Need help with with macbook pro 820-3115-B

[MadMaxDog]

Re: Need help with with macbook pro 820-3115-B

Here's another question I don't know quite how to phrase. To kill boredom I sometimes just put the multimeter in diode mode and go around measuring some caps for shorts (red probe on ground). Most caps will test ok, having one side with a good connection to ground and the other with none, but every once in a while, quite randomly, I get a very quick continuity beep from the multimeter when I first touch the cap's side that shouldn't be connected to ground, then the brief short goes away. (hope that made sense, it's the best I can explain it).
It happens sporadically, and after it does, then it normally wont manifest again until I've measured another 10 to 15 components.

So my question is: Is this normal?, is the fact that I'm measuring around in diode mode is putting tiny amounts of current through the system and is it getting stored by some caps, and then gets released when I briefly touch the cap, or is this a symptom of a problem? (some shorted cap, maybe?)

[BlueMidnight]

Re: Need help with with macbook pro 820-3115-B

I would like to use (or learn to use) the functional test points featured on page 7

Page 7 appears to just be describing the test jigs for use in testing boards at the factory. In some cases, it specifies that the jig should have a certain number of pogo pins to contact ground or one of the signals under test. It does not describe any one place on the board where you can measure all these signals right next to each other, or anything like that.

Additionally, the tests they do at the factory are simple pass/fail tests. To find out why something isn't working requires testing more than just the signals listed there.

Most caps will test ok... ...but every once in a while, quite randomly, I get a very quick continuity beep from the multimeter when I first touch the cap's side that shouldn't be connected to ground, then the brief short goes away.

Resistance/diode tests will apply a small voltage and current to the circuit. Empty capacitors look like a short circuit at first because they just absorb all the energy being applied. But as they charge up, they start looking like a bigger and bigger resistance. Eventually, they charge up high enough that the multimeter will see it as an open circuit. And of course, if you go and test capacitors that are connected in parallel, they will all charge up the first time you test any of them. After that, you will just read all the others as an open circuit.

[MadMaxDog]

Re: Need help with with macbook pro 820-3115-B

I just connected the charger again to re-test the voltage on all the power rails, and on the very first fuse in the system (F6905) I'm getting a fluctuation I didn't have before. It jumps from 14.0v to 14.7v, then back to 14.0v etc...

Anyone know what that's about?

EDIT: I'm getting all sorts of erratic behaviour now. Near R7011 which I just measured, I had 8.2v on one side and 1.2v on the other, then suddenly it went up. Now it measures 14.7v on one side and 3.4v on the other.

[MadMaxDog]

Re: Need help with with macbook pro 820-3115-B

Another symptom, not sure I mentioned this before: right after I disconnect the charger, I get a short to ground on PPBUS_G3H. If I keep measuring, the multimeter will beep for 10 seconds then the short will go away. If I connect the charger again, then disconnect it, the short comes back for another 10 seconds.

[MadMaxDog]

Re: Need help with with macbook pro 820-3115-B

Here's where I'm at now: Still confused with the _L names, and here's why:



See the image above? First I thought U7941 wasn't working, and that replacing it would fix my board, because it has voltages going in left and right, and nothing coming out on P3V3S5_EN_L_R (I can't find a datasheet for that bad boy). But now I'm not so sure anymore, since P3V3S5_EN_L_R has an _L in the name, is it supposed to be LOW?


P.S. Red voltages are from the first time I measured things, then something changed. I don't know what, but before I could boot up the board only when the battery was connected, and fan would spin for 20 seconds, now when I press power button fan only jerks a bit but does not spin (also only with battery connected), so I measured everything again just to be sure. Those second measurements are the Orange ones, so around U7941 nothing really changed.

Is U7941 dead?


Thanks

[BlueMidnight]

Re: Need help with with macbook pro 820-3115-B

...the multimeter will beep for 10 seconds then the short will go away.
...since P3V3S5_EN_L_R has an _L in the name, is it supposed to be LOW?
Is U7941 dead?

Already answered these questions here and here.

For right now, measure the things listed in the power sequence on page 3 of the schematic. Nothing else. And don't desolder things. The danger is that you may cause more problems on top of the original problem. This may make it unrepairable (if it isn't already).

[MadMaxDog]

Re: Need help with with macbook pro 820-3115-B

Sorry, I wasn't trying to be repetitive, (even though I do feel I'm going round in circles - something still escapes me).
I did read the explanation about the 10 second short, but, when you say: That is normal. Resistance mode on your meter will send a tiny current through the circuit. It then measures voltage on each probe and uses Ohm's Law to figure out the resistance of the thing you are testing I don't understand why that short would go away after 10 seconds. if I remove the probes, turn off the multimeter, then back on again and measure once more, should it not send that tiny current through again and give me the same reading? For 10 seconds after I disconnect the power, it reads as short, after that, no short, no matter how many times I measure.

I've re-measured quite a few things on page 3 (posted below). Near AC ADAPTER I get a fluctuation between 14.0v and 14.7v, and I'm also getting some strange readings around U7200 (photo 2). At the bottom of the page, on Q7221, where it says 0v I sometimes get 3.4v... I've tried connecting power, disconnecting, connecting battery... I can't figure out the pattern (if there is one) but every once in a while I get voltage, most of the time I don't. Any idea how that could happen?

I also don't feel too confident about the two resistors at the top marked with orange circles (on photo 2). Shouldn't some of those 2.15v be getting past R7269?

So I know (I think I know) I need something that should be coming out of U7200 (PP5V_S3_REG), but I don't know if it's not coming out because U7200 is missing an input or because it's malfunctioning.

[BlueMidnight]

Re: Need help with with macbook pro 820-3115-B

I don't understand why that short would go away after 10 seconds.

Ohm's Law is Voltage(V) = Current(I) x Resistance(R)

Your meter applies as much voltage as required to force a specific amount of current through the circuit. It then reads the voltage that was required. Since it knows the V and the I, it can calculate the R. If the meter reaches the max voltage it is designed to provide before the current gets to the magic number, then it considers the resistance to be out-of-range, infinite, or as no continuity.

But if the circuit already has a voltage on it, then the meter will think that the resistance is really small. It doesn't know that the voltage is not coming from itself. It is just blindly using Ohm's Law. As the circuit powers down, the meter will be able to read the true resistance.

Shouldn't some of those 2.15v be getting past R7269?

The other side of R7269 is ground. But it's a somewhat confusing diagram. Instead of drawing the line, they should have just put another ground symbol there. See the attached picture.


Regarding your pictures... step 12 doesn't look right. Please confirm measurements all around Q7800. According to what you wrote, it is getting the enable signal, but is still turned off. Actually, you should probably measure/confirm all voltages for steps 10-13.

[pyry]

Re: Need help with with macbook pro 820-3115-B

you are missing the acivasion of the 5v PM_SLP_S4_L that's why you have 3.44v on pin 2 of Q7221

[BlueMidnight]

Re: Need help with with macbook pro 820-3115-B

Hmm, now that I read it again, I think my explanation for the 10 second short could be a bit confusing and was missing the second possibility.

Just think of it like this: The meter has a max voltage that it will apply to the circuit to do its test. This voltage is not very high. If the voltage of the power rail is 0, then the meter will charge the capacitors in the circuit to the meter's max voltage. If the voltage on the power rail is higher, then it may discharge down to the meter's max voltage.

Either way, it may look like a short circuit until the voltage stabilizes. Or it may look like an open circuit. I depends on the meter design and which probe is where. Only after the voltage stabilizes will you get an accurate reading.

And the capacitors in your circuit will hold that voltage for a while. Trying to read the resistance again will just show an open circuit.

[MadMaxDog]

Re: Need help with with macbook pro 820-3115-B

Thank you BlueMidnight. Your previous explanation made a lot of sense to me, and while we're at it, I have a certain confusion about the use of capacitors in certain places. I understand that capacitors filter out AC and are usually connected to ground as near as possible to sensitive chips to help stabilise the power. But in the schematics there are several places where capacitors are placed between two voltages. In this case, what exactly are they doing? If they only let AC through, are they simply letting the AC through from one power line to another so it can eventually reach ground (through some other capacitor along the way)?... But being AC, can it not travel backwards and forward through the same capacitor if there is voltage on both sides? I'm not really understanding how they work in those cases.

Here are the new voltage readings. I re-measured steps 10 to 14. I also measured around
Q7800.

The third photo baffles me. As usual, the red readings are the original ones, and the orange ones are what I get now. Seems strange that with identical voltages as previously, now I don't have an output from Q7820. Could it be dead?

[BlueMidnight]

Re: Need help with with macbook pro 820-3115-B

I'm sorry. I forgot to say you should measure things with a charged battery and the AC adapter disconnected. That's still the wrong AC adapter. It looks like it's failing to get past the same place it was before. And that will forever be the case when that adapter is plugged in.

[BlueMidnight]

Re: Need help with with macbook pro 820-3115-B

...I have a certain confusion about the use of capacitors in certain places.

I can think of a few main uses (there may be more) on a typical motherboard.

1. Bypassing or differential noise suppression. This is the one you are familiar with. It smooths out the power, usually right before a sensitive chip or at the input of a board or power supply. The natural capacitance of the traces (or noise on the power rail) will prevent power from reaching the chip instantly when it is needed. Having a capacitor nearby provides a reservoir it can use until the rest of the circuit catches up. You can also think of this as the AC ripple/noise being sent through the capacitor to ground. They are different ways of describing the same thing.

There are also capacitors like C7820/C7821 that don't seem too useful right now, but when the FET gets the 0V signal to turn on, one side of the capacitors will also be connected to that 0V.

And then there are capacitors like C7050 that are filtering out noise. R7050 is used to create a small voltage drop that used to figure out the current flowing through the circuit by using Ohm's Law. By using C7050, that circuit is averaging/smoothing out its measurements so they don't jump around too wildly.

Here is a nice document here about noise filtering.

2. In-rush limiting. This works in a similar way to the first use. But it happens when the power rail is first receiving power. The energy is sucked up by the capacitor on the rail until it is charged. This makes the voltage on the rail rise more gradually than it would otherwise. That can be more gentle for sensitive components connected to that rail. After the rail is at its full voltage, the capacitor will then act like a bypass capacitor.

3. Boosting. By charging a capacitor and then placing it in series with another voltage, those voltages get added together. It's the same thing as when you put batteries in series. For example, C7220 and C7260 allow U7200 to have higher voltages available to activate the gates of the 2 high side FETs. They are charged and then used to boost voltages thousands of times each second.

[MadMaxDog]

Re: Need help with with macbook pro 820-3115-B

Ok, I have good news and bad news, but first, let me apologise for the delay in posting. I just haven't had much time lately.

The good news is I have finally managed to get my hands on the correct charger, and I re-measured everything. So all measurements in Pink are brand new.

Bad news is I haven't progressed anymore, and because the new charger may have changed the values I feel I should start over from scratch.

I also have a few more *confusions* to sort out.

So here's the story:

After re-measuring everything, I took a deeper look at page 3 (below), and followed the number sequence starting at number 1, checking for correct voltages.



It seems I'm missing PM_SLP_S5_L, PM_SLP_S4_L and PM_SLP_S3_L, but then again, they do have '_L' on them, so maybe they should be 0v ? (steps 13, 14). Even though I seem to also be missing something around step 7, which comes first, I decided to investigate the PM_SLP's first, because pyry had mentioned that could be my problem.

Starting with the one he mentioned (PM_SLP_S4_L), I follow it and see it's created by U1800, and so are PM_SLP_S5_L and PM_SLP_S3_L. So how do I figure out if my problem is U1800 or if it's missing some other activation? I can only guess, but all the opposite pins that say SOMETHING_PWROK seem to be important (cause I think I want my POWER] to be OK]).




I followed SYS_PWROK (or PM_PCH_SYS_PWROK) first, which seems to come from nowhere, and just connects to HOOK3 of J2500, so I guess 0v there is just fine for now and not part of my problem.

Then I followed PM_PCH_PWROK which seems to come out of something that expects me to have voltage on ALL_SYS_PWRGD and CPUIMVP_PGOOD. I have 0v on the first, and oddly 109mV on CPUIMVP_PGOOD, but now I'm back at page 3, on number 25-1, when I was trying to follow number 14 backwards...

...but I didn't let that discourage me. It became kind of obvious I was doing something wrong, but I figured, what the hell, and just kept going.

I followed ALL_SYS_PWRGD to U7960, which seems to expect a whole lot more than it's getting.




PVCCSA_PGOOD seems to come out of U7100, so I'm not going to mess with that right now.

CPUVCCIOS0_PGOOD comes from U7600 and
P5V3V3_PGOOD comes from U7200... and I know I've been poking around this guy before. So I re-measured everything around U7200 (when they don't have IN or OUT or some sort of arrow showing the direction on the pins, I get a bit lost).



VIN seems fine, EN0 too. VO1 (that I interpret as Voltage Out 1) seems wrong. So what's missing on the left side? I guessed PP5V_S3_REG, and followed that. Then I find out I'm going the wrong way, and that PP5V_S3_REG is actually created by U7200 and not the other way around. So what's missing to create it? My best guess is that P5VS3_EN_L, where I have 3.45v should instead be 0, so I followed that to see where it came from.

I end up with this, on page 73:



then I notice this:



And here's where the real confusion starts: I have voltage on SMC_PM_G2_ENABLE, and I have 0v on all the others, but since they have _L, it means that 0v makes them True (or 1) right?

So technically I have everything I need to be in S0 state, but I'm stuck in S5? What gives?

Then I think maybe the only way to get to S0 state is by following the entire sequence, and going from S5 to S4, then S3, and only then am I allowed to get to S0?

[BlueMidnight]

Re: Need help with with macbook pro 820-3115-B

Okay, here is how the Sx signals work. The typically used states are S0, S3, S4, and S5...

S0 = On
S3 = Suspend to RAM
S4 = Suspend to disk / Hibernating
S5 = Soft Off (off but plugged in)

Don't let the numbers confuse you. The S0 state is referred to as the "highest" state because the board is powered up the most in this state. The S5 state is the "lowest" state because the board is the most powered off in this state.

After some first steps, the board enters the S5 state. Then it passes quickly through each of the states up to S0. The Sx signals on the board will be "TRUE" for the current state and every higher state. Not every Sx state gets its own PM_SLP_Sx_L signal though. There normally only exists a PM_SLP_S5_L, PM_SLP_S4_L, and PM_SLP_S3_L signal.

In other words, if it is off (not hibernating), it will start with all the signals = TRUE. Then go to the S3 state, which will set PM_SLP_S5_L and PM_SLP_S4_L = FALSE. Then it will go to the S0 state and set PM_SLP_S3_L = FALSE also. And since these are active-low "_L" signals, they are 0V when they = TRUE.

Some other signals may have the Sx state in their names. But be careful, because that doesn't always mean they are supposed to be powered/TRUE in that state. Sometimes yes, but sometimes no. Sometimes they do things needed before entering that state, etc.

That "State" table in the schematic is a little confusing. In the table, "1" means logic high (3V), and "0" means logic low (0V). Knowing that, you will see that it matches my description above exactly.

It may also be helpful to know that there are Gx power states too. S0-S4 are in the G1 state and S5 is in the G2 state. There is also a G3 state. You will see these mentioned in some of the signals as well. See the Wiki article here.


So here is what I see in your pictures:

- Your step 7 is good.

- Step 10-2 is not good. I see that Q7840 appears to be off (you wrote PP5V_SUS_FET = 0V) even though that circuit is getting the enable signal (P5V_3V3_SUS_EN = 5V). Try measuring around Q7840.

- You are stuck in the S5 state. It doesn't leave the S5 state until step 11 (but you never get past step 10-2). That is why you are not getting S3 power rails like PP5V_S3_REG. That's expected behavior.


Also, I want to make clear that the power sequence diagram doesn't explain what it does in each step. It just shows you where something will happen. You are just supposed to know what that something is. I think sometimes it can be a little confusing. For example, step 11 points to PM_SLP_S5 L. But what happens to that signal in that step? The answer is that step 11 is the step where it leaves S5 and enters S4, so PM_SLP_S5 L will go to 3V. The default/beginning voltage for that signal is 0v. Step 11 is where it finally changes. I hope that makes sense.

[MadMaxDog]

Re: Need help with with macbook pro 820-3115-B

Once again, Thank you BlueMidnight for that detailed explanation.

As suggested, I measured around Q7840. Here's the result.



Looking at this, I'm getting two different results from 2 components that seem almost identical. If Q7820 is getting 0v on it's Gate and outputting 3.31v from it's Source to it's Drain, I figure Q7840, which also has 0v on it's Gate, should be similar, and should be outputting the 5v it has on Source.

Although when I read P-Type, for some reason I imagine I need Positive voltage on the Gate for it to open. This seems to be the opposite. I found a datasheet for SiA427DJ here (pdf), but it didn't mean much to me.

I don't know enough to confidently decide if Q7840 is dead without some sort of confirmation. The only way I can think of testing this is by running a jumper from Q7840's Source to it's Drain, to mimic the functionality. Would that be a terribly silly thing to do?

[BlueMidnight]

Re: Need help with with macbook pro 820-3115-B

A P-Channel enhancement mode MOSFET will remain off unless the gate voltage gets low enough below the source voltage. With 0V on the gate and 5V on the source, Q7840 should be on. So it does appear like it might be bad.

Your plan isn't terrible. But before trying to short out Q7840, check PP5V_SUS_FET for a short to ground (just to be safe). With the board unpowered, set your meter to diode mode and put red probe on ground and black probe on the power rail. In diode mode, the display will read in volts. A short will read as close to zero. Less than 0.1V probably.

[MadMaxDog]

Re: Need help with with macbook pro 820-3115-B

I don't believe it's shorted. I get 0.36v on PP5V_SUS_FET. It's low, but there's still something there. I tried shorting the pins on Q7840, but still no joy. When I short the power button pin to ground (on keyboard connector), the fan jerks a bit. But that's it.

[BlueMidnight]

Re: Need help with with macbook pro 820-3115-B

The board obviously has more than 1 problem. The Q7840 issue wasn't even there at the beginning. So you need to fix each problem and then continue the same troubleshooting process to find the next.

In this case, you will just start at step 10 in the power sequence and continue on from there. Keep a jumper across Q7840. There should be 5V on PP5V_SUS_FET. Continue measuring the next steps on the power sequence diagram to see if anything has changed.