MacBook Air A1466 820-00165-A (no power, no light)

[mon2]

In your thermal pic, who is heating up to almost 50 degrees C?

SMC_RESET_L should be a logic '1' as a normal parked state. When this line is logic '0' = logic low, then the board is in RESET state.

What is the resistance to ground of L7630?

Review this thread for other suggestions:

820-00165 - No green light. SMC_RESET_L is 1.15v | Rossmanngroup Forums

[reformatt]

Yeah, I expected the lines to be oscillating but didn't know whether the voltage range was acceptable.

I'm going to check the other steps in the flow chart and report back. The SMC was sold as "new"... no idea where it actually comes from.

Don't know where you sourced it, but I wouldn't trust anything other than from a donor board. From the readings you have, I'd strongly suspect this chip is blank. We have no way currently of programming these from blank. The board needs to be able to boot first for us to reprogram it using the SMCtool.

At the moment, we know that SMC's from 820-3437 boards can be used on 820-00165 boards. They will have an issue with audio IIRC until the SMC is reprogrammed with Muerto's SMC tool.

Wasting your time until you get a known good SMC from a donor board.

[mcplslg123]

Recently has a very similar case on 820-00165A where PPBUS_G3H was stuck at 8.21V. Checked all steps till smbus comminication and found nothing wrong. Refloed the SMC but still same.
Finally managed a smc from same donor board and replaced the smc and it booted.
So it could be bad SMC chip itself.

[kevingill]

What does the area around Y5110 look like?
Also, is U5110 nice and clean?

[davidebeatrici]

Sorry for the delay.

In your thermal pic, who is heating up to almost 50 degrees C?

SMC.

SMC_RESET_L should be a logic '1' as a normal parked state. When this line is logic '0' = logic low, then the board is in RESET state.

Yeah, but what is the voltage threshold that determines logic 1 or 0? I'm concerned about R5100 being out of spec.

What is the resistance to ground of L7630?

P1V05S0_LL: 178Ω
PP1V05_S0_REG_R: 178Ω

Don't know where you sourced it, but I wouldn't trust anything other than from a donor board. From the readings you have, I'd strongly suspect this chip is blank. We have no way currently of programming these from blank. The board needs to be able to boot first for us to reprogram it using the SMCtool.

At the moment, we know that SMC's from 820-3437 boards can be used on 820-00165 boards. They will have an issue with audio IIRC until the SMC is reprogrammed with Muerto's SMC tool.

Wasting your time until you get a known good SMC from a donor board.

I considered buying a donor board as suggested here, but I quickly realized that there is no way to be 100% sure that the SMC is going to be good, even if there is no corrosion around the area.

Since the chip from AliExpress is very cheap (€7.36 for 2 pieces) I decided to give it a try. I can leave here the link if allowed. Interestingly, there are other sellers who provide the chip already programmed for specific boards but the conversation with the one I bought from is as follows:

Me: I successfully soldered the chip. Is there anything else that is required to make it work?
Seller: What's the problem? This requires direct programming
Me: How can I program it? Do I have to connect SCL and SDA to some kind of hardware programmer?
Seller: Friend, don't need programming, you can use it directly on the machine
Me: Oh, really? I'll have to check if there is some other issue on the board then, because PP3BUS_G3H is not boosted to ~8.65V by the SMC (it stays at ~8.20V). For reference, it's a 820-00165-A.

Waiting for another reply.

What does the area around Y5110 look like?

Also, is U5110 nice and clean?

Only missing C5126.

[mon2]

To me, regardless of the ball logic states, the part should not be so hot on the temp. Do you feel comfortable with the soldering? Alignment?

[davidebeatrici]

Yup, the result can be seen in the pictures I posted in the previous page.

[mon2]

Personally, at this stage, I would consider to flux and reflow this BGA again. 'Make it dance' as Louis would say to align the balls like soldiers.

Was this (potential) short removed before the replacement was soldered onto the pads?

[davidebeatrici]

It's actually not solder, but lint left behind by a Q-Tip. I used a brush later.

[reformatt]

Just to be clear, there isn't a publicly available method to reprogram these SMC's outside of Muerto's SMCtool (which requires the board to be able to boot and run that program off a USB stick, since it's using Apple's own tools to do the reprogramming).

The only way you're going to get one pre-programmed is if it's come off a donor board. Vendors on Aliexpress are a bit of a lottery, especially with IC's. I usually look at the store I'm buying off and see how long they've been on Ali before I order these days - a lot disappear pretty quick after you made the purchase.

Donor boards (especially 820-3437 versions) are usually available on eBay as defective, as most people have no idea how to fix them (faults are usually quite minor in my experience).

This store on Ali has been around since 2012, and the listing for the SMC is for each model of board. The cost of these is about the same what I get defective boards for usually off eBay.

https://www.aliexpress.com/item/1005...Cquery_from%3A

[davidebeatrici]

Thank you. The price is actually honest considering you receive the chip ready to be soldered and somewhat guaranteed to work.

[davidebeatrici]

The seller confirmed the SMC is blank and has to be programmed manually in order for it to work on this board, as we expected. For obvious reasons no sellers are willing to give up the method they use, it could be as simple as a tool that writes the firmware over the SCL and SDA lines.

I ended up buying 3 820-3437-B boards with untouched SMC for ~€15 each. I'll let you know when I receive them.

[reformatt]

For obvious reasons no sellers are willing to give up the method they use

Because they don't have one. They are bullshitting if they claim to have.

SMC's off Aliexpress just not worth it. At least with the donor boards, you get other things like connectors etc, you can even steal the RAM if it's 8GB to upgrade the 165 boards. These are more valuable since they natively run Monterey and have generally better speeds.

[davidebeatrici]

Because they don't have one. They are bullshitting if they claim to have.

Then I guess they repurposed some booting boards strictly for SMC programming purposes.

SMC's off Aliexpress just not worth it. At least with the donor boards, you get other things like connectors etc, you can even steal the RAM if it's 8GB to upgrade the 165 boards. These are more valuable since they natively run Monterey and have generally better speeds.

Agreed, especially as I already need a capacitor and a resistor (possibly 2 if R5100 turns out to be unreliable). If you can find non-cannibalized boards (no missing components, untouched SMC area) at a good price like I did, there is no reason to get a single chip that is only slightly cheaper.

Luckily the board I'm attempting to repair already has 8 GB. As for the boards I bought, 2 have 4 GB and 1 has 8 GB according to the pictures:



1. MLB 1.4G HY 4G
2. MLB 1.3G EL 4G
3. MLB 1.7G EL 8G

The sticker on the board at the middle says "POWER PROBLEMS (TURNS ON SOMETIMES)".

[davidebeatrici]

I apologize for the delay, it's been a busy month.

Can you believe all three motherboards came with missing SMC_XTAL and SMC_XTAL_R? Let me elaborate...

---

MLB 1.4G HY 4G

Very clean board, no visible corrosion.



Reflowed the SMC, no luck.

---

MLB 1.3G EL 4G

Visible corrosion and a hint ("turns on sometimes"). Attempted several times to turn it on, without success.



Cleaned the corrosion, reflowed the affected areas (SMC included) and the board seemed to work fine. Until it suddenly didn't want to turn on again a few hours later.

I didn't investigate further because the board is fully installed in the machine right now.

---

MLB 1.7G EL 8G

Clean board, no visible corrosion. However, there is a large darker area on the top.



Reflowed the SMC and the board turned on, but on a second attempt it was dead again. Turns out it only works if I heat up the SMC area a little bit beforehand.

This leads me to think the chip itself is defective, because the balls underneath don't get melted unless I set my hot air station to ~370°C. In short: it's not a reflow.

---

Basically, I'm not confident in taking any of the 3 chips to install on the 820-00165-A...

[davidebeatrici]

Now, let's talk about the machine itself.

The light on the charger never turned on, which leads me to think it's a defective DC board since I tested it with two different motherboards (MLB 1.3G EL 4G and MLB 1.7G EL 8G).



The keyboard appears to be working fine, except for the left-side ALT, which does absolutely nothing when pressed. I cleaned the corrosion under the battery, but unfortunately there was no change.

The screen appears to be in good condition, but sometimes the backlight didn't work. The first time it happened was when there was an automatic attempt to boot macOS from the original SSD.
I don't think it's a screen problem because the issue always manifested itself right when the board started and never afterwards.

Finally, the battery, which was manufactured in 2017 according to the label on the back. It's detected but with a SoC of 0%. I quickly realized why: ~0.330V between positive and ground and didn't take current.



Turns out the cells are not dead, which is expected since the battery is not that old. ~5.4V go into the battery board.

The fuse (10AH2) is fine and the drop in voltage happens before that, I believe one of the two unidentified chips at the right is the culprit. I bet BQ20Z451 locked both charge and discharge and as far as I know there is no way to fix it without an expensive tool.

To be fair, pinpointing the exact issue on the motherboard is what I cared the most about because it helps improving my skills and experience in electronics, but it's disappointing to spend a lot of hours and still have a broken machine.
Not sure how I should proceed at this point. I'm open to suggestions, as always. And thank you very much for your support so far, of course.

[reformatt]

Are those photo's prior to reflow of the SMC? Edge bonding needs to be removed before reflow and 370 degrees is way too low for these boards (in fact I only use that temp to remove the SMC edge bonding). They have a large ground plane that sucks up a lot of heat. Once you remove the edge bonding, you need to eyeball each corner under a microscope to make sure that all the bonding is gone. I use one of the Qianli underfill removal tools to pick at each corner while using some hot air.

Working out what temp is best depends on your hot air station and your nozzle diameter. For example, I use 460 degrees with an air flow of 50 on my Quick TR1300 station with a 10mm nozzle for SMC's. Get the board up to temp first, then apply some flux, then go back in and do the reflow. Flux should be sucked in under the chip, and you should be able to nudge the chip in a corner and see it snap back. That's a proper SMC reflow.

With Magsafe, there are a few issues that can stop a board booting. With some aftermarket chargers, the 165 boards will not boot without a working battery. This is because PM_BATLOW_L is held low. Genuine Magsafe 2 chargers don't have this issue. So with the one wire circuit, no charger light shouldn't stop the board booting with a genuine charger, however it will prevent the battery from being charged.

[davidebeatrici]

Are those photo's prior to reflow of the SMC?

Yes.

Edge bonding needs to be removed before reflow and 370 degrees is way too low for these boards (in fact I only use that temp to remove the SMC edge bonding). They have a large ground plane that sucks up a lot of heat. Once you remove the edge bonding, you need to eyeball each corner under a microscope to make sure that all the bonding is gone.

I'm pretty sure I did that the correct way as I could see the chip moving around when I touched it with my tweezers. I can confirm that working on the SMC required a very high temperature compared to what I generally use, ~320°C.

I use one of the Qianli underfill removal tools to pick at each corner while using some hot air.

Just bought one for convenience, I successfully used tweezers on these boards.

Working out what temp is best depends on your hot air station and your nozzle diameter. For example, I use 460 degrees with an air flow of 50 on my Quick TR1300 station with a 10mm nozzle for SMC's. Get the board up to temp first, then apply some flux, then go back in and do the reflow. Flux should be sucked in under the chip, and you should be able to nudge the chip in a corner and see it snap back. That's a proper SMC reflow.

The method is the same I used, however with ~370°C and ~70% airflow on a Yihua 706. Not sure about the nozzle diameter, but it's pretty small as I rarely work on big SMD components.

With Magsafe, there are a few issues that can stop a board booting. With some aftermarket chargers, the 165 boards will not boot without a working battery. This is because PM_BATLOW_L is held low. Genuine Magsafe 2 chargers don't have this issue. So with the one wire circuit, no charger light shouldn't stop the board booting with a genuine charger, however it will prevent the battery from being charged.

I had a suspicion the missing light was due to the faulty battery, thank you for confirming. In my case the behavior was the same with both a genuine charger and an aftermarket PD -> MagSafe adapter.

Right now I'm waiting for a proper chip holder to arrive, so that I can reball the SMC properly and with a bit of luck repair at least 2 out of 4 boards. Fingers crossed.