Killed my Hub?

NakedRabbit

No idea why. i2cdetect -y 1 does not show the Hub exists. Did it burn out? What can I do?
What I was doing at the time:
I had calibrated all my sensors and I was testing them, one at a time, with the Blender plugin and the KCeptor model. Only one cable plugged into J1. I ran ./notochord --scan so I could connect to Blender, and I had three Kceptors attached in a line, double-checked jumpers and polarities (ins to outs). Notochord showed only two Kceptors and I wiggled the cable, wondering if it were simply not connected well. It's then I noticed the Hub was dead, but I cannot say whether it was alive when I hit the initial notochord command.

Power goes through the Hub and I can see the Pi very easily, but no blue light and no i2cdetect -y 1 on the Hub. So it seems I have somehow ruined it, or so it would seem. Is there a way to get another?

NakedRabbit

Oh yeah Hub must have been alive or I would never have been able to see that two Kceptors were attached at all. So somewhere along the line with me handling the cable I must have caused some kind of short or something? Or is it possible I got a failed Hub that just went out and I am imagining something I did made it die? In any event, I have a dead Hub, I'm fairly sure.

daylanKifky

NakedRabbit

Hi.
The signs you describe seem to indicate that the Hub is indeed damaged. Chances are only the Buck converter is burned, so if everything goes well you should be able to continue using it with the 3v3 power rail from the raspberry.

In order to confirm the issue:

Basic test

Plug just the Hub (no KCs, no Raspberry) to the regular cellphone charger with the USB cable. Does the led light? If it doesn't the power source is probably down.

Complete test

To better troubleshoot it you would need a tester. Please give me the following lectures

Measure voltage at the test point 10 (TP10 near the mini USB).
Take off the plastic Jumper at JP1 and measure voltage at the uppermost pin (be careful not to touch others with the testprobe)
Also measure the other two pins on JP1

If you have never used a multimeter before: you have to select the "DC voltage mode", use one of the probes (normally the red one) to touch the indicated points while you hold the other (normally the black one) on a reference to ground.
In the Hub you will find a reference to ground at the test point 7 TP7 close to the pin header

Once we confirm the problem I will give you indications to use the raspberry power rail.

I would love to identify the exact action that caused this issue. We have experienced damage in the buck converter while plugging a KC while the Hub is powered, but just wiggling a cable shouldn't be enough to cause it. Have you plugged the KCs while the Hub was on?

By the way, this problem is caused by an underrated Buck converter. As you know the hardware you received is the same prototype as our team uses to develop the system so it lacks many of the protections consumer grade electronic normally have. This is something that will change with our upcoming official release.

NakedRabbit

OK, I know how to use a multimeter!

Voltage at test point 10 is 5.53
Voltage at JP1 pin 1 is a big 0.
Voltage at JP1 pin 2 is 0.8
Voltage at JP1 pin 3 is also 0.

Curiously, the blue light flickered when I took the jumper off JP1 and connected power. No idea what that means but it was interesting.

What I know of electronics you could fit in a bug's nose, but I'm pretty sure this confirms what I thought. Somehow I "live" connected the KCeptors to the Hub and this caused some kind of short. And that's consistent with what you already know about the underperforming buck converter.

I have already set the configuration up for the 3.3v rail as per the wiki, with a cable from the 3.3v pin on the Pi to pin 1 on the Hub, and have set the JP1 jumper to pins 2 and 3 in expectation. But you know what? Despite my eagerness to try this, I'm NOT going to do so just yet until I get confirmation!

daylanKifky

Hi!

NakedRabbit OK, I know how to use a multimeter!

Great! this will help a lot!
By the way, I normally try to give answers taking the less possible technical knowledge for granted. That way other users can also understand it. (or at least I try)

The voltages you read should indeed confirm the burned buck converter hypothesis.

NakedRabbit I'm NOT going to do so just yet until I get confirmation!

Wise man!!
In this case I suggest you to go one step at the time. The two power sources are completely isolated from one another, but you never know.

First of all attach just the 3v3 and GND wire from the RPI. Test the voltage is 3V3 at the pin 3 of JP1.
If that's the case put the jumper on JP1 as shown in the wiki and connect the rest of the cables. The led should turn on.
do the i2cdetect test. It shold return a positive 73 for the Hub

If everything does well shut it down, and proceed to calibrate the remaining sensors, always making sure everything is powered off before plugging the KCs.
It's tedious, I know.

NakedRabbit

daylanKifky By the way, I normally try to give answers taking the less possible technical knowledge for granted.

Of course! I apologize if that sounded in any way sarcastic. I was just excited that my meager knowledge of technical things was actually going to be of benefit in this case.

And YES, HOORAY! Plugging it all in as suggested gave me an immediate blue light on the Hub and the KCeptor. i2cdetect gives me the 73 and I'm on my way!

daylanKifky It's tedious, I know.

Not at all. It is the process and what must be done, so it shall be done. Let us consider this case CLOSED and I'll keep updating my build log.

prm3d

I have a similar problem. I originally connected the Hub to the Rasp, using a dedicated power source like the vimeo tutorial. It was working properly and calibrated 4 Kceptors. Then when calibrating the 5th I decided not to turn off the power before disconnecting and connecting the new Kceptor. When I tried to, the blue light turned off. I tried powering only the hub by itself and the blue light doesn't come on and I'm worried I might have shorted the hub or something. I don't have a voltage meter yet to be able to check for further problems. I tried the other method of powering, by the 3v3 rail from the SBC, but still am not getting any indication the hub working. The raspberry turns on no problem and I tried "i2cdetect -y 1" but i get a response with a bunch of dashes, which I assume means nothing is responding. Did I kill by hub? What can I do to fix this issue? I've also tried the various jumper cable pin configs but still no blue light.
My biggest worry was that it was working fine, and now it isn't.

daylanKifky

prm3d you were not referring to this single pin were you?

The black plastic jumper on JP1 should be placed just as in that photo. The jumper cables from the RPi should all go to the J9 header.

If you do have the setup like this your situation is quite strange. We have experienced issues with the buck converter on the Hub when live plugging a KCeptor (That's why we included a warning message in the kit package), but that part of the circuit is completely isolated from the rest at JP1, so the other powering method should not be affected.

In order to keep conversations tidy, could you open a new thread and post photos of your wiring?
If you are able to find a multimeter it will help a lot to troubleshoot the problem.

prm3d I've also tried the various jumper cable pin configs but still no blue light

What other jumper cable configurations have you tried?

daylanKifky

When trying the raspberry power source did you put the jumper on JP1 on the lowermost position?

prm3d

daylanKifky I believe I have. Just to
make sure, you were not referring to this single pin were you?

daylanKifky

NOTE: the resolution of @prm3d's issue is here https://forum.chordata.cc/d/111-beta-test-hub-led-not-working

If anyone else is having a similar issue please open a new thread!

Important note for the participants of the 2nd Betatesting program: REMEMBER NOT TO LIVE-PLUG KCEPTORS TO THE HUB doing it might burn the Buck converter on the Hub.