Nordic PPK2 - Updated Power Profiler

[Kean]

Nordic have released their new Power Profiler Kit II at about US$90.  https://www.nordicsemi.com/Software-and-tools/Development-Tools/Power-Profiler-Kit-2
DigiKey have some stock, but are selling out fast.  Elsewhere it is backordered anywhere from a week to 3 months.

It can measure in 4 auto-switching ranges from 100's of nA (resolution, not FS) up to 1A, sampling at 100 kSa/s.  It can act as a programmable source 0.8V to 5V (with boosting from USB power), or just as an in-line current sensor (still limited to 5V 1A).

One neat thing is that it have 8 digital inputs that can be used as a simple logic anaylser for correlating system state with changes in current draw.

It is worth noting that it cannot measure reverse current (i.e. battery charging), and also is only specified for +/-20% accuracy.  That is worst case I guess, and should generally be better.  It even has some internal calibration shunts.  And while the firmware is apparently closed, the host application is open source and thus so is the communications protocol if you want to extend/rewrite it.

Full user guide at https://infocenter.nordicsemi.com/index.jsp?topic=%2Fug_ppk2%2FUG%2Fppk%2FPPK_user_guide_Intro.html

It is not quite a Joulescope in functionality (actually a long way off), but reasonably low cost and quite compact.  I've ordered one to carry in the backpack.

[Kean]

Oh, and the schematic and Altium design files of the PCB (Nordic PCA63100) are also available at https://www.nordicsemi.com/Software-and-tools/Development-Tools/Power-Profiler-Kit-2/Download#infotabs

[exe]

Interesting, U2 doesn't have type/name on the schematic. I'm on the edge if I want to buy it. I'd like to play with it, but not now. But leading time of 20 weekds and bad availability make me thinking if I should buy it now.

[nali]

Interesting, U2 doesn't have type/name on the schematic. I'm on the edge if I want to buy it. I'd like to play with it, but not now. But leading time of 20 weekds and bad availability make me thinking if I should buy it now.

If you view the schematic in Adobe Reader you can click on a part to get the full details including Digikey & Mouser part numbers - it's an ADA4622.

[Kean]

Well Digikey has now sold out of their initial stock, and Mouser website says they expect stock arriving mid March.
Nordic did say during the launch webinar that European distributors should be getting stock soon, so presumably there is more stock working its way through shipping channels.

[exe]

Blame me, at the last moment I pulled the trigger I bought one of very few remaining kits.

[whims]

Was able to get one before they sold out the first time. This thing is selling out fast, but I see there are more in stock now.
From my initial testing, it works really good for a tiny 100$ kit. The 20% is most definitely a minimum accuracy. Testing with some resistors and a bench supply, it was accurate down to half a microamp at around 1% accuracy. Haven't gotten to test more dynamic measurements, but I expect that this switching of ranges might cause some errors. Hoping they will release a CLI or similar for it, because this javascript stuff is not for me :/ Still, definitely worth the money you pay for it.

[Kean]

Good to see more stock is available.  Those lead times were obviously quite conservative... same as the accuracy specs...
I had a bit of a play with mine last week when it arrived, but I've not done any actual testing yet.  I will probably use it properly over the holidays when things are quieter in the office.

[Weston]

I also impulsively bought one (people are talking about them a lot on twitter).

The software provided by Nordic is pretty meh, but it looks like someone has written a python API for it, which should make it a much more useful tool: https://github.com/IRNAS/ppk2-api-python

[Radiohead]

Has anyone found some time to play with the PPK2? Is it an improvement over the PPK1?

I always had problems connecting with the PPK1 and the Power profiler software from Nordic. Quite often I would press the start button and no data would be displayed or gathered. Then I would need to disconnect and reconnect the PPK1 until it worked.  I hope this is resolved now that it's a stand alone tool.

The PPK1 is still quite a useful tool despite it's limitations. It also worked wonders for testing the timing of certain actions as I didn't need to add some GPIO toggling to determine the timing with a logic analyser, the power consumption would often tell the same without having to add anything to my code. Though I did quite often run into it's current limit with LoRa.

Sadly on Mouser the next batch is due to arrive at the end of March (28-04-2021), so is it worth the wait or not?
Are the better (affordable) alternatives? Didn't dave's current PSU project also have some power consumption measurement capabilities? Dave's quite silent about that.

[exe]

My ppk2 collects dust. I bought it just because I was worried about availability in the future. I can give you mine for a month, if you want.

[Radiohead]

Thanks for the offer, but I have a PPK1 available for testing.

I'm just curious what people's experiences are and whether it's an improvement to pick-up later on.

[whims]

For the price, it's worth the wait. If you don't care about pricing, I guess there are better tools around but I am using my PPK2 all the time. Not too much for reviewing battery lifetime but for timing and having an adjustable power source.

[justanothername]

Hello everybody.
Sorry for reactivating this old thread. The hardware is practically open source. Do you think it is feasable to modify it to have a range of a decade lower?
I only need 100mA for the highest current but 10nA resolution in the lowest range.

[Tweepy]

Sure, the hardware schematics are available (not open source), but not the device's firmware source code. I don't see how you could change theses ranges (resistor values?).
The desktop app is available here: https://github.com/NordicSemiconductor/pc-nrfconnect-ppk
That's definitely a pity the device is not more open, either the interface or the firmware, at this price point (80€), with the 0.8V-5V power supply, it's definitely a great tool for any embedded developer, and a must have for any battery project!

Look like NRF recently updated the Power Profiler app (from nRF Connect) with Version 3.1.1 and new features (screenshot attached):
- CSV export
- real time vs data logger mode
- trigger mode (on current, not logic input) with pre and post trigger buffer
- adjustable sample rate

Now that most distributor have it in stock as from Q2 2021, has anyone anything to say about it, used in real projects?

(edit, link to desktop app source)

[pallav.aggarwal@gmail.com]

Contact their customer support on dev zone and they can help you if there is a way to change the range.

There are a few more other power profilers, see here:  https://pallavaggarwal.in/dc-current-analyzer-for-embedded-iot-product-development/

In my opinion, Joulescope is the best and then I like Nordic Power Profiler Kit 2 the next.

[justanothername]

In my opinion, Joulescope is the best and then I like Nordic Power Profiler Kit 2 the next.

We got a Joulescope in the company as well as an Otii. Joulescope is superior in my opinion. Better resolution, higher voltage range and free API. The Otii has a PSU built in, but for current sinking or battery emulation you must pay a pricey license.
The battery model is crap imho. I sent the guys at Otii an extensive review years ago, but the model hasn't change since. The joulescope support from matt liberty is one of the fastest I encountered, so another point for Joulescope.

Anyway, these devices are expensive, and there is demand for a lot more for each firmware programmer. Since many of the projects are with tiny cells (like coin cells), the lower range from the PPK2 is just not sufficient.

[martin2019]

Full disclosure, I am the designer of a product in this space, the P1125.

I recently evaluated the Nordic PPK2 and, like some of the other products mentioned, there are artefacts in the measurement profile when the "auto-ranging" feature thresholds are triggered.  Basically these designs use FETs to switch in different sense resistors.  But the FET gate control signal will inject into the measurement path because of large Cgs of the FET. 

How do you distinguish those artefacts from your target's current profile?
Some of the other products additionally don't have the BW or sampling rate to effectively track the current usage of the target, under sampling, resulting in misleading battery life predictions.

Consider the P1125, www.sistemi.ca/p1125

It doesn't use a "auto-ranging" circuit/algorithm - there is a proprietary circuit that can track 7 decades of current magnitude, 1uA to 3.2 Amps.  The P1125 GUI presents a Log scale for current, allowing one to see the targets full operation.

The holy grail of current measurement (IMO) is the Agilent N6705B.  Its very expensive, >$20k, but it is amazing (although a bit tricky to use).  I have one of these beasts and use it to qualify P1125 and competitve products.

The P1125 is positioned to provide as good performance as the N6705B but at a cheaper price point and ease of operation for the software developer to use at her desk.

[Swake]

You write your product is better than the Nordic PPK2. Well it better be, it is 23x more expensive too.
The N6705C, the newer version of the Agilent is 'only' 7,5k€

[hary]

Hi

I'm not clear with the PPK2 spec.

It says it can  act as Source meter mode and Ampere meter mode.

In Source meter mode, it can deliver between 0.8 to 5V, supposely thanks from the 5V USB cable the PPK2 is powered from.

In Ampere meter mode, you need to power the DUT by yourself.

So can I give the DUT more than 5V ?  I guess so, but could somebody confirm please.

[Venturi962]

According to the Block diagram here: https://infocenter.nordicsemi.com/index.jsp?topic=%2Fug_ppk2%2FUG%2Fppk%2Fppk_block_diagram.html

Power In is labeled as 'Max 5V'

My guess is that the purpose here is for a lower noise supply than using USB Voltage.

[palpurul]

Hello everybody.
Sorry for reactivating this old thread. The hardware is practically open source. Do you think it is feasable to modify it to have a range of a decade lower?
I only need 100mA for the highest current but 10nA resolution in the lowest range.

May I ask what's the application for such range?

[justanothername]

May I ask what's the application for such range?

Actually, last summer we had two interns in our company do the modification. One did the hardware modification, the other did an extension for the ppk2 software for entering new calibration values. It was shown that it is indeed possible to reach the desired resolution. We need such a high dynamic range for profiling ultra low power applications with sporadic RF uplinks. A multimeter would not allow the uplinks to complete, because of the high burden voltage and the slow autoranging. For everyday use, the Joulescope is the tool of choice.
Actually I got a ticket to write a follow up article and to publish the modified sources for the PPK2. I will try to get it done the next weeks.

Edit: one part of the modification can be found here: https://github.com/ThoMint/PPK2-Mod. As said it is uncomplete and I will consolidate the work done in near future.

[palpurul]

May I ask what's the application for such range?

Actually, last summer we had two interns in our company do the modification. One did the hardware modification, the other did an extension for the ppk2 software for entering new calibration values. It was shown that it is indeed possible to reach the desired resolution. We need such a high dynamic range for profiling ultra low power applications with sporadic RF uplinks. A multimeter would not allow the uplinks to complete, because of the high burden voltage and the slow autoranging. For everyday use, the Joulescope is the tool of choice.
Actually I got a ticket to write a follow up article and to publish the modified sources for the PPK2. I will try to get it done the next weeks.

Edit: one part of the modification can be found here: https://github.com/ThoMint/PPK2-Mod. As said it is uncomplete and I will consolidate the work done in near future.

Thanks for the answer! I actually know about how burden voltage causes the DUT to reset, so that's where the auto-shunt switching comes into play.

I was wondering what type of low-power application goes down to 10nA of sleep current.

Thanks!

[justanothername]

I was wondering what type of low-power application goes down to 10nA of sleep current.

Thanks!

Two quick examples:

1) Coin Cell and Ambiq AM1815 RTC + MCU. The AM1815 has an integrated load switch that cuts the power of the subsequent circuit unless a wakeup alarm is reached. Quiescent current: 14nA (with RC Oscillator) or 55nA (with XTAL)!

2) LDO  TPS7A0233PDQNR (Iq: 25nA) + STM32F011 MCU (Iq in Stop: 344nA)

With a resolution of 100nA you are basically blind for these cases.