One LPF vs two - noise vs sharpness differences?

The place for all discussion on gaming hardware
Post Reply
thebigcheese
Posts: 707
Joined: Sun Aug 21, 2016 5:18 pm

One LPF vs two - noise vs sharpness differences?

Post by thebigcheese »

Before I go cutting traces and running wires to experiment, curious what other people's findings have been. I am using a Mickris RGB board in a CoreGrafx that goes into an OSSC. OSSC has the default LPF on, but the Mickris board has the LPF disabled with no convenient pad to switch between on and off like Voultar's board. The image is razor sharp, but there is a tiny bit of noise with some games and it would be nice if I could get rid of it. I can still enable the LPF on the board, but it would mean cutting a trace and connecting the relevant pin to ground.

All that aside, I am wondering if having the signal go through two LPFs would even help and, if so, just how much loss of sharpness there would be. If it fixes the noise at the cost of just a little decreased sharpness, then I don't mind. But if it ends up looking more like the AV Driver I have in my TG16... no thanks. That is still noisy and significantly less sharp. Anyone try this or have any thoughts on it?
syboxez
Posts: 99
Joined: Sun Jul 24, 2016 4:33 am

Re: One LPF vs two - noise vs sharpness differences?

Post by syboxez »

An LPF will not solve your issue here. It will only work to make things slightly less sharp. Instead of trying to mask the issue, you should attack it at its root.

Video noise in analog video can be caused by several things. Chroma interference can pop up as jailbars or checkerboarding depending on the console. Some consoles also have jailbars as a result of not enough decoupling capacitance on the ICs. Dirty power rails can cause several issues, most commonly when using a bad quality SMPS, you see bars of noise going up or down the screen. Dirty power rails can also be caused by bad capacitors. Noise can also be caused by unshielded or poorly shielded cables.

Can you be more specific with your issues and show pictures?
thebigcheese
Posts: 707
Joined: Sun Aug 21, 2016 5:18 pm

Re: One LPF vs two - noise vs sharpness differences?

Post by thebigcheese »

I've already done the jailbar fix and routed the cabling several different ways until the noise was minimized. I can post pictures of the mod work if you like, but that part has mostly been diagnosed in the recent CoreGrafx jailbar thread.

I don't have a capture card, or at least not one that would work for this, so I can't really shoe the noise very well. It consists of very thin, very faint vertical lines and a very faint static. But it seems to occur only in some places/games while being completely absent in others. And it's not even visible from the couch, so I am really nitpicking at this point.

My question was mainly about the type of LPF employed by the 7374. For example, if the OSSC cuts 3 dB starting at 9 MHz and the 7374 does the same, then I wind up with more noise reduction. But it looks like the 7374 starts at 9.5 MHz, so maybe not so useful in this case.

I have not tried recapping the console, though they seem to be fine from a visual inspection. I've also not replaced the 7805. I am fairly certain it's just that the console is noisy as that seems to be my experience with the PCE family in general. My TG16 is way worse than this even after a recap. So that's mainly why I was trying to cover it rather than fix it. Not sure it can be fixed outside of moving everything to a new, better designed PCB...

Should also mention that I am using fully shielded coax cable, so although I am using sync on composite, I shouldn't be getting any noise from that.
syboxez
Posts: 99
Joined: Sun Jul 24, 2016 4:33 am

Re: One LPF vs two - noise vs sharpness differences?

Post by syboxez »

What power supply are you using? That can make all the difference. Also you should disable the LPF on the 7374 and just stick to the one in the OSSC.

What type of wiring are you using inside the console for the mod? Are any wires run close to signals which could cause interference? If it is the internal wiring causing it, you could try using mini-coax for the mod if you think it will fit in the console.
thebigcheese
Posts: 707
Joined: Sun Aug 21, 2016 5:18 pm

Re: One LPF vs two - noise vs sharpness differences?

Post by thebigcheese »

The LPF on the mod is always disabled unless I cut a trace/lift a pin and run some wire. That's why I am asking about this before doing that. Don't particularly want to rehash everything that went in the other thread, but the short version is that putting pretty much anything on the top side of the PCB inside the CoreGrafx results in noise. No coax wire (that I have, anyway) would fit and it would just be a nightmare trying to properly ground all the shields, so instead I ran the wire on the underside outside of a shield, as you can see here: https://photos.app.goo.gl/HKE864n5uYAjJZHV8. The mod board itself is mounted just behind the DIN to avoid interference from the various 62xx chips: https://photos.app.goo.gl/ySfpptcP3cn3R9uq7. This is not a perfect solution by any means as we are now subject to noise picked up by the traces, but it is noticeably less noise than running the wires over or even near any of the 62xx chips in my testing. The ribbon cable is also running outside the flexible shield (very carefully routed to stay outside of it, as you can see) but inside the metal shield on the underside of the console, so it is about as well-shielded as it can be.

It also didn't seem to make any difference where I grabbed power from, so I am grabbing it from a via right by the board and ground from the expansion port. Grabbing power here allows me to avoid running it too close to the video wires, which could also add noise. Finally, in that last picture I actually wound up reversing the wires for R and B because I wired up the ribbon cable backward, so that wire isn't going over the THS7374 like that anymore.

So, suffice it to say that I don't think there is much else I can do in terms of routing/wiring to reduce the noise. I have tested with two different PSUs, an official Genesis PSU and the official NEC IFU PSU. I primarily use the NEC one as the CoreGrafx is hooked up to the IFU all the time, but the noise is present with both PSUs and with the console by itself or attached to the IFU.
syboxez
Posts: 99
Joined: Sun Jul 24, 2016 4:33 am

Re: One LPF vs two - noise vs sharpness differences?

Post by syboxez »

thebigcheese wrote:The LPF on the mod is always disabled unless I cut a trace/lift a pin and run some wire. That's why I am asking about this before doing that. Don't particularly want to rehash everything that went in the other thread, but the short version is that putting pretty much anything on the top side of the PCB inside the CoreGrafx results in noise. No coax wire (that I have, anyway) would fit and it would just be a nightmare trying to properly ground all the shields, so instead I ran the wire on the underside outside of a shield, as you can see here: https://photos.app.goo.gl/HKE864n5uYAjJZHV8. The mod board itself is mounted just behind the DIN to avoid interference from the various 62xx chips: https://photos.app.goo.gl/ySfpptcP3cn3R9uq7. This is not a perfect solution by any means as we are now subject to noise picked up by the traces, but it is noticeably less noise than running the wires over or even near any of the 62xx chips in my testing. The ribbon cable is also running outside the flexible shield (very carefully routed to stay outside of it, as you can see) but inside the metal shield on the underside of the console, so it is about as well-shielded as it can be.

It also didn't seem to make any difference where I grabbed power from, so I am grabbing it from a via right by the board and ground from the expansion port. Grabbing power here allows me to avoid running it too close to the video wires, which could also add noise. Finally, in that last picture I actually wound up reversing the wires for R and B because I wired up the ribbon cable backward, so that wire isn't going over the THS7374 like that anymore.

So, suffice it to say that I don't think there is much else I can do in terms of routing/wiring to reduce the noise. I have tested with two different PSUs, an official Genesis PSU and the official NEC IFU PSU. I primarily use the NEC one as the CoreGrafx is hooked up to the IFU all the time, but the noise is present with both PSUs and with the console by itself or attached to the IFU.
Might I recommend some copper tape? Basically adds another PCB layer that's exclusively ground plane.
thebigcheese
Posts: 707
Joined: Sun Aug 21, 2016 5:18 pm

Re: One LPF vs two - noise vs sharpness differences?

Post by thebigcheese »

syboxez wrote:Might I recommend some copper tape? Basically adds another PCB layer that's exclusively ground plane.
Where would you put it? The wire is sandwiched between two ground places at the moment, so I'm not sure what good adding more would do. Putting it over the chips would allow me to run wires over them (assuming I connect the copper tape to ground), but boy does that run some bad risks of components shorting. Plus closing things off from proper ventilation. Assuming all of that checked out, the video signals would still have the same contamination from trace routing whether I get the signal direct off the chip or from the expansion port unless I lift the pins/cut the traces, which I don't really want to do.
User avatar
FBX
Posts: 2348
Joined: Wed Feb 18, 2015 10:18 am
Location: DFW area, Texas
Contact:

Re: One LPF vs two - noise vs sharpness differences?

Post by FBX »

I've been hearing a lot lately that the Core Grafx is notorious for image noise issues. It might be worth switching to a PCE instead. Of course I myself get the best results from the Super Grafx, but those cost an arm and a leg these days.
thebigcheese
Posts: 707
Joined: Sun Aug 21, 2016 5:18 pm

Re: One LPF vs two - noise vs sharpness differences?

Post by thebigcheese »

PCE has it's own issues, namely that then I have to deal with that RF modulator. So I dunno. Plus I just got this :p I can certainly live with the noise, was mostly just curious if there was anything else to try.
Post Reply