Diagonal lines in video - wiggle power switch to make them dance

[ckrtech]

OK. Got one for you guys.

 

4 switch revision 14.

 

Changes:

Refreshed with new filter cap + the other two electrolytics

New voltage regulator

New C241 near voltage regulator

Longhorn Engineer (and eventually) UAD Revision D composite and S-video mod

 

I see diagonal lines in the video that shift around when I wiggle the power switch. Problem was there prior to any component refreshing (with Longhorn Engineer mod in place. Put it there in 2009)

 

I got them to almost/mostly disappear when testing with Sea Quest by wiggling the power switch.

 

I can never get them to completely disappear with Stay Frosty 2 - they are present in the ground and the sky.

 

Despite cleaning the switch, the problem remains.

 

Flipping the channel selector or hitting RF box makes no difference. Also note the shield that surrounded the cart port and chips is long gone.

 

I swapped the Atari AC adapter for another Atari AC adapter I have. Seems pretty much the same.

 

I pulled another switch (color/b&w) off a non-working unit and used it as a replacement power switch (also using contact cleaner and working it a bit). No change.

 

Just for the heck of it - pulled TIA from another unit and dropped it in. No change.

 

Not quite sure where to go next. The sensitivity to playing with the power switch is weird. I welcome any thoughts. Would like this video to be rock solid (which looks like it should be provided I get rid of these lines)

[ckrtech]

Replaced C204. No change. Apparently also just flexing the cartridge in the slot also causes them to move around. Do I have some sort of grounding issue? Hmm.

[Rom Collector]

I'm thinking this could possibly be RF related. Before I switched to composite, I would have similar effects on my video. Especially if I used the Supercharger, or Cuttle Cart.

Problem was solved by about 98%, when I used a few different ferrite chokes on the power, and audio cords.

BTW, I'm a Ham (KJ6JJK), so I have had experience dealing with RFI.

[ckrtech]

Hmm. Not sure if I have any ferrite chokes to try, but I am not opposed to using them. Do you think there is a specific area that is more prone to picking up the noise? The RF box itself is still in circuit (but not used), but it sounds like noise people encounter is typically from an external source.

 

Did you choke your ac adapter close to the power input to the Atari? Also - how did you wire your composite output internally? Did you use shielded cable?

[Rom Collector]

I just reread your first post and see now I mistakenly thought you were still using RF. (I'm getting old and senile)

So now I think your problem is less likely RFI related.

That said, I think I would still try some ferrite anyway. I love the stuff. I use it liberally. My favorite is MFJ-701, but I use others also.

The easiest, and first thing I would try is to choke any RFI or EMI near the power input. Like you said.

Ferrite is available in different mixes. I would try whats available or most common first. Ferrite beads on component or jumper leads may be the answer.

Unfortunately, I'm an analog guy, and usually deal with radios or audio equipment. Which means I don't know the Atari very well, and you should get a second opinion.

 

Come on guys, lets get more input here.

[ckrtech]

Just to port over a couple of things I posted in the UAV thread, I have attached a video to this post. It shows a PVM displaying composite video from the UAV in the Atari 2600. "Blue only" is enabled on the PVM in order to make the lines easier to see. I toggle color on for a second in the middle of the video.

IMG_4988.MOV

 

In addition, I scoped the power at TIA pin 20 as well as the output from the new 7805 regulator. There is (was/still is) serious noise on in those spots.

 

TIA pin 20:

 

7805:

 

Playing with the power switch did cause the waveform on the 7805 to change a little bit -

 

I had a bunch of 220uf caps lying around, so I placed them across TIA pins 20 and 1 as well as the 7805. It seemed to reduce the noise quite a bit on the scope - and I believe it has reduced the strength of the noise in the image.

 

It could be that I need a little bit better caps or additional caps or something else to get rid of it. I am not quite sure why the waveform changed so much just from jiggling the power switch. The 5 volt line is much more closer to a straight line - or at least with a much lower peak to peak value - than it was prior to the caps. There are still a few spikes in the line, though. I don't have a new waveform image demonstrating this at the moment.

 

Open to additional suggestions. I suppose I could remove the RF modulator in case it is what is producing the interference/noise (in other words - the noise is generated by the Atari itself).

[SignGuy81]

I know you've already tried cleaning/swapping power switches but could you temporarily solder in a jumper across the power switch and see what happens?

[ckrtech]

After the capacitors were added to the 7805, TIA, and to the 4.7uf (in parallel) near the RF module, the power switch does not seem to have as much of an effect when wiggled.

 

There is definitely still noise present at TIA pin 20, and I am fairly certain the noise in the power is what is getting picked up and displayed on video output. Fortunately, the noise is not as wild as in the waveform posted above.

 

I am using the term noise rather loosely here. I believe there is both ripple and high frequency noise on the +5v line at the TIA chip. Adding the 220uf cap across TIA Vcc and Vss reduced both the crazy voltage swing in the TIA pin 20 waveform in the post above as well as diminished the strength of what I am seeing in the composite image on the CRT.

 

Eliminating or significantly reducing the high frequency noise seems like the next step... and I assume that means throwing some ceramic decoupling caps at it.

[+-^CrossBow^-]

This is a 4 switch unit right? Do you still have the foil tape between the switches? I seem to recall that some 4 switch units still use this. You state wiggling the power switch changes the condition, so I'd be curious as it does sound like some sort of grouding issue.

[ckrtech]

I don't have the foil tape anymore. I believe those static strips that were included with revision 14+ were also available to place on revisions prior as indicated by the service manual. The intention was to cut down on static build up that could discharge and damage the ICs.

 

That said - Again, the power switch doesn't seem to cause much of a difference when wiggled now. The additional capacitors seemed to have at least tamed the wild responses originally observed by wiggling the power switch. Prior to the change, I could also see the waveform change by wiggling the cartridge a little bit as well.

[ScooberDiver]

Composite mod? Make sure you lifted one end of R209 1.8KOhm resistor. It feeds audio into the video line and when that goes into your composite video circuit it messes with your picture quality.

On my rev 14 and rev 16 boards I was seeing wavy lines that kept changing. It was most obvious with the ET title screen, on the large blue solid area.