I’ve been happy with my Ring Doorbell camera, and when one of our motion lights stopped working, I decided I wanted to use the Ring motion detecting Floodlight Camera to replace it. The only problem is that the Ring Floodlight Camera is designed to be wall mounted (about 8′ high) and Ring specifically says it can’t be mounted under an eave. Challenge accepted.

As It turns out, you CAN mount a ring floodlight cam under an eave, but the camera part doesn’t have quite enough play in the provided ball joint.  To fix this, you need to loosen the retaining screw, pop the camera unit out of the ball joint, and then grind a notch that will allow it to swivel upwards (formerly downwards) just a bit more.

The end result looks like this:

Here you can see the notch I ground out of the ball joint:

I used an angle grinder with a grinding wheel, but the plastic is soft, so you could do it with a rotary tool or even with a file by hand if you had a lot of extra time. Note the masking tape to make sure the camera cable stayed well out of the way of the grinding wheel.

There is an internal square tab inside the ball joint, which also has to be filed down (I used a hand file for this one):

After this small modification to the ball joint, there was plenty of flexibility to aim the camera exactly where I wanted it and have the bottom of the motion sensing pod level with the ground. Of course, the lights are “upside down”, but you can still point them in any direction you need. The “rain shields” normally on the top are now on the underside, but no water will get caught in them (because they are angled down, and because the entire light and fixture is protected under the eave anyways). Most people won’t even notice.  If I was willing to disconnect the light wires, swap the two and re-connect them, I could have made them be “upright” again, but I felt that this was too much work for just an aesthetic change. [Note that the camera part is shipped “upside down” in the box, and normally you would need to flip it over when wall mounting, so you can omit that step.]

I am lucky, in that I have a low roof, so that the angle of the camera is still right around where it should be for capturing good images of faces. If you had a two story house, mounting a camera under the eave wouldn’t give you a very good angle.

Does this modification void the warranty? Possibly. If the device fails due to this modification, it would certainly void the warranty. [For example, if the camera unit falls out and breaks after I modify the ball joint designed to hold it.]  However, if the camera unit were to fail due due to an electronic or software problem completely unrelated to the modified ball joint, the Magnuson-Moss Warranty Act could give me legal standing to insist that Ring replace/repair the camera unit because its failure was unrelated to my modification.  [I’m hoping the situation doesn’t come up….I made sure to test the floodlight camera before I broke out my angle grinder to make sure everything was working right before I started hacking on the ball joint.]

I broke my PWS4721 power supply by connecting it to a battery and driving current back into it. (I don’t remember if I reversed the polarity, or just had the output voltage lower than the battery voltage.) The end result was that the power supply had it’s output shorted internally, so that the output voltage was always near zero, and the current was right at the maximum current limit but no power was coming out of the front (or back) connections.

The broken component turned out to be Diode D408 on the main circuit board right in front of the power output header. It appears to be a reverse polarity protection diode, so it’s likely that I accidentally reversed the leads when connecting to the battery (all I remember is the spark). This diode is a 1N5408 (general purpose 1000 volt 3 amp diode) which I was able to replace for $0.40 (Although my total cost was closer to $10, as I ended up buying 10 of them just to have a few more sitting around if needed, and shipping cost me $4.33 from Digikey.) The only specialized tool I needed was a Trox-10 (t-10) screwdriver to remove the security (star) screws from the factory maintenance port so I could remove the back panel.

Of course, I had to disassemble the entire unit to get to the bottom of this circuit board to make de-soldering and diagnosis easier.  The output power rails were shorted before I removed the diode, and were NOT shorted after I removed it, and the diode had failed shorted, conducting in both directions.

In retrospect, I could have probably desoldered the diode in place from the top of the circuit board (the long lead would have been easy, and since the diode was already ruined, I could have heated up the body and pulled the whole thing out from the top and then replaced it without removing the entire circuit board. But, at the time, I wanted access to the rest of the circuit board just in case the diode wasn’t the (only) problem.

You know it’s a high quality piece of equipment because in addition to checking that it worked, they let it burn in to find out if any parts were going to fail quickly, and then calibrated it!

If you want to see how to tear down a PWS4721 and what is inside, here is the video:

And a few photos of the main board with heatsync and top logic board:

A local artist asked me if I could build a mechatronic system that would cause “something” to emerge from a piece of art “every so often”. (I’m being vague as it’s not my idea to share.)

I suggested the addition of an PIR (passive infrared)  motion sensor so that the “something” could also react to the approach of people.

I am using a SparkFun Arduino compatible Micro-Pro board, due to its small size and built in USB port which can be used for both programming and power.  In the video and photos, I’m using a small 9g hobby servo, but I have also tested the same circuit with a standard 3003 size servo.  I am powering the SERVO directly from the RAW (red) and GND (black) pins on the MicroPro board (so basically, right from the USB power), and the control line (white) is wired to pin 3 on the board.

The PIR module has three pins. +5V, ground and signal. You provide it with power, and when it detects motion it raises the signal line (it even has a built in pull down resistor on the signal line).  I connected it’s (yellow) signal line to pin 5, and ran it’s black line to GND and red line to the VCC connection on the Micro-pro. (Which is actually tied directly to the USB power supply and doesn’t make use of the onboard regulator when the pro-micro is powered via the USB port.)

The code is a small arduino sketch that will sweep the servo when it detects motion, or every X seconds if it doesn’t detect motion, with a configurable “cool down” period after activating. You can download it here:  mechatronic.ino.zip

The entire circuit can be powered by a USB cell phone charger. I wanted to be able to make it battery powered simply by plugging in a USB powerbank, but unfortunately, the circuit doesn’t draw enough current when not moving the servo to keep most USB Powerbanks from shutting down due to their “low current auto power off” features.
There are a few USB powerbanks on the market which have an “always on” feature for use with low current devices such as webcams, but you can’t just power it with any USB powerbank.

The last hurricane knocked out our electricity for 3.5 days. (Other people in our neighborhood were without power for 6+ days). I have a 120 volt inverter that I used to power our fridge, microwave and internet, but our well pump requires 240 volts.  It was really annoying to have to transfer water from our tub to the toilet tank to flush, and not be able to wash our hands at the sink or get water from the tap when we wanted to.

My wife has given me a $1,000 budget to expand our backup system so that we can run the well pump in power outages, and the first step is to figure out the actual power draw of the well pump.

The pump is powered off of a 20 amp circuit, so the maximum draw is 240v * 20A = 4,800 watts. I measured the actual continuous draw when the pump is running at just under 10 amps, or 10 * 240 = 2,400 watts continuous.  I will probably need to purchase an inverter with a surge rating of twice that to support the inductive surge the pump motor is likely to draw when it first turns on.

I hooked up an energy monitor to the pump circuit and monitored it for a few weeks. The energy monitor estimates that the well pump uses between 6-8 kWh of power over the course of a month, so I will only need 1-2 kWh of battery capacity to be able to run it for several days.

Continue reading →