Archives for the month of: July, 2012

Tonight I did another test of the pump with control and sensing. The flow sensor worked right out of the box. The voltage monitor I made using two resistors and an analog pin took a lot of fine-tuning, but after that it matches my multimeter’s opinion. The water level sensor doesn’t work at all yet, and the water temperature sensor looks like it’s going to be complicated to set up.

On the data logging shield, the real time clock (RTC) worked immediately, but the SD card gives an ominous message (“Initializing SD card…initialization failed.”). On the Adafruit forums, the main culprit seems to be crappy soldering. Totally likely, so here are photos:

Here’s hoping the ever-helpful Adafruit folks can tell me what I did wrong. PS: in beginning hobbyist electronics, there’s a hell of a lot of value in paying a bit more to get responsive, high-quality customer service, and wow, does Adafruit deliver.

Edit 17 July: yep, crappy soldering. Adafruit support even pointed me to the specific board parts (the digital pin headers) that were probably at fault. Fixed.

While I wait for the various bits of electronics to arrive, I decided to do a slightly more formal test of the pump.

I pumped a pretty much full 5g container up 74″. The pump was sucking air in the reservoir at 2:15, however: there was about 1″ left in the bottom, and the top bucket wasn’t quite full; call it 10/14 * 5g, say 3.6g. I’d say at 74″ of lift, conservatively, this pump does about 1.6 gal/min.

There’s good reason to make sure the pump is fixed standing upright on the bottom of the bucket. That will take a bit of design.

I’ll repeat the test to see if the pump shuts off when it finds no more water, but I’m not sure it does. That’s a bit worrying. All the more reason to have a ton of intelligence (i.e., only let the pump run for 2 min or until there’s 2″ of water in the reservoir. Good thing that the circuit design is well along the way,

The evapotron has a series of challenges. First, I can’t let the charger run while the battery is under load. That will fry the charger pretty quickly. OTOH, lead-acid batteries die when they overstressed or overcharged.  This all means  means that either I have to constantly check the battery load, or I have to automate it. Also the pump doesn’t seem to know when to quit, in two ways. A) it keeps pumping when there’s nothing more to move; and B) in practical terms, it doesn’t really need to run constantly. It could run every 10 or 15 mins, max.

So if I’m gong to control charging and timing, an Arduino is going to be involved. At that point, I kind of went nuts. Why not keep track of the battery voltage? The pump current? The water level in the reservoir? The water temp? The flow into the reservoir (i.e., total input)? And once all that’s happening, why not log it all to an SD card? And it’s going to need a minimal UI to watch it all. And the next thing I know, I had this.

The part on the left is pretty stable. There will be a bunch of wires from the sensors in the reservoir and the pump brought together in a junction box with a 4′ cable to the control box where the Arduino, battery, and charger will live. I’ve just started thinking through how the junction box will work.

The junction box has to be mostly weatherproof (it will be somewhat sheltered under the evapotron) but the control box needs to be sealed. Dust, rain, spilled grey water, etc., need to be kept out. Good thing I bought some Sugru.

I decided to test the 7xF cell NiMH batteries in a realistic way. I charged them fully, attached the Arduino and the LEDs, and I ran a test program adapted from dbu’s LED Globe idea and code. Here’s what the setup looks like:

I expected it to run for about 2 hours, and I hoped for three. Anyway, I put a voltage meter on the power bus, and measured every 15 mins or so. Results in the graph below.

Holy cats. The battery kept 200 LEDs blinking for 477 minutes, nearly 8 hours! So nominal 8.4v batteries supplying 13 Ah are more than enough.  The power curve is nice and flat until it fails, which is excellent.

A couple of observations here. First, the Polulu voltage regulators totally rock, hella efficient. At no point were they even slightly warm to the touch, so nothing wasted as heat. Second, in real use, the LEDs draw way, way less than their rated max current. Third, the giant, ridiculously heavy batteries are more than enough. No Li-Ion polymer for me, at least in this round.


I’ve mentioned before that I’m building an Evapotron, a device for making grey water (i.e., water used for washing dishes and people) evaporate so it doesn’t have to be packed out.

Today I tested a mechanism for driving grey water up and over the pumped cascade which is the model for this implementation. In short, I need to get dirty water out of a reservoir, lift it about 6′, and let it dribble over a snow saucer and thence down a cylinder to be exposed to sun and wind. A friend suggested that a submersible bilge pump would be a good choice.

I got a couple of Rule 25s automatic pumps. These guys have an internal circuit to check every 150 seconds if there’s water present. If so, they run ’til they suck dry, then go back to waiting. That’s perfect for this application. I’ll power them with a cheap 12v lead-acid battery. In the first test, it works.

The battery will live in a sealed (against dust and water) clear plastic box next to the bucket where the pump sits. Inside the box will be the battery and a charger. The charger will connect to the (highly flaky) camp AC. The charger is “smart” in the sense that it knows only to charge when the battery’s voltage is low. Still, I’d like it to turn off periodically, and I’d like to turn the pump off at about 7pm through about 8am in the morning. I’m thinking of building a real time clock + Arduino that could control the charger via a PowerSwitch tail and the pump via a relay.

I can think of many other uses for such a circuit. For example, it would be nice to turn on a power strip that charges all my rechargables one night per month (they usually sit dormant for months and die). It might also send me email about the status of the batteries.

Anyway, my project partner P has a bunch of good ideas for the actual mechanism of the Evapotron. I think this is going to be excellent.

I’m hardly an expert in anything, but I’ve accumulated a few ideas and a little experience as I’ve gotten cranked up in the last couple of months. I put together this list of resources and stuff for starting in Arduino hacking.

The first thing you need is a bunch of tools, and Lady Ada’s toolkit helped me. One tool that was missing but has drastically improved my work is a self-adjusting wire stripper. You’ll need an Arduino, of course, and a nice kit with a few supplies is a good start. You’ll need an intro to the programming language (Wiring), and this book is a gentle intro. This book from O’Reilly is a more advanced programming book but with more useful stuff to do. For all the deep reference material needed for any programming language, see the Arduino site itself.

If this page helps you,

please buy this shirt. 😉