No Signal Output From 7800

[rasch_C]

A couple obvious reworks, and others are questionable.

 

Sent from my Pixel 2 using Tapatalk

[ChildOfCv]

I was going to ask about the French of that. I was reading up today on the constriction of an NTSC composite signal and then though about my reading and wondered about the mismatch. So, what I should be reading at this location is in-fact the horizontal frequency?

 

My plan is to composite mod this. So, if the RF modulator is bad, I don't really care. I'm just hoping to verify a working console before modding.

 

-Allen

 

Sent from my Pixel 2 using Tapatalk

 

Yeah a composite signal is, in fact, a front porch, sync pulse, back porch with color burst, then video signal, and there should be about 15.7K per second (525 lines 30 times per second). About 10% of that will also be vertical retrace.

 

Since the timing is established by TIA or MARIA's internal counters and they're getting a proper signal, any output in the ballpark would likely mean that they're giving you a good signal. But the only way to be absolutely sure before doing the AV mod would be to use a proper oscilloscope to look at the signal there.

 

I thought there might be a possibility that the frequency meter should count 4.5MHz. That's because that's what the sound should be modulated to, and that is also mixed in just before final RF modulation. But maybe that's at a relatively low level so that the voltmeter only sees the composite video. That, or the sound oscillator is also dead. But then it's not necessary if you do the AV mod either.

[+-^CrossBow^-]

A couple obvious reworks, and others are questionable.

 

Sent from my Pixel 2 using Tapatalk

 

Yes those are reworked for sure. But they also look to be the correct components perhaps taken off other Atari boards. I've used parts of a 5200 to bring 7800s back to life as well as more than a few 2600 TIA and RIOT chips. I don't have the board layout and schems in front of me, but I can certainly compare those shots to my parts board when I get home later. My parts board 7800 is largely intact minus the Maria, TIA, and a few banded capacitors that were all used to keep other 7800s alive.

[rasch_C]

Looks like at this point I'll get my scope up and/or just do the A/V mod. I've been having a heck of a time locating some of the precision caps in the correct specs. I'll spend my night tonight hopefully tracking them down and ordering them now that my laptop is back up, and maybe order an A/V mod.

 

Give some time for parts to come in and I'll report back with either results from a working oscilloscope, or a working console after A/V mod.

 

Thanks again for all your help! It's been a great learning experience.

 

-Allen

 

Sent from my Pixel 2 using Tapatalk

[rasch_C]

So, now that we're finally out of the -30°F temps .... brrr, that cold caused a lot of issues including taking my internet out for multiple days!

 

After locating all the caps for my old scope and realizing that it would be ~$90, I decided it is not worth it to recap it. I did only spend $10 for it at a hamfest and thought it would be an easy way to get my feet wet. I am planning on buying a Rigol DS1054Z shortly.

 

I decided to breadboard the easy composite mod and temporarily test that to see if I get picture. Still not getting a picture

 

I have attached pics of the signal reading on my scope - it works roughly, but I don't know how well or accurate.

 

-Allen

 

 

Sent from my Pixel 2 using Tapatalk

[ChildOfCv]

Rigol? Okay Mr Big Spender Oh well, that would probably be about a gillion times better than recapping an old scope.

 

Anyway, to show a full horizontal sweep you should be set on about 10us with a multiplier of 1. That will spread the sweep across a little more than 6 divisions on the screen, assuming I got my calculations straight...

 

Also, try scoping U3 pin 3 directly (careful not to short pins). That is where the raw timing sync pulses are generated and before they are mixed into the composite signal. It should look mostly high with a low dip 15.7K times per second. With the above settings, once every (just above) 6 divisions.

Edited February 4, 2019 by ChildOfCv

[rasch_C]

Rigol? Okay Mr Big Spender Oh well, that would probably be about a gillion times better than recapping an old scope.

 

Anyway, to show a full horizontal sweep you should be set on about 10us with a multiplier of 1. That will spread the sweep across a little more than 6 divisions on the screen, assuming I got my calculations straight...

 

Also, try scoping U3 pin 3 directly (careful not to short pins). That is where the raw timing sync pulses are generated and before they are mixed into the composite signal. It should look mostly high with a low dip 15.7K times per second. With the above settings, once every (just above) 6 divisions.

 

I don't know about big spender, but I suppose it is more than the $10 I have into this old Heathkit The Rigol seems like the best value entry-level scope. I could get a used Tek scope or something for less, but I determined that I want to have a modern scope for the kids in the Makerspace and it is more compact. The space is very limited.

 

I tried going to 10us, but the trace was not persistent. You would just start to see the trace go from the left to the right of the screen. At the lower setting you would just see the dot moving from the left to the right. I believe the last pic is at the 1 multiplier. Am I seeing the color-burst at the start of that signal? We are definitely beginners using a scope.

 

I'll see about firing everything up tonight and testing on U3.

 

 

Thanks again,

-Allen

[ChildOfCv]

Do you mean you saw a single trace and then a blank screen? That typically means it didn't see anything to trigger on. The trigger switch can be set to "NORMAL" or "AUTO". On "NORMAL" it typically waits for a trigger before doing the next sweep. On "AUTO", it is free-running if no trigger is present, but will stabilize on the signal when it finds a trigger.

 

In order for a scope to show you a consistent signal without the trace looking like a fuzz factory, it needs a unique point on the signal to trigger on. You should set the "trigger" input for negative slope since you're looking at a signal that goes to nearly 0 during sync pulses. You may also want to set it for DC, to look for a trigger level just above 0V. Adjust the trigger level knob until you see something stable. For trigger source, use LINE (looks like it's already set there anyway). Also, make sure your input amplitude is at a decent level.

 

Now, if a setting can't be made to work, it's possible that either the switch position or the capacitor or resistor connected to that setting is dead.

 

BTW, I have an even older Heathkit in storage. It's constructed with about 7 vacuum tubes and has even fewer features than what you have. I learned what a 2000V jolt feels like while running through the calibration procedures. Fortunately it only left a couple of bite marks on my arm where it brushed against the CRT plug. If you do calibrate your scope, wear long sleeves (maybe even a jacket)

 

For my hobbyist use I currently have a USB scope, which lately seems to be rather noisy, but at least gives me a usable signal display.

Edited February 4, 2019 by ChildOfCv

[rasch_C]

Do you mean you saw a single trace and then a blank screen? That typically means it didn't see anything to trigger on. The trigger switch can be set to "NORMAL" or "AUTO". On "NORMAL" it typically waits for a trigger before doing the next sweep. On "AUTO", it is free-running if no trigger is present, but will stabilize on the signal when it finds a trigger.

 

I had my timescale backwards in my head. At .1 sec (all the way counter-clockwise) is when it shows just the sweeping dot and then 10 Msec is when I see the sweeping signal (the complete trace does not stay lit). It goes dimmer as I continue to go clockwise making it really hard to see.

 

I did just play around some more with it tonight - getting a better feel for the scope. I set to Multiplier 1, 10uS timebase, DC, Negative trigger. Put my Volts/Div at 1 and reading pin 3 of U3 while power is on the trace went up 3 reticles and was a flat line for the complete width - I believe this would be indicating 3 volts?

 

-Allen

[ChildOfCv]

Okay that makes sense. The time scale tells you how fast the dot sweeps horizontally across the screen. Obviously when it's slow it's going to make a bright dot, and the faster it traces across, the dimmer, especially if you don't have a good trigger source or if your trigger level is out of the signal's range.

 

3V sounds about right. Those 74LS chips output just about 3.2V on highs.

 

I expected it to go to 0 on its pulses though. Looks like it goes to 3V instead. It must invert the signal while it's modulating it. So if you set the trigger on "+" instead, do you see two pulses per sweep? They should be a little over 6 divisions apart.