#245 Deep-Sleep Current: Which is better? µCurrent or Current Ranger? (ESP32, ESP8266)

during deep sleep microcontrollers consume very low currents and batteries live longer unfortunately it is not easy to measure these small currents today we will check out two small boxes which promise to make this possible and we will use a $10,000 professional power analyzer to establish the reference gritty youtubers here is the guy with the Swiss accent with a new episode and fresh ideas around sensors and microcontrollers the first contender is the well known microcurrent called which is made by Dave Jones from ee-vie block it is the mother of this device category and costs around $80 the other is called current Ranger and is built by Felix Russo from low-power lab it just entered the market and is an extension of Dave's concept and costs around 120 dollars the reference is keysight's n6705a DC power analyzer I asked the swiss keysight dealer compute controls if they would be willing to help us with this project and they kindly agreed but what is so critical about the matter why can't we just use our multimeters to measure this current let's quickly analyze the setup we have a power source usually a battery and a microcontroller in our case an ESP 32 it runs a small sketch which connects to Wi-Fi calls open weather maps and goes to sleep for 10 seconds the n6705a shows what happens in this diagram the ESP 32 wakes up connects to Wi-Fi gets data and goes to sleep again the peak current during Wi-Fi operation is around 530 million per and the deep sleep current around 5 micro ampere fortunately the n6705a can cope with this extreme range of 1 to 100,000 and shows us this curve now we can compare this diagram with the curves produced by our small boxes but first we have to talk about another problem it is called Burton voltage to measure the current flowing from the battery to the ESP here called device on the test or D UT we insert a current meter into the flow inside such meters you find the resistor and the voltmeter the meter can calculate the current with Ohm's law the voltage needed by the meter unfortunately reduces the voltage for the device on the test so it should be as small as possible this fluke multimeter as an example has a resistor of 2 ohms for its milli ampere range 2 times 500 milli amperes equals 1 Walt loss the ESP 32 only gets 2 point 3 volts and creates a brownout reset this is why multimeters cannot be used for that purpose by the way the unity meters are even worse because a small Burton voltage also means we need a sensitive instrument we have a dilemma which was solved when they've introduced this little box called you current gold it contains 2 precision op amps in series in a low noise setup which provide an amplification of 100 so this box can use a 10 milli ohm resistor in the milli ampere range which is two thousand times better than the fluke let's do the math again with this resistor 500 milli ampere times 10 milli ohm equals 5 millivolt the micro current gold has only a burden voltage of 5 millivolt and after the amplification it shows 500 millivolts at the output which can be displayed by an oscilloscope or measured by a multimeter if we connect the micro current gold to our es P 32 we see a similar curve like the one shown by the keysight instrument cool to check the microcurrent gold I focus at the few microseconds at boot we see that the ESP starts and switches Wi-Fi on the peak current is more than 500 milli ampere the microcurrent gold shows only 440 milliampere if we look closer we see that these Peaks are very short that is probably the reason for the difference the current range by the way shows similar results than the microcurrent gold I did not take more into this matter later on we see that the ESP 32 consumes around 113 milli ampere the microcurrent gold shows 110 this is very close and with a burden voltage of maximum 3 millivolt with this setup we can measure the current consumption of the ESP 32 when it is working summarized the microcurrent gold is good enough for the girls I go out with as a ve would say after a while the ESP 32 starts to deep sleep during this phase the 10 milli ohm resistor would only show 50 nano volt which is close to nothing even if we amplify it by the factor of 100 we only get 5 micro volt still too low for our multimeter or oscilloscope this is why Dave uses three different resistors for the three different ranges which are manually selected using this switch the resistor used for deep sleep measurements is 10 ohms the Burton voltage therefore is 0.05 millivolt and the meter shows 100 times more 5 millivolt both are OK by the way the micro current gold has no built-in meter you only have two banana plugs as an output to connect either to a multimeter or an oscilloscope because cables with banana plugs easily pick up noise we also see it in the oscilloscope but it measures their correct value of around 5 millivolt risk effectively five micro ampere next problem the burden voltage of the 10 ohm resistors at 500 milli volt peak is 5 volts which eats up all our 3.3 volt and the ESP 32 starts to reset all the time we have to manually switch the two ranges between the two states of the MCU very fast of course now let's have a quick look at the reference the keysight n6705b power analyzer this device not only has a precise current meter which by the way works the same way as our two small devices with different resistors for the different ranges but also has a built-in power source which could be used to simulate all sorts of unusual situations today we only use it as a simple 3.3 volt power supply and a current meter it also has a high-speed analog to digital converter and a fast switch for the ranges so we do not recognize that it changes ranges during measurement we just get all the details in one chart and if we lock the data on a stick we can analyze it later on a PC and how does it solve the issue with a Burton voltage because it has an internal power supply it does not stabilize the voltage at this point it maintains the voltage here so it has no burden voltage at all what else did you expect from a $10,000 meter and now comes felix current ranger it has a very similar analog setup as the micro current called but added an Atmel chip to switch the resistors in auto mode the chip continuously measures the output voltage and automatically selects the range if the voltage is too small or too high we see this if we watch what happens if an ESP 32 is connected the range regularly changes from milli ampair to micro ampere with this box the ESP 32 does no more crash and we can measure both states without man--who intervention the curve looks very similar to the diagram of the key side there are two differences during switching from the micro ampere to the milli ampere range we see a wrong peak which is not visible on the key side but because it has a very high value it is easy to detect its origin I assume is the switching circuit the second thing is that the value during sleep sleep is automatically amplified by a factor of 1,000 because it is so small it still shows us nearly zero on the oscilloscope but if we select a 100 millivolt range we see that it measures around 5 millivolt respectively 5 micro ampere which is correct again and because we can connect to the probe directly to the current ranger it also has less noise than the micro current called you can add a small elite display to the current ranger it shows the current in big letters unfortunately all my oh let's have a different pin sequence so I had to create a workaround but definitely this OLED is a nice addition the nano ampere range is quite delicate and I did not want that the auto mode accidentally selects this range so I went to felix arduino sketch and commented a line now my current ranger precisely does what I want the nano ampere range can only be selected manually this is the advantage of open source software one last difference the microcurrent gold is powered by a 3 volt coin cell and can display minus 1 point 5 to plus 1.5 volts this is ideal for analog circuits or if you want to analyze charging cycles because then current flows in both directions but it reduces the positive range by a factor of 2 the current ranger also offers this biased mode the default mode however measures only positive current and so we get a range of about 3 volt also a plus if we deal with microcontrollers as mentioned the microcurrent gold uses a coin cell for power such a cell last for many hours of operation no problem unfortunately if you forget to switch it off the battery might be dead next time you want to use it and you have to remove all four screws to replace it the current Ranger uses auto power off and a lipo battery which is another critical difference for a forgetful guy like me the lipo however does not come with a kit nor comes the jst connector and you have to rework the case to fit the USB connector this is why I printed my little white box summarized multimeters impose high burden voltage and cannot be used to accurately measure current for microcontrollers especially if Wi-Fi or deep sleep is used the microcurrent gold solves the issue of Burton voltages for relatively constant currents it is a simple and handy device for measuring these small currents however its operation is entirely manual it's plus minus 1.5 volt range sometimes is quite small measuring deep sleep scenarios create a new problem which cannot easily be solved with the microcurrent gold because ranges have to be adjusted manually the ESP 32 regularly crashes when it tries to wake up and the microcurrent gold is still in the micro ampere range the current Ranger is an enhanced model which inherits the good traits of the microcurrent gold it adds many useful features like Auto ranging or a light or battery including a charging circuit both are accurate enough for characterizing current consumption of microcontrollers for a hobbyist the auto-ranging of the current Ranger is fast enough to solve the issues of Burton voltage the ESP 32 does not crush the results of the measurements are mostly comparable with what was measured by the keysight meter one exception is short spikes but they do not contain a lot of energy and therefore are not very important for makers both are relatively pricey devices compared with a standard multimeter I tend towards the current Ranger if you can afford the higher price it has no apparent disadvantages and a lot of advantages over the microcurrent gold I was glad to use a professional reference for this comparison like that we saw that our devices are usable for makers but we also saw that there is quite a difference to the n6705a from keysight not only in price one last thing I did not test the Bluetooth capabilities of the current Ranger it would enable wireless data locking Wireless here is important it is easy to create ground loops or even short circuits mainly if you work with oscilloscopes which have a ground reference for that reason it is not advisable to use the USB connection on the current Ranger during measuring I also used a USB isolator for the monitoring of the ESP 32 I hope this video was useful or at least interesting for you if true please consider supporting the channel to secure its future existence you find the links in the description thank you bye you

Channel: Andreas Spiess

Views: 74,926

Keywords: electronics, ucurrent gold, diy, nodemcu, deep sleep, esp8266 project, ESP32 project, greatscott, arduino, arduino project, esp8266, current ranger, currentranger, esp32 tutorial, guide, how to, simple, esp32 project, hack, smart home, lowpowerlab, wemos, LORAWAN, hobby, ucurrent eevblog, project, eevblog, N6705C, Keysight N6705C, iot, esp32 deep sleep, esp32, ttgo, wifi, do-it-yourself, beginners, esp32 micropython

Id: HmXfyLyN38c

Length: 15min 1sec (901 seconds)

Published: Sun Dec 09 2018