Atari 400 power on successfully only now and then

[slaanesh]

I have an Atari 400 computer which I recently upgraded with a 48/52K RAM card and the UAV s-video  and sound module.

Everything looks/sounds great on it except it doesn't boot all the time.

In fact in mostly doesn't boot. Now and then it powers up fine, but other times it flickers (yellow?) very briefly and then just sits at a black screen.

 

I've cleaned the RAM and CPU card contacts as well as the cartridge port but doesn't seem to make any difference.

I've measured the voltage on the 7805 (actually something else) and it was giving 11.8V.

I've measured exactly 5V on the UAV board itself on the two pads provided.

 

Anyone help diagnose this further? Thanks!

[slaanesh]

Correction to the above:

7812 was 11.8v

7805 was 11.95v

[Sugarland]

12v on the 7805? I thought it converted 9v AC to  5v DC.

[Rybags]

I'd say if the 5V reg was outputting 12V then most of the computer would be boom, shanker! and you'd not even get a black screen.

[slaanesh]

I won't try it anymore until it's resolved.

I tried replacing the 7805 but I still get the same result.

From my understanding a 7805 takes a variable DC (~9-20v) on the input and should produce 5V on the output pin.

 

I have a complete set of replacement caps, would this resolve the issue? Many of the modern caps I have are physically smaller but have the same voltage and capacitance. Should be okay, right?

Are there any other reference points for measuring voltages on the power board that I can test?

 

[slaanesh]

So I've replaced all the caps on the power board.

There doesn't seem to be any difference so I don't think it's the caps

 

I didn't replace 10uF on the motherboard (yet) but I may as well do that for completeness.

 

On first power up it did turn on... and I got the memo notepad.

 

After that just black again.

 

Just to recap, I'm using the Atari UAV video mode and the 48/52K RAM card.

[TGB1718]

what pin are you measuring the 5V on, picture attached for the pins, the input pin may be around 12V DC

but obviously the output should be 5V DC

 

as @Rybags says, if its outputting 12V DC it would likely blow every chip in the 400.

 

 

[Sugarland]

Try cleaning and reseating the chips. I'd try the CPU and GTIA first. If that doesn't work, do you have another good/working A8? Swap the chips from the non-working to the working machine to see if the symptom moves with the chips.

 

Check the chip sockets to be sure they are clean and not broken.

[slaanesh]

So I'd previously re-seated all the main components that I could on the CPU board. It's a Rev D. so there were 7 chips on it.

I cleaned the pins with Isopropyl but not the sockets... how do you clean them? Rub a needle into each pin socket entry? Spray contact cleaner into it?

 

On the power board I measured the voltages on the J201 and I got ~+12V on pin 2 and ~-5V on pin 18. In line with the Atari service guide.

The guide didn't say to measure two other pins 6 and 7 which are marked +5a and +5b. But I did measure these, they had about 11v if I recall... is that correct?

 

As for measuring the 7805, I measure both pairs - the input + ground and the output + ground. The input was around 13v or so and the output was as I mentioned earlier close to 12v. And this is a replacement 7805 I just put in. Doesn't make sense...? Yet the above voltages seem to be okay?

 

And just a note, the power board is no longer connected to the main board. The measurements have been made with it removed.

 

I don't have any spare easily accessible components... Well I have a couple Atari 800XLs which work perfectly, one PAL and one NTSC. But if they don't have socketed components then I don't want to mess with them.

 

I think I have to undo the mods.. surely one of these may be the issue? Both are completely reversable... painful as I don't enjoy doing these kind of things.

 

Edited January 11, 2020 by slaanesh

[Sugarland]

Yes contact cleaner with maybe a cotton swab. Then I think reseating the chip ten or twenty times can scrape away corrosion left behind. Just be careful with the chip (or use another generic 40 pin chip).  Then try swapping the chips to another working machine to see if the symptom moves. Of course the CPU will not swap to an XL and the CPU is suspect here.

[TGB1718]

Acording to the circuit diagram 

Pin   Voltage

18     -5

2       +12

6       +5

5       +5

7       +5

So if thats what your getting on J201, it look ok

 

[slaanesh]

Thanks for that... I am finding that I have ~11V on lines 5, 6 and 7 on J201... these should be 5V. 

 

What could be causing this? I just replace the 7805... surely the new one isn't also broken?

 

Are there any other components that could be causing this?

 

What about R202, that big resistor... IF that was kaput, would that be causing issues?

I could successfully measure R203 floating at ~6.73k (suppose to be 6.8k) but couldn't get a reading on R202...

Edited January 13, 2020 by slaanesh

[jimerson]

If you did any soldering to install the mods, I would double-check carefully for any solder bridges, blobs, stray wires or anything that could be jumpering between components or traces. Especially in places like J202 where it appears from the diagram 5v and 12v are in close proximity. I always use a strong magnifier or microscope when inspecting boards.

[slaanesh]

I didn't do any work on the power board until after the issue was already happening.

 

At this point all I have done is replace all the caps and the 7805 on the board and the symptoms remain identical.

 

One interesting point is the traces on pin 2 and 3 of the 7805... There seems to be continuity between these two pins. Is that normal? Or would component failure cause this?

This continuity seemed to be present before I replaced the 7805.

 

 

[TGB1718]

I would remove the 7805 (possibly this has also blown due to whatever is causing the problem), hence continuety between 2 and 3

and check across those pins with it out.

 

Then check all the rectifier diodes, you may have to unsolder one leg of each diode to get a sensible reading. CR206,7,8 & 9

 

While the 7805 is out check R203 (33 ohms, might be R202 for you). I notice 2 resistors labeled R203 on the circuit I have, the 6.8K one you mention is on the UHF modulator

and has nothing to do with the power circuit.

 

Also check across the pins on the 7805 while its out, Unfortunately I don't have one here the let you know what to expect, but you

shouldn't have any 'shorts', but might get some conductivity one way, but not the other.

 

Let us know your progress

 

 

 

[slaanesh]

Thanks for your suggestion. I'm not really sure what to try now but willing to learn about what to do here.

---------------------------

 

Okay I've removed the 7805.

 

Without desoldering one of the legs I checked the diodes CR205,6,7,8,9. They all give a voltage drop of around ~700 - I'm assuming this is 700/1000 of a volt or 0.7v

Repeat: this is without removing any legs from the board

 

I also checked R202 which is the big 33 ohms resistor and it reads as 33.1.

 

I checked the traces on the board where the 7805 was without powering it up.

 

In continuity/diode mode on my MM, when I connect pin 1 + 2 it reads a voltage drop ticking over until it gets to 2000 whereupon it says "1" - assuming 2.000 volts is the most it will do.

The same thing happens whith pins 2 + 3.

 

If I go to resistance mode (is this useful?)  and I set it to 20k it just ticks over with increasing resistance... *shrugs* on both 1+2 and 2+3.

 

[TGB1718]

From the voltages measured across the diodes they look right, the forward volts drop of a silicone diode is ~0.6-0.7V.

But it still may be one is allowing current both ways, hard to tell in-circuit.

 

If you want to be completely sure you will have to lift one leg

 

To test on the diodes you need to confirm they conduct one way and reversing the meter connections and get no conductivity

If you get conductivity both ways or no conductivity both ways, then its duff.

 

 

The end of the diode with a white bar should be -ve to conduct(black lead on meter) with +ve one the other end

and obviously if you connect +ve to the white bar and -ve to the other end it shouldn't conduct.

 

If I had a 7805, I would confirm what you are measuring, I will trawl through my box of spares and see if I have one somewhere

 

[slaanesh]

I lifted one leg off the CR209 and CR207.

Both seem to be correct providing a reduction in voltage one way and not the other.

I also put the 7805 back in and can confirm that pin 3 still is providing ~11.5v.

 

So I dunno ... what else can I try?

 

The 7812 seems to be correct... the input is ~27v and the output is ~11.8v.

 

I also measured the old 7805 by connecting it to a 9v input source. It's input was 9.5v and output ~5v, so it's still operational... 

 

Something seems to be "leaking" 11.8v on the 7805's pin-3 output somehow... somewhere...

 

The board has brand new caps - they are all smaller than the originals and have been soldered in a different orientation as these were the only ones I could get.

Edited January 16, 2020 by slaanesh
Added photos

[TGB1718]

I think the next thing to do is check continuety between components to see if you have a broken track.

 

Also look at the solder joints, maybe resolder any that you haven't yet touched.

Look for solder splashes that may be shorting tracks together, looking at the photo's there are a few suspect places, but hard to confirm.

 

Near the centre of the bottom of ther board photo, there are 5 vertical pins, the top 2 look to be shorted, this may be correct, but worth checking

in case its a solder bridge.

 

 

If thats ok, then check for any short circuits between tracks.

[slaanesh]

The bridge of the top two pins at the 5 vertical pins is just trick of the light - there is no continuity between them.

 

Originally I didn't do any soldering on this board at all, so there shouldn't have been any blobs or tracks.

 

The Atari 400 was working before I performed the mods and disassembled it. Maybe I guess something bent? cracked (reflowing may solved this if I find anything).

 

Are there any sections of the board which I should concentrate on?

 

 

[slaanesh]

Or any other components worth replacing? I'm starting to lose hope with this board ...

[jimerson]

I wonder if you may be measuring phantom voltages. I would connect appropriate-sized resistors to the DC power rails to simulate the normal computer load, then check the voltages under load.

[TGB1718]

On 1/17/2020 at 8:29 PM, jimerson said:

I wonder if you may be measuring phantom voltages. I would connect appropriate-sized resistors to the DC power rails to simulate the normal computer load, then check the voltages under load.

I had wondered about that too, but any sort of load should do and the circuit has the RF Modulator powered by the 5V line

so it should be ok, but...  at this stage it's worth trying anything.

[Mclaneinc]

Did the 400 work ok before the recent mods?

[slaanesh]

Yes worked perfectly fine before the mod.

it still worked intermittently after but I haven’t tried it for a week now as I don’t want to connect the power board to the rest of the computer anymore.

i’m hoping the rest isn’t damaged... but I dunno.
i have ordered a new power board and it’s on the way.