Power Issues. JuiceBitz Quality Control

[Vic]

When I upgraded to a Rosetta and a Pi3, I bought a batch of JuiceBitz cables to try out various installation ideas.

The one I preferred in the end used a shorter 0.5m cable, but with that I have been plagued with 0x50005 issues and more recently, Wifi Losses.

I flew today and it happened again just after landing. Having time to investigate, I noticed that the Micro Usb plug was warm. Pulling the plug and touching it, it was hot!

Getting back home I tried another longer JuiceBitz cable and no such issues and the Throttled value went away.

Close examination of the suspect plug and the socket in the Pi revealed no issues so I decided to dissect, carefully removing hot melt glue from underneath the casing. The first image reveals all!



You can see a thread of braid is loose and able to touch the end of the green wire's connection, The green wire is data I beleive but is possibly also connected with a sensing circuit, usually a resistor.  When I connect the USB-A of the cable into the power pack and touch this loose wire to the Grreen wire, the power pack activates so it senses a load.

Moving the loose wire away and plugging the lead into the Rosetta, the plug still gets hot and the device throttles. This, I beleive, to be due to poor soldering of the too short Red wire where solder is used as a bridge.

Hopefully a one off but it shows a bit of a quality control issue even with these cables!  So, even if you are using a good power pack and a (clearly good spec) JuiceBitz cable and still experiencing power issues I suggest you try another cable as a first resort.

***Edited to add 3rd Pic, the main culprit, really bad (dry) soldering on the +ve lead which is where the heat is emanating from!

Vic

[Ian Melville]

I have had 2 PilotAware supplied 1m JuiceBitz cable fail with 0x50005 issues. I put it down to 1m being too long if the cable was on the fringe of tolerance.
I tempted to strip them back to see if I have a similar cause.

[JCurtis]

My first thought was fake cable, there are 1000's out there, but the power leads do actually look like 20AWG. 

Still not great construction, surprised they are not machine made.

[Vic]

I have made up a small rig comprising a SMD Micro USB socket on a piece of breadboard to allow me to reliably measure the resistance of the cable cores.

I've got five JuiceBitz (All genuine, bought from their Amazon Store) ranging from 0.5 m to 2m in length. Only the shortest one gives the throttling issue.

All the good ones measure a resistance of the Positive lead at between 0.1Ω to  0.19Ω  ..The -Ve is even better as it also uses the braid.

The bad one detailed in my first post measures 2Ω. This doesn't sound like a lot, but at 5Volts, with the PI needing to draw around 600mA, this could be knocking the voltage down to 4 volts or less!   

[exfirepro]

Good spot Vic,

I have to say, I haven’t ‘til now come across any issues that would have led me to suspect poor quality with any JuicEBitz cables (and I’ve used lots of them), but I have come across several instances of bad connectivity between the microUSB plug and the Raspberry Pi socket - due probably to repeated disconnection and reconnection of the power cable. The first time I came across this, I tried a new cable, but to my surprise found that the problem was that the socket had become ‘stretched’ - and so no longer properly held the cable. This surprised me as I had always believed that normal design principles for plug / socket connections were to design the plug to be the weaker part as it is always easier / cheaper to replace a lead than a board. Apparently not so with the Raspberry Pi 

I will certainly bear this in mind in the future.

Regards

Peter