monitoring current fluctuations

Hello everyone,
We'd like to monitor the fluctuations through time of an entire pump system, the electrical system is quite complex and needs a few amp/watt probes.
The idea is to have all the probes hook up to a recorder, which could be off the shelf or an arduino custom thingie, and keep recording on a memory chip. When a problem happens we pop out the chip and read it on the computer and read the graph.
Does anyone have experience with that?
Thank you.

I have some experience in this area.

What type of current are you trying to measure, AC or DC?  How much current?  What voltage level?  How many points?

I'm partial to the Arduino Mini Pro, it's small, draws very little power and is available in 5V (5-20V supply) and 3.3V models.  You can buy them on ebay (or direct from China) for about $2 each.  

For something with a bit more horsepower I really like the NanoPi Neo, it has a quad-core Arm processor that runs at up to 1.2 GHZ and draws about 1 watt of power.  I'm using one right now as a combination Web Server, Email Server, DNS, file and media server (music an video), as well as the central data collector for my network of sensors.  With 512MB ram, the optional heatsink (highly recommended), and shipping it costs about $20.  

Or you can buy it with an aluminum case and OLED display for about $40 delivered.



The advantage of having a server online constantly is that you can connect to it at any time using a smartphone, tablet, computer, etc. and check the status right then.  This might allow you to catch a problem before it develops into a failure.
This is an example of part of my power monitoring system display:

I tried Arduino years ago and loved that it was programmed in pure C. I don't need a server as I'm a few hundred feet from the community pump and server would require installation of an internet line.
The graph is what I'm looking for, what's the max frequency of capture? How do you input this many points and what's the cost for the measuring tools? I was thinking of using induction coils because quite a bit of current spikes through.
Did you write your capture code yourself?

What I'm trying to measure is 3 phase 240V, maybe 60 Amp of current max when the pumps startup. my guess is 7-10 points of measurement, I'm kinda Jerry rigging a diagnostic tool here.

For measuring 7 points I use a Brultech ECM-1240.   It uses current transformers to measure current, and a normal transformer to measure voltage.  It takes 100s of samples per second and calculates "true power" on each channel then stores the results in a 32 bit 'watt second' counter.  It spits out packets of data at preset intervals with the current value of the watt second counter for each channel.

You can set the intervals as low as 1 second, but I have them set them at 10 seconds.  By subtracting the current watt seconds from the previous reading I can tell how many watt seconds have been used in the last 10 seconds, divide that by 10 and I get the average power level for those 10 seconds.

Other than the cumulative counters, it doesn't store readings, so you need a computer to store the info or forward it to one of the cloud based data collectors.  If you use Wifi, then you don't need to run wires out the computer.
You can buy the ECM-1240 and 14 current transformers for $180.  http://www.brultech.com/store/index.php?id_product=35&controller=product

Most of the current transformers are rated for 40 amps. They'll work at higher current, but the readings wont be as accurate.

If you want to build it yourself you can find information on building a similar system using Arduino's on https://openenergymonitor.org/
However, the measurements aren't as precise and in the end it will cost almost as much, possibly more.

Thank you for the link, I contacted Brultech.

We are trying to debug the pump house so I'd guess that the time accuracy of the curve matters more than the current accuracy. I don't know if a system exists that can measure such rapid chain of events: one circuit closes another circuit causing another switch to close the pump circuit. 1 second sampling rate is too low but I'm thinking in terms of discrete events and what might be happening is something else altogether.

Oh I see.  Well if you just want to measure timing and aren't concerned about power that's pretty easy.  

Even an Arduino can measure up to a million or so events per second, although adding timers and recording the data to an SD card, etc. would slow it down a bit, perhaps only a few hundred thousand events per second.

A Pi could easily handle billions of events per second, and many of them have built in Wifi interfaces.

Do you have a schematic of the pump systems?

A few 100K record/second should be enough no? How fast are spikes from a relay closing and filling up a motor startup capacitor? My hunch that it's where we're having problem, both capacitor on both pumps burnt again!
Also where do I start in term of programming this? Are there sources that are no crazy badly programmed (like most open source are the link you gave look like a shop and I don't see anywhere a thing that stores electrical measurement at high frequency like that, seem more like consumption monitors...