Solar cell + supercap to power a Teensy LC

Status
Not open for further replies.

Joegi

Well-known member
I want to charge a 1F(5V) supercap with a small solar cell(6V) and when a voltage of ~2V has been reached a Teensy LC should be powered with that voltage. The teensy then regularly is measuring the voltage(using a z-diode to reduce the voltage) and charging speed. When there is enough sunlight and the charging speed is high the teensy is waiting until the supercap has entirely charged (5V) and then is starting a motor via a motor driver. Both, teensy and motor driver (pololu drv8835) are working down to 1.7V. With little sunlight the teensy is only waiting until 5V-xV until it is starting the motor to not have a too big interval between the motor runs. The Teensy is running from 1.7V on stable enough (for my taste). The problem is that when I connect the Tennsy directly to the solar cell + supercap it stops charging at 1.62V. It seems that at that voltage there is a point where the Teensy is drawing relatively much current, but does not start properly, because if I manually raise the voltage to 1.7V the Teensy is starting properly and when set to hibernate mode there is no noticeable difference in charging speed with or without connected teensy so the solar cell is sourcing much more current than the teensy draws in hibernate mode. My problem is to get past that 1.62V. I tried it with different z-diodes, a z-diode plus a standard diode, standard diodes in series, but it always starts with low current and so a voltage below 1.6V! What to do?
Btw. when the cap has charged to about 1.7V and the teensy is up and running it bypasses the z-diode(currently used, but not working as expected) with one channel of the two channel motor driver to be able to run down to 1.7V.
Thanks for the answers in advance.
 
The right answer is probably digging into the reference manual for the LC to see if you can setup the LC startup logic to wait until well after 1.7V min in the cap to kick in, since you need enough stored power to get it through the initial startup before it heads off into sleep mode. Brownout detect may do what you want, though depends on current draw of the powered but held in reset condition.

Less clever option may be to have a separate FET based power on logic and leave the teensy to only kick in once power is up and stable. That will be less efficient, but depending on what you are doing may mean you can just ditch the teensy and do whatever you need with an op amp or two and maybe a 555 chip.
 
I found these STM1061 ICs which seemed to be perfect so I ordered a few as I was sure that the Teensy LC has a reset pin, but now after the ICs arrived I had to recognize that there is no reset pin or pad!
 
Challenge accepted :)

This will let you use the STM1061 to determine when the voltage to the Teensy gets turned on, just find yourself a PChan Mosfet with a low VGSon and RDSon.
Ive never done this with such low voltages, I recommend you breadboard it and make sure it works as well as it does in my simulation.
It does not provide reverse polarity protection so keep that in mind. I can show you how to add a second PChan to do Reverse Polarity protection.
Also keep in mind the STM1061 only works down to about 0.6V.

UV LO using STM1061.JPG
 
A 1F supercap can drive 1A of current for 1s while losing 1V.
Let's say your Teensy draws 50 mA -- it can run 20 seconds per V of drop, and the Teensy will probably want at least 3.0V, so at most you can run for 40 seconds.
If you add a very small motor that draws, say, 100 mA, then you will cut the possible runtime by the same amount.

Also, supercaps really don't like drawing too much current out of them.
30 Ohm internal resistance is not uncommon: https://www.digikey.com/product-detail/en/eaton/KR-5R5C105-R/283-2810-ND/1556238
Although with different design trade-offs, you can get down to half an Ohm or so: http://www.illinoiscapacitor.com/pdf/seriesDocuments/DGH series.pdf

So, the question is: what value does the supercap give you, that the plain solar cell won't give you on its own?
Do you really need that 10-30 second accumulation of runtime?
If you need more runtime, perhaps what you want is something like a LiPo Rider instead (a Seeed studio implementation of solar charge controller / DC regulator for a LiPo battery.)
 
Challenge accepted :)

This will let you use the STM1061 to determine when the voltage to the Teensy gets turned on, just find yourself a PChan Mosfet with a low VGSon and RDSon.
Ive never done this with such low voltages, I recommend you breadboard it and make sure it works as well as it does in my simulation.
It does not provide reverse polarity protection so keep that in mind. I can show you how to add a second PChan to do Reverse Polarity protection.
Also keep in mind the STM1061 only works down to about 0.6V.

View attachment 11444

Thank you very much for your effort (sorry for the late response), but I found this: NCP304A, a voltage detector with active-high output that can source up to 70mA, so I could power the Teensy + motor driver with it! The problem is, I didn't find any source where I could buy it! Is there any similar IC with better availability (for private persons)?
 
A 1F supercap can drive 1A of current for 1s while losing 1V.
Let's say your Teensy draws 50 mA -- it can run 20 seconds per V of drop, and the Teensy will probably want at least 3.0V, so at most you can run for 40 seconds.
If you add a very small motor that draws, say, 100 mA, then you will cut the possible runtime by the same amount.

The Teensy (@24MHz) runs down to 1.7V (together with the motor driver) and will be set to hibernate mode most of the time and even in normal mode it does not draw more than 20mA!
The motor draws 50mA@5V and runs for about 2min from 5V down to 1.7V! To increase the time I could add another capacitor or use a bigger one!

Also, supercaps really don't like drawing too much current out of them.
30 Ohm internal resistance is not uncommon: https://www.digikey.com/product-detail/en/eaton/KR-5R5C105-R/283-2810-ND/1556238
Although with different design trade-offs, you can get down to half an Ohm or so: http://www.illinoiscapacitor.com/pdf/seriesDocuments/DGH series.pdf

So, the question is: what value does the supercap give you, that the plain solar cell won't give you on its own?
Do you really need that 10-30 second accumulation of runtime?
If you need more runtime, perhaps what you want is something like a LiPo Rider instead (a Seeed studio implementation of solar charge controller / DC regulator for a LiPo battery.)

It's not about accumulation of runtime, but about having runtime at all. I will have to see if I will be pleased in the end, but I'm also thinking about a (entirely new design) with a bigger or two solar cells (but still with supercap)!

Thanks for your input!
 
Last edited:
Status
Not open for further replies.
Back
Top