7800 UAV install without case drilling

[+x=usr(1536)]

A while back, I got the idea into my head that it might be possible to install a UAV into a 5200 without also drilling the case to install A/V jacks.  As it turned out, that hypothesis was correct.  With that installation having gone well, it seemed like it would be worth giving the same idea a shot on the 7800.

 

How to install a UAV won't be directly covered here; there are a ton of other resources out there that explain the process better than I could.

 

As with the 5200, the goals were as follows:

• No case drilling of any kind
• RF shields to remain in place
• Composite, S-Video, and audio outputs to be fully-functional
• RF to remain intact and functional as backup video output
• Achievable with off-the-shelf parts

The 7800 has one massive advantage over the 2600, 5200, and A8s in terms of keeping RF output functional: it doesn't use a 4050.  This means that you don't have to figure out how to accommodate or repackage that IC if you do want to keep RF.  Frankly, conscious effort would be required in order to disable or remove RF output from a 7800 given how it implements its video output.

 

Parts used are:

• UAV.  Supplied by The Brewing Academy.  Definitely order the kit version and build it from scratch; the reasons for this will become apparent.  Be sure to order it with right-angled Dupont connectors for the video outputs instead of the standard terminal block.
• 8-Bit Classics' Atari XL/XE 5-Pin DIN S-Video & Composite AV Cable.  These can also be found elsewhere without the DIN-5 monitor connector used on the A8 machines.
• Jumper wires with Dupont connectors.  I used these ones from Adafruit because they were what I had available.
• Gorilla Tough & Clear Mounting Tape.

 

After rewatching installation videos from a couple of different sources and doing some test-fitting of the UAV in various places, it looked as though the top of the RF modulator was usable as a mounting location with some minor modification of the RF shield.  As an added bonus, this leaves the UAV's colour timing adjustment pot accessible with the RF shield in place.

 

Looking at the following picture of the RF shield from the rear, there's normally a small lip running across the tall part of the opening.  This shows the shield with that lip removed, which provides clearance for the UAV:

 

 

With that taken care of, the UAV (already built) and audio output (not shown) were installed per the normal methods and the UAV adhered in place on top of the RF modulator.  Note that it's set about 1-2mm over the rear edge of the modulator:

 

 

Having it hang over the end a bit is necessary for a margin of safety related to clearance under the RF shield.  Here's a view showing more clearly how much it juts out from the modulator:

 

 

Next up, prepare the A/V cable for installation.  No photos specific to this process to show, but there are things to be aware of:

• The following assumes use of the recommended cable from 8-Bit Classics.  If not, the general principles still apply but it may be necessary to figure a few things out for yourself.
• Don't cut the 5-pin DIN connector off.
  • Press in on the metal tab visible on the outer sheath, then slide the sheath down the cable and away from the connector.
  • Discard both of the metal outer halves of the connector; one will be crimped to the outer part of the cable.
  • Carefully desolder each wire end from the DIN-5 connector, taking care to not separate the grounds (thick bare wire bundle with no insulation) or audio output (red and white wires soldered together).
  • Discard the DIN-5 connector once everything is removed from it, and do the same with the outer sheath.
• Depending on the Dupont jumper wire set being used, it will likely be possible to colour-code the Dupont wires to the A/V cable.  This is strongly recommended as they correspond to pins on the cables and markings on the UAV's PCB.
• Run a continuity check on the cabling to verify that the following is correct before soldering anything - cable markings may change without warning.
  • Yellow: Composite Video
  • Green: Chroma
  • Blue: Luma
  • White: Left Audio Channel
  • Red: Right Audio Channel
  • Black: Grounds (all signals, including audio)

 

Once the cable is complete, carefully make a notch in the outer insulation in order to allow it to fit in the cable channel in the 7800's chassis.

• After cutting the notch, twist the cable so that the wires inside twist together.  Take care to not overdo the twisting: this could cause internal wiring breaks.
• With the cable twisted, heatshrink - as tightly as possible - the notched section.  Hold the twist on it for a couple of minutes to allow the heatshrink to fully harden.

 

Feed the cable in from underneath the 7800, placing the notch in the cable channel.  The photo below shows the location of the channel.   Dupont wires are not connected:

 

 

Next, route the A/V wiring along the rear of the 7800:

 

 

That's pretty much it.  Test that everything is working, and sew it back up if it looks good.  Take care when reassembling to make sure that none of the A/V wiring running across the rear of the unit is interfering with the screw holes used to clamp the case halves together.

 

From underneath, this is the end result:

 

 

And completed at the rear:

 

 

Doing this probably sounds a lot more involved than it really is.  The main things are to take your time when assembling and routing the A/V wiring, heatshrink the wire connections well, and check continuity before doing anything that would be a pain to revert if needed.

 

One change I would have made: adding a male Dupont pin to the audio output where it comes off of the resistors at the PCB; this would have allowed for all wiring connections to be unplugged as part of a teardown.  As it stands, I'm currently out of male Dupont pins, so hardwired that connection.  It's not the end of the world, but will likely be sorted out one day.

 

It's worth noting that the A/V cable can only make about two full loops around the cable retainers underneath the console before there's no more room for it; this will leave a couple feet of it sticking out.  Not something that bothered me since I had no plans to use them for cable storage, but probably worth mentioning just to set expectations.

 

Wrapping the UAV in heat shrink: optional, but if the appropriate size of heat shrink had been on hand, it would have been used.  Your call.

 

To use the RF output, just connect it to a tuner as per usual.

 

Other than that, it certainly seems to have been worth the effort.  Drilling the cases becomes less and less attractive the older these machines become, and not having to do so is definitely preferable.  Any questions, feel free to ask

[+-^CrossBow^-]

@x=usr(1536) No chroma fix?

 

[+x=usr(1536)]

28 minutes ago, -^CrossBow^- said:

@x=usr(1536) No chroma fix?

 

Not yet.  Going to do that one after Thanksgiving when I can sit down and do some proper comparison testing.  Squeezed this one in while waiting for the 2600's CPU to get here.

[+Muddyfunster]

Circuit pr0n. I don't understand it, but it's so pretty!

[+x=usr(1536)]

Well, here's a fun one.

 

The 7800 stopped outputting video yesterday.  It had managed a couple of video blackouts a day or two before it happened (flashes of no video followed by recovering), but it's now total.

 

The 2600 side plays blind: I can start a game of Pac-Man and hear sounds that match joystick movements, but there's no video.  The 7800 side seems to be totally dead.  No video, no audio.

 

FWIW, there is nothing coming out of the RF modulator.  This shouldn't affect UAV operation since the UAV is taking its power and video from before it enters the modulator, so there should logically be signal on the modulator.  It's also taking its video from the MARIA and TIA separately, so at least one should work in a single-dead-chip scenario.

 

Nothing is getting overly-hot, and the only change made since the UAV was installed was to install a Console5 cap kit about a week ago.  Everything had been working perfectly after that point.

 

I'm low on ideas - initial probing hasn't turned up anything amiss, at least not that I can find without a scope (which I do not currently own).   Voltages look good at the ICs and UAV.

 

Thoughts?

[+-^CrossBow^-]

@x=usr(1536) I assume you have tried to remove or otherwise disable the UAV at this point?

 

WAIT!!!... give me a closeup pic of where you are tapping the gnd and +5 to the UAV from as I might see something.

 

[+x=usr(1536)]

2 hours ago, -^CrossBow^- said:

@x=usr(1536) I assume you have tried to remove or otherwise disable the UAV at this point?

Yep.  No change.

2 hours ago, -^CrossBow^- said:

 

WAIT!!!... give me a closeup pic of where you are tapping the gnd and +5 to the UAV from as I might see something.

 

Here you go:

 

 

That's where I've been pulling power from since the day of the installation.

[+x=usr(1536)]

Aaaaarggghhhh....  For giggles, I turned it on with 2600 Pac-Man inserted.  It worked.  Then I plugged in the Concerto, and it's now happily playing Beef Drop.

 

Looks like I have a crappy connection somewhere.  Everything checked OK for continuity before reassembly, so my guess is that something is getting nudged when it's all being put back together.

 

False alarm, folks.  Apologies for thinking that there was a falling sky around here

[+x=usr(1536)]

After some more fiddling, it appears as though the issue is that the UAV was moved slightly out of place from where it was originally installed.  This (I think) is causing the RF shield to contact it somewhere and cause a short.  The fix will be to a) move it back to where it should be and b) wrap it in heatshrink, thus insulating it.  Don't have any heatshrink of the appropriate diameter, so will order some and leave the top shield off in the interim.

 

On another note, I took the opportunity to make some changes to the power wiring and add in one of @-^CrossBow^-'s Chromafix boards, developed in conjunction with @juansolo and @marauder666.  The Chromafix is a nice addition: the change to the video output is subtle, but definitely noticeable.  Colours are better-defined, and much closer to what you might expect them to be.  On my CRT TV, the picture is also a little brighter than before.  Again, not by a huge amount, but definitely to an improved degree.

 

Photos of the changes and Chromafix install follow:

 

First change: the power wiring was relocated to the rear of the RF modulator PCB, and heatshrink added to where it's soldered to the board.  The heatshrink was just an oversight on my behalf during the initial installation, but the wiring really should have been on this side to begin with:

 

 

The MARIA colour output (white wire) was originally soldered directly to the Colour In via on the UAV; this was due to not being certain if it would clear the screwpost to the left of it or not when the case was reassembled.  As it turned out, there is enough space for a right-angled pin header:

 

 

I found out fairly quickly that with the UAV on top of the RF modulator, there's insufficient clearance for the Chromafix board with the top RF shield installed.  This meant remote-mounting it on the side of the RF modulator.  First order of business was installing right-angle header pins:

 

 

+5VDC (red) and ground (brown) for the Chromafix were taken from pins 1 and 8 on the UAV's IC header (disregard the filth; the board was cleaned up with alcohol after photos had been taken):

 

 

This is how it all ends up fitting together.  The white wire is MARIA colour input and output.

 

 

A view of both the UAV and Chromafix installed:

 

 

And that's pretty much it for this installation.

 

Had I thought about installing the Chromafix up front, I wouldn't have done it this way.  There are just too many fiddly bits to have to deal with if you add it into the equation to make it worth the aggro; just placing the UAV with Chromafix added on top of the OS ROM as usual would be a lot easier.  Having said that, this is still a workable location for a standalone UAV, and either approach will manage to retain RF output as well.

[+-^CrossBow^-]

Not quite done...

 

Try switching the channel select switch and see what happens to your output...

 

You should be pulling the +5 from the pin directly next to the GND pin off the RF modulator. The pin you have it coming from is for the channel select switch to provide +5 to the modulator to select one channel or the other. So the only reliable spot to get +5 in that area is from the pin directly next to the GND. Or...just run new wires to grab it from center large traces down the center of the 7800.

 

 

[+x=usr(1536)]

18 hours ago, -^CrossBow^- said:

Try switching the channel select switch and see what happens to your output...

Thank you for the suggestion; that really helped to point out a problem I didn't know existed.  More:

18 hours ago, -^CrossBow^- said:

You should be pulling the +5 from the pin directly next to the GND pin off the RF modulator. The pin you have it coming from is for the channel select switch to provide +5 to the modulator to select one channel or the other. So the only reliable spot to get +5 in that area is from the pin directly next to the GND. Or...just run new wires to grab it from center large traces down the center of the 7800.

Today I learned that the channel select switch on this 7800 can actually change settings depending on how the console is handled.  The switch doesn't have to fully slide to one side or the other to change the output channel; just a small movement can do it.  And, as you mention, one side is +5VDC and the other side is 0VDC, so depending on which side the switch is set to, the UAV may not be getting power based on the pin that UAV +5VDC is connected to.

 

In troubleshooting all of this, it never occurred to me to check the switch (or measure voltage at the RF modulator PCB for it) because it was theoretically being circumvented.  And in the 20-odd years I've owned this 7800, I can't recall ever having to change the channel on it.

 

I'll test stuff and move the +5VDC two pins closer to the ground wire tomorrow.  Pretty sure I have a couple of good channel switches laying around somewhere, so will test them & swap one in at the same time.

[+x=usr(1536)]

+5VDC to the Chromafix board has been relocated per @-^CrossBow^-'s recommendation.  Left the channel switch alone: it's still functional, albeit touchy.  If it ever becomes a serious issue, I'll just remove it altogether and live with the RF modulator being permanently-defaulted to channel 3.

 

Calling this one done for now.  It was a really good exercise in terms of understanding what the limits of installation reasonably are on a 7800.

[+-^CrossBow^-]

20 minutes ago, x=usr(1536) said:

+5VDC to the Chromafix board has been relocated per @-^CrossBow^-'s recommendation.  Left the channel switch alone: it's still functional, albeit touchy.  If it ever becomes a serious issue, I'll just remove it altogether and live with the RF modulator being permanently-defaulted to channel 3.

 

Calling this one done for now.  It was a really good exercise in terms of understanding what the limits of installation reasonably are on a 7800.

You mean both the UAV and C-Fix are attached at a different point now for +5 right? Again, the RF pin directly next to the gnd is always +5 but there are several other places near there as well, including one of the unused vias for the expansion slot.

 

[+x=usr(1536)]

51 minutes ago, -^CrossBow^- said:

You mean both the UAV and C-Fix are attached at a different point now for +5 right?

Yep.  The Chromafix still pulls power and ground from pins 1 and 8 on the UAV's header for the 4050, but +5VDC to the UAV has been moved to the RF pin next to ground.

51 minutes ago, -^CrossBow^- said:

Again, the RF pin directly next to the gnd is always +5 but there are several other places near there as well, including one of the unused vias for the expansion slot.

Yup.  I had considered using the expansion slot vias, but decided that I want to keep them clear for future experiments.

 

In any event, there's still some adjustment of the colour pots to do.  Initial settings were done with the console cold; the colour match drifted out as it warmed up.

 

Disregard the dark capture - I'm using an $18 S-Video / composite to HDMI upscaler, and it's not exactly what you might call professional quality.  This screenshot was taken from S-Video straight into OBS.

 

[+-^CrossBow^-]

35 minutes ago, x=usr(1536) said:

Yep.  The Chromafix still pulls power and ground from pins 1 and 8 on the UAV's header for the 4050, but +5VDC to the UAV has been moved to the RF pin next to ground.

Yup.  I had considered using the expansion slot vias, but decided that I want to keep them clear for future experiments.

 

In any event, there's still some adjustment of the colour pots to do.  Initial settings were done with the console cold; the colour match drifted out as it warmed up.

 

Disregard the dark capture - I'm using an $18 S-Video / composite to HDMI upscaler, and it's not exactly what you might call professional quality.  This screenshot was taken from S-Video straight into OBS.

 

Odd that you are still getting some artifacting color? On all of my displays through s-video both CRT and my LCDs, those white striped bars are all just white stripes and I only see the colors from them when I use the composite output.

 

Then again, I have all of my s-video sources going into my Extron 7sc and then the output from that goes into my OSSC in the game room. In the lab I just have the s-video connected directly to my LCD and CRT s-video inputs but don't have a way to capture them from up there.

 

 

[+x=usr(1536)]

9 minutes ago, -^CrossBow^- said:

Odd that you are still getting some artifacting color? On all of my displays through s-video both CRT and my LCDs, those white striped bars are all just white stripes and I only see the colors from them when I use the composite output.

Unfortunately, I don't have any displays that can natively handle S-Video; that's being run through a Tendak S-Video / composite upscaler followed by capture on a generic USB-C to HDMI adapter.  Composite runs straight to the CRT and looks significantly different compared to S-Video, which leads me to believe that at least one of the devices in the capture chain is doing something with the video before it reaches OBS.

 

For comparison, here's a photo of composite output on the CRT, also pre-color adjustment:

 

 

My suspicion is that one or both of the cheap capture devices I'm using is the root of the problem.

 

Thinking about it, the USB-C to HDMI adapter being used here is one that my wife had been using and swapped out with my original one.  Why she did that, I don't recall, but that's probably worth looking into.

9 minutes ago, -^CrossBow^- said:

Then again, I have all of my s-video sources going into my Extron 7sc and then the output from that goes into my OSSC in the game room. In the lab I just have the s-video connected directly to my LCD and CRT s-video inputs but don't have a way to capture them from up there.

Yeah, I can see a great tidying of my video setup coming next year.  Right now everything is switched through a 2000-era Radio Shack 4-way pushbutton manual source selection switch, and while it's working fine the idea of combining that functionality into something that can also handle HDMI conversion in the same box is appealing.

[+x=usr(1536)]

Confirmed: the problem is in the Tendak Upscaler, not the USB-C capture card.

 

Colour-adjusted composite video, straight from the 7800 to the CRT:

 

 

Colour-adjusted S-Video, sent via HDMI by way of the Tendak to a Samsung LCD TV:

 

 

Interestingly, this appears on all of the Tendak's supported resolutions (480i, 480p, 720p, 1080i, 1080p).  What's odd is that I don't recall it doing this when I was using it with the A8, but that doesn't necessarily mean that it wasn't and I just didn't notice.

 

Going to see how the 5200 compares; it's the most conveniently-accessible UAV-equipped device around here right now.

[+-^CrossBow^-]

Hmm...those lower hues are still way fuzzier than I'm used to seeing with a C-fix board installed. But at least the artifacting bars are pretty close to what I'm used to seeing, although I can still a tinge of the orange and teal off those bars still an they really should be pure white.

 

My AV setup wasn't all that expensive and it has been really great for my uses. The OSSC can be had for a little over $100 I think these days new. My Extron 7sc ran me another $45 shipped from Ebay. I did have to buy BNC to RCA adapters but those are like a buck each. So really for under $200 you should be able to get a good setup with an OSSC in the mix. If you wanted to simplify even more, then spent another $100 for the Tink5x an way to split the signals to it.

 

[+x=usr(1536)]

3 hours ago, -^CrossBow^- said:

Hmm...those lower hues are still way fuzzier than I'm used to seeing with a C-fix board installed. But at least the artifacting bars are pretty close to what I'm used to seeing, although I can still a tinge of the orange and teal off those bars still an they really should be pure white.

Glad to hear you mention that you see it, because I wasn't completely certain if I was seeing it or not.

3 hours ago, -^CrossBow^- said:

My AV setup wasn't all that expensive and it has been really great for my uses. The OSSC can be had for a little over $100 I think these days new. My Extron 7sc ran me another $45 shipped from Ebay. I did have to buy BNC to RCA adapters but those are like a buck each. So really for under $200 you should be able to get a good setup with an OSSC in the mix. If you wanted to simplify even more, then spent another $100 for the Tink5x an way to split the signals to it.

The Retrotink 5X is exactly the unit I'm planning on going for.  Since the Rat Shack A/V switcher doesn't seem to be introducing any real degradation into the video or audio signals, it can hang off of the 5X.

 

While the Extron looks promising, the only reason I'm not going for one is space.  There just isn't a good spot for a rackmount device in here.