Question re: UAV installation in heavy sixer

[+x=usr(1536)]

Earlier this evening, I installed a UAV in a heavy sixer.  I've done eight or nine UAV installs at this point, so am reasonably familiar with them, but this was my first in one of these machines.  Because of this, I'm not sure if what I'm seeing is expected behaviour or not.

 

The issue is really bad colour that can't be adjusted back to normal with the stock colour pot when viewed over the RF output.  That's not to say that the colour can't be adjusted, but rather that there's no way to dial it back to where it should be.  Pac-Man has a fascinating range of psychedelic colours, the Harmony Encore menu is almost unreadable, and other games that I'm familiar with couldn't be brought back to their normal colour palettes.

 

This machine had amazingly good RF output, so I'm a bit surprised by this.  While I realise that the goal here is to not use RF, I'm keeping it intact a) because there's no good reason not to, and b) it's a useful backup for troubleshooting.

 

The installation itself went fine.  Someone has been in there before me, given rework done to one of the joystick ports' solder joints and an already-socketed 4050 being installed.  Not surprising for a 45-year-old machine

 

I did a bit of a variation on the installation instructions by soldering the 4050 to the top of the machine pin leads instead of going with the suggested piggybacked socket.  This was twinned with cutting the 5V pad and soldering to only its left side for power; continuity testing shows no connection between the two sides of the pad.  The 4050 also checks out from its location on the UAV to the underside of the 2600's PCB.  Pin 6 on the TIA is lifted per the instructions.

 

Note that I do not currently have cables attached to the UAV that will let me compare its output with RF.  Those are in the mail and should be here in the next day or two.

 

Is what I'm seeing over RF expected behaviour for this machine when a UAV is installed?

[+-^CrossBow^-]

It is the 820Ω resistor for color saturation. You need to remove it from circuit. It is the resistor just below where Pin 9 of the TIA is located. It should be the rightmost upper resistor just below the TIA, I removed one of the legs on one I did and then grabbed the color from a different point off the board. Let me try and find my pic...

 

Okay..so I can't find a pic per se...

 

But it is R213 just below the TIA. I lifted the rightmost leg from it off the board. I then grabbed my color signal from the right most leg of R212 a bit up from there and to the right further. But you could also just grab the color from that same via that the right leg of R213 was soldered into.

 

[+x=usr(1536)]

1 hour ago, -^CrossBow^- said:

It is the 820Ω resistor for color saturation. You need to remove it from circuit. It is the resistor just below where Pin 9 of the TIA is located. It should be the rightmost upper resistor just below the TIA, I removed one of the legs on one I did and then grabbed the color from a different point off the board.

Ah, OK, and thank you.  The version of the instructions that I'm working from read as though this was only necessary if the TIA was soldered in, but mine's socketed; I now see that this was an "Eats, shoots & leaves" moment on my behalf.  Guess what I didn't desolder as a result.

1 hour ago, -^CrossBow^- said:

But it is R213 just below the TIA. I lifted the rightmost leg from it off the board. I then grabbed my color signal from the right most leg of R212 a bit up from there and to the right further. But you could also just grab the color from that same via that the right leg of R213 was soldered into.

Yep, I'm getting colour from the rightmost leg of R212 as well, and power from the leftmost leg of R231.  I'll lift R213 when I'm back at the house and see how it goes

[+-^CrossBow^-]

I'm honestly not sure how the resistor being needed or not was done. But I think all of the original heavies have it in place in that same spot and I know later revisions of the 4 switchers have it while some other 4 switchers would have it soldered on the bottom side of the PCB as a factory bodge or it might not be present at all. But yeah anytime the colors are looking wonky from a UAV on a 2600 and the color trimmers doesn't make it right. Check to see if you are getting 800ish ohms between pins 6 and 9 off the TIA and if you find it does, then start looking and tracing the board to see where it is coming from.

 

[+x=usr(1536)]

Lifted R213, and that fixed the colour problem on RF.  Thanks

 

It looks like I'll need to order another UAV, though.  This one was used for some experiments prior to installation, and I appear to have damaged at least the traces used for sync.  Kinda thought that might be the case, but figured it had a 50:50 chance of working.

 

That one is very much on me, and the lesson is learned: buy the kit version, not the plug-in, in the future.

[+-^CrossBow^-]

5 hours ago, x=usr(1536) said:

Lifted R213, and that fixed the colour problem on RF.  Thanks

 

It looks like I'll need to order another UAV, though.  This one was used for some experiments prior to installation, and I appear to have damaged at least the traces used for sync.  Kinda thought that might be the case, but figured it had a 50:50 chance of working.

 

That one is very much on me, and the lesson is learned: buy the kit version, not the plug-in, in the future.

Yeah I always buy the kit version and use what I want for the 5200 installs. I just use the basic for the 2600 and 7800 though and never use the plug in version for the 2600. Just as easy for me to wire it up manually and place the UAV where I want without messing with the 4050 on the board. If you poke around a bit, I'm pretty sure you can fine other locations on the UAV to attach the sync to to get you going. I had to do that with one of the UAVs I got not long ago that was only showing black n white on 2600 games. Turned out the TCol trace apparently didn't get etched fully and it was partially missing on the UAV. I found another spot on the UAV and just make a quick jumper using part of a cut lead.

 

[+x=usr(1536)]

1 hour ago, -^CrossBow^- said:

Yeah I always buy the kit version and use what I want for the 5200 installs.

I'll admit to having bought the plug-in version for an upcoming installation in one of my 5200s, but that was mainly out of curiosity: I wanted to see what one would look like from the factory, so to speak.  IMHO, there's nothing wrong with the plug-in versions, but if you're doing anything that deviates from the norm in terms of the installation, they just don't have the flexibility of the kit versions.

1 hour ago, -^CrossBow^- said:

I just use the basic for the 2600 and 7800 though and never use the plug in version for the 2600. Just as easy for me to wire it up manually and place the UAV where I want without messing with the 4050 on the board.

Yep, and that's how I'm thinking of redoing this installation.  I'd like to use the 4050's socket, largely because it opens up otherwise-unused real estate for mounting the UAV, but then there's the problem of what to do with the 4050.  I'll figure something out

1 hour ago, -^CrossBow^- said:

If you poke around a bit, I'm pretty sure you can fine other locations on the UAV to attach the sync to to get you going. I had to do that with one of the UAVs I got not long ago that was only showing black n white on 2600 games. Turned out the TCol trace apparently didn't get etched fully and it was partially missing on the UAV. I found another spot on the UAV and just make a quick jumper using part of a cut lead.

The resistors just south of the TIA are looking like possibilities.  Ideally everything would be run back to the UAV and plugged into it with Dupont connectors, but I may save that experiment for when I'm doing this on one of my machines.

[+x=usr(1536)]

OK...  Two steps forward and several steps back.

 

Got the UAV kit, built it, installed it.  Also installed an IAB at the same time.  Powering on gives...  Nothing.  Not even a flicker on the screen.

 

Double-checked the UAV: jumpering is correct and continuity is good on both them as well as from the 4050 (socketed on top of the UAV) through to the 2600.  Continuity is also OK on the A/V cable back to the UAV.

 

Replaced the 7805 on the switchboard; no change.  TIA pins 1 & 20 show no voltage across them.  The right leg of R213 is still lifted from the previous attempt at UAV installation.

 

The only obvious thing I can see is that the ribbon cable between the switchboard and PCB has split at both sides and may consequently have internal trace damage.  Continuity says all the pins are fine, but I'm getting a bridged reading from pins 1 & 11 when testing in-circuit.  Not sure if this is intentional or not.

 

Anyone have any ideas for where to start?  It was fully-working over RF before I started in on it, so the machine wasn't a basket case or similar.  I'm not familiar with the early 2600s, so could use a hand on this one.

 

Some photos for reference:

 

General overview:

 

 

A somewhat closer look at the top:

 

 

The underside, showing the IAB connections:

 

[+-^CrossBow^-]

What output voltage are you showing from the 7805 when you power it on? I actually have had this happen a few times but only with sixers. For some odd reason, most of the GNDs and voltages will show shorted in quite a few places on the switch board until the actual switch is in the ON position. With one h6 I had to remove the large Allen Bradley resistor that was above the power switch as it appeared to have been shorted out somehow. Just removing it was enough to get the Atari working back up again. I didn't realize the IAB was so large? I actually didn't use the 4050 socket on my H6 install and wired everything to the UAV manually. I installed the UAV in place where you put the IAB in your install. But it allowed me to connect everything up to alternate points off the legs of components or empty vias instead of having to run any wire under the PCB or attach directly to the TIA itself. 

 

This picture was taken before I found out R213 needed to be lifted on the right leg. That was done afterwards and then shrink tubing applied to the end of the lead so it would short out on anything and can be reused in the future if everything were to be undone.

 

[+x=usr(1536)]

1 hour ago, -^CrossBow^- said:

What output voltage are you showing from the 7805 when you power it on?

It was showing anywhere from 0-19V fluctuating on input, which was what led me to replace it.  There is one spare Traco 2-2450 sitting here, but I'd rather not use it unless absolutely necessary as it's slated to go into the 7800.

 

The original PSU was putting out 12-14V fluctuating, so it was swapped with one from a 2600 Jr.; it also showed voltage in the same range.  I did manage to find my multivolt adapter with both the correct plug for the 2600 and a 700mA output, so may try that one later.  Two bad original PSUs is unlikely, but it's not like it's never happened before.

1 hour ago, -^CrossBow^- said:

I actually have had this happen a few times but only with sixers. For some odd reason, most of the GNDs and voltages will show shorted in quite a few places on the switch board until the actual switch is in the ON position.

Interesting.  I'm wondering if removing and cleaning the power switch might not be needed - all of them were hit with contact cleaner and were working fine, but it would make sense for that one (or Reset) to be the most gummed up.

1 hour ago, -^CrossBow^- said:

With one h6 I had to remove the large Allen Bradley resistor that was above the power switch as it appeared to have been shorted out somehow. Just removing it was enough to get the Atari working back up again.

Will give that a shot later today.  The workbench has been taken over by my wife for a couple of projects she's working on this morning.

1 hour ago, -^CrossBow^- said:

I didn't realize the IAB was so large?

It's definitely not small.  While I wouldn't necessarily call it massive, an SMD version would make for a much more compact package.  For reference:

 

 

1 hour ago, -^CrossBow^- said:

I actually didn't use the 4050 socket on my H6 install and wired everything to the UAV manually. I installed the UAV in place where you put the IAB in your install. But it allowed me to connect everything up to alternate points off the legs of components or empty vias instead of having to run any wire under the PCB or attach directly to the TIA itself. 

Funnily enough, I had run across a couple of threads where you detailed doing that and was going to go the same route myself, except that the IAB dictated using that space for it.

 

Running under the PCB wasn't my first choice either, but given a choice between soldering directly onto the TIA or just using one of its socket's pins, the latter was more appealing.  Ran the IAB power that way as well because I wasn't finding anywhere good to tap on the top side.  Not ideal, but it works, and with the gauge of wiring I'm using there's no jamming / rubbing between the PCB and RF shield wall.

1 hour ago, -^CrossBow^- said:

This picture was taken before I found out R213 needed to be lifted on the right leg. That was done afterwards and then shrink tubing applied to the end of the lead so it would short out on anything and can be reused in the future if everything were to be undone.

Thanks for the reminder - the lifted leg on R213 still needs to be heat-shrinked on my board

[+x=usr(1536)]

4 hours ago, -^CrossBow^- said:

With one h6 I had to remove the large Allen Bradley resistor that was above the power switch as it appeared to have been shorted out somehow.

And having now tested it out of circuit, it reads anywhere between 0Ω and 55Ω when it feels like it.  Field service manual says it should be 680Ω.  It's toast.

 

Is it OK to just bridge the connection, or should I order a 2W 680Ω resistor to replace it with?

[+-^CrossBow^-]

3 hours ago, x=usr(1536) said:

And having now tested it out of circuit, it reads anywhere between 0Ω and 55Ω when it feels like it.  Field service manual says it should be 680Ω.  It's toast.

 

Is it OK to just bridge the connection, or should I order a 2W 680Ω resistor to replace it with?

I actually left it out completely as the later model 6 switches don't even have it. 

[+x=usr(1536)]

With R101 removed, the system is now showing 5V at the regulator and voltage is making its way to the UAV, IAB, and ICs.  This is good.

 

The 6507, however, is getting hot enough to fry an egg on.  A replacement has been ordered.

 

If, by any chance my eBay purchase turns out to be a dud, does anyone have a couple of C010745 CPUs they could send my way?  There isn't enough other stuff that I'm looking for right now to make up a suitably-sizeable order from Best.

[+-^CrossBow^-]

That now makes at least 2 instances where that old allen bradley is now to blame for this and appear to be on their last legs. I thought the original 6507s did get pretty hot to begin with? On the one I serviced that had the same issue, I didn't replace any of the ICs but I didn't exactly feel for it they were getting unusually hot or not either. I do know I left the console doing burn in testing for several hours without any issues.

 

[+x=usr(1536)]

11 minutes ago, -^CrossBow^- said:

That now makes at least 2 instances where that old allen bradley is now to blame for this and appear to be on their last legs.

Hopefully that's not the case, because if so then every 2600 with one of these on there needs to replace or remove it ASAP.

11 minutes ago, -^CrossBow^- said:

I thought the original 6507s did get pretty hot to begin with?

They can, from what I understand, but this one left me with a tiny burn blister from where it was touched originally.

11 minutes ago, -^CrossBow^- said:

On the one I serviced that had the same issue, I didn't replace any of the ICs but I didn't exactly feel for it they were getting unusually hot or not either. I do know I left the console doing burn in testing for several hours without any issues.

My concern is that this may have taken out other stuff that remains to be discovered.  I'm really hoping that the TIA, RIOT, UAV, IAB, and 4050 survived or this is going to become a really expensive machine.

[+x=usr(1536)]

This has not been a fun afternoon of troubleshooting.

 

The 6507 arrived in the mail.  Went to replace it and the frame for the original socket came out with the dead IC.  Didn't have any 28-pin sockets of the correct dimensions, so had to cut down a 40-pin.  Replaced them, verified continuity, all was well.

 

Managed to break a pin off of one side of the UAV's socket headers.  Repaired that.  Verified that continuity at the UAV on all pins was good; it was.

 

Turned on the machine with Pac-Man inserted.  Blank screen (not totally dead; signal is being sent down the cable), but with a brief crackle from the speaker.  The game plays blind, but from what I can tell the intro 'music' is playing faster than it should.  Audio quality is fine.

 

Controls work, as best I can tell.  Turning corners, etc. results in more dots being eaten, so it's likely responding OK.  Fire button starts the game as expected.

 

The CPU I'm using is an Atari one, but it's a Mexican-made Rockwell.  1983 datecode.

 

At this point, I'm low on ideas and am debating putting the 4050 back in its socket, removing the header pins from the UAV, and wiring directly to it from the luma / chroma / sync sources just to see if anything changes.

[+x=usr(1536)]

Did some more poking around this morning after having had a chance to lose the impulse to set the thing on fire and throw it out the window.

 

RF is working, but there appears to an issue with sync - it's as if vertical hold doesn't exist, but without the fine scrolling that usually entails.  The picture just sort of vertically jumps into the wrong place as it follows the electron beam from top to bottom.  Also, from what I can see of how Pac-Man is playing, the ghosts appear to be jumping around the maze rather than moving around it fluidly, though that may be the sync issue causing an optical illusion.

 

Metered out pins 1-20 on the TIA to the group of five resistors just South of it.  The majority of pins 2 through about 17 or 18 show continuity to all of the resistors.  TIA is almost certainly shot, so didn't bother with pins 21-40.  No damage, solder bridges, etc. are visible, so the IC is probably just toast internally.  This may also go some way towards explaining why Pac-Man was playing fast.

 

Next step: order a TIA, swap it in, see what happens.  Might also do the RIOT at the same time, since that'll at least give the board a full main chip swap.

[Matt300ZXT]

Ugh, I do NOT like modding the older 2600s.  I had NOTHING but trouble trying to do work on a Tele Games light 6er.  I eventually just said to hell with it and pirated everything off of it I could use and tossed the board.  I hope you have better luck than I did.

[+x=usr(1536)]

This is, hands down, the pickiest UAV installation I've done.  No other machine has fought me like this one has.

 

To be fair, it's a 45-year-old system, and monkeying with it is uncovering other issues that likely would have remained masked until something actually went kaput.  Sure, it's younger than I am, but knowing how my body bounced off the pavement at 45 vs. 35, or 25, or 15 I can forgive it most of the pain it's causing.

 

Current plan is the Console5 recap plus TIA, RIOT, and CPU replacement (which will leave me with a spare working 6507, something I don't currently have).  Probably also going to replace the voltage regulator that I put in there with a Traco.  This one is supposed to be a Christmas present, so I figure that bulletproofing it up front is better for everyone involved.

[Ecernosoft]

On 11/27/2022 at 10:32 AM, x=usr(1536) said:

Did some more poking around this morning after having had a chance to lose the impulse to set the thing on fire and throw it out the window.

 

RF is working, but there appears to an issue with sync - it's as if vertical hold doesn't exist, but without the fine scrolling that usually entails.  The picture just sort of vertically jumps into the wrong place as it follows the electron beam from top to bottom.  Also, from what I can see of how Pac-Man is playing, the ghosts appear to be jumping around the maze rather than moving around it fluidly, though that may be the sync issue causing an optical illusion.

 

Metered out pins 1-20 on the TIA to the group of five resistors just South of it.  The majority of pins 2 through about 17 or 18 show continuity to all of the resistors.  TIA is almost certainly shot, so didn't bother with pins 21-40.  No damage, solder bridges, etc. are visible, so the IC is probably just toast internally.  This may also go some way towards explaining why Pac-Man was playing fast.

 

Next step: order a TIA, swap it in, see what happens.  Might also do the RIOT at the same time, since that'll at least give the board a full main chip swap.

I hope you get your 5200 fixed! Never knew you were a hardware guy.

 

(P.S. What is your experience with the 5200, and would you reccomend it to someone else who's into this stuff? I thought about getting one when I get older.)

Edited November 28, 2022 by Ecernosoft

[+x=usr(1536)]

Well, this has taken an annoying turn.

 

The TIA, RIOT, and CPU arrived yesterday.  Tested the TIA for continuity before installing it: it seemed to be as expected.  Swapped it into the unit, saw the same continuity problems as before which had led me to believe that the TIA was bad.  Did not power the unit up since there's clearly a problem somewhere else.

 

Doing continuity testing from the resistor block South of the TIA shows continuity all over the place on the PCB.  The results are at least consistent with the UAV removed or installed, so that's not the culprit.

 

My best guess is a shorted component or components somewhere, and a lot of time with the meter in my future.  Cap kit hopefully arrives today, so that will at least eliminate some possibilities.

[+-^CrossBow^-]

You know I have a 4 switch vader that I just remembered has similar issues. It refuses to power up although all the chips tested good and the VR appears to be working. I was also able to confirm that I'm getting power to all the ICs so I didn't want to bother with it at that time and set it away to work on another. I need to get back to that sometime when I've got more free time. But it is the only 2600 to give me any problems like this aside from the H6 I mentioned earlier.

 

[+x=usr(1536)]

12 minutes ago, -^CrossBow^- said:

You know I have a 4 switch vader that I just remembered has similar issues. It refuses to power up although all the chips tested good and the VR appears to be working. I was also able to confirm that I'm getting power to all the ICs

That sounds familiar

13 minutes ago, -^CrossBow^- said:

But it is the only 2600 to give me any problems like this aside from the H6 I mentioned earlier.

Based on nothing more than gut feeling, my suspicion is that there's a short in one or more of the transistors.  How current would be getting to the places that it is otherwise utterly baffles me.

 

Having said that, this is certainly a multi-layered problem.  Today's task: finding a $0.59 part (the 2W, 680Ω resistor) that doesn't come with a $10 shipping price tag attached.  At this point, I'd like to get it reinstalled just to have everything back to stock.

[+scorpio_ny]

Hi @x=usr(1536),

I am sorry that you are having issues  issues with your install. I hope the cap kit you ordered resolves the issue.

 

I am following this topic with keen interest  since I am looking to do the same install on two light sixers that I own. I have already installed Console 5 cap kits to both to reduce the possibility of future headaches. I also added a sockets for the 4050 and installed the joystick port static protection upgrade for this model. Looking to order the UAV’s this week.

 

Maybe you or someone else can answer: Do I need to install the audio audio board offered by the brewing factory?

 

 

[+x=usr(1536)]

On 12/6/2022 at 1:45 PM, scorpio_ny said:

Maybe you or someone else can answer: Do I need to install the audio audio board offered by the brewing factory?

It's not 100% necessary, but it does make for a tidy way to pull the audio out.  Having said that, the main reason I went for it was because this machine is intended to be a Christmas present, so decided to go the whole hog on modifying it.

 

@-^CrossBow^- made a blog post on installing a UAV into a heavy sixer that describes the process for picking up audio without an IAB.  It's fundamentally the same as how you'd connect an IAB, but with using only direct audio connections to the board.