The ESP32 has become a mainstream Maker platform. This is why we got many different boards to
choose from. Which one fits your project? Let s get an overview and test the boards. And of course, we will add the information
to our comparison table from the last videos. Gr ezi YouTubers. Here is the guy with the Swiss accent. With a new episode and fresh ideas around
sensors and microcontrollers. Remember: If you subscribe, you will always
sit in the first row. Here are the contenders: Eight boards from
number 27 to 35. Most of them are available on Aliexpress,
and some also on Banggood. Only one is from Sparkfun. In this comparison, I did not include boards
Made in Europe and only one Made in the USA. Maybe this will change in the future. At the end of the video, you will know why. All boards use the standard ESP32 chip. Maybe I will do an introduction to ESP32-S2
boards in a future video. For my last ESP32 board videos #193 and #201,
I had clear criteria. Most of them are still valid. Some like chip version or Does the board crash
if you disconnect USB? are no more needed because these problems do not exist anymore. The rest of the criteria are still valid:
- Does the board use a shielded module or just an ESP chip soldered on the PCB? - How many pins are broken out to pin headers? - Does the board come with an antenna connector
for an external antenna? - Is the board breadboard-friendly? - How readable is the pin labeling? - Does the board have an additional flash
button? - Is a battery switch available? - Which battery connector is needed? - What type of voltage regulator and Li-Ion
management chips are used? - How much current does the board consume
from the battery and USB? - How fast can it charge the battery? Some boards have additional components like:
- Displays - SD card slot
- PS- RAM (WROOVER module) - I2C connectors
Two boards have no battery connector and no charging chip. What are the results of the investigation? The first one is obvious: LilyGO, with its
TTGO brand, manufactures many boards, and only a few are from other manufacturers. One is Wemos with its LOLIN boards. The next observation: There are two form factors:
One with two pin rows, usually also narrow enough to leave one row of pins on each side
on a breadboard. Or the much smaller boards with four rows
of pins. They cannot be used on breadboards, but this
form factor is my favorite for projects. A third finding is: Only two boards use the
WROOVER module with additional PS RAM. Fortunately, one for each form factor. This might be important if you use microPython
for your projects. Only one board has a display. For some projects, small displays are convenient. I suppose you remember the small finder for
Weather Balloons in video #360 or the Meshtastic communicator in video #337 where the T-Beams'
small OLED displays were helpful. For most projects, such small displays are
too small, or you want to have more freedom where to mount it in your box. The Lolin D32 Pro is probably made with such
a project in mind: It has an additional display connector. Too many boards have SD card slots. I say this because I never used an SD card
on a connected device. I always transfer the data to a server. Two boards have a built-in I2C connector. This might be an advantage for a development
system if you standardize your sensors on a particular connector. I usually solder the wires directly to the
deployed boards. So I do not need a connector. And as a development platform, I use this
Keyesstudio board with this baseplate. For each pin, it also offers a ground and
a 3.3 volts pin. Very handy because for most projects, I need
more ground pins than available on the standard boards. This board, BTW, also has three 5-volt pins. It has no battery, which is not a problem
for a development board. The only disadvantage I found is that I have
to push the boot button to upload. Maybe you know the correct board definition
to avoid that? Power consumption is essential for battery-operated
devices. This is why I did my own measurements and
did not believe the datasheets. How did I test the current consumption? I distinguished between the ESP32 chip or
module and the rest of the circuitry. ESP32 chips and modules do not differ a lot
in consumption. So the difference must come from the parts
added by the board manufacturer. This is why I measure when the MCU is idle. Like that, we can easily see the differences
coming from the added components. I wrote a sketch with two phases: Running
idle and deep-sleep. Like that, I was able to read the values during
the two phases. Before we look at the results, I want to focus
you on something interesting: In the previous tests, boards used different charging and
regulator components. This time we nearly always find the same chips:
The LTC4054 as a charger and the ME6211 as a voltage regulator. So I would not expect significantly different
behaviors of the boards. Will this become true? The first tests are connected to USB power. I insert a simple USB ampere meter to measure
idle current without a connected battery. As expected, the idle currents are very similar
at around 40mA for all boards. I did not measure deep-sleep current because
it is not essential if we have USB power available. Next, I connect a small Li-Ion battery and
measure the maximum charging current to a discharged battery using a multimeter. This shows us how fast a battery can be charged,
for example, with solar power. Or how long we have to wait while charging. All boards use either the bigger JST-PH or
the small GH connector. It is always good to have a bunch of such
prefabricated cables in your lab. Most boards charge with around 500 mA, which
is ok because more current could also become dangerous for small batteries. The T1 charges at 350mA. Which is probably still ok for most scenarios. Next: How much current do they use during
battery operation? Because this measurement is delicate, particularly
if you want to measure the deep-sleep current precisely, I use this OTII. It replaces the battery and provides exactly
3.7 volts to the battery connector of the board. Here you see the current curve: First, a peak
when the chip starts and all capacitors have to be charged. Then a lengthy delay where I can measure the
idle current. Then deep-sleep. I select a range and read the average current. You can watch video #47 if you are interested
in deep-sleep and how to calculate battery run-times. All boards consume around 40mA when the CPU
is idle. Very similar to our previous measurement with
USB power. No surprise because all boards use linear
regulators. These regulators reduce the voltage, but nearly
the same current flows at the output and the input. Let s continue with the deep-sleep, measured
in Micro-Amperes. Traditionally we find a lot of differences
in this discipline. Also this time: The Sparkfun board consumes
6240 A. I have no clue where this power is used. It even has no always-on LED. The best is the LoLin D32 with 70 A. A good
value because this small battery would power such a board in deep-sleep for more than 500
days. Then come the D32 Pro and the T-Display. Both also have no always-on LED but consume
200 and 300 A. The D32 Pro has a WROVER module with an additional PSRAM. I did not find information on how much the
PSRAM consumes during deep-sleep. The infamous leaders in this discipline are
the little T7 V1.3 and the T1. Both with a whopping 1500-1600 A. Why is that? Because of stupidity. Both have an LED, which is always on. Also, during battery operation, as we see
here in the diagram. But even if I desolder the LED, the T7 V1.3
consumes 870 A. Not good. Up till now, this was my favorite board for
battery operated projects because of its size. I still will use it with a small 6 volts solar
panel attached to it as I did in my light sensor project. As we see on this chart, the solar panel can
keep the battery voltage constant even during winter in Switzerland. We will have the shortest day around Christmas,
and the battery still is fully charged. The sensor transmits its values every 10 minutes,
BTW. Fortunately, the T7 V1.5 with the WROVER module
has a similar form factor and, after desoldering, the power LED only consumes 78 A. Much better. I think, after this video, this will become
my favorite board for battery operated projects. Till I get one of the boards from the superpower
project of video #351, of course. Anything else to mention? The documentation for the TTGO and LoLin boards
is good. Usually, they even provide a diagram. I added the links to the table. Like that, you find details like where is
the LED or the buttons connected. And the pins of the boards are clearly labeled. Sometimes even on both sides. For breadboard usage, top labeling is best. For soldering wires, it depends on which side
you solder the wires. I prefer from the top to stick the boards
with double-sided tape. So top labels are my favorite. The Live D1 mini is my favorite board for
USB powered projects, BTW. You find a link to this table in the video
description. I did not delete the older boards. Maybe you have one lying around and want to
know how it compares with the newer boards. Did I forget anything? Yes, the prices. After the shutdown of civil aviation last
spring, shipping costs increased considerably. Now, shipping is sometimes more expensive
than the board. Often it does also not help to order more
than one because shipping cost increases fast. Sometimes it is even cheaper to place two
orders instead of one with the needed number of parts. When I recently bought products from Digikey
or Mouser, I learned that they deliver free of charge if I order for more than 60 dollars. All customs hassle avoided, delivered at my
front door in two to three days. Like that, products from Adafruit suddenly
become more competitive. I will have a close look at this situation
and at Amazon's role in distributing parts for our needs. Maybe I will create a video about that topic. Summarized:
- We got new and improved ESP32 boards, mainly from LilyGO and Wemos
- They come in two form factors: Small with two double pin rows and bigger with two single
pin rows - A lot of them have a battery connector and
a Li-Ion charger on board. Very convenient for battery operated projects
- The power consumption if powered from USB or during regular operation was the same for
all boards - Unfortunately, I still found significant
differences in current consumption for deep-sleep - Sometimes also because of stupidity: Two
boards have power LEDs also during battery operation
- If we remove these LEDs, at least one board of each form factor has an acceptable deep-sleep
current consumption - The other boards with batteries still can
be used for solar-powered projects - All boards have lots of pins. It should not be a problem to connect sensors
and displays - Many boards offer an SD card slot. If you need something like that
- The documentation of the boards is much better than in the past. Usually, we get an official diagram
- Prices, mostly shipping costs, went up considerably since the last videos
That is all for today. As always, you find all the relevant links
in the description. I hope this video was useful or at least interesting
for you. If true, please consider supporting the channel
to secure its future existence. Thank you! Bye
