Very strange Sega CD video issue

[+-^CrossBow^-]

This will be very difficult to describe and I will try and get some video up later if I haven't resolved it later tonight. But here goes..

 

I went to play some of the homebrew Escape 2042 last night and was curious to try and use an audio CD again while playing the game. However, I have a model 1 sega CD and apparently the game isn't compatible with that model. Pier Solar also gives me issues with the Sega CD model 1 but works fine on model 2s and CDX units.

 

Anyway, I first powered on the system without a cart and immediately noticed that the graphics only within the window on the bios screen was displaying a snow like pattern that was moving seemingly random all over the screen?! But ONLY in the window section where the Sega CD logo is moving around. The window frame, and the text along the bottom etc were fine. I then took the CD and forced it into cd player mode. And that screen looked just fine. No issues at all. I then put a game into the genesis cart slot, and that too looked just fine and normal. I then put in Sega CD and that also looked normal?! I take everything back out and just fire up the system without a game and noticed that the snow pattern was now even worse on the bios screen?!

 

So for as near as I can tell, this is only happening with video like FMV or video that is directly being generated by the Sega CD.

 

Now for what I have done to troubleshoot this further...

 

First I noticed that my genesis is now running ridiculously hot temperature wise inside it. I also have a 32x attached and noticed when I was pulling the genesis 1 to 32x video cable out that the metal ground around the genesis 1 connection was too hot to touch?! Same results for the RF shielding and especially the default heatsink! Somewhere a short has been made causing the components to heat up way more than they should!

 

So to verify if the issue with this video problem was coming from the SEga CD itself, I then plugged in my spare model 1 va6 genesis into the sega CD and powered it on. No video issues...

 

So the problem appears to be coming from the sega genesis itself but only on video input from the sega CD. I'm suspecting that whatever or where the short is, is apparently interfering with the VDP in some way. I did find that one of my ground wires came loose from my s-video out jack and was likely shorting against the chroma out, but I don't think that would cause what I'm seeing.

 

Any ideas on this? I would normally post some thing like at the Sega-16 forums, but their forums system isn't working properly right now because I've sent multiple requests for account activation and resets with no replies...

 

[+-^CrossBow^-]

Okay...an update.

 

I spent a few hours last night essentially undoing and redoing both the original s-video mod I did on this thing years ago and redid my YM3438 add on mod. I wasn't able to find any shorts other than the one loose ground wire off my s-video out connector. I also replaced the s-video connector as my soldering work on that one (It being my first s-video mod ever), wasn't very pretty or even done very well. New connector gives me more room to solder easier, I tied both grounds together on the new connector for a single wire solution and was able to get actual heat shrink tubing applied to prevent such shorts in the future.

 

Anyway...all of that to say. That the video issue has been fixed when using a Sega CD. I'm not assuming that the ground short was causing the issues, but still odd that it only affected the video input from the Sega CD and not everything else as well?

 

However, even after reflowing the solder joints and checking both 7805c regulators, they are still getting way to hot and eventually go into thermal shutdown.

 

I'm not assuming the only cause of this since the voltages all seem to check, is that my converted gamecube power supply into a Trio power supply is the cause? I checked the voltages coming off all three points and the voltage is a steady 12.5. However, the amps is reading 4.85 steady being pushed to the regulators. My understanding is that the supply should only be providing the current that is actually being needed? So why there is nearly 5amps being sent down to the regulators doesn't make sense. Granted that was a ready without a load on it so maybe that is why. I can tell you that my other Genesis model 1 doesn't exhibit these heat issues and is using the same GC switching supply for it as well.

[Osgeld]

supply will provide as much current as the circuit is using (or until it shuts off or burns out) , sounds like your's is using 5 amps

 

no current is ever sent, its sucked

[+-^CrossBow^-]

supply will provide as much current as the circuit is using (or until it shuts off or burns out) , sounds like your's is using 5 amps

 

no current is ever sent, its sucked

Correct.

 

The readings I took, was by unplugging all three connections so that the power supply isn't actually providing any power to anything. I take my meter probes and first measured the DC voltage. 12.5 steady on all three adapter connections. I then set my meter to DC current mode and my Fluke (ancient series 77) was showing a full 4.85A from all three connections. So, since I didn't have anything plugged into this, there shouldn't have been much of a reading at all from the amps correct? In fact to really know how many amps are being drawn, I would need to make the measurements at the 7805c input voltage and ground points.

 

Maybe the 7805c regulators are going bad and I should just pop in two new ones? Could it be due to the mods? I also noticed that the YM3438 chip is also really warm. But that could be due to the proximity of it right near and just under the heatsink for the regulators and that entire area is getting crazy hot. I really need to get a newer meter with a temp probe on it for things like this.

 

The original 7805c are +5 1.5A output regulators. Maybe I should try going with a larger amperage output regulator? Recommendations?

Edited June 22, 2017 by -^Cro§Bow^-

[Osgeld]

measure the resistance of the 7805's pin 1 to 2 and 3 to 2

 

1&2 should be in the meg ohm range, 2&3 should be in the 15ish k range

 

 

pure guess here, but if it is on the 5 volt supply side you would see a pretty darn big voltage drop or if they are crapping out a scary rise, what's the voltage tween pin 2&3 (2 is ground 3 should be +5) a good 7805 should start freaking out and oscillating when it gets close to its current limit

Edited June 22, 2017 by Osgeld

[+-^CrossBow^-]

measure the resistance of the 7805's pin 1 to 2 and 3 to 2

 

1&2 should be in the meg ohm range, 2&3 should be in the 15ish k range

 

 

pure guess here, but if it is on the 5 volt supply side you would see a pretty darn big voltage drop or if they are crapping out a scary rise, what's the voltage tween pin 2&3 (2 is ground 3 should be +5) a good 7805 should start freaking out and oscillating when it gets close to its current limit

 

Interesting, I just picked up some IC probes for my ancient Fluke 77 to make stuff that like easier without me having to carefully hold the probes manually all the time. I will see about looking into measuring this and keeping it attached as it heats up to see what happens. I actually did measure them last night and one is outputting 4.7 volts and the other 4.95 as I recall. I remember when I was playing with the DC to DC converters that I thought it was odd that they had slightly different outputs so I swapped them when I reinstalled them and the output voltage stayed at the IC location on the board and didn't follow the 7805s. So the slightly lower voltage reading from the one is due to something else in that circuit I'm guessing?

[Osgeld]

could be

[phoenixdownita]

Dunno if it matters but I had somewhat similar experience wrt getting a Genny 2 superhot and thermal shutdown eventually with the added need to wait 20 mins to turn it back on (no video issues that I could tell though). You should see the rig I put together (a dual 7805 with small balancing resistors in the out line) to try and power the thing beside thermal shutdown of the onboard one how I didn't get zapped it's a mistery.

 

In my case I did the SVideo mod swapping out an SKA2195D for a CXA2075 ... while doing so I missed that one of the pins of the old SKA was used as a voltage reference for the audio circuit, usually in less than 5 minutes of Outrun I will get an extremely hot Genny and the audio will fade to zero .... I fixed it with a voltage divider (two 10K Ohm) as suggested in a god forsaken website (the CXA2075 has that pin NC anyway). Note that I have another Genny 2 in which I performed the same exact mod with no issues and no need to fix the audio/heat (it too had the SKA2195D but the motherboard was different, I think one was a VA1.8 the other a VA3).

 

 

I am not saying you have the same issue (likely you do not as I venture you are using a Genny 1 as there's real no point to use a Sega CD model 1 with a Genny 2) but maybe, just maybe given you fidgeted with the audio subsystem you hit something along those lines .... or maybe it is something totally different. For sure 5A is way too much current.

 

On a side note I replaced all the 7805 in my consoles with modern switching replacements that can source up to 3A (guaranteed 2.5A) and stay basically stone cold ... as I was at it on many systems I was able to remove the heat sinks when I could with no ill effect (the new VRs are really cold).

[+-^CrossBow^-]

I am not saying you have the same issue (likely you do not as I venture you are using a Genny 1 as there's real no point to use a Sega CD model 1 with a Genny 2) but maybe, just maybe given you fidgeted with the audio subsystem you hit something along those lines .... or maybe it is something totally different. For sure 5A is way too much current.

 

On a side note I replaced all the 7805 in my consoles with modern switching replacements that can source up to 3A (guaranteed 2.5A) and stay basically stone cold ... as I was at it on many systems I was able to remove the heat sinks when I could with no ill effect (the new VRs are really cold).

 

Yes I am using an va2 HDG model 1 genesis for all of this. The only other odd thing on the audio side with adding in the 3438 ontop of the 2612, is that you tie the +5 from one of the pins off the chips to pin 16 for some reason. I've never understood why this is. But it was to help fix the timing issues when trying to use a discrete 3438 in a model 1 genesis. I've done that same mod now on about a half dozen different model 1s and haven't seen that section get hot at all.

 

Now last night I was able to play the genesis and leave it running on a game's demo mode for about 5 hours last night solid. While it was still very warm, it didn't shut down and no video anomalies at all.

 

Oh..I did try changing out the 7805s with DC-DC converters. It worked, however.... there was very noticeable audio crackling, odd audio tones being played, and horrendous floor noise generated by using them. I took them out and put the regulators back in. The Genesis is the only console I've not had good luck changing the regulators out with DC-DC versions. All my other classic consoles have been converted over. I was using the Traco ones that include filtering and don't require additional caps, but it didn't make a difference.

 

The other thing I've done differently with my genesis is that instead of the usual white heatsink compound, I am using artic silver and had wet sanded the areas on the regulators and heatsink for better heat transfer. It could be the damn thing is getting so hot on the heatsink because it is actually sinking more heat than it was able to previously. I hadn't thought of that until just now.

 

I'm still looking to maybe change out the regulators with 2A or 2.5A output variants. But it is getting difficult to find 220 type through hole linears that can do that. Mouser only stocks 1.5A versions..ugh... I'm thinking with the added 3438 and hand wired s-video mod boards, plus a few other broken trace patches here and there on this old girl, that it might be drawing more than 1.5A current from the regulators. I need to crack it open tonight and measure this again.

 

Can you tell I love working on this stuff!!! LOL

[+-^CrossBow^-]

Update:

 

Just did some testing using my IC probes so I could more easily test things and this is what I'm getting from at the 7805s. They both have more or less the same readings with only the input amperage being different on them.

 

Both output 4.96 volts steady.

Both show 11.25 volts on the input lines.

The 7805 closest to the YM2612 section is showing an input Amps of 4.31

The 7805 closest to the power jack is showing input Amps of 3.89

 

Keep in mind all of this is with a 32x and Sega CD model 1 attached and powered as well.

 

I did the same tests with just the Genesis and a single genesis power supply. The results were:

 

Both output 4.96 volts steady.

Both show 10.42 volts on the input lines

Same amperage readings which, I find odd given that the original power supply can only output 1.2A max. So how are the 7805s showing such a large input amperage reading if the power supply they are using isn't even able to provide that much amperage in the first place?

[Osgeld]

they should be shutting themselves off well before that

[SignGuy81]

Update:

 

Just did some testing using my IC probes so I could more easily test things and this is what I'm getting from at the 7805s. They both have more or less the same readings with only the input amperage being different on them.

 

Both output 4.96 volts steady.

Both show 11.25 volts on the input lines.

The 7805 closest to the YM2612 section is showing an input Amps of 4.31

The 7805 closest to the power jack is showing input Amps of 3.89

 

Keep in mind all of this is with a 32x and Sega CD model 1 attached and powered as well.

 

I did the same tests with just the Genesis and a single genesis power supply. The results were:

 

Both output 4.96 volts steady.

Both show 10.42 volts on the input lines

Same amperage readings which, I find odd given that the original power supply can only output 1.2A max. So how are the 7805s showing such a large input amperage reading if the power supply they are using isn't even able to provide that much amperage in the first place?

 

Can you describe how you are checking the input amps? Where are you breaking the circuit or are you using a clamp type meter? If you are using the amps setting on a multimeter you would need to break the circuit somewhere and place the leads in series. For example take an AC adapter you don't care about and cut one of the wires and connect one test lead to one of them and the other to the other one, so the meter is in series. If you don't have an AC adapter you want to cut up you can remove a component lead and put in jumper wires to connect to instead in series. Or if clamp meter instead of cutting one of the wires would be to separate the two so you can get the clamp around one of them. Not saying you haven't done done it right but those readings seem odd for both regulators that are still outputting as they should.

 

 

EDIT:

The readings I took, was by unplugging all three connections so that the power supply isn't actually providing any power to anything. I take my meter probes and first measured the DC voltage. 12.5 steady on all three adapter connections. I then set my meter to DC current mode and my Fluke (ancient series 77) was showing a full 4.85A from all three connections. So, since I didn't have anything plugged into this, there shouldn't have been much of a reading at all from the amps correct? In fact to really know how many amps are being drawn, I would need to make the measurements at the 7805c input voltage and ground points.

I understand how you were doing it now. You were first checking the the connections at the AC adapter connectors on amps setting not plugged into the consoles in parallel, which is wrong, and I'm assuming you done the same with the voltage regulator inputs. There would have been no current flow at all on the output side of the AC adapters not plugged into a console because there was no load. You got the reading you did because you checked wrong. Check current in series, voltage in parallel.

Edited June 25, 2017 by SignGuy81

[SignGuy81]

 

You were first checking the the connections at the AC adapter connectors on amps setting not plugged into the consoles in parallel, which is wrong,

 

To long for me to go back and edit this but I left out an important comma.

 

You were first checking the the connections at the AC adapter connectors on amps setting not plugged into the consoles, in parallel, which is wrong,

 

Without that comma I realized it may come off as me saying the problem was not checking it in parallel on amps but was trying to say that you were checking it with the consoles not plugged in, and then that you were also checking them in parallel. I think people got what I meant plus I think I made it clear in another paragraph about which way to check but I just wanted to clarify.

[+-^CrossBow^-]

Well...here is the odd thing. I had the meter in Amp DC mode and those are the numbers it gives. But if I then move my red probe to the 10amp connection point and try to test again, this actually causes the switching gamecube supply to shut down and then no voltage at all is being provided? Unplugging the gamecube switching supply and plugging it back in, it reset it and it starts working again.

 

Just realized, you might not know what I'm talking about. I use a modified gamecube power supply that another AA member here modified for me to use as a trio adapter to power the Genesis setup with. I understood why I was getting the high readings without a load, but didn't understand why I was still getting those readings when measured from the regulator input voltage lines?

 

I also just realized I never told you what the output amps reading was from the output voltage side. Both regulators will fluctuate on this as different things are displayed..etc. But they average to only about .9A on both. I guess my main concern is why the regulators are generating so much heat to begin with as compared to other systems I have including another model 1 HDG genesis that doesn't get nearly as hot?! Whatever was causing the issue before seems to have settled out because over the past two nights, the genesis hasn't reset itself or anything else odd since I redid the s-video mod connections and tidied up the 3438 mod. My personal Genesis is always the first for new mods. So the 3438 mod looked pretty ugly and had too much exposed wire leads.. And as I stated above, the s-video connector's chroma out had a loose ground wire that was shorting against lord knows what.

[SignGuy81]

Well...here is the odd thing. I had the meter in Amp DC mode and those are the numbers it gives. But if I then move my red probe to the 10amp connection point and try to test again, this actually causes the switching gamecube supply to shut down and then no voltage at all is being provided? Unplugging the gamecube switching supply and plugging it back in, it reset it and it starts working again.

 

 

Once again this is because you are trying to check amps in parallel. The meter has very low impedance on the amps scale and basically you were shorting the power supply out when you placed it across the power.

 

You have to break the circuit to test amps with test leads from a multimeter.

 

 

. I understood why I was getting the high readings without a load, but didn't understand why I was still getting those readings when measured from the regulator input voltage lines?

 

It doesn't matter what type of power supply was used, the output of them, not hooked to up to a load(Genesis, Sega CD, 32X, etc.) will have no current flow if not plugged in. And like I said if you wanted to test it while plugged in you would have to cut on of the wires and put the test leads in series.

 

 

[SignGuy81]

[+-^CrossBow^-]

Gotcha. Well, I'm not cutting back on the insulation on my power supply lines just to test that then. For one given the location of where the power supply is, I can't even get to it that easily only to the supply adapter leads and they are pretty short in all honesty. What I can do, though not with the gamecube supply, because again I can't get back to it very easily. Is to connect my kill-a-watt off a standard genesis supply and connect that up. But that would only test the Genesis itself and not the entire amperage draw off of everything. Ehh...it is working it seems at this point and if it isn't shutting off or reseting, then let it be. LOL

 

Thank you for explaining that better to me.

[SignGuy81]

Well, you maybe able to find some ac adapter extension cable like this http://www.ebay.com/itm/Extension-Cable-Cord-for-Foscam-IP-Cameara-Power-AC-Adapter-5-FT-/272266912055?epid=1573033851&hash=item3f645f6137:g:8XAAAOxy7MtRr6QKthat you can cut a wire, plug into the end of the ac adapter and into the genesis, check the amps, then move it to the sega cd, check the amps, then 32x.

 

The one I linked is not the correct one I'm sure though just showing as an example, you would have to find one that matches the plug size(I think you would just need one as they are all the same plug size right?)

Edited June 25, 2017 by SignGuy81

[+-^CrossBow^-]

Close, but the 32x uses the Genesis model 2 plug on it. My Kill-a-watt is able to measure amperage. What I could do though not totally accurate, it just plug all the separate power adapters into a surge strip by themselves. Then plug that strip into the kill-a-watt. But that is only going to measure what the adapters as a whole are drawing off the AC. Again, it really isn't that critical at this point.

[SignGuy81]

Close, but the 32x uses the Genesis model 2 plug on it. My Kill-a-watt is able to measure amperage. What I could do though not totally accurate, it just plug all the separate power adapters into a surge strip by themselves. Then plug that strip into the kill-a-watt. But that is only going to measure what the adapters as a whole are drawing off the AC. Again, it really isn't that critical at this point.

 

I know you aren't that concerned about it at this point but just wanted to add one more thing. I thought you didn't have 3 separate power supplies and only had the converted gamecube power supply which was why you couldn't use the kill-a-watt meter individually. If you have 3 separate power supplies also instead of plugging the kill-a-watt meter into wall outlet, then power strip into that, instead plug the power strip into the wall outlet, plug the kill-a-watt meter into one of the strip outlets, then plug one of the adapters into that, and the other two in any other spots on the strip, turn everything on, and check what you have on one(check with everything going, using CD player, running CD game, etc), turn everything off, then swap the ac adapter with another into the kill-a-watt meter, turn back on again and check that one, then do same for the last one and you will have the individual reading on all of them.

 

EDIT:

 

I just thought of something that might explain why you said that. I might be wrong but I just got to thinking you may have meant the way the slots are aligned on your power strip that you cannot plug everything in to it plus the kill-a-watt meter. I know some strips have them aligned the way that it makes it difficult to do so because they are aligned with the length of the strip and you have to mount the power adapters in a way that they cover a bunch of the outlets. And even the ones where the slots are aligned perpendicular to the sides of the strip can be so close together it can be hard to plug adapters in side by side. May have to use two separate strips if that is the case.

Edited June 25, 2017 by SignGuy81

[+-^CrossBow^-]

Correct I do have a combined Gamecube to trio adapter setup. But that doesn't mean I also don't have the actual individual power supplies LOL! A single GC power supply only needs one plug on the surge strip and since it is a switching supply, it is more efficient vs the 3 separate wall warts. And yes, I could measure each separately etc. I've actually got quite a few large server style strips where I could plug all three at once quite easily. This was originally to try and figure out why the 7805s generate so much heat. In the end, now that I've corrected at least one short in some very early mod work I did plus cleaned up and prevented any other possible shorts, this simply isn't critical at this point.

 

In fact to get back on topic and to provide the solution to my original problem:

 

It appears that the ground wire that came loose from my older and crappy soldered s-video connector was causing some issues with the VDP. Additionally, there could have been some possible shorts within the 3438 mod as my resistor and capacitor leads were way to long and unprotected. All of that has been corrected, and the Genesis now works just fine.

[Osgeld]

super

[+-^CrossBow^-]

LOL....naturally after playing on the Genesis for several hours the other night the issue came back!? But I now know where the cause is...or rather what triggers it. Apparently the 32x. Yeap.. I have a 32x attached along with the Sega CD. If I take the 32x off, then the video anomaly is not present. Put the 32x back on, and the video from the Sega CD is all digital snowy again. It appears to be a loose connection at the cart port as slightly moving the 32x around can lessen the issue and in some cases correct it, but then with reduced colors as the image then takes on a sepia like tone to it. Guess I need to reform the pins on the Genesis some more along with some de-ox-id to my contacts on both 32x and Genesis.

[+-^CrossBow^-]

Further update on this very odd issue. It is most definitely a heat related issue and only appears to affect something through the 32x itself. The video anomaly with the Sega CD attached is confirmed to only happen with the 32x attached and after the system has been playing a game for a while. Like being on for more than an hour. I took my new meter with a temp probe and just stuck into the case for an ambient temp rating. The ambient are temp by the rear expansion port area has a peak of about 112° F. A good portion of that temp appears to be coming not only from the two VRs inside but also from the 68K itself. I'm not sure I have an adequate way to get the probe onto the actual surface of the cpu for an accurate measure there, but that seems like a lot of heat inside a stock genesis doesn't it?

 

And if I just let it sit for about 10min to cool the hell off and power it all back on, then the sega CD bios screen is normal again. So at this point, I can't quite figure out if the heat issue is because of the 32x somehow or because of something else. It might honestly be tied to a single game I've been playing lately and it is a game that really pushes the stock genesis hardware to some extreme limits. Hmm...

[+-^CrossBow^-]

Another update. The 32x itself is the culprit to all of this. Even if I'm not playing a 32x game the entire system gets very very hot with the 32x installed. Last night I did something different. I kept the 32x installed and plugged in, but no game at all inserted just left the system running on the Sega-CD bios screen. The other change here, was that I decided to just use the AV out from the Genesis itself and not from the 32x. So essentially the 32x is just sitting there plugged in but I'm not watching the sega-cd bios through the 32x video encoder. Well I came back an hour later and sure as anything, the digital speckles were there on the sega-CD bios screen. I took the 32x off and turned it right back on and the speckles were no where to be found and I had a clean sega-cd bios screen again. Plug the 32x back in and the speckles come back.

 

I then tried something I usually don't care about. I went ahead and found and installed the little metal RF clips that came with the 32x originally. Plugged it all back in and ran the experiment the same as before. 32x plugged in with power but no game cart inserted and video and audio from the genesis itself and NOT from the 32x. Left it running about an hour and checked on it. The speckles were still there but much much much reduced from before. So...it would see that those RF clips are doing "something". But I'm still trying to figure out what the heck is up with the 32x. I did recap it about a year ago so I don't think that has anything to do with it. This is a fairly new development or at least new that I've noticed. Trying to source another 32x base unit on the cheap to see if this is a 32x thing with my genesis or...