MB-6582 No 5V at PSU J3

[otto1]

OK so I carefully tested before plugging in the powersupply. I know the power supply is working fine but I am not getting 5V @ J3. I am getting a connection between 5V and all the IC sockets. So I guess I need to backtrack and find where I've lost a connection between the power supply plug in and J3. Any help would be appreciated. oh, this is rev 2 PCB if that matters.

Thanks!

[Wilba]

Maybe you forgot the bridge at J73.

[otto1]

Maybe you forgot the bridge at J73.

I didn't forget the bridge. I even resoldered the bridge just in case it was a cold solder joint. :( Other suggeestions?

[Wilba]

Maybe this picture will help you work out what's wrong.

Just to be clear, are you testing if pins on IC sockets have 5V?

And you say the pins on IC sockets have 5V, but the top pin of J3 does not?

I suggest you check for 5V for all the pads shown in the diagram.

Edit the image and mark which ones have 5V and which do not.

It would be a lot quicker and accurate than describing it in words.

[otto1]

Thanks for your help Wilba. OK, starting from the power supply jack, testing each pad, I get 5V up until power switch pad 2. This is with the power switch flipped "off". If I flip it "on" then I don't get 5V anywhere, which seems weird...

[Wilba]

That might mean you have a short between 5V and ground.

Don't test with power connected anymore.

Test continuity only, with a multimeter (i.e. beeps when probes are touching).

With the switch on, you can test if the 5V pads are connected to each other, and if they are connected to ground somehow.

[otto1]

Yes it appears that everything connected to pad 2 is also connected to ground. Up to pad one none of that is shorted to ground. It's not obvious to me where I have shorted. Any hints on how I might track down the short? Should i assume the short to ground is at either J2, J3 or J4?

Edited February 25, 2010 by otto1

[Wilba]

Since you say S1 pin 1 has 5V when switch is off, and not 5V when switch is on, this means there might be a short somewhere AFTER S1 pin 2, i.e. not in the power socket pads.

Basically the short can be anywhere 5V and GND come close together, such as J3, J4, in the "bankstick" IC sockets, all the 100nF capacitors, etc.

Check all the pads which should be connected to 5V. See picture. The last time this happened to somebody, it was in the "bankstick" IC sockets - that's where there are a lot of 5V and GND pads close to each other.

If you can't find an obvious solder blob shorting pads, then you can move onto more drastic measures - cutting the 5V track to discover which "half" of the track is shorting.

[otto1]

I'm getting myself confused here. I've looked for any areas where solder might be touching multiple times and with a magnifying glass and I see none. My friend is also building the kit, were trying to do this together. His tests 5v fine at J3 with power supply on. So I used his board to A&B mine with to see what the differences are. With the power switch off, just checking continuity (with an analog multimeter "OHM X1K" setting, touch the tips together the pin goes to the right...). What surprised me was this his board does show coninuity in the same places as mine between 5v and ground, however the needle slowly drops in these areas on his board while the needle stays steady to the right on mine. So, it's like mine has constant continuity between 5v and ground and his shows initial continuity then drops off. Could that indicate something else going on or is that normal?

If this still indicates a soldering short, I'm at a bit of a loss as to where to look. I realize cutting the 5V is an option but that still leaves me looking at half a board with no obvious soldering "blobs"...

[Wilba]

Since there are several capacitors between 5V and ground on the PCB, testing resistance between 5V and ground tracks sometimes causes unexpected results, as the capacitors get charged by the current coming out of the multimeter.

On my multimeter, there's a "continuity tester" mode - it beeps if theres a very low resistance between the probes. Sometimes testing between 5V and ground will beep for less than a second, as the capacitor is being charged by the multimeter, which senses this as high conductance. Once the capacitor is charged, the beep stops.

Perhaps what you are observing is the same thing. On your friend's one, testing resistance is showing a non-constant resistance because the capacitors are charging. On yours, the tracks are (we suspect) shorted, and so the capacitors are also shorted and don't get charged.

Here's a picture where I think you should break tracks to discover which section of the 5V tracks are shorting to ground.

Start with the locations marked with green circles. Then you can decide which of the other locations (marked with pink circles) you can use to narrow it down. This is only a suggestion... you want to cut tracks until you get it down to a few pads that must have the short, then you only need to desolder a few components and check if that fixes the problem. i.e. let's say it's one capacitor between the 5V/ground tracks that is shorting (the capacitor itself, not solder blobs) then you'll never find out which one just by looking at solder blobs. Similarly for all we know there's a short in the copper between a pad and the ground plane. Finding it will be very hard if you don't narrow it down to a smaller area to look and start removing components.

I wish I could suggest an easier way but there really isn't, and I think you've given up on looking for solder blobs shorting... it probably is something else.

Let me know what your results are after cutting tracks... draw on the same picture showing where you cut and which section is shorting.

[otto1]

Thanks for your help and patience Wilba!

What is the best way to actually cut the tracks? Just cut into the board with a knife or???

[Wilba]

I suggest starting the cut with a craft knife, then using the corner of a thin "precision" screwdriver to scratch a good groove (i.e. through the copper and into the substrate). When you want to rejoin it, lightly scrape the solder mask away from both sides of the track, add some solder and place a cut resistor lead or other solid core wire across the groove.

You can see a good example of cut tracks :blush:

[otto1]

I have made 2 cuts per your instructions. The first cut showed that the short is in the upper section of the PCB. So I made the second cut to seperate the sid area. when I test sid pin 25 (5v) to another 25 there is polarity as expected and there is no polarity at ground, yay. However, when I test pin 14 (gnd) it still shows polarity to ground as expected but it also shows polarity to 5V at J3 or J4, however there is no polarity between pins 14 and 25. So should I just ignore that and focus on the section above the first cut where J2,J3,j4 etc are?

[Wilba]

You've cut only the 5V tracks... all the pins that should be connected to ground still are connected.

So when you test pin 14 of the SID socket (ground) to 5V pins at J3, it's showing as connected. That's OK.

You can now test between any of the pins connected to 5V (like the picture of the 5V tracks) and any one of the ground pins (i.e. at J3 or J4).

You'll then notice that there is no connection in the SID sockets, and hopefully nowhere else lower than your cut track at point #1.

Perhaps you should now refer to the first picture I posted, showing the 5V tracks connected to the power socket and the switch.

Start desoldering the joints connected to these 5V tracks. You don't need to remove parts yet... just take off the solder. Just use desoldering wick/braid to remove the solder and check there are no shorts between the pad and the ground plane. Retest if 5V/ground are still connected after each joint is desoldered.

All the areas around the pads on the bottom of the PCB is connected to ground. Perhaps there is a fault in the PCB somewhere around those pads... a tiny bit of copper connecting the pad to the ground plane. Use a magnifying glass to see if there is anything there.

[otto1]

Start desoldering the joints connected to these 5V tracks. You don't need to remove parts yet... just take off the solder. Just use desoldering wick/braid to remove the solder and check there are no shorts between the pad and the ground plane. Retest if 5V/ground are still connected after each joint is desoldered.

All the areas around the pads on the bottom of the PCB is connected to ground. Perhaps there is a fault in the PCB somewhere around those pads... a tiny bit of copper connecting the pad to the ground plane. Use a magnifying glass to see if there is anything there.

Using desoldering wick I have desoldered the 5v connections, still have the short. I'm not sure why desoldering without actually removing the component would work though. I mean if I am just desolderign the the component can still touch the copper until I actually remove it, right?

It's really difficult for me to tell if there is a fault in the board. I mean there are areas where the red stuff is slightly off-set from the pads but it will have a black circle or square around the pad but still I could see how one could easily overlap in the offset areas... Is my next step to actually remove the components one at a time? I can't see narrowing it down until I do that but I want to make sure I'm on the right track.... ugh...