Thank you for contacting us. The difference between the two voltages shown is that one is for the system power of the module and the other one is the voltage needed to power the expansion board. If you need to power the expansion board, use a 12V power supply as explained in https://software.intel.com/en-us/getting-started-on-joule
You can see more information on the power signals of the Joule here https://software.intel.com/en-us/articles/intel-joule-module-datasheet
Actually, the voltage into the expansion board can handle a wider range than what is specified (Sergio, can you dig up the exact specification? It would be useful to have it documented somewhere...) which allows the expansion board to be run off a battery... in particular, a 3S LiPo works fine (I use one in my Joule robot), even though the voltage varies from 12.4V down to 11.1V as the battery discharges.
No problem, thank you for letting us know this is what you were looking for. Don't hesitate to contact us again if help is needed.
The ranges of the different power signals in the Joule can be found in the following chart:
The signal that comes from the DC jack of the expansion board is called +VDC_IN.
It is interesting to see that the module can sense VDC_IN over the range of 4V to 20V. But while useful, this is only sensing, it does not indicate what the acceptable range of supply voltages are for the *expansion board* (I would expect them to be somewhat less, but not outside these ranges). I looked all over and could not find a specification for the allowable input voltage ranges for the expansion board. All the docs just say "12V" but a bit more detail than that would be useful for people wanting to run their systems off a battery or some other alternative power source.
I agree it would be useful to have this information in the community. I personally haven't tried to use a power supply different from the recommended 12V. I'll try to find more information on the matter and post it here.
Thank you for your patience.
From the board schematic, it can be seen that the DC Jack input is connected to a power mux which has a voltage input range of 4.0V to 12.4V. The datasheet of the power mux can be checked in this link: https://www.fairchildsemi.com/datasheets/FP/FPF3042.pdf . This component can be checked in page 7 of the board schematic: http://www.intel.com/content/dam/support/us/en/documents/joule-products/intel-joule-expansion-board-schematic.pdf
Thanks. Based on the power management on page 4 of the same schematic, the mux then routes input power to a buck convertor which brings it down to VSYS, which then goes to a boost convertor (to go back up to 5V), to a buck-boost (to go to 3.3V) and to a buck convertor to 1.8V. I could not find a specification for VSYS on the expansion board but according to the Joule module data sheet, VSYS should be around 4V and no lower than 3.6V (and not more than 5V). If we assume VSYS is 4V, then the input probably needs to be a *little* higher than that to provide room for the buck converter to operate, say around 4.6V at least.
Therefore, it looks like the board can probably be powered from around 4.6V (up to around 12.4V... I have confirmed the upper bound when using fully-charged 3S LiPos).
So, in *theory*... it should be possible to power the board from something as low as 4.6V, and from a 5V supply in particular (although one with a relatively high amperage, solve for x in 12*3 = 5*x, so x = 7.2A). I'll have to try that to confirm the lower bound... although in general, it's a lot easier to find a 3A 12V supply than a 7.2A 5V supply. I would expect 5V to work though since the mux also routes 5V from the USB-C input as an alternative power source, although current may be relatively limited (at 5V) from that source.
One power source this should enable, BTW, are 7.4V NiMH batteries (which is handy, as they are easier to charge in-place than LiPos).
As for current requirements, I did do a basic power consumption test; see attached image, where I show an ammeter measuring the overall average current to a running system. This was while running a benchmark where I ensured the core clocks were all juiced up to 2.4GHz (eg in Turbo). As you can see from the ammeter, the (average) supply current used is 0.636A at 12V. Note I am also running a fan and a Transcend 12V powered USB 3.0 hub (which is in turn connected to a keyboard and mouse).
HOWEVER, I have also noted that if I use a 1A 12V power supply the board may fail to boot, so I think it needs a little more current during boot for some reason, and it may have some "spikes" at other times my ammeter was not able to capture. You of course may also need more if you are running a RealSense camera, other peripherals, etc etc.
This should not be considered an official number, just a data point. Also, the current consumption did fluctuate a little, but generally was under 0.8A.
Thank you for taking the time to run these test and sharing your results. These posts will help as a valuable addition to what’s already in the power section of the datasheet of the Joule.
We encourage you to remain involved in the community.
Have a nice day.
I sincerely think Intel can and should do more in being more specific about acceptable voltage range and options.
Why doesn't Intel take one Joule and test it by feeding it with 5V, 9V, 12V and even 16V and 20V and put it in black and white in the specification? including actual numbers that may cause it not to boot properly.
In one side the documentation suggests it goes from 4V to 20V, but in the other side, the support personnel just can't actually say "Yes, this thing also runs on 4V and on 20V" and confirm that.
This is a pretty expensive and state of the art device. Support and Documentation should behave as that too.
"So, in *theory*... it should be possible to power the board from something as low as 4.6V, and from a 5V supply in particular (although one with a relatively high amperage, solve for x in 12*3 = 5*x, so x = 7.2A). I'll have to try that to confirm the lower bound... although in general, it's a lot easier to find a 3A 12V supply than a 7.2A 5V supply. I would expect 5V to work though since the mux also routes 5V from the USB-C input as an alternative power source, although current may be relatively limited (at 5V) from that source."
Sorry to cut in here so late, but I just noticed that thread, having some similar issues with the power supply on joule
(Power Supply, again)
Anyway, you missed the fuse F1 on the extension board, which is rated only at 3.5A, so the calculation of 7.2A seems to be a little high for the normal use.
Also, the coil on the VSYS Regulator is only rated at 4.1A, and most of the voltages are derived from there, it seems that your calculation was very pessimistic?
Did I miss something?
To reply to crmakers, the development carrier board was designed and tested for a specific use case, running off a 12V 3A power source as a desktop development system, and that's what is in the docs. I agree that it would be convenient if Intel could provide some additional detail on the acceptable voltage ranges for extra use cases, but in the meantime, given that the community does have access to the schematic, there's no reason we can't figure out some "bonus" use cases ourselves.
I have already confirmed at least one useful additional use case: running directly off a 3S LiPo battery (11.1V to 12.4V). This is only with my own personal testing... but I've been running a robot with that configuration with no problems for a while now. This is still pretty close to the original spec though. It would be nice to know how far we can go to unlock some additional use cases.
To e2k8, good point about the fuse. While the board probably does not actually need 7.2A at 5V (just like it does not need 3A at 12V, at least not all the time) running the board off 5V is probably a stretch. I suspect 5V is just too low a voltage to give good power density, so I would not recommend it. (Update: I know of at least one case noted here on the forum where a constant-rebooting problem seems to have been due to trying to use a 5V power supply).
However, the 7.2V use case would be useful (I'm already working with another NiMH-based robot where that would be convenient) so that's what I'm going to be focusing my testing on.
Of course you can always just run using an external regulator (and I am testing some options there as well). The question is under what situations an external regulator is or is not needed (with this carrier: of course, with another carrier, you can do what you want in terms of power management).