New build trouble - OK standalone but LCD fails with RPi

Hi!

I've just build my 4.2 kit but the unit behaves very erratic, especially the LCD, with the RPi connected. The unit seems fine in stand-alone, ie without the RPi attached and seems to be mainly working OK internally with the RPi attached. Obviously the connectiong between the HM and the RPi have worked/is working since it has flashed the mega. I can also re-flash from the web ui and everything seems ok from the flash programming output. I also get temperature readings and fan level readouts the seems correct on the web page with flames.

But, as soon as I hook up the RPi, the LCD on the HM fails. Mostly it just displays a blinking block cursor in one corner and nothing else. Sometimes, by randomly pressing the 4-way button in different direction, I get some very strange screens on the LCD, see below. The LEDs still seem to show the expected output, ie lit yellow for cover open mode when going all the way to the left.

I tried to find any logs on the RPi which could indicate where the error lies but it seems like there are no logs generated from the communication between the HM and the RPi, right?

My second thought is that there be any electrical error where the RPi attaches to the HM that only affects the HM:s communication with the LCD. Any suggestions on what to do to confirm or reject that?

Any comments and suggestions appreciated

/Thomas

First thing I would want to do is check all of your solder points. Start with the 26 pin connector that the rPi plugs into. Make sure there are no cold solder joints (all the solder should be shiny). Next, I would check the solder connections on the processor and the 8-bit counter shift register. Make sure they are all good solder joints with no bridging between solder points.

Also, be sure that you are powering the unit from only one power source when you are testing. I use the 2.1 barrel connector on the HeaterMeter to provide power to everything.

Just from the picture you have, it looks like the bs170 may be shorted or looks like it could give you issues in the future. Post some close up pictures of the boards front and back.

Thanks for the support. I used the HM live for the first time today! Yessss! Without the benefit of web access, though.

Here are some additional pictures.... Can't say I'm fluent with a soldering iron but I do examine every solder with a lope, to see if primarily there is a cold solder and secondarily if there is a risk for shorts.

Is there a way to measure the function of the bs170 in isolation?

Here are some additional pictures, let me know if any other angels could help:





As per the RPi A+ addendum instructions I mounted the buzzer and the BS170 on the opposite side. Took some thinking to decide that the BS170 is mounted the right way, had it the other way around first. Is the BS170 related to the RPi and LCD communication?

Also, the buzzer don't work currently, when I am in the web ui, on the alarms page and press the buzzer icon. Might be related?

/Thomas

I think you have it backwards still. The flat edge is suppose to go up towards the LCD. Also, the speaker won't activate by hitting the test alarm button. It's only for testing scripts, email etc.

You have to flip it upside down otherwise the (D)rain and (S)ource legs would be reversed.

Steve_M said:

You have to flip it upside down otherwise the (D)rain and (S)ource legs would be reversed.

Click to expand...

Ah ur right, sorry nevermind.

Are you sure that the display isn't touching the RPI somewhere?

Does the web interface work normally when the HM and RPI are connected? IE do all the probes work, can you switch the LEDs on and off, does the blower work? If the answer is yes then you know that the problem is not caused by the ATmega 328 or the shift register.

As pointed out, the long exposed legs on the buzzer could possibly touch the rPi board and cause a problem, you'd have to look very carefully at that to know for sure. I would move the buzzer to the other side of the board as suggested.

On the transistor, when you move it to the other side of the board you basically have to bend the center leg in the opposite direction as it is in when soldered to the top side of the board, then it can only fit in the board in the proper orientation (on the back side of the board)

The HM LCD will flake out like that if you have poor connections, either where the LCD meets the HM board, or on the shift register, or on the ATMega pins associated with the shift register. Attaching the rPi could make or brake a flaky connection, so I suggest that you reflow the solder on the LCD, Shift Register and ATMega before you go any further. Of coarse make sure the power is off first, and I would also remove the ATMega from its socket to protect it from heat before I solder its socket.

Thanks for all the suggestions. I found the problem, It was related to the the RPi shorting a few connections when pushed fully into its connector. More precisely, it was the legs of the leftmost (as seen when looking down at the LCD and button) LCD daugtherboard pins that came in touch with the shielding case of the USB connector in the RPi.
Cut them as close to the board as possible and then some electrical tejp to ensure isolation and everything runs fine.
At least until the min-SD card get touched and pops out, which has happened a few times...

Thank you all for the support! Great project!

/Thomas