worst solderer.
It disables colour through composite and S-video yes.
And pin5 is indeed not connected to anything. Neither is pin11 (composite sync out).
edit: forum upload resized the picture... here's the full resolution one.
Last edited by Druid II; 04-13-2011 at 01:19 PM.
1HCg6o6zJkxtjNfyzGHtwZbXgCC7Kdf231
Just did this to the other Megadrive I have. Worked fine there as well.
Strange both EU ones? Both Fudoh's and mine are JAP.
Anyway after a lot of trial and error I think I have found the best combination of resistors to use in the RGB amp (if your MegaDrive doesn't benefit from cutting pin 6). This gives much better colors than before without colorbleeding or color changing :)
R,G,B should be tapped directly from the SEGA 315-5313 chip.
This gives a rocksolid picture on my XRGB-3 setup.
Again default > amp > default... :
Last edited by Bearking; 04-13-2011 at 02:56 PM.
Konsolkongen
Yes, both EU. Is Fudoh having the same problems? Can you link to where he's discussing it?
Looking good, anyway. I'd say brightness is a touch high this time but honestly if you'd have shown me the pictures of the RGB amp in isolation from the originals I wouldn't have known.
We have been discussing it in private messages :)
Konsolkongen
Start talking, Fudoh ;)
Your input is always valuable. It was that page on scanlines that really kicked me into high gear on the whole RGB thing.
Quick question:
What are the model numbers of your machines - so we can figure out whether there's a difference there that could be affecting the success of this.
Megadrive PAL-l
Model no. 1600-05
serial 120113353
The other one is the same, serial 120029447. Both have the EXT port.
Motherboard version would be useful as well. They're all different even if the cases and features are identical.
I'll post mine later, can't be bothered to walk upstairs either.
My JAP says:
5402: VA5
5433: VA6
Yes it says both.
Konsolkongen
Damnit, this thread makes me want to check this out too. But for that, I need to make an RGB cable for my MD1.
One question: Most schematics say that I need a 220uf capacitor on the r/g/b pins, is this necessary or can I omit them? What difference will it make?
Last edited by Druid II; 04-13-2011 at 07:04 PM.
1HCg6o6zJkxtjNfyzGHtwZbXgCC7Kdf231
Posting Permissions
Reply With Quote
Bookmarks