I can relate. I've really been curious to test some buck converter's quiescent current but the price of such a tool is too much to bear at first. Eventually I get over it and just buy it...
I have a similar drawer caled "I'm sure that I'll never need this, but I bought them because measuring devices are like Pokemons: 'gotta catch'em all' "
@@jtreg There is also a great maker community! We have annual maker conference focused on open source software/hardware called Tuxcon. Andreas should come as a speaker some time!
Thank you for the recommendation, Andreas! I am using the PPK2 now for evaluating a CubeCell board. One hint: by pressing ctrl-shift-alt-A you can open the "expert menu" and increase the buffer size from 200 MB to e.g. 1000 MB. This increases the total sample interval you can use.
Nice video! Thank you! We already got a preview in the beginning of the video of the upcoming boards. Looking forward to those DUATs (Devices Under Andreas Test) :) Have a nice Sunday!
Thank you for this vídeo. It's very useful, not only for makers but for professional engineers like me. Some times is nice to have simple equipment to do simple test without the need of complex configurations but with good results.
Excellent video. Thank you Andreas! It added all the missing tiny details to my knowledge on the subject and ... created a window to spend extra 100 dollars on aparently excellent kit from NordicSemi :)
Amazing video, just what I needed. I was given the task to evaluate and see if we should go with the PPK2 or the Otii, considering the difference in price being almost 10-fold. This video gave a good overview of different characteristics.
Once again - I have learnt so much from you, Herr Spiess - thank you for your excellent videos. I have ordered a Nordic PPK II from Farnell in the UK. It seems very popular as I have to wait until February 2022 to see it - but the price is very good - now £75. I'm 53 and have been dabbling with electronics since I was maybe 8. If only we had youtube in the 1970s - I would have been so much more advanced by now 😄 - PS I love Switzerland: I have been twice. Most beautiful country. And unlike Great Britain, the trains work 😂
Was looking for this kind of kit for so long! Have been lucky to wait before buying µCurrent or so. Thank you, i got the last available nRF kit from adafruit :D
I bought a PPKII some months back to work on a solar project, I've found it very good. If you are working on small projects away from your electronics bench it's kind of like having a tiny bench PSU that you control from a computer. Very handy.
Thanks for the excellent and very useful video. I was just searching for an inexpensive alternative to measure low currents and power. You saved my day :-)
@@AndreasSpiess For all who are interested: Adafruit has still some of the kits available, but you need to consider the delivery costs etc. Other shops may have up to 52 (!!!) weeks delivery time. The funny fact: I ordered mine yesterday, and today Nordic offered me to send me a kit for review.
Yes Andreas, the video was useful. Your demonstration is convincing. Not just anyone can measure! And even when you feel comfortable, you can still be fooled. Many thanks with a Breton accent. ;) ;)
I'm back home and waiting for me was this and other videos - like usually, great details and well done video and suggestions! You are the BEST Andreas! Thank you very much. (PS: I have a new granddaughter in Virginia, where I was past 2 weeks!).
@@AndreasSpiess Sonner or later something will come for you! I have 3 grandkids and one is 15 years old, so sooner I can become grand-grandpa! You are blessed and I'm also very blessed, that is more than we need!
Thanks for another informative video. Minor point: you often say “comfortable” when “convenient” would be a better word. When you say “not very comfortable” it makes me think that you are in pain!
One of the "false friends" of German and English. "komfortabel" is often used in the sense of "easy to use" or "convenient". It also does have the meaning as in "comfy chair", but now that I think about it I almost never hear someone used it in that way. I don't know about the Swiss though. Their "German" is weird sometimes ;) .
@@JeffreyBaitis Very true and a lovely fellow (even though he continues to deny murdering his wife - look for the clues). I really only, impudently, add a correction when it really matters - otherwise they are a joy! :)
Thx, very informative as always :) I use a simple trick: Just wire a switch in parallel with your multimeter. This way you can bypass the burden resistor whenever you want. Just close the switch when the ESP boots up and open it when it is in deep sleep. This way you can measure deep sleep with the mA range of your meter without any issues, and that's what you care most. I've got a 1P2T toggle switch connected to a NO/NC microswitch for this. It allows me to select if pressing the microsowtch opens or closes the bypass ;)
I also just shortcut the multimeter with a stacked cable and/or change ranges. Same works with 2 shunts in series (one cut short for high currents). You can also measure in front of a low quiescent current LDO to get some rough indication. Nevertheless I now plan to upgrade to this one. Looks very convinient.
Yes, that's how I do it too. My esp32 is in deep sleep most of the time so I am mostly interested in deep sleep current so the multimeter and switch is my cheap solution. That's how I discovered that having wake on touch in micropython uses 200uA, not helpful 🤣
As always a very interesting video. I've been playing with the Adafruit MagTag and Pimoroni Inky Frame recently and finally bought myself a PPK2 to properly delve into the low power world. I wrote an article on Instructables looking at NeoPixels with the PPK2. I included the table (with credit) from your video at 13:55, I hope you don't mind. BTW, the Otii specs say 5V max if it's locked to high current range. I'll include a link to it in a reply to this comment as RU-vid often hides comments with links.
Yeah, I moved from my CurrentRanger to the PPK-II and it's really awesome to use. It's great that I can use it to also power my device I'm testing and it's super quick to get a board setup and results captured. It's a crazy price for what you get.
@@AndreasSpiess Yes I did! I'm interested to see if/what you use them for. Keep in mind the FeatherS2 is not designed for the "lowest" of deep sleep current, though my latest FeatherS2 revisions has reduced it another 30%. Keep up the awesome videos - every one is a wealth of knowledge!
Texas Instruments has such power meters built in many of their development kits, also for a very good price, because you not only get the power meter, but also a programmer, and a nice programmable microcontroller onboard. I own two: MSP-EXP430G2ET development kit which is able to measure milliwatts and features MSP430 microcontroller, and LAUNCHXL-CC1352R1 development kit which is able to measure microampers (I never tried nanoampers) and features CC1252 microcontroller with ZigBee, Thread, Bluetooth LE support and some more radio standards, all at a price of 40$ + shipping, available from the official website. The only downside: in order to use power meter, you have to install huge Eclipse-based SDK, and sometimes it behaves weirdly. As for everything else, I find it good value for the price.
@@AndreasSpiess TI make the most exquisite data sheets. I find myself using their components, not because they are the cheapest, or most capable, but because I know how they are supposed to be used. I particularly like thier example circuits and layouts, which give ideas about what you can and an not get away with. As far as I'm concerned, TI are very maker friendly.
I worked with some of their DSPs a few years ago, and their demo code was atrocious. They didn’t have consistent names for their registers. One of their demos had a shared init function with a generic name, and about 20 different samples that called it. When I looked inside it, it was not actually very generic at all - it initialized things in very oddly specific ways. Then most amusingly frustrating experience was a board that had an EEPROM with demo code, and a JTAG interface. You could use the JTAG to load your demo code into RAM to try it out. They did not provide clear instructions for uploading your code to the EEPROM. If you uploaded code to the EEPROM without including magic undocumented initialization routines, bad things would happen. It would boot and immediately start executing code from the EEPROM which would scramble the JTAG setup seriously enough to prevent you from using JTAG to rewrite the EEPROM! This is code that worked from SRAM just fine. You needed to lift a pin from the EEPROM to recover.
@@gf2e That kind of demonstrates my point as well. I've found missing dimensions in many datasheets, but TI have always been happy to hear from me, and have whipped up a revised datasheet within 48hrs. Other manufacturers have completely ignored my requests for a certain physical dimension, or information about whether the IC is current / marked to be discontinued. I do the odd thing with uCs, but I tend to stick closer to the analogue side of things. Based on my previous dealings with TI, I feel like if I were to struggle with example code, they would be more than happy to help.
I think TI is the best choice when looking for a PMIC, in particular when you need low power. When looking for a DCDC converter I don't even bother with other manufacturers and do straight to their web and use their comfortable product selector.
I like your work! I specialised in energy harvesting and micropower radio sensors. As you say, current is never constant, eg 25mA for 5 or 10mS radio pulse etc. normal multimeter no good. I used voltmeter (with serial/usb logging) and capacitor to do all important work. eg 4700uF cap and 4 second sample period, (but can connect supercap in parallel to slow discharge rate etc) just switch off main psu when ready and put log file into spreadsheet. easy, accurate permanent record for your documentation, with graphs and analysis. (my self powered energy monitor has been running for 10 years now! my battery powered monitors are still being sold!!) Keep up the good work.
its a tiny battery! you are usually trying to determine battery life!!! (can even add series resistor to match real battery ESR).but a real battery may last for years. this method gives reliable result in seconds to minutes. it even matches typical battery voltage discharge eg my energy harvesting electronics had to work over 1v6 to 5v5 but lithium may be 2v7 to 4.2v like my gps module. I tried everything before this simple fast accurate automatic method. they then used it to sample my units during the mass production runs. only stopwatch needed for that. (thanks for replying. I am working through your other videos now. good stuff.)(my latest project is a very very low cost solder reflow oven)@@AndreasSpiess
I wouldn't say that all oscopes are connected to ground, my owon USB scope is isolated so that way people don't fry their expensive computers. But other than that a good video, i never even considered that the shunt for micro amps in the amp meter could be a problem when measuring higher currents before the device goes to sleep, this will probably save me some future debugging time.
I never had a USB oscilloscope. Interesting. AFAIK they have to do that to prevent 220 volt on the case if you connect the GND of your probe the wrong way round. What happens if you connect it via USB to your PC?
This looks like an excellent tool. I was using an Agilent meter with some custom software before, but the max sampling rate was 10KHz and it wasn't nearly as easy to use. It did go to to 3A which was useful for centralist cellular modems but it also cost about 1000€
So it was quite cheap for an Agilent ;-) In my last test I was able to use a Keysight Instrument for comparison. This was more than 10'000 dollars. This one here is probably not comparable in quality but good enough if you use it from time.to time.
A small (cheap) Tip if you want to meassure small currents less often: limit the burden voltage by using a Diode to avoid the brownout: Just put the Diode parallel to your Shunt-Resistor (/Multimer), it will limit the burden voltage to ~800mV (Si-Diode) and will not have much influence in the lower regions you might be able to switch the shunt or range while the diode takes the current.
As usual: wherever I go, Andreas was there already. But that is ok. I just purchased PPK2 and I am astonished by its capabilities. I already managed to get my ESP32 to get into deep sleep with average 28uA during sleep. Considering my sensors wake up to measure temperature, humidity and light (and also checking its battery voltage) and they go to deep sleep - the whole procedure (from sleep to sleep) takes anything from 1.1s (if light is measured with TEPT4400) to 1.7s (if I use IC2 lux sensor). With device sleeping for 180s it gives me super battery life. Btw I replaced voltage divider resistors to 240k - PPK2 does not lie. Fast wifi connection tricks: delay between retries 10ms, fixed WiFi channel. It gives unbelievable savings on time to connect to the router. Also, hibernation instead of deep sleep - that will be tested next. :-) EDIT: hibernate did not bring anything better - so probably out of this 28uA of deep sleep current, 27uA belongs to components rather than to ESP32 itself
You were lucky to get a PPK2. Many viewers complained that they did not fin one. And thank you for the tricks for a fast connection. 1.1 seconds is really fast!
@@AndreasSpiess I ordered from DiGiKey - they have full stock ;-). Btw: after more tests done, bare ESP32 in deep sleep mode... 5uA - not even hibernated ;-). With regards to 1.1s: out of that, 0.3s is... "wait 300ms because maybe there is MQTT coming for you" - otherwise it would be 0.8s. But then, when MQTT message comes, device would not have enough time to react and/or respond. And I am not able to make it shorter. Without that 0.3s (non blocking) delay implemented, device works, but is without remote control. But since you are MQTT master: maybe you know how to do it? ;-)
I dunno about "maybe as a christmas present". I had my credit card out and another window open as soon as you mentioned high speed autoranging and a half decent desktop app. Thanks! I have Dave Jones' , and the Current Ranger, and they've both paid for themselves, but.. this looks a great addition.
Thank you for your review of useful solutions for us makers. However, I would like to point out that Power Ranger can be used with a cheap Bluetooth module HC-05 thanks to your Video #245. Bluetooth serial sends current measurement that Arduino GUI monitor tool can display without the need for oscilloscope.
You are right. I heard about that from one of my Patreons. If I remember right, this was not the case when I made my initial review. Still, the author of the software writes that you still need an oscilloscope. Maybe the sampling rate is not very high.
These are the kind of videos that I am subscribed for! Small question, why didn't you try an oscilloscope and a current clamp connected to it? I have a cheap current clamp and it measures in the range of mA DC ...
Thank you! I have a current probe for the oscilloscope. But it was more expensive than the Power Profiler kit and it only goes down to mA. For those small currents and voltages, a resistor ist probably good enough.
Andreas , as usual a perfect video. You just forgot the new CX3300 from Keysight, dedicated to this task, for entry price of 40 000$ without probes! Certainly nice to have 😂
Dear Master Andreas, as always, your analysis are perfects. I have a request for your next video : could you explore the making of a low energy blé or zygbee sensor leveraging nRF 51 or similar MCU and check if this can works on battery powered devices (cr2032) beyond temperature sensors like CO2, lux, pm ?
Yesterday Mouser had more than 100 pcs on stock of the Power Profiler Kit II. I did not order, thought I had time to think. Today: 0. I hope you get percentages, Andreas! 😅
Hi Andreas. Have been a supporter for quite a while now. Have you checked out Espressif's rainmaker? I find it extremely easy to use and very versatile. It even works with Google Home for a few device types if you're interested in that. It works on the esp32s2 and esp32c3 if I'm not mistaken. They are extremely cheap! Especially when buying the module only.
Ooh. Nordic Semi Power Profiler Kit seems to have Linux support too. Suddenly I am interested. Nothing to do with the pile of 18650 and tinypico / ezsbc in my "round tuit" drawer... Ooh, a new Andreas Spiess video on battery powered ESP32? I know what I'm doing next..
Hello Professor! I mentioned you on the Radxa forum on the topic of their new board. My username is capedra. Looking forward a great review from you related to this board. Best Regards!!
Den gemittelten Stromverbrauch kann man meistens auch ohne ein spezielles Gerät messen: Die Pulsströme liefert ein dicker Elko (100-1000 uF) direkt am Verbraucher. Der Strom wird über nen Widerstand zugeführt über dem eine R-C Reihenschaltung liegt. Die Spannung über dem C ergibt dann den Mittelwert des Stroms. Man kann z.B. 1 Ohm nehmen (oder 0,1 Ohm oder 10 Ohm) über dem 10 K in Reihe zu 100 uF liegt. Größere Werte ergeben längere Zeitkonstanten. Das Ganze ist sehr einfach und funktioniert ziemlich gut..
Das stimmt um einen Mittelwert analog zu bestimmen. Dann muss man die Unterscheidung von "deep-sleep" und "operation" vermutlich von Hand machen. Ich bin mir auch nicht sicher ob man verschiedene Widerstände für die beiden Phasen braucht. Vielleicht ein Thema für ein Video? Für die Optimierung ist allerdings ein Mittelwert oft nicht genügend und man muss tiefer in die Tasche greifen. Ich bin übrigens seit Langem ein Subscriber deines Kanals und ich finde deinen "trockenen" Stil höchst unterhaltsam.
Usually the deep sleep current is the most critical. The simplest (affordable) way is to use the low range and short the multimeter during esp32 boot (Wi-Fi Connect). No brownsouts. Of Course less data/information.
What is the impact of a few capacitors before and after the 3.3v regulator to help smooth out the current spikes? Maybe you can avoid brownouts with them? I've been putting them on my esp32s, and they "seem" to help, but i haven't done any sort of quantitative measurements of it.
In the electronic goldmine (no affiliation with me) they sell power resistors (5Watts) and 0.02 ohms. Their offer is 3 resistors for 1.99 plus shipping.
It looks like you have quite an influence on the sales of the products you are presenting ;) I tried to buy the excellent PPK II you presented and all online shops (Mouser, Digikey, etc..) report no stock until at least 26 week. You should get a percentage on the sales ;) Great video. Looking forward to receive and try my PPK II.
Guess what Andreas, yes I finally found 3 piece of gold this weekend. Wait till you see my new batteries very big. Also thanks for your last comment mate on my channel.
Thank you for this video. Going to buy the PPK2 as I have no other way to test low current power consumption for my projects. Looks like Nordic has done an excellent job with the software and hardware. Any thoughts on using different connectors to reduce noise? Possible to use something like Ethernet twisted pair, or shielded TP wire plus short leeds? I'm don't have any real RF knowledge. Great job as always!
Did you set all the floating gpios to low before enabling the sleep mode ? my lesson with arduino, sleep mode current improved a lot after putting all the floating pins to low.
Would it be possible to have a matrix of features between those tested in the video and the N6705C? Lots of work, I know but it would be immensely helpful when justifying cost in professional environments.
the price to measure < µCurrent is way to restrictive for the average hobbyist... there has to be a cheaper way... i mean i dont wanna pay around 100$ to just read some low amps occasionally.. my multimeter isnt even worth 100$!... id upgrade that before purchasing one of these... very nice video yet again Andreas!.. covered everything and explained it in terms an idiot like myself can understand!
I agree about the comparison with the multimeter. So far I did not see a solution for less than the 100 dollars other than using the multimeter for the two separate phases and let the MCU crash in the microampere range.
@@AndreasSpiess couldn't you just connect a suitable low resistor over the multi-meter after measuring how it changes the ranges to get a conversion factor? For myself the brownout was less of an issue because for my projects the "idle/sleep" portion always dwarfs the active times. Switching from mA to µA works as well if you are quick about it or add a bulk cap :)
I like the simplicity of a manual bypass switch. If you’re going to switch between different states more frequently, could you use a MOSFET on a GPIO pin? Have your ESP code switch ranges automatically before it does something high power.
I used original PPK for many years - it was very useful, but the current limit of 80 mA made it usable for BLE/Sigfox/LoRa only. And the software was unstable. PPK II is much better, both current limit and software. I combine PPK with Otii, because Otii can display UART data (I don't need to create extra signals as Andreas did in the video) and Otii can measure higher currents. Once you add motors or other actuators to your project, other tools cannot measure the consumption properly.
Thanks for the informative video. Do you know of any current sensors that would accurately work in the pico amp to micro amp range for transistor measurements.
I'm looking at the Power Profiler but my device has a Vin (min) of 6V which is then stepped down to 3.3V for all devices on the board. Is it possible to use the Power Profiler in Ampere mode with a Source above 5V?
I'm trying to use the ppk2 to measure the current of an ESP32 board. But I'm struggling with the esp's USB connection. How did you adapted the USB cable to be powered by the ppk2? I'm planning to use the ppk2 in source meter mode.
The ESP32 has to be powered by the PPK if you want to measure its current draw. Often, USB cables provide also power and your PPK will not show correct values. So, no USB cable during PPK measurements.
I use a 100mA moving coil meter in series with a 100micro amp meter with a N/C shorting button across it . When I see the current is very low I can press the button to enable the micro ammeter without breaking the circuit , not so good but cheap and enough to test sleep modes
I went with the lo-fi way of "divide the length of time it takes to drain a battery by the voltage". It's a shame nobody invented an arduino based device yet like the oscilloscopes.
@@TheRainHarvester I meant the self assembly one you can now buy just about anywhere like banggood but yeah basically same thing. You would think you could make something similar for measuring low currents somehow. I will check out your channel!
All thee digital meters are not much more than an amplifier, ADC and an MCU. The Current Ranger even uses an "Arduino" chip and is programmed with the IDE. Only the Joulescope uses an FPGA for range switching AFAIK.
I have been frustrated because the Power Profiler Kit 2 has been nearly impossible to buy through the usual channels over the last 12 months. . I can now report that as of 31/08/2022 that Digi-Key has a stock level of 896 of these Power Profiler Kits. . So hopefully this information might help those who were languishing in the abyss waiting to get their hands onto a Power Profiler Kit 2. . I just placed my order a couple of hours ago after waiting for over 12 months !!!
Very useful tool. I was buying a Current Ranger, but probably this is a better solution for me. One question: is the 100kS/s rate related also to the logic analizer part? Or does the digital part have an higher sampling frequency? I ask this because I was figuring if I could monitor an SPI connection with this device as bonus
I assume it has the same sampling rate. I would go for a cheap Chinese clone for a logic analyzer because there you get also the decoding of the messages. Very handy and really cheap (and much faster).
Dang. I came across this video way too soon, even though I'm subscribed. It's not possible to source the PPK II anywhere at the moment. Expected delivery dates are late February :(
Quite interesting, how about checking the efficiency of switching power modules. Seeing if they meet their efficiency at extremely low power consumption compared to linear regulators. I know I keep pounding over the same idea. 😂💡
I understand your point. For low currents (and low voltage differences) I use a linear regulator because most switchers have a higher "quiescent" current (the minimum current to make them work). I only use switchers for higher voltage differences. Which never happen if I design a low power device.
sleep mode current measurement of a device: An 1mv FS A/D will be used to measure the DCvoltage drop across the Rgnd where R]gnd : a 100cm 20awg wire that supplies Gnd current to the device under test. 20awg wire => 52.96^milli-ohm/meter example: the A/D indicates 75uvDC voltage drop across the R]gnd wire I]sleep = Vmeas/R]gnd = 75*10^-6/53*10-3 = 1.41*10^-3 = 1.41ma
Interesting method. Do you not have problems with noise on this long ground wire and the small voltages? Why do you not use a much smaller (dimension) resistor with the same value? This is what these meters shown in the video do (they even use SMD resistors to reduce noise). They use a relatively small resistor and an amplifier + standard ADC (mV meter)
Thank you for this helpful Video! I found no detailed measurement data for ESP32 current consumption on the internet. I'm writing my master thesis and your video is the only helpful source for real ESP32 current measurements (especially for WiFi peaks). Is there any way you could provide me your measurement data so I can cite them in my thesis? Unfortunately youtube videos are not a valid source. Thanks in advance!
very nice video! Thanks. How to avoid the ground loop with the nordic? If you had to connect it to usb, and if computer is NOT a battery powered (laptop)? when you use an oscilloscope too, it's a mess. Need usb insulator (like one with Linear LTM2884) or, better, an isolation transformer i gues! Correct?
Hello sir, I have started working on esp32. I have one doubt about the circuit. I have two different sensors with serial data and serial clock pins in both. How to connect them with one esp32 at the same time ??
An interesting question. Here is the video I just watched: ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-qosClvn9GqU.html I will have a look at it during my next video.
@@AndreasSpiess I love some of this low cost kit that is available today. The video is fine as is his site - another rabbit hole to investigate! He also seems to have his own vocabulary when he mentions 'non-idealities'. ('Faff' - we can all get away with).
I think this is not in our hands because the load defines the current with its internal resistor. If we would try to restrict it we would create a voltage drop and a brownout. Espressif (or the WiFi standard decides on that.
Too bad I don't have a use for measuring power consumption so accurately. I would've definitely bought the "Power Profiler Kit II". I usually use a precision 1 Ohm resistor and see what's on the scope, and call it good enough.
That is for sure good enough for "large" currents. For deep-sleep (below 100 uA you probably need a bigger resistor. Some viewers use a switch across the bigger resistor which they press when the MPU is about to boot again.
The fuse in the multimeter has a much higher resistance than the shunt .... so if you still want to go with a multimeter : do the fobidden thing and use a much higher current rating (with a lower resistance :-)
If you're just interested in sleep mode consumption: Just bridge the shunt of your multimeter until your microcontroller enters sleep mode and then remove the bridge.