Archive for the 'Control' Category
As a fan of taking my dihydrogen monoxide in straight, uncut, uncorrupted liquid form, I purchased a six stage reverse osmosis filter some years ago.
My filter is a few years old now, and the super calcium enriched water we have in Michigan has taken its toll. The filter used to shut off when the production tank was full. However, the filter has recently developed a leak in the auto shutoff valve, so it’s constantly consuming water, even if the pressure tank can’t hold any more. This leak started slow, and I could live with it. In a few months, it had sped up to wasting a considerable amount of water. So, rather than replace the auto shutoff valve, which would have been too easy, I decided to add a microcontroller instead. The other problem my filter has, I discharge the brine water out to storage tanks next to my house, which hold the waste water, for watering plants and other uses. Problem is, when the temperature gets too low, the discharge line can freeze, which is very bad for the filter. So my microcontroller solves this problem as well.
The mcu is performing a few simple logic operations, with a few twists. First, it checks to see if the pressure tank reads “high” or “low”, via a pressure sensitive switch. If the tank reads “low”, that is the first “1″ in my logic comparison – a simple AND operation. Second, the mcu checks to see if the outdoor ambient is above 38°F. That is my second “1″ in the AND operation. If the pressure is low and the temp is good, a flag is set to enable the boost pump and water supply valve. A timer detects this flag and begins a countdown of 30 seconds (roughly). I chose to use a timer to buffer any false readings that may occur for whatever reason. After the 30 seconds has elapsed, the mcu turns on two mosfet switches, controlling the current for a solenoid valve and the boost pump. The mcu continues to monitor the two variables of the AND operation, if either changes, the “make water” flag is cleared, and the mcu turns off the water supply valve and the boost pump.
(click for super-sized version)
Two N-Channel mosfets control the current for the boost pump and the solenoid valve. The valve is a 24VAC lawn sprinkler valve, but I have tested it and it works well on 12vdc. The boost pump is a 100 gpd, 100 psi diaphragm pump, which delivers slow but steady high pressure water for optimal membrane operation. Pull-down resistors R3 and R7 are provided to prevent mosfet self-destruction should something happen to the microcontroller. The connectors TEMP, PRES and DISABLE are pin headers for connecting to external transducers. For reading temperatures, I chose the Dallas DS18B20 high precision digital thermometer with 1-Wire interface. The pressure switch is a simple normally closed pressure sensitive switch that opens around 45 psi. It is wired to present the PIC with a logical “1″ when the pressure is low. The disable connector will be connected to a shut-off switch, allowing me to temporarily halt production. The LEDs are just for indication of various operational conditions and modes. A 7805 regulator provides the pic with 5v from the 12v supply for the pump and valve.
Rather than cobble this together on protoboard, I made up some PCBs real quick. Here is a top view of the controller, without any connections, except for the thermometer. When installed, the thermometer will be installed at the end of a ~10ft shielded cable, for sensing the ambient temperature near my water storage tanks.
Here is the messy solder side of my controller. I haven’t hosed it down with alcohol yet to clean off the flux. I waited till it was good and late before soldering anything, so some of those joints are pretty nasty – I apologize.
(click to download full res)
The PCB layout was pretty easy – I had a lot of room to work with, as the enclosure I have selected is pretty big. This image shows the parts layout, along with the bottom copper artwork. The top layer artwork (in red) is for two jumper wires. Clicking on the above layout will get you a 300 DPI monochrome TIFF of the bottom layer artwork. Print this out at 300DPI and you should have a 1:1 scale of the pcb, incase anyone wants to make their own. Hit the contact justDIY link on the side menu if you want the firmware source code – it is written in Proton Plus basic.
One of these days, I’ll get around to re-wiring my house, however there are a number of obstacles in the way … the largest being the 2ft of clearance between my crawlspace and my floor joists.
Along with plenty of 20a rated outlets to go around and ethernet glaore, one thing I’d like to have is music in every room, well, nearly every room. I’d especially like to have audio in the kitchen, work shed (attached), bedroom and dining room. The bedroom would be a special case, I’d like to have the audio system, and an additional thin/fat client for full AV usage.
My idea for the audio system revolves around using a linux server. Here’s the basic idea … each room would have a small LCD screen and some buttons (or maybe old pda’s?). So I’ll call that the interface, whatever it ends up being. Each interface will be able to select the source (perhaps four or five source channels, whatever I can get away with, plugging cheap sound cards into my linux box). If the source is digital, i.e. mp3, the interface should be able to pause playback, next track, previous track, random track. The interface would also need to present a pre-programmed list of Internet radio stations (my main source of music, I have a very small cd/mp3 collection). In the case of internet radio, the interface should let me select the previous station, next station, random station, play, pause, etc. When using a live source, cable, tivo, etc, I dont want to mess with an inter-device interface, so, the only option would be mute. For any source, the interface will also provide a local volume control.
MPG321 and the linux kernel modules apparently have support for multiple DSP audio devices, so a rather simple script should allow for a specific input (file or live) to be routed to a specific output. For mixing of sources, that is, two rooms want the same source, I plan to use a series of analog multiplexors, designed for just this application. These simple ICs can be controlled by the server, and connect specific analog ouputs to specific analog inputs. For example, a 8 channel mux arranged as 2×4 will allow a single stereo output to select from a choice of four inputs. So one IC is required for each stereo output, allowing each source channel to select any of the four sound cards as inputs. A similar IC setup would be used to allow the sound cards to share live-output sources (tivo, cable) as inputs, connecting their line-input to any live source available.
I’m not sure about distribution. I could either go with matching transformers, and send line level signals over cable to the room, and use local amplification… or, I could go with central amplification and just use some heavy speaker wire to drive speaker jacks in the rooms. My house isn’t huge, so there wouldn’t be any runs more than 50-70 ft, so inexpensive 14ga speaker wire should serve nicely … I’m not looking for killer sound while I’m stir-frying … just want to be able to hear the tv or shoutcast stream without having to blast my stereo at the opposite end of the house. The advantages I see of central amps, I can build up a big power supply using old computer psu’s, and drive inexpensive car-audio DC amps, one for each source. The server could control the power to the amps, turning off unused sources. Secondly, line level signals are really low voltage… trying to send them over a long run of wire seems to be asking for trouble … unless I use something expensive like RG58 or RG6. Local amplification also means I need to locate an amplifier somewhere in the room, and provide power for it … not that big of deal I suppose.
This project is a long ways off, but its something that keeps crossing my mind, so I figured I should write it down.
Later!