Teensy 4.1 flashing, very hot and non responsive (cannot be detected by computer)

General:
Answers to your questions:
D1 is a Transzorb and has a voltage rating similar to a Zener diode, except it is super fast, will clamp any over voltage in about 1 nanosecond.
L1 is an inductor, which is part of the Pi filter (Or low pass filter) that comprises C1, L1, and C2. It has to have a current rating to handle all the
current of the teensy.
J3 is the input power from whatever the source of power is to the Teensy.
In this schematic the power flows from left to right.
Hope that helps,
Ed
 
Thank you SOOO MUCH @sbfreddie. can you please outline what those components are specifically so I do not accidentally mix anything up and so I can use the exact components.

main questions are:
  • what type of diode is D1? would an ultra fast or normal diode work in its place?
  • What component is L1
  • Is J3 representing the teensy or the power source to the 5V line?
(apologies as I am not well versed in electronic symbols)
Also I cannot find the TSV diode or the 200mH inhibitor (does the inhibitor have to be 200mH because I cannot find any that big, I can only find maximum 6200uH RF Inductor) anywhere in store near me. Is there any substitute I can use for these parts such as normal diodes or lower inhibitors? however I will DEFIANTLY include this exact circuit in our PCB design thank you very much
 
Last edited:
Also I cannot find the TSV diode or the 200mH inhibitor (does the inhibitor have to be 200mH because I cannot find any that big, I can only find maximum 6200uH RF Inductor) anywhere in store near me. Is there any substitute I can use for these parts such as normal diodes or lower inhibitors? however I will DEFIANTLY include this exact circuit in our PCB design thank you very much
Plenty of the diodes available here.
 
Plenty of the diodes available here.
I know they are online however I need them in store as I need them tomorrow and there is no stores which offer those specific diodes in store. I also have the same issue with the 200mH inductor however I believe that value to be quite high given the research I have done so:
Can I create this circuit either without the Transorb or with another diode substitute?

Are you sure that 200mH is the exact rating inductor I need for my requirements of protecting a teensy 4.1 (because again no store has 200mH the highest they sell if 6800uH)?

Can I get away with using a lower rated inductor? and if yes what rating can I use/buy:
in store options are:


Description
Ordering
L7032 68uh RF Inductor / Choke

68uh RF Inductor / Choke

L7032
L7036 150uh RF Inductor / Choke

150uh RF Inductor / Choke

L7036
L7038 220uh RF Inductor / Choke

220uh RF Inductor / Choke

L7038
L7038A 220uh RF Inductor / Choke

220uh RF Inductor / Choke

L7038A
L7040 330uh RF Inductor / Choke

330uh RF Inductor / Choke

L7040
L7040A 330uh RF Inductor / Choke

330uh RF Inductor / Choke

L7040A
L7042 470uh RF Inductor / Choke

470uh RF Inductor / Choke

L7042
L7042A 470uh RF Inductor / Choke

470uh RF Inductor / Choke

L7042A
L7044 680uh RF Inductor / Choke

680uh RF Inductor / Choke

L7044
L7046 1000uh RF Inductor / Choke

1000uh RF Inductor / Choke

L7046
L7046A 1000uh RF Inductor / Choke

1000uh RF Inductor / Choke

L7046A
L7048 1500uh RF Inductor / Choke

1500uh RF Inductor / Choke

L7048
L7050 2200uh RF Inductor / Choke

2200uh RF Inductor / Choke

L7050
L7052 3300uh RF Inductor / Choke

3300uh RF Inductor / Choke

L7052
L7054 4700uh RF Inductor / Choke

4700uh RF Inductor / Choke

L7054
L7056 6800uh RF Inductor / Choke

6800uh RF Inductor / Choke

L7056
 
That circuit certainly will give you amazing filtering from power supply voltage spikes. But if the 12 volts is from a battery (another guess...) and Teensy gets its power from that LM2596-based DCDC step down converter which is powered by a 12V battery, seems unlikely you'd be suffering from those sorts of problems on the power supply.

Adding that sort of protection won't hurt, other than the size and cost of the parts. You might lose a small amount of power in the inductor, so pay attention to its DC resistance spec.

BTW, I've often wondered if a Transzorb has anything more inside than just zener diodes, with special specs and marketing from Vishay? Does anyone know what's really inside?
 
That circuit certainly will give you amazing filtering from power supply voltage spikes. But if the 12 volts is from a battery (another guess...) and Teensy gets its power from that LM2596-based DCDC step down converter which is powered by a 12V battery, seems unlikely you'd be suffering from those sorts of problems on the power supply.
To answer your first question yes, we are using a 12V battery as our main power supply.

my question is that by this do you mean that there is no point in me implementing this circuit or that it will still protect the teensy just in more rare cases? Because I thought this circuit was to protect the teensy from over voltage/current in general not just from power supply spikes no?
(never mind I just read your edit)

however in that case @PaulStoffregen which of the inductors I have listed above would work for this circuit and can I complete the circuit with a different diode or no diode?
 
Last edited:
Can I get away with using a lower rated inductor? and if yes what rating can I use/buy:
in store options are:

That circuit will still give some filtering, though less, if you use a lower value inductor. And just to be realistic, 200 mH is huge! The largest on that list is 6.8 mH. But it (probably) can't work. The max current spec is 100mA, and Teensy alone uses about that much. It also has DC resistance of 30 ohms. So if you have 100 mA current, it will also lose 3 volts.

Edit: but the filter might be effective before the DCDC converter. The current wil be lower, and the converter will still work if you lose some of the input voltage due to the resistance of the inductor. Still, 30 ohms resistance is pretty marginal.


my question is that by this do you mean that there is no point in me implementing this circuit or that it will still protect the teensy just in more rare cases?

Even with a full schematic diagram of your entire project and complete specs on all the parts used, trying to figure out the cause of failures involves guesswork and speculation. Without even knowing what parts you've used and how everything is connected, how could anyone way whether there is no point to any particular type of protection circuit?

Best anyone can do is blind guessing. So with that in mind, my guess is the sorts of spikes that type of circuit protects against are pretty unlikely from a battery and DCDC converter. Usually those sorts of spikes come from things like power lines that can get struck by lightning. They also tend to happen in cars due if something goes wrong with the battery and the ingition system is firing the spark plugs. Certain types of industrial machinery is also notorious. Noisy input voltage is a very real concern for certain types of applications. But when your power source is only a battery, you probably don't have voltage spikes unless something in your system is creating them. Not having fast clamping diodes on DC motors would be one likely source.

Truth is, I just don't know. Even if I did see a schematic or clear diagram, I'd still be guessing. And there are many more complicated ways to protect circuitry, like snubber circuits on the motor switching, clamping or zener diodes on the signals, ferrite beads, and so on. But based on everything you've said so far, with the caveat I still haven't seen any photo or diagram, my guess is you'd get better returns from your effort by adding series resistors to some or all of the I/O signals. I know you have limited time and that's not practical to do for all of them before your deadline... but doing at least some even with 100 ohm resistors might be better than nothing. If you're worried about the power, you might get much of the benefit of that circuit by only adding 1 zener diode, rated at 5.1 or 5.6 volts. Whether any of these will be effective, I can't say.
 
Ok that makes sense, should i replace my current fuse and ultra fast diode that I have placed between the 5V line and the VIN pin of the teensy with 1 zener diode or is there no need for a zener diode with those two or should i add the zener diode on top of those two ( fuse and ultra fast diode)
 
should i replace my current fuse and ultra fast diode that I have placed between the 5V line and the VIN pin of the teensy

Normally you would connect the ultra fast diode in parallel with the motor.

I can't answer any further without seeing a diagram or clear photos. Sorry.
 
Normally you would connect the ultra fast diode in parallel with the motor.

I can't answer any further without seeing a diagram or clear photos. Sorry.
I have already put the ultra fast diodes on the motors too in parallel. I will buy a 5V6 Zener diode then and replace my fuse with that. As for the resistors on the IO pins since I do not want to have to solder the wires to the resistors to be able to connect them in series to the small breadboard so I will be unable to add them in this circuit however I will certainly add them in our PCB design and I will post the schematic for that here when it finishes to see if I have done it correctly.

Thank you again for all your help and I will let you know how it goes this weekend.

Edit: after research I noticed i need to put zener diode in parallel, also I will try to keep the fuse with the ultra fast diode in the circuit in series
 
Last edited:
Thank you for sending info on motor. It's the smoking gun!!!

Have you connected FG Signal Output (green wire) from motor to Teensy? This signal level is from 0V to 6Volts maximum. Teensy is not 5Volt tolerant which means, this signal will destroy Teensy.

There are methods to connect this signal to Teensy safely.
1) use 2 resistors as a voltage divider
2) use a 5V to 3.3V logic level converter.

Let us know if you're using this motor signal.
 
Last edited:
Thank you for sending info on motor. It's the smoking gun!!!

Have you connected FG Signal Output (green wire) from motor to Teensy? This signal level is from 0V to 6Volts maximum. Teensy is not 5Volt tolerant which means, this signal will destroy Teensy.

There are methods to connect this signal to Teensy safely.
1) use 2 resistors as a voltage divider
2) use a 5V to 3.3V logic level converter.

Let us know if you're using this motor signal.

To be honest I wish I was because then we would know the problem. However no, I am not using the FM Signal Output wire as currently we do not have a use for it so it is just tucked away and not plugged into anything. We are only using the Ground (connected to ground line), Power (connected to 12V line), PWM (connected to teensy PWM pins) and CW/CCW (direction wire) (Connected to teensy IO pins). However I have now also added the super fast diode in parallel between the ground and power wires of the motor so that should hopefully add some protection for the back EMF.

Are the rest of the wires ok?
 
General:
This circuit was specifically designed to perform 2 functions:
Remove any noise from a switching Power supply (DC to DC Converter).
Protect from large voltage transients, such as start up problems of DC to DC converters, or high voltage spikes from what ever source.
The components were chosen specifically for the circuit in which they exist.
This is commonly referred to as a custom design, where all the components are researched for there proper performance under the voltages and currents of the circuit they are used in.

The general circuit component values will have to be changed to meet your circuit requirements.

Regards,
Ed
 
Signals from Teensy to motor, PWM and CW/CCW, look okay. Teensy meets the minimum 2V logic HIGH level.

Motor datasheet does not provide input current specification for these signals. As others have mentioned, adding a resistor in series with these signals might be adding some protection against the motor drawing too much current on these signals.

At any time, do you have Teensy powered by PC USB cable simultaneously with 4.7V step down power supply?
If yes, these 2 power sources should be "diode OR'd" so they don't fight each other. With the PC putting out 5V, it will win the battle over the 4.7V step-down power supply. The step-down might(?) behave badly.
If Teensy is always connected to step-down, an alternative to diode-OR'ing, is to cut the USB cable wire supplying the USB 5V. Carefully remove some of the outer cable insulation to find the internal cable wires. In almost all cases, the USB 5V wire is red. Cut it and put some tape on the free ends so they don't touch something they shouldn't.
Better still, if you cut the 5V USB wire, now you could add a diode into the path of the USB 5V wire and do the diode-OR'ing inside the cable. All that's left is to cut the pcb track on backside of Teensy and add a diode between the 2 pcb pads. Have a look at Teensy schematic, upper left corner. https://www.pjrc.com/teensy/schematic.html
See photo from msg#17 (it's Teensy 4.0 in photo, but same applies to Teensy 4.1) Worthwhile to read all of msg#17.
Alternative to cutting Teensy track (assuming diode in USB cable), is to add diode between step-down output and Teensy Vin pin.
I hope that wasn't too confusing. Read slowly.

BTW, you don't happen to have a Teensy analog pin reading the 12V battery do you?
If yes, what voltage dividing resistor values are being used?
 
Last edited:
Signals from Teensy to motor, PWM and CW/CCW, look okay. Teensy meets the minimum 2V logic HIGH level.

Motor datasheet does not provide input current specification for these signals. As others have mentioned, adding a resistor in series with these signals might be adding some protection against the motor drawing too much current on these signals.

At any time, do you have Teensy powered by PC USB cable simultaneously with 4.7V step down power supply?
If yes, these 2 power sources should be "diode OR'd" so they don't fight each other. With the PC putting out 5V, it will win the battle over the 4.7V step-down power supply. The step-down might(?) behave badly.
If Teensy is always connected to step-down, an alternative to diode-OR'ing, is to cut the USB cable wire supplying the USB 5V. Carefully remove some of the outer cable insulation to find the internal cable wires. In almost all cases, the USB 5V wire is red. Cut it and put some tape on the free ends so they don't touch something they shouldn't.
Better still, if you cut the 5V USB wire, now you could add a diode into the path of the USB 5V wire and do the diode-OR'ing inside the cable. All that's left is to cut the pcb track on backside of Teensy and add a diode between the 2 pcb pads. Have a look at Teensy schematic, upper left corner. https://www.pjrc.com/teensy/schematic.html
See photo from msg#17 (it's Teensy 4.0 in photo, but same applies to Teensy 4.1) Worthwhile to read all of msg#17.
Alternative to cutting Teensy track (assuming diode in USB cable), is to add diode between step-down output and Teensy Vin pin.
I hope that wasn't too confusing. Read slowly.

BTW, you don't happen to have a Teensy analog pin reading the 12V battery do you?
If yes, what voltage dividing resistor values are being used?
To answer your first question no, after burning the first teensy months ago and doing extensive research on why it could have happened one of the things I learned was to never power the teensy with the computer and the power source at the same time and we have an indicator light which indicates when the battery is connected meaning we should not connect it to the computer so that is something I am sure we did not do.

As for the second question no, the teensy is in no way connected to the 12V line including any analog pins. The only way they are "connected" in the circuit is that the 12V line goes to a voltage step down which decreases it to around 4.7V and then that is used to power the teensy's VIN pin.
 
I will post the schematic for that here when it finishes to see if I have done it correctly.
On this note, we have finished one of our two PCBs and the image of the PCB and the schematic are below:

For context this is a secondary PCB which includes a Raspberry pi pico, 16 TCRT 5000 colour sensors and 1 multiplexer. The purpose of this is that the raspberry pi pico will be programmed to use the multiplexer to flip through the sensors and find which of them has the highest reading of brightness/reflectiveness. Then that sensor number and the sensor reading is sent to the teensy 4.1 on the other PCB (Main PCB which is in progress) using UART (RX and TX). This way we save processing power and time on the teensy while also getting the required sensor checks done.

image 2.PNG
image.PNG


My questions are simply:
is this circuit correct?
Do you think I require any protection for the raspberry pi pico or should it be ok? (protection does not need to be as extensive as the protection intended to add to the teensy 4.1 as raspberry pi pico is very cheap)

The "Main" PCB design including the teensy is still in progress but I will also send that here once it is completed.

Regards,
Ryan
 
A couple of thoughts on schematic and layout...

74HC4067 mux Vcc is connected to 5V. It's control inputs are just above 3V. RPi_pico will just barely meet this. You could change to 74HCT4067 (2V HIGH logic level threshold) and RPi-pico will fully meet this logic HIGH.
OR,
Change the Vcc voltage to 3.3V to 74HC4067.
As an option, place a couple of zero Ohm resistors so you can select either 3.3V or 5V to supply Vcc to mux.

RPi_pico and mux could use with decoupling capacitor, 0.1uF ceramic cap's from Vcc to GND, located close to the Vcc pin(s).

The TCRT5000's (16 total) are drawing 0.36Amps of current from 5V supply. The layout should have very wide tracks to handle this current (both 5V and GND). Depending on how long the battery will last, you might consider to power them ON-OFF when a measurement is needed.
 
Hi Everyone,

An update on how the competition this weekend went:

After a difficult competition our team was successful in becoming the State champions and earning our place in the national competition. However some things to note with the safety circuits:

At the beginning of the match we were having some problems with our teensy robot and further investigation lead to these findings
  • We thought that the zener diode was causing the teensy 4.1 to flicker on and off with power however this has not been confirmed and I will confirm this with further testing in the future
  • The fuse with a diode soldered to the bottom to it was causing a voltage drop which we did not account for. This was determined to be the main cause of the teensy not being powered as instead of receiving 4.7V it was receiving around 2-3V.
  • Thus we made the careful decision to remove the zener diode, the fuse and the diode connected to the teensy.
  • After these safety features were removed the robot performed beautifully with no issues and no damage to the teensy.
  • Thus we came to the conclusion that the most plausible cause of burning for the teensy was the reverse EMF from the motors which we did not previously know about but now that we added the fast acting diodes between the terminals that issue seems to be resolved.
  • The three teensys we ordered a week ago arrived today during the finals and of course we were unable to use them but it is good they finally arrived so we have 2 spares just in case.

As for PCBs:
  • Thank you @BillFM for your reply on our colour sensor PCB. The errors you pointed out are being fixed and the PCB should be ordered this week
  • As for the "Main" PCB with the teensy, that design is still in progress however if it is ok I will post the final schematic and PCB here for a final check before ordering.
Thank you for all your help and none of this would have been possible without you.
 
Hi all,

Thank you very much @BillFM for your advice on our colour sensor PCB, we made the changes you said and ordered them. We received and tested them recently and they work beautifully.

Today we also finished designing our "Main" PCB so it would be a big help if you could have a look and see if this pcb using a teensy is correct.

We divided it into two layers, a "top" layer:

toppcb.png

top.schmatic.png



And a "Bottom" Layer:
bottompcb.png

bottomschematic.png


Besides checking if in general it is correct could you please also confirm whether we have put the fast acting diode between the motors' terminals correctly? Because we planned to take off the ones we currently soldered on as they are big and don't allow for crimping, so we decided to put them on the actual PCB. However we are not 100% sure if that is the correct method of doing it as we're using a power plane so we cut out the sections underneath the "ground" of the motors but I am unsure if that is correct.

Regards,
Ryan Beikrasouli
 
Are those two layer PCBs? Is so your ground flood on the "bottom" board is poor, it has far too may large holes in. At the very least move the traces so there is some ground going past the end of the teensy. Even better reduce the clearance setting used so the flood can go between the pins in the teensy (Obviously check your manufacturers capabilities first but this should be easy to do).

You have a number of traces going from one side of the processor to connectors on the other side. It depends on which pins are using special features and which are being used as GPIO but take a look and see if you can swap things around to use pins close to where the signal goes. You can often simplify the board with a little bit of creative re-pinning. The aim isn't necessarily to shorten the traces, it's to reduce the number of signals that have to cross over. Every time you have to switch layer you're adding more gaps in the ground plane.

It does depends on the LED but 22 ohms in series with 3v3 seems a little low. Keep in mind you don't need to hit an LED with its maximum rated current, most can take 20mA but to the eye the difference between 5mA and 20 mA is normally barely visible. For most situations all you gain from hitting them harder than a few mA is circuit that takes more power and runs hotter.

I don't see a single capacitor anywhere on the board. Having a few on the power rails, especially close to connectors where the current draw could change rapidly is normally a good idea.
 
@AndyA Thank you for the feedback we are looking into those points now. Is it possible for you to recommend a value for the capacitors as we are unsure as to what value(s) would suffice?

Regards,
Ryan Beikrasouli
 
Ryan:
Most people over the last 50 yrs I have seen use a .1 uF ceramic and something like 5->100uF Electrolytic (Tantalum) right next to every IC on their boards. Now some people only use the .1uF ceramic, which works well for high speed switching parts I personally prefer an Electrolytic as well.

Regards,
Ed
 
Hi all,

I am one of the teammates designing the PCBs with @General_Hex.

For the capacitors, I've added 4 ceramic capacitors on the bottom board and 2 on the top board - all of which are 0.1uF. I am currently looking into Electrolytic capacitors but I haven't added any yet.

Bottom Board:
1725975664773.png

C1: Between 5V and GND near the Teensy 4.1
C2: Between 5V and GND near the wire powering the line sensor and the wire going to a motor controller
C3: Between 3V3 and GND near an IR sensor
C4: Between 5V and GND straight after the voltage regulator that steps down 12V

Top Board:
1725975872925.png

C1: Between 3V3 and GND for the BNO055 (IMU)
C2: Between 5V and GND near an LCD and HC-05

As for the wiring issues I swapped a couple pins on the right side of the Teensy.
1725976451224.png


It's a bit hard to swap the pins on the left but I did try to move the traces around to free up more space.
1725976700260.png


However, I'm not quite sure how to change the clearance settings so the flood can go between the pins in the teensy. Would that be in the Board Setup menu of KiCAD?

Regards,
Isabelle
 
However, I'm not quite sure how to change the clearance settings so the flood can go between the pins in the teensy. Would that be in the Board Setup menu of KiCAD?

Yes (in KiCAD 7), the clearances can be adjusted in the "Board Setup" menu, specifically the settings under "Design Rules" then "Net Classes", either by adjusting the "Default" (for all traces, if that's all that you are using), or by adjusting any custom classes that you may have defined as appropriate.

Mark J Culross
KD5RXT
 
Back
Top