Teensy 4.1 dead?

T

tcurdt

Member
It seems like my Teensy just died and TBH I have no idea why.
It was working just fine for a while.

IMG_4035.jpeg


Either way:
Plugging it into USB, there is no sign of a LED on the Teensy at all.
Holding the program button down for more than 20s does nothing.

Is there any documentation (that I didn't find yet) on what to test on the board?
Based on this thread the first step is to check the regulators.


I only see one - shouldn't there be two? 5V and 3.3V?

teensy41_4b.jpg


The 5V rail is fine. The 3.3V is only at 0.3V.


I powered on a motor - but that was on a different power rail of the PSU.
And any connections to the servo drivers should be optically isolated and the cables are shielded.
Could EMI cause this? I don't really see how a power spike could reach the Teensy.

Any guidance would be appreciated.
Thanks!
 
I can't really tell from your photo but is it possible to remove the Teensy 4.1 from the carrierboard?
Yeah, the 3.3V pin showing 0.3V is not good. Perhaps something on your board is shorting the 3V3 line?
Is regulator U4 running hot? U4 is the chip on the right of your yellow circle above.

Paul
 
PaulS said:
I can't really tell from your photo but is it possible to remove the Teensy 4.1 from the carrierboard?
Yeah, the 3.3V pin showing 0.3V is not good. Perhaps something on your board is shorting the 3V3 line?
Is regulator U4 running hot? U4 is the chip on the right of your yellow circle above.

Paul
Click to expand...
Thanks, Paul.

Unfortunately I acted against my better instincts and followed the instructions to solder it on.
I guess the only way to really tell what is going on is now to remove the Teensy.
With that many pins it looks like a painful evening with a slim chance of success.

I don't have a thermal camera, but measuring the temp I didn't get anything higher than 28C.
The LED being the hottest.

I am just puzzled where that sudden short could be coming from.
Maybe something is dead on the carrier board. Maybe an opto coupler on the input side.

I am still wondering if EMI could be the cause.
Because if so, I need to do some re-thinking.
 
Before desoldering anything, check out "Power Up Sequence" on the T4 bootloader chip page. It has important info you should understand about the power up process.

Try to check the whether the power on request (step #5) is sending at least 1 volt to the enable pin on the regulator. If so, then you might try removing the regulator. Before soldering a new part, probably a good first step would be to apply 3.3V from a current limited power supply. Set the limit around 150 to 200 mA. If the main chip is requesting 3.3V and actually applying 3.3V causes your board to work, then you can know that adding a regulator will solve the problem.

You might however discover bad news. Hopefully this isn't the case, but sometimes when these sorts of problems happen the end result is 1 or more chips powered by 3.3V get damaged internally. That's why you should test with a current limited power supply. If the board draws too much, it's probably a sign something else is damaged and replacing on the regulator won't be enough.
 
Try to check the whether the power on request (step #5) is sending at least 1 volt to the enable pin on the regulator.
Click to expand...
Puh - the U4 is really quite tiny. Measuring almost seems harder than replacing it (assuming the hot air soldering works out).

The test point seems to be on the back - which I cannot reach without a desolder of the Teensy.
I am not sure I have the equipment to measure PMIC_ON_REQ on the front.

Or is there another spot I could test that I am missing?

parts_placement_teensy41_top.png

test_points_teensy41.png


I am still wondering how a EMI/over-voltage spike could have reached the board.

github.com

grblhal_2000_PrintNC/PrintNC_GRBLHAL_Breakout_Schematic_A8_FULL_ASSEMBLY.PDF at master · Expatria-Technologies/grblhal_2000_PrintNC

Expatria Technologies GRBLHAL Breakout board for PrintNC - Expatria-Technologies/grblhal_2000_PrintNC
github.com github.com

Maybe the problem is with the carrier board.
Then it might not even be worth trying to fix it.
 
These 5 to 3V3 LDOs fail when:
1) Vin is >5.5V, or,
2) Vout is driven while Vin is lower than Vout (reverse current scenario), and current out of Vin can sink to GND.

You may have 5V logic outputs feeding into Teensy inputs. With that you may have been pushing >3.3V onto the 3V3 power rail. Doing so while Teensy is in sleep mode (or pre-boot) and LDO is not (yet) enabled, is deadly for the LDO.

A protection diode from Teensy 3V3 rail back to 5V does help.

A 5V relay on the 5V rail may also get you >5V spikes on the 5V rail when relay switches off.

Your 24->5V power supply may end up fighting with 5V from USB when that’s plugged in?
 
sicco said:
These 5 to 3V3 LDOs fail when:
1) Vin is >5.5V, or,
2) Vout is driven while Vin is lower than Vout (reverse current scenario), and current out of Vin can sink to GND.

You may have 5V logic outputs feeding into Teensy inputs. With that you may have been pushing >3.3V onto the 3V3 power rail. Doing so while Teensy is in sleep mode (or pre-boot) and LDO is not (yet) enabled, is deadly for the LDO.
Click to expand...

Hm. I don't yet see how that could have happened in my case. (see below)

sicco said:
A protection diode from Teensy 3V3 rail back to 5V does help.
Click to expand...
back? Just like this?


teensy41_4.jpg

sicco said:
A 5V relay on the 5V rail may also get you >5V spikes on the 5V rail when relay switches off.

Your 24->5V power supply may end up fighting with 5V from USB when that’s plugged in?
Click to expand...

I have a (big! 1200W) cisco (server) PSU that has 3 rails. (Way overkill)
42V (for the servo motors), 12V (powering the carrier board) and 5V (unused)

So far I don't see where the 5V spike could have occurred.

The board was running when I turned on one of the motors.
This could have causes a spike on the 42V. Maybe also on the 12V? Not sure.

It could have caused an EMI spike on the input/output like that are somewhat parallel to the servo power lines.

But all the inputs/outputs should be opto isolated.

Hence me being puzzled.
 
Yes, diode like that. Schottky.

Is your T4 USB 5V still wired through to your 5V onboard regulator (TPS54331DR, easily pushes out 5V at multiple amps...)? That's potentially damaging the USB hub / PC where 5V USB normally comes from (if one source says 5.01V and the other says 4.99 is what it shall be, then...).

On where 5V may trickle through to Teensy I/O pins:

Possibly here:
1704470963126.png



And here maybe ( (8k2 / (8k2+2k)*5V = 4V... = too high )
1704471094560.png



And here, that's far-fetched and 51k should protect enough (5V - V_fwd_Led - 0.6V) Maybe emitter and collector swapped here?:

1704470564731.png
 
tcurdt said:
Alright. EN measures 1.0-1.1V.
So if I can source a NCV8186 somewhere it might be worth a try.
Click to expand...

Yes, that's a good sign. It means the power management in the main chip is still working and requesting the 3.3V power to turn on.

Next step is to desolder the dead regulator and after you see the EN signal, just (very carefully) apply external 3.3V power. Best to use a lab bench current limited power supply, so you can set a low limit and see if the whole board tries to draw too much current when you apply 3.3V power.

If it works with external 3.3V power, then order a replacement regulator chip. Looks like Mouser and Digikey probably have them in stock. But if you can't find it (or anyone much later finds this conversation by search) I probably have some left over here. PJRC used the NCV8186 part only during the 2022 chip shortages when the TI parts were unavailable.
 
I'm not yet convinced that the regulator is the culpritt.
Before desoldering the regulator from the board, I would try to cut/desolder both 3V3 pins from Teensy to carrierboard.

Paul
 
sicco said:
Yes, diode like that. Schottky.

Is your T4 USB 5V still wired through to your 5V onboard regulator (TPS54331DR, easily pushes out 5V at multiple amps...)? That's potentially damaging the USB hub / PC where 5V USB normally comes from (if one source says 5.01V and the other says 4.99 is what it shall be, then...).

On where 5V may trickle through to Teensy I/O pins:
Click to expand...

I didn't cut the connection but the USB it's only ever getting power from one - USB or the carrier board.
In fact I only used the USB to flash and never again since then.


sicco said:
Possibly here:
View attachment 32819
Click to expand...

This one isn't connected. So I think it's safe to rule out.


sicco said:
And here maybe ( (8k2 / (8k2+2k)*5V = 4V... = too high )
View attachment 32821
Click to expand...

Now - that is an intriguing one!
This looks like it's the RS484 connection to the VFD - which has cause much grief before.

A previously working VFD stopped talking via RS484 (I suspect the RS484 side of the VFD is dead now).

I got a new one - which started to transmit RPMs wrongly (sometimes, by a constant factor, weird).
Which still is a mystery as there should be a checksum over Modbus RTU.

Now you make me wonder if this could be related.

sicco said:
And here, that's far-fetched and 51k should protect enough (5V - V_fwd_Led - 0.6V) Maybe emitter and collector swapped here?:

View attachment 32818
Click to expand...

Those are the RS484 LEDs. TBH I am not entirely sure. Worth checking!
 
PaulS said:
I'm not yet convinced that the regulator is the culpritt.
Before desoldering the regulator from the board, I would try to cut/desolder both 3V3 pins from Teensy to carrierboard.

Paul
Click to expand...

Puh, that sounds like a tricky one.
Cutting is one one thing - repairing that afterward another.
Any practical soldering advice here? How would you approach that?
 
PaulStoffregen said:
Yes, that's a good sign. It means the power management in the main chip is still working and requesting the 3.3V power to turn on.
Click to expand...

❤️
PaulStoffregen said:
Next step is to desolder the dead regulator and after you see the EN signal, just (very carefully) apply external 3.3V power. Best to use a lab bench current limited power supply, so you can set a low limit and see if the whole board tries to draw too much current when you apply 3.3V power.
Click to expand...

I got a lab bench PSU. As you said I would start with 150mA.

PaulStoffregen said:
If it works with external 3.3V power, then order a replacement regulator chip. Looks like Mouser and Digikey probably have them in stock. But if you can't find it (or anyone much later finds this conversation by search) I probably have some left over here. PJRC used the NCV8186 part only during the 2022 chip shortages when the TI parts were unavailable.
Click to expand...

The problem with digikey and mouser are always the shipping costs for small orders.
Paying 20 USD + 0.70 USD suddenly makes buying a new Teensy more suitable.
I am just not sure how to unsolder a part with that many pins without loosing my mind.

And I haven't found the NCV8166 somewhere else yet.

I am not sure how much shipping to the EU would be if you send it - but I very much appreciate the offer.
 
I can't help with the shipping costs. For anywhere outside the USA, PJRC ends up paying that much or more just for postal air mail.

But if you do buy another Teensy 4.1 direct from PJRC, I'd be happy to include a NCV8186 chip for free. If replacing the chip works, you'll have a spare Teensy 4.1 to use for another project, and if it doesn't work out at least you'll have the Teensy to build another board.
 
tcurdt said:
How would you approach that?
Click to expand...
Assuming you used 4-5mm high headers to mount the Teensy, I would (and actually have done) melt the plastic away with a soldering iron and then use small side-cutting pliers to cut the pin. The 2 pin halves are then easily desoldered from Teensy and carrierboard.
For repairing the connection, just insert a bare header-pin and solder it on both boards.

Hope this helps,
Paul
 
PaulS said:
Assuming you used 4-5mm high headers to mount the Teensy, I would (and actually have done) melt the plastic away with a soldering iron and then use small side-cutting pliers to cut the pin. The 2 pin halves are then easily desoldered from Teensy and carrierboard.
For repairing the connection, just insert a bare header-pin and solder it on both boards.
Click to expand...

sounds good ... but I fear the ethernet jack will be in the way 😔

Screenshot 2024-01-05 at 22.26.17.png
 
PaulStoffregen said:
I can't help with the shipping costs. For anywhere outside the USA, PJRC ends up paying that much or more just for postal air mail.

But if you do buy another Teensy 4.1 direct from PJRC, I'd be happy to include a NCV8186 chip for free. If replacing the chip works, you'll have a spare Teensy 4.1 to use for another project, and if it doesn't work out at least you'll have the Teensy to build another board.
Click to expand...
❤️

Shipping costs are just crazy internationally. Thanks for the offer.
 
tcurdt said:
sounds good ... but I fear the ethernet jack will be in the way 😔
Click to expand...
Another method I tried succesfully a few times: using a powerful soldering iron, heat the pin from one side (in your case the carrierboard side) until the solder on the Teensy side also starts to melt, then push the floating pin to the carrierboard side. Keep on heating (adding enough solder to ensure good heat transfer) and use tweezers to pull the pin out from the carrierboard side.

Paul
 
Alright. Regulator arrived.

Since removing the Teensy wasn't really an option, I tried to unsolder the 3V pin an press it through while hot - but had to give up.

Then I tried to unsolder the regulator - but it refuses to budge.
I tried air and regular (which both work fantastic on other SMD parts).
I had the air on 450C and regular on 360C.

Is it using some high temp solder or glued on or something? It's still solid as a rock. I don't get it.

Though I ask before ruin this thing completely.
 
You must log in or register to reply here.
Back
Top