Identify a bad MAX77620A


I’m new here and i have a question. I look for a way to check if the MAX77620A is good or not. I know it’s possible have some caps shorted around the Ic, like for the M92T36 and the P13USB but i would like to know if have an other way…

Can someone tell me the symptoms of a MAX77620A faulty ? The switch not start ? the switch stuck on logo ?

Thank’s a lot

The best way to determine if it is doing it’s job is to measure the various power rails coming off it. If you are sure you’ve identified and removed all shorts on the power rails, then any rails not measuring properly could be a sign of damage.

From my notes, you should have 3.3v, 1.8v, 1.35v, 1.15v, and 1.1v with a ~4v supply feed from the battery. There are a few pictures floating around of where to take your measurements, Google will be your friend here.

Good luck!

You are correct, though 3V3PDR is supplied elsewhere by way of the “ENXX” IC and afaict is only going to the MAX PMIC in order for it to give/generate the PG signal (power good) as i guess all other primary rails produced by the PMIC ultimately derive from the ~4v input.

The ~1v one (two inductors paralleled) is presumed CPU VCORE… though maybe it’s the 1.15V rail, have you found anything in the avaliabe Tegra datasheets that confirms/denies this? (though i appreciate the datasheets available aren’t a direct match)

Checking the rails can simply be done by measuring the voltage on the inductor “outputs” surrounding the PMIC, and checking for faults here just entails measuring resistance relative to ground on these same inductors

*edited to correct error/mistakes

Thank you for the clarification. I’ve not done much digging on the Tegra datasheets. I presume you are aware of the two bucks on the other side of the board? One drives the CPU and gets activated after the bootstrap processor hands over execution control to the main CPU complex. The other is for the GPU and turns on when the Horizon home screen appears.

Superb work identifying the ENXX, just goes to show how easy it is to mis-identify things. I need to go back through some of my gash boards with a new set of eyes!

I’ll ping FXDX and encourage him to correct the headphone amp reference.


Sheriff, could you help me identify the bucks you mention on the front side, just a brief description so I could located and test it on my board. Thanks

Yeah these two have me somewhat confused, My first impression was they are SW governed as a result of power demands… possibly the other CPU cores (?) as from memory these rails were not present (and again from memory) were not present even after HOS was booted on some boards which signified a FW version difference… the next thought was possibly one of these IC’s was dediicated to GPU power rails, but i couldn’t find anything conclusive in the avaliable datasheets without a correct pinout to verify. Then, from memory again, i recall measuring approx 1.4V from one of these IC’s (depending on load) which is uncharacteristic for a CPU VCORE.

The only reason for those thoughts is the assumption that primary CPU core would be present and ready to go after bootloader stage and the only example of that I’ve found would be at the max PMIC afaict.

Thanks man, i suppose it’s like chinese whispers and things spread and are morphed, but tbh i should have realised the IC sooner, the fact that the output was on the wrong side (left) should have been a clear indication this was a Japanese manufactured part :wink: but i got my R’s mixed up and thinking i already checked all availiable manufacturers (Rohm for example)…

btw i saw your post over on GBATemp ages ago and it definately made me smile, regarding your magic soldering iron, definately had a few situations like that in the past causing scratching of the head… would be curious to know if you identified the root cause of those failures?

He’s referring to the two Max IC’s on front of the board, which you know about :slight_smile:

Thanks Severence,

Ya I measure that chip prob 10 plus times now,

And almost whole board in diode. Got a feeling my board gonna be a faulty emmc.

Gonna wait til I could swap it over another console to do a test in Hecate like you brought up. Last resort would be RE all I g it before giving up

I meant measured components and test point around that Chip( diode , resistant, n voltage)

Working through the source code for sept and lockpick_rcm, I can see that before handover to CCPLEX (main CPU complex) from the BPMP (boot processor), the T210 power subsystem is configured in software, including enabling the MAX77621 buck and ~0.8v to the CPU (among other rails).

I’ve got a board here that was refusing to get to stage 2 of key derivation (using sept) in LockPick_RCM and it looks like my CPU MAX77621 has died as it doesn’t fire up and sept black-screens after hand-off from Lockpick.

1 Like

Very interesting, thanks for the info!

For anyone not following,

BPMP = primary boot CPU (aka the arm7 Boot and Power Management Processor) powered by way of the MAX77620A

CCPLEX = Main CPU (Cortex-A57) powered by way of the MAX77621

Out of interest, does this bad IC of yours identify in Ubuntu in the terminal stats window? never thought to check on a switch which had one of these that was bad, but if it doesn’t ID then that might be a really simple way to help beginners check these two IC’s

So I’m not sure. I can certainly check. Does Ubuntu not require the CCPLEX to boot?

I’m frankly tempted to code a new RCM payload that focuses on hardware debug for testers and fixers…

Sorry I should clarify what I said above. As your quite right, Ubuntu almost certainly wouldn’t boot without CCPLEX.

As I recall seeing stage 2 key derivation failure which was caused elsewhere (there’s at least two other possible causes for this)

If the surrounding components test fine with respect to their resistance measurements relative to ground, then the issue is likely elsewhere and as a result, Ubuntu will still likely boot. Though as you say, if the IC is electrically or physically bad and/or the surrounding measurements aren’t good, then yeah Ubuntu most definitely will not boot.

I had the same thought, an RCM payload would certainly be very useful, but my programming skills were/are seriously lacking and I very slowly backed away after seeing documentation and the related code snippets (hekate)

I’ll run some measurements around the IC tonight when I’m done with the day job and report back. Thanks.

I’ll have a play when I get some time and will aim to get it up on GIT for others to contribute to if I can get something working.

would you mind listing the symptoms for MAX77620A?

So I’ve replaced the MAX77621 tonight and the unit is back up and running again. The new buck fires up as expected on boot unlike the one I removed. There is nothing quite like the feeling of getting a dead board resurrected!


1 Like