@ Nicolas - Great to know the MS is supporting the spec not just with software releases!
A module like GMod/MS would have been helpful to me a while back. I had a Spider board, T35 display, SD module Seeed GPS and a Seeed GPRS modem all powered by a GHI DP module. If all but the GPRS were active I was fine, but once I started up the cell, my system would restart - this was the case with USB and a 9V/2A wall wart.
I believe GHI is working on more Red modules to increase (inject) power, but a diagnostic board would also be helpful. I imagine that It would also help to determine how much battery run time one could achieve.
I was thinking like how the CurrentMod sits between a socket and a module, this module would as well.
The one thing that needs to be addressed for this is that if a Gadgeteer app is expecting a module to be there and itās not, some times apps will throw exceptions on start-up.
**
On second thought, would a board with several corresponding in & out sockets with one control socket make more sense. This way you could turn on/off several different modules with only 1 control socket.
Iām not sure a full Gadgeteer module would be the best solution. A board like an extender module with the power lines brought out to terminals would let you insert a current measuing device in the circuit with ease. It would also be very inexpensive and transparent to the module be monitored.
Transparency is the idea. My module is as you describe, except that it already has the shunt resistor and an amplifier added. There are test points where you can connect a meter is you want, but the readings are going to be a bit oddā¦
You can always remove the shunt resistor from my module then you can just connect an ammeter between the test pointsā¦
The testing Iām doing is around @ ransomhall new MakeGPS module (way better then Seeedās) and how the different power supplies affect itās performance at capturing a fix. So Gadgeteer to the rescue as I built a device that cycles the power x minutes on then y minutes off and the GPS device records capture time (GPS Time), and such then I can take those results and compare the differences in performance for each power supply. Now of course it would be the catās meow to be able to measure and record power requirements, but until that module is available, we make do with what we have. I have to run each setup multiple times as satellite position do affect capture times and so to eliminate that factor I run tests over a period of time and use the average for comparing.
Gadgeteer rocks and this is why I buy just about every module I can find as there is no limit to what you can do with Gadgeteer and the most frustrating thing about Gadgeteer is how long it takes to ship modules here. @ Nicolas I would be very interested in those modules and tools, so please let me know what you need from me to help make that happen.
In my opinion that still makes a module more complicated than needed to do the job as you then have to implement some sort of measurement device external to the board and worry about calibrating the on board shunt/amp and the scaling factors that they will add to the mix. Everyone interested in measuring voltage and current should have a multimeter. Using a multimeter to measure what it is designed to measure gives you properly calibrated and scaled readings with a measurement tool that adds no additional load or complication to what your trying to do. Over engineering a solution just adds complication. Again, this is just my opinion (based on several years of experience).
NOTE PRELIMINARY RESULTS - Donāt sell the farm based on these numbers, your milage etc may vary
Based on a sample of about 20 tests per board using USB connection via a USB wall wart.
DualUSB average 3:33
SingleUSB average 3:01
Love Module average 1:55
Now I have to admit this is somewhat surprising as I thought the Love module would been a bit faster given its linear vs the others being switching, but I wouldnāt have guessed that much of a difference. So perhaps I need to be a litte more rigorous and run some bigger samples and see if those differences remain and how much it changes for a given setup over a longer period (also how much satellite position impact performance and what does that look like). There could be a horse in the pony race, so Iām going to look into it a bit deeper.
As for the debug module, yep we all have meters (I even got a new one recently as people were laughing at my old analog meter from 40 years ago, but then again Iām a software guy and old enough to remember when computers had real bugs in the hardware), but Gadgeteer for me is about building stuff and so Iāll definitely be getting some of these modules as Iāll use them for testing and for prototyping device health sub-systems because Gadgeteer is just so dam cool and fun.
That is easy enough to say, but harder to implement, especially for someone that is new to electronics.
[ol]There are two rails, 5V and 3V3. Most people should have one multimeter, but not everyone has two.
To measure current the multimeter must form part of the circuit. So if you remove the multimeter then your circuit dies. Now you sit with a module that only works when there is a multimeter connected, or a link installed. You canāt use one meter to first test the 5V then the 3V3 as you loose the other when you measure the one. Your module resets when you install the meter into the circuit etcā¦[/ol]
If you want it simple then there should have been 0.1 ohm shunt resistors installed, and no amplifier. That would give you 0.1V per amp and your module would still work even with no multimeter connected.
But that will not be able to give you the avarage current over an hour or day, as might be needed when testing a gps or cell module, as their currents vary wildly depending of what they are doing. A cell module for instance can draw 50mA most of the time, and spike to 2A during comm bursts.
This module that I did does not need calibration, if you can accept Ā±1% error, and can give you ossiliscope type current readings. I must still do my cost calculations, but they should be around $15, without cables.
Under engineering just means that you must do it over later. If you are going to do something, do it properly the first time. This is my experience.
It is my understanding that the DualUSB introduces more than a little noise on everything in the general area. Your results donāt surprise me much, except for SingleUSB. If you have a scope, you ought to put it on DualUSBās output.
Ran the Love module again with some changes (power on 12 minutes, power off 3 minutes) for 33 cycles:
Average 0:02:24
Min 0:00:46
Max 0:08:46
Dev 0:01:49
So now Iāll do the same for others (running Dual now). Ultimately what I should do is get another one or two more GPS modules so I can run these tests concurrently.
Now I should say I have a number of these power modules and I didnāt pay any attention to which version of each I grabbed (meaning they might not be the newest versions) so the Dual is a v1.3, the Love is v1.2 and the single is a v1.2
@ godefroi being a software guy, I donāt have a scope, but if I were to get one, what would you recommend?
And the second chart is the whole test (ie all 3 modules) to see if there is a day/time relationship that jumps out. Now Iām curious as to why the Single USB had such a large performance delta between the early trial and this run (it did however have more non connect cycles 4, then the dual and love modules which both failed to get a fix on 1 cycle each). Likely there are some other factors influencing the results, but that the dual seems to be consistently (as far as two trials can be consistent) the slowest, but at best I would say its a tweak level improvement (ie I wouldnāt trade a big bag of gold for that level of improvement).
If you use USB power on the DP then the switching regulator is not being used, only the linear regulator, so all your power modules use only linear regulators.
Your discrepencies might only be related to the number of satellites in view or the exact switch on time etc.
Have you considered writing a program that times the gps lock, then does a cold reboot, and repeats? That can log all start up times, or extract the max,mean,div and only store, or display, those. then leave it on for 24 hoursā¦
