New Teensy 4.0 Suddenly Not Recognized

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!
Hi h4yn0nnym0u5e,

That is very helpful. I really appreciate you sharing your experience, and the link. I'm blown away at how awesome you all are, and how great the forum is. Teensy is much more than a product - it's a legend, with caring people behind it.

So cool!

Thanks!

Mike
 
Hi, just an update - I thoroughly cleaned the boards with flux remover, scrubbed them, and let them dry well. They couldn't be more clean.

Behavior is unchanged. 3 boards fire up and run the sketch, but do not work correctly with attached button.

I took same working button from working unit and tried it in the 3 units - so it's not the button hardware, or the cable. I think the 3 units are just fried in some way. But they sort of work. Not OK for a product though, obvs.

The 4th board flashes but behaves strangely, and is effectively defunct.

5th board, I had desoldered and unfortunately ruined in the desoldering process.

Next step is to put a brand new Teensy 4.0 into a socket on one of the existing carrier boards.

I will do that in the next few days or week and post results.

As I said, I HAVE to solve this, because it goes into a product I make. Apologies for the slowness of my replies; I work another job and have been a bit busy.

Thank you all again for your tremendous support. Let's hope a new teensy with little to no high quality no-clean flux in soldering, works.
 
For these latest tests, did you monitor the 3.3V power with your multimeter?

Do you have any way to monitor the USB current? If not, please buy (and use) one of these inexpensive USB power monitoring products.

usbtester.jpg
 
For these latest tests, did you monitor the 3.3V power with your multimeter?

Do you have any way to monitor the USB current? If not, please buy (and use) one of these inexpensive USB power monitoring products.

View attachment 34333
I will buy one and use it.

I do have a multimeter; I guess I thought the point was moot if the board wasn't working right, since from same USB on brand new M3 MacBook Pro, other units work fine.

I will check 3.3V power on one of the defunct boards and report back. I'm assuming you just mean 3.3V to GND on the top of the Teensy? Or did you mean to check the computer USB power? In that case I will need to wait until I get the item you showed above, which I'll order today.

That product looks really neat; thank you for posting it.

And thank you again kindly for the amazing support.

EDIT: unit ordered. Will arrive in a few days. In the meantime, I'm going to test a carrier board with a new Teensy that is known to work. I will also check the Teensy voltage with a multimeter. Thank you again.

EDIT 2: Tested the new Teensy with multimeter installed in previous carrier board. 3.294V between 3.3v pin to GND, and 5.237 between VIN and GND. "Blink" has been running for a few mins without issue.
 
Last edited:
For these latest tests, did you monitor the 3.3V power with your multimeter?

Do you have any way to monitor the USB current? If not, please buy (and use) one of these inexpensive USB power monitoring products.
That looks like a nice device - with an important but rare to use feature: It will also show when current is flowing INTO USB Host.

That can make the Teensy act funny based on a T_3.x issue seen here with externally powered display matrix that wasn't grounded 'enough' to external power so the Teensy was carrying the current to the Host and that was making AUTO uploads unreliable. Don't recall how that was discovered (without this device) - but adding another GND wire from Matrix to the powered base board ribbon connector made it right.
 
I can cautiously say it looks like the issue has been in fact as suspected: unclean, "no clean" flux.

The previous units that I reported as "working," but not working with buttons as expected - this was user error. I didn't think about the fact that since faders were not connected, there were floating pins, which of course threw off button behavior. After thorough cleaning and connecting faders to floating pins, I have two completely working units now, that were previously defunct, and a third brand new one - a new Teensy installed in an previous carrier main board, working.

I suspect I can reinstall two more removed Teensys and at least one may work fine; I will try.

One carrier board and Teensy appear unsalvalgable and behave strangely no matter how clean I get them.

Also, in other news, I am now getting failures in the field, customers returning units behaving bizarrely, the strange behavior showing up after a few hours of use. Of course they were thoroughly tested before I shipped them.

Dirty boards is the only explanation I can come up with, as the designs are tested and I have units working for years without issue. Software is well tested and again, working for years without issue.

Why some boards have worked and tested fine, then failed in the field, I cannot say - humidity? Heat from the unit spreading the flux further? I don't know

This has been a very terrible and expensive lesson in flux. I have processed and used at least 100 Teensys in a similar manner, leaving the "no clean" flux on the board and I've had no problems at all - until now. This was a new flux, a "freebie" I got with new 63/37, and I'll never do that again. On closer inspection, it didn't actually say "no clean." It said everything else - neutral PH 7, non-corrosion, etc:

IMG_0509.jpeg

I am also leery of using the solder from the same company now, since it contains the same flux:

IMG_0510.jpeg

I'd rather bin it than have more failures, as this is getting expensive and humiliating.

I am not 100% feeling good about all this yet, but after some thorough testing I will post more...It's been incredibly stressful as I sell units to customers and now some are failing. But at least I think I know why, and how to fix the issue.

Thank you again everyone. I appreciate the immense support. I'm sorry to bother you all with such a stupid final problem as dirty boards. What a joke. I just didn't know any better. Until now.

I will post further again after thoroughly testing the units, as well after receiving the USB sniffer Paul recommended. I am hopeful that this is, indeed, the end of the issue.

All the best,

Mike

EDIT: after all this, I thought you might like to see one of the products I make with the Teensys:


IMG_0511.jpeg
 
Hi Mike, glad to hear that things are looking better. I hope the amount of field returns does not increase.
About the solder you are using: I wonder if I can talk you into trying a low flux content solder? You stated a few messages back that you briefly tried without success. Perhaps you should try it another time? Please buy solder from reputable brands. Bohools is an unknown ShenZhen brand.
This is the solder I use for more then 10 years now: https://www.rapidonline.com/multicore-loctite-288322-362-60en-5c-solder-wire-0-7mm-0-5kg-85-6215. I never clean the parts after soldering...

Paul
 
Thank you so much for the recommendation Paul. I really appreciate it.

I haven't actually used that Bohool solder yet - only the flux - I've been using Kester 63/37 the whole time.

I will check out what you shared. I have 3 lbs of the Kester at the moment - Kester should be OK though, right?

Thank you again,

Mike
 
I'm trying to find the datasheet for that Kester solder but there are many variants. Kester partnumbers look like 24–6337-0010. What is the exact part# you are using?
Just to be sure that I understand you fully: you apply a separate NC flux before soldering with the Kester solder?

Paul
 
I'm trying to find the datasheet for that Kester solder but there are many variants. Kester partnumbers look like 24–6337-0010. What is the exact part# you are using?
Just to be sure that I understand you fully: you apply a separate NC flux before soldering with the Kester solder?

Paul
Hi Paul, yes, that's exactly right. I have tried soldering without any additional flux, but it just doesn't seem to flow well, and I don't get good joints. And I heat both the component lead and the pad simultaneously...

Here's a pic of the Kester I've been using the last 3 years (1lb roll lasts a LOOOONG TIME!):


IMG_0514.jpeg
 
That's interesting: that solder has 3.3% rosin flux. So it should flow nicely. Is the temperature of your Metcal station set high enough?

Paul
 
If you're able to water wash and bake dry your circuit board, using Kester 331 solder and Kester 2331-ZX liquid flux gives amazing results. I personally use that combo for PCB prototypes. Organic acid flux like Kester 2331-ZX is amazing stuff which makes hand soldering so much easier with consistently excellent results. But it absolutely is corrosive and must be washed off within a few hours. I've heard many success stories for different ways to dry the water, but I prefer a low temperature oven because I just don't have the patience for ways that take hours or days.

I also use Kester 44 (same as msg #60) but usually only when I need to solder wires or add parts or make repairs that can't be water washed.
 
Last edited:
That's interesting: that solder has 3.3% rosin flux. So it should flow nicely. Is the temperature of your Metcal station set high enough?

Paul
It doesn't have temperature control...it's one of the older ones, and uses some kind of induction heating (I forget how it works). I can say, with a large tip, it gets very hot, and with a smaller one, hot enough...


IMG_0516.jpeg

IMG_0515.jpeg

Thanks so much for your expertise!

Mike
 
If you're able to water wash and bake dry your circuit board, using Kester 331 solder and Kester 2331-ZX liquid flux gives amazing results. I personally use that combo for PCB prototypes. Organic acid flux like Kester 2331-ZX is absolutely amazing stuff. But it absolutely is corrosive and must be washed off within a few hours. I've heard many success stories for different ways to dry the water, but I prefer a low temperature oven because I just don't have the patience for ways that take hours or days.

I also use Kester 44 (same as msg #60) but usually only when I need to solder wires or add parts or make repairs that can't be water washed.
Oh! You just helped me figure out what's gone on with those boards EXACTLY!

The flux was corroding over several hours/days! OMG that makes so much sense! Well...wow...that's also why they were working fine before I sent them out to customers, and also why they went bad after a short period of time...there was nothing wrong with them initially, until the flux corroded them. That's my hypothesis, anyway, which I think is correct.

Geez I feel such a neophyte with all this. I never claimed to be an expert, but I didn't think I was THAT green!

Clearly I am.

Also, I appreciate you sharing exactly what you use too. What do you use solder-wise if it's not a prototype? Or do you always use the above combos?

Thank you kindly Paul!

Mike
 
I'm using a heavy duty step up/down transformer...it does 230-115 and vice versa, although right now in the photo the wall wart is covering the 115V plugs, which are hidden below it.

I'm an American but I live in the UK :)


IMG_0517.jpeg
 
Last edited:
It doesn't have temperature control...it's one of the older ones, and uses some kind of induction heating (I forget how it works). I can say, with a large tip, it gets very hot, and with a smaller one, hot enough...


View attachment 34374

View attachment 34375

Thanks so much for your expertise!

Mike
If your METCAL is like the one we used to use in our development lab (when I was working . . . now retired), the soldering temperature was determined by the specific tip that was plugged in (IIRC, there were 600 degrees F tips & 700 degrees F tips that we used). You may want to verify the exact specs on the tip that you are using.

Hope that helps . . .

Mark J Culross
KD5RXT
 
If your METCAL is like the one we used to use in our development lab (when I was working . . . now retired), the soldering temperature was determined by the specific tip that was plugged in (IIRC, there were 600 degrees F tips & 700 degrees F tips that we used). You may want to verify the exact specs on the tip that you are using.

Hope that helps . . .

Mark J Culross
KD5RXT
Yes that does help! That's fantastic information!

Thank you!
 
Hi, I just wanted to follow up with this thread. I bought the USB sniffer Paul suggested and used it.

For the one board that never revived, it shows power consumption of 0.497A with nothing hooked up besides the Teensy. No LEDs, no screen, no faders, nothing.

So I think this tells the tale: there's some kind of short or something fried on the Teensy due to the dirty flux. A Teensy should never use 500mA at idle - I think more like less than 100mA.

Another units that was defunct, that revived after thorough cleaning with a flux cleaner, shows power consumption of ~ 0.120A, WITH the 12 LEDs hooked up AND an LCD screen:

IMG_0560.jpeg

Incidentally, LEDs are run with PWM, so they are not at full brightness most of the time; hence the lower than expected total power consumption.

@Paul, are there other measurements you feel would be good to provide? Can we consider this issue resolved? From my standpoint, it is.

Thank you all again. I can't tell you how much I appreciate the incredible wisdom and support you all have provided. I wouldn't have been able to sus this out without you!

Especially thank you to @PaulS, who pointed out the flux issue that led to the solution :)

Mike
 
Back
Top