Dear Nick,

Yes, I have. Actually the usual way to run them, here in my country, is via dedicated adapter as we have discussed before.
Free USB port on routers, etc is a rare option here. I have even the hosts put them on surge protectors and UPS's if handy.
Did that myself as well. Yet, dead probes come up every few weeks.

Thanks for sharing the useful links, (the failure rate on that graph is a disaster) that still brings up two more issues:

1- Though that's an insight, what good can that do for the problem of physically unreachable probes (including those with distant non-tech hosts)?

2- If there's an ongoing research about such an important issue, why didn't we ambassadors get a serious notification on this list, for God's sake? We could have been informing people in advance. Just saying...

I have distributed both Verbatim & SanDisk probes and 4/5 of my batch is down already. I think that is on the Atlas Monitors as well, yet the atlas servers feed me with single "your probe #xyx is down" e-mails. Still here to do more, I am also really disappointed with the energy and time I spent for dead probes, let alone the thousands of kilometers of land travel to spread them.

Soon we will probably be replacing every 3rd gen. probe on the planet. Am I the only one panicking? Bringing the white "black box" we gave people up every now and then, and expecting yet another favor, won't help us in the long run. They have trusted us and we gave them some cheap toys.

If the project wants to reach every corner, not just healthy and wealthy spots, then the probes should be as stable and standalone as possible. Yes, failures happen, but the right piece of hardware for this project, I believe, is what works with the minimum failure rate as long as the power grid is on and the router is connected, and the host forgets it's even there. End of story.

Yours,
M. Tajbakhsh


Nick Hilliard  wrote:
                
> Budiwijaya wrote:
> > This, I cannot guarantee the probes will be connected to a stable power source.
>
>you need to fix your power supply problems to the probes, then  Like
>anything else, if a probe doesn't have reliable power, it is unlikely to
> work properly.

> Nick


Nick Hilliard  wrote:
>                 M. Tajbakhsh wrote:
> > Months into the project, I wish I have hit the 50 or even 100 probes by now.
> > Yet, I'm stuck with faulty probes here and there and I am asking too
> > much from non-technical and geographically dispersed hosts to literally
> > repair their uninvited out-of-my-reach pets. Buying a new USB memory is
> > a whole new story.
> > 
> > Guys, please either fix the software or do logging with probe's internal
> > memory, or whatever.
> > This is becoming a huge drawback.
> 
> The RIPE NCC people already gave information on this:
> 
> > https://labs.ripe.net/Members/philip_homburg/further-analysis-of-ripe-atlas-version-3-probe
> 
> > https://labs.ripe.net/Members/philip_homburg/troubleshooting-ripe-atlas-probes-usb-sticks
> 
> 
> In other words, this may be caused by the USB power source.  Have you
> tried using a dedicated mains socket to USB converter for the units
> which are showing problems?
> 
> Nick
>