New Teensy 4.0 Suddenly Not Recognized

Thundercat said:
Thank you Paul. Yes it is a relatively new laptop (since beginning of this year). How could I check what you are asking?
Click to expand...
By measuring with a scope on the Teensy Vin pin...
Thundercat said:
But wouldn't the regulator I'm using, well, regulate the input? I'm using it literally out of the data sheet example to make sure I am using it correctly...
Click to expand...
I would expect the Mac M3 to have protected USB ports [with respect to short/overloads] on the USB power.
Do you by any change know the powercomsumption of your board?

Paul
 
Other questions that came to mind:
1. were any different components assembled on these latest batch boards? like different supplier/different revision?
2. how much power does the LCD draw from the Teensy 3V3?

Paul
(don't ask me why I ask these questions...)
 
Hope you don't mind me saying, but from the photo I see a lot of residual flux on the soldered Teensy pins.
I have read threads on this forum where residual flux mattered/caused issues.
This is the cleaner I always use: Kontackt Chemie PCC.

Paul
 
PaulS said:
Other questions that came to mind:
1. were any different components assembled on these latest batch boards? like different supplier/different revision?
2. how much power does the LCD draw from the Teensy 3V3?

Paul
(don't ask me why I ask these questions...)
Click to expand...
Thank you Paul.

1. Same supplier. Same parts I’ve used for some time. Good Q.

2. About 100mA I believe.

3. Total Teensy Power consumption:

12 LEDs * ~20mA = 160mA. These are powered via regulator since they take a lot of power. Regulator is 2A.

Plus the LCD which I believe is around 100mA.

So perhaps 260mA, a good portion handled via the regulator.

Plus whatever the Teensy 4.0 uses.

Thank you so much.
 
PaulS said:
Hope you don't mind me saying, but from the photo I see a lot of residual flux on the soldered Teensy pins.
I have read threads on this forum where residual flux mattered/caused issues.
This is the cleaner I always use: Kontackt Chemie PCC.

Paul
Click to expand...
Thank you so much. It’s a no clean flux and I’m nervous to spray the Teensy with anything. But I’m willing to try anything to get this working.

Thank you for the product link and I’ll check it out.
 
PaulS said:
Don't want to spoil the fun, but the regulator [VR1] is a 300mA device.

My last question for today: does your board/device run without LCD? Perhaps it's any idea to check the failing boards without LCD connected?

Good night,
Paul
Click to expand...
Thanks Paul.

Regulator powers LEDs only, and they come in at 240mA or less.

Could that be a problem?

Same regulator works fine on other boards too…

I’m open to all suggestions. I can change it out if need be.

To your question about using the board without the LCD - well I never got far enough to attach it! Nor did I get far enough to attach the LED buttons either; they are separate modules. The units died without being hooked up to the LCD or any LEDs!

I guess that answers the regulator question too, but I can still beef it up if you think that's a good idea, for future runs of the PCB.

Thank you, I really appreciate your help.

Mike
 
Last edited:
I have run out of quick ideas - I guess it's time to do some serious measuring/debugging/testing on the broken boards.

Paul
 
PaulS said:
I have run out of quick ideas - I guess it's time to do some serious measuring/debugging/testing on the broken boards.

Paul
Click to expand...
Yes, in process. First step, to follow on your flux comment, is a thorough clean of the "no clean" flux. I suppose this could the simplest answer, as I did change the flux a few months ago, and that's about when the issues started. I had a couple instances of strange behavior and I remember scratching my head over it. Didn't think about the NC flux. I'll think about it now.

I'll report back findings.

Thanks for your kind support.

Mike
 
Maybe this is stating the overly obvious, but until this problem is fully resolved, please put sockets on any more PCBs you build so you can easily remove the Teensy.
 
PaulStoffregen said:
Maybe this is stating the overly obvious, but until this problem is fully resolved, please put sockets on any more PCBs you build so you can easily remove the Teensy.
Click to expand...
100%!

Previous ones had been soldered on a few weeks ago in advanced prep of fulfilling orders, but new ones going forward, yes!

Thank you for the wisdom. Even the stupidly obvious is helpful at times.
 
I'm just re-read this entire thread, looking for anything I missed and just generally trying to get a clear picture of this difficult situation. I still don't have a clear picture, so I'm hoping you might compose 1 good message to summarize all the available info. Hopefully that can give us a fresh start that leads to ultimately figuring out why you boards are failing.

I see you shared a photo of the PCB in msg #8 and a schematic in msg #17, and answered followup questions on the VR1 part. While better than no info, I want you to understand this still doesn't give a complete enough picture. In particular, the schematic shows 3 connectors. One is to a display, which seems unlikely to be the issue. But the other 2 connect to stuff we can't see. I should also mention your photo has 4 connectors visible, not 3. If Teensy is dying from some issue with the connected circuitry, everything you connect is effectively part of the design whether it's part of the schematic or not. I understand this is probably a commercial product where you're not eager to share all details, but in the interest of helping you look for anything that might be damaging your boards, the more clearly and fully you can show us everything which connects, the better the odds are someone here might notice an issue.

But the ultimate problem might not be your PCB and other circuitry. Very difficult to know. Of all the messages you've written so far, these words in msg #16 really stand out (to my eyes)

I tried a different USB cable, and port, but no joy.

The last one that "died" or stopped responding, I plugged it in, and it was recognized as a USB port. It had already been flash fused locked with all 3 sketches.

It was visible for about 3 minutes, then stopped showing up in the port monitor of Arduino. Subsequent disconnecting/reconnecting did not cause it to show up again, ever.
Click to expand...

My best guess about a situation where a previously thought-to-be-dead USB powered board starts working again, but only briefly, may be a problem with USB power delivery. @PaulS already asked several questions about your USB power, but I want to also echo that concern and emphasize further investigation with some specific tips...

Start by monitoring the USB voltage and current. There are many ways you might do this. If you don't have traditional gear like a scope or good multimeter, at the very least please use an inexpensive device like one of these so you can see the actual (approximate) voltage and current at all times:




USB hosts and hubs are supposed to implement over-current detection. But how specifcally they do this and how they recover from a fault condition isn't required by the USB spec. Even for things the USB specs require, how well implemented they are in any specific hardware varies quite a lot.

Some USB hosts and hubs will temporarily disable a USB port if the connected USB device misbehaves, especially related to drawing too much current. They may detect a problem quickly, or it may take time, which could be a limitation of the hardware or an intentional design decision to allow some types of USB devices to work. One possible theory (or guess) about what's happening could be the USB port on your Mac might be silently shutting itself off.

That's why I'm urging you to connect a voltmeter and ampmeter, or at the very least get a couple of those cheap monitoring products which do it, so you can keep an eye on the voltage and current at all times while trying to diagnose this problem.

You might also try connecting with a couple different powered USB hub. Don't use an unpowered hub... but one (or 2 or 3 for the sake of testing) with power supplies, so the hub provides the power rather than your Mac's USB port. The hub takes over the job of detecting too much current flow, and of recognizing when a USB device is connected and unplugged.

If you're going to make a PCB rev, I'd also highly recommend adding a connector or test points to monitor the 5V input and Teensy's 3.3V power. Or for the PCBs you've already made, just solder some wires. If you don't already have a couple voltmeters, buying even the cheapest models can give you a way to always watch the 5V USB and 3.3V from Teensy. Like monitoring the USB current, while you're troubleshooting I would recommend always having those visible. Maybe nothing anomoluous will appear, but if you do see them change in any unexpected way, that extra info and seeing it happen when you try specific things could be the clue that leads to a breaktrhough in understanding this problem.

But again, this is a lot of guesswork from only a very small number of words and a rather incomplete picture of the hardware really connected to Teensy. I know you're probably not wanting to share so much detail, but if you could better show us ALL of the stuff that's connected (not just the PCB but anywhere an electron could go) maybe one of us might notice something that's been so far overlooked.
 
PaulStoffregen said:
I'm just re-read this entire thread, looking for anything I missed and just generally trying to get a clear picture of this difficult situation. I still don't have a clear picture, so I'm hoping you might compose 1 good message to summarize all the available info. Hopefully that can give us a fresh start that leads to ultimately figuring out why you boards are failing.

I see you shared a photo of the PCB in msg #8 and a schematic in msg #17, and answered followup questions on the VR1 part. While better than no info, I want you to understand this still doesn't give a complete enough picture. In particular, the schematic shows 3 connectors. One is to a display, which seems unlikely to be the issue. But the other 2 connect to stuff we can't see. I should also mention your photo has 4 connectors visible, not 3. If Teensy is dying from some issue with the connected circuitry, everything you connect is effectively part of the design whether it's part of the schematic or not. I understand this is probably a commercial product where you're not eager to share all details, but in the interest of helping you look for anything that might be damaging your boards, the more clearly and fully you can show us everything which connects, the better the odds are someone here might notice an issue.

But the ultimate problem might not be your PCB and other circuitry. Very difficult to know. Of all the messages you've written so far, these words in msg #16 really stand out (to my eyes)



My best guess about a situation where a previously thought-to-be-dead USB powered board starts working again, but only briefly, may be a problem with USB power delivery. @PaulS already asked several questions about your USB power, but I want to also echo that concern and emphasize further investigation with some specific tips...

Start by monitoring the USB voltage and current. There are many ways you might do this. If you don't have traditional gear like a scope or good multimeter, at the very least please use an inexpensive device like one of these so you can see the actual (approximate) voltage and current at all times:




USB hosts and hubs are supposed to implement over-current detection. But how specifcally they do this and how they recover from a fault condition isn't required by the USB spec. Even for things the USB specs require, how well implemented they are in any specific hardware varies quite a lot.

Some USB hosts and hubs will temporarily disable a USB port if the connected USB device misbehaves, especially related to drawing too much current. They may detect a problem quickly, or it may take time, which could be a limitation of the hardware or an intentional design decision to allow some types of USB devices to work. One possible theory (or guess) about what's happening could be the USB port on your Mac might be silently shutting itself off.

That's why I'm urging you to connect a voltmeter and ampmeter, or at the very least get a couple of those cheap monitoring products which do it, so you can keep an eye on the voltage and current at all times while trying to diagnose this problem.

You might also try connecting with a couple different powered USB hub. Don't use an unpowered hub... but one (or 2 or 3 for the sake of testing) with power supplies, so the hub provides the power rather than your Mac's USB port. The hub takes over the job of detecting too much current flow, and of recognizing when a USB device is connected and unplugged.

If you're going to make a PCB rev, I'd also highly recommend adding a connector or test points to monitor the 5V input and Teensy's 3.3V power. Or for the PCBs you've already made, just solder some wires. If you don't already have a couple voltmeters, buying even the cheapest models can give you a way to always watch the 5V USB and 3.3V from Teensy. Like monitoring the USB current, while you're troubleshooting I would recommend always having those visible. Maybe nothing anomoluous will appear, but if you do see them change in any unexpected way, that extra info and seeing it happen when you try specific things could be the clue that leads to a breaktrhough in understanding this problem.

But again, this is a lot of guesswork from only a very small number of words and a rather incomplete picture of the hardware really connected to Teensy. I know you're probably not wanting to share so much detail, but if you could better show us ALL of the stuff that's connected (not just the PCB but anywhere an electron could go) maybe one of us might notice something that's been so far overlooked.
Click to expand...
Paul, I humbly appreciate your incredibly in-depth answer, and all the time it took you to share. I can't tell you how much I appreciate you and the folks here. Do you ever sleep?

I will follow up with your suggestions and report back. I know I/we will get to the bottom of this. It's probably something really stupid, like the change in solder flux and it's not really pure as I thought - I've been loathe to spray the teensy with anything for fear of damaging the programming button or ruining it somehow; otherwise I clean my PCBs really well. I will be able to test further in the next day or two and I will post back. At this point, there is no product, if it cannot be made to work. So I don't mind sharing more.

Pix below. Only thing missing is a program tactile switch.

Thank you again everyone, I will report back with more details and findings.

More pix:

[photos deleted]
 
Last edited by a moderator:
15 sec Restore can work from a USB Power Bank - it does not rely on Host Communications.
> this will not work on a Locked Teensy as the stored factory code is not acceptably .eHex encoded.
> IIRC at least one went bad before final LOCK'ing?

When plugged into the power bank the 5V and 3.3V should appear as normal?

Press and Hold the Button for the 13s-17s interval watching the bootloader RED LED Flash to release the Button ASAP.

On an unlocked Teensy try more than once if the first attempt fails.
Perhaps Holding the Button during initial Plugin and starting the clock then.
 
Just to repeat once more, while trying the 15 sec restore or anything else, please get into a practice of continuously monitoring the 3.3V voltage, VUSB voltage, and USB current.

Also remember this restore process only works on not-yet-locked hardware. It does work if you've run the 1st Fuse Write program, but does not work if you've run the 3rd Lock Security program.
 
defragster said:
15 sec Restore can work from a USB Power Bank - it does not rely on Host Communications.
> this will not work on a Locked Teensy as the stored factory code is not acceptably .eHex encoded.
> IIRC at least one went bad before final LOCK'ing?

When plugged into the power bank the 5V and 3.3V should appear as normal?

Press and Hold the Button for the 13s-17s interval watching the bootloader RED LED Flash to release the Button ASAP.

On an unlocked Teensy try more than once if the first attempt fails.
Perhaps Holding the Button during initial Plugin and starting the clock the
Click to expand...
I cleaned the boards really well to remove any suspicious "no clean" flux from them. @defragster, I did your suggestion of first holding the button in prior to inserting the USB cable, and then counting about 16 seconds. No flashing red light whatsoever. But I released and waited.

On two units, they became recognized again after about a minute! Cautiously, they look to be normal again, but I haven't hooked up anything else like the buttons or the LCD or anything. Two others are dead as a doornail however.

Before setting off the fireworks, I have to say I've had them come in and out of operation, so I will further test to see if they are really working.

I will follow up in more detail in the coming days, as you have all generously helped me and deserve to know the outcome. I am hoping it's just the "no clean" flux was actually conductive and fried the boards in question - that would be the best outcome because it's repeatable, and avoidable, and would be the resolution. And it explains all the weirdness neatly and perfectly. So let's hope that's what caused this.

But I'm not sure yet. I'll keep you all posted, and thank you all so much for bearing with me through this. I'll post again this week with updates.

Oh and of course, the real test will be to install a new teensy in a new board, making sure no flux remains, and using a socket, and seeing if it works as expected. That will tell the tale I think.

Thank you all again.

All the best,

Mike
 
Last edited:
Wonderful thank you Paul. No I didn't monitor them yet. I have a nice Fluke meter; I'll take some readings and report back.

Many thanks!
 
Hiya, just an update.

One of the previously "dead" units has now successfully flashed, making 3 flashed units out of the 5 that went "bad." One of them is beyond salvation as I desoldered it with a Hakko desoldering tool, but it didn't survive the process. Too many pins, and pads lifted.

So out of the original 5, there's 4 extant, and 3 "working" after a fashion.

However, the "working" ones all behave erratically the same way - not processing button inputs normally. If I didn't know better I'd say it was a software issue, as they all behave exactly the same - recognizing button clicks but not button holds. But the other "good" working units have exactly the same firmware and the buttons behave correctly - I even removed the button from a working unit and tried it in the "bad" units - same issue. So the hardware buttons are fine, and the software is fine. I literally flashed a "good" unit with the same firmware I flashed these units with, one after the other. The "good" units behave normally, and these don't. So it ain't the software. I spent months by the way getting click, double click, click and hold etc working, so this is hardened code.

Which suggests they all got fried in the same way, or some other mystery.

The fourth unit, after holding the button prior to insertion of the USB cable, and then continuing to hold it for 16 seconds, started flashing the orange light - not the blink program; much faster.

But that's all it does. It is not recognized in the serial port, whereas the other 3 are.

I've got some heavy duty flux remover coming in a few days, and I'll spray the heck out of all of them and see if they come back.

I'm just a little bit worried that this is not actually what's going on, so when I get more time next week I'm going to assemble a unit with a new Teensy in a socket, in one of the existing main boards (I have no more of the main boards at this point, so I'll have to use one of the ones from this problem situation. But, again, these main boards, same batch, have worked reliably in other units, so I don't necessarily suspect them). In any case we'll get more data points to this mystery.

Of course I will flash and test the Teensy prior to insertion.

This is looking more and more like dirty boards due to low quality "no clean" flux. I'd love that to be the issue; so easily solvable. But I will continue to update this thread with findings.

Thank you again for your patience everyone.

Mike
 
Last edited:
Thundercat said:
I've got some heavy duty flux remover coming in a few days, and I'll spray the heck out of all of them and see if they come back.
Click to expand...
Yes, please do. It may well be that by rinsing the pins with flux remover, some residual flux has entered under the BGA. Perhaps some compressed air will help?

It begs the question though why you used flux in the first place. Did you feel that to be necessary? Personnaly I have never used additional flux when soldering pins to a Teensy - no problem with solder flow whatsoever.
Although the solder I'm using (Ersin Multicore 362) has a tiny amount of flux in it: <0.5% halide, non-corrosive flux.

Anyway, I'm cautiously optimistic after your testing up till now.

Paul
 
PaulS said:
Yes, please do. It may well be that by rinsing the pins with flux remover, some residual flux has entered under the BGA. Perhaps some compressed air will help?

It begs the question though why you used flux in the first place. Did you feel that to be necessary? Personnaly I have never used additional flux when soldering pins to a Teensy - no problem with solder flow whatsoever.
Although the solder I'm using (Ersin Multicore 362) has a tiny amount of flux in it: <0.5% halide, non-corrosive flux.

Anyway, I'm cautiously optimistic after your testing up till now.

Paul
Click to expand...
Thank you Paul.

I never soldered without flux. When I've tried briefly, it never seems to make a good joint. I have a nice Metcal station and good quality tips, but seems to require some flux. But I'll take your advice under consideration. Maybe I can do the top soldering onto the Teensy without it. But there will still be rosin or flux in the solder anyway. I'm using leaded Kester 63/37 since it's eutectic and seems to make good quality, forgiving joints.

Another data point: I have made other units with this exact same no-clean flux, without issue. Same procedure, same lack of cleaning the Teensy top pins. That's why this has come up as such a huge puzzler to me. I never connected the current issue with it. And still, maybe it's not that. But I hope it is.

In any case, again, I'll report back findings.

Thank you all again,

Mike
 
@Thundercat - Awesome having those Teensy units return to some degree of function - so not burned up or totally destroyed. @KurtE recently reposted the Pin_Test we worked out that allows seeing each pin tapped to GND or 3.3V in turn to verify proper digital function. Not an exhaustive sketch but will give some baseline.

Never trained in soldering here - have gone to finer gauge flux core and had good luck on pins. Heating the join area is fun but touching the solder to iron gets liquid solder ( and expresses flux ) for better heat transfer and liquidity so then can feed the join area (not the iron) - with no added flux.

Have solder paste used for rework/touchups (usually on QSPI pad items) to get heat transfer and liquidity and cleaner joints without cooling solder strings following the iron. Of course solder paste is handy for those too. But that all leaves the PCB slimed.

Had a QSPI Flash chip go not working (it worked before sitting on shelf) - (a couple repeats) [clean/reflow/clean] failed to restore function it had before sitting. Got chip off - maybe with new hotair unit - and there was a solid wet flux area under the Flash chip. So 91% alcohol cleaning was not done well enough - and the rework flux just pulled under. Not sure why that went bad - many other QSPI PSRAM and Flash soldered and cleaned to work.
 
By the way, I can't seem to edit previous posts. Since the group has seen the pix and schematics, I'd like to remove them as they are sensitive info for me. But I can't seem to find a button to edit my posts - I swear I've been able to edit previous posts before!

Can anyone help?

Thanks,

Mike
PaulStoffregen said:
As requested, I've deleted the schematic and photos.

Did you watch the voltage and current while working with the unresponsive hardware? Make any progress?
Click to expand...
Hi Paul, thanks for the assistance with the posts.

I’m getting some flux remover spay in a couple days off eBay. I thought it would be prudent to do a deep clean before further tests.

I will post results here in a few days after a thorough clean.

I really appreciate all the support, more than I can say.

I’m still very nervous this won’t solve the issue, but as they say, one step at a time.

Also I’ve got some much higher quality flux for future endeavors, although I will try soldering top Teensy pins without flux in future.

Thanks again; I totally appreciate everyone’s time.

Mike
 
One further suggestion for your "new Teensy old carrier PCBA" test unit, is not to lock the Teensy. You then at least have the option of being able to do the 15 second restore, which I understand you can't do with a locked device. Given you're going to socket the Teensy, if it fails after a burn-in period, and can only be restored if unplugged, then that helps to narrow down the places to look.

There was a longish thread to do with a "startup bug", resolved in 1.59 but it seemed only to occur in limited circumstances - maybe it or something similar can still happen, with slightly different circumstances? I certainly encountered it in the Audio+TeensyDebug form, and the 15s restore appeared to be the only way out at the time.

EDIT: obviously, this is not a solution for production units - you want to lock those!
 
You must log in or register to reply here.
Back
Top