Ldanielrosa

Allegro has hall effect current sensors for several ranges.

Have you looked at http://focus.ti.com/lit/ds/symlink/tle2426.pdf

The rectifier is only conducting when the input voltage is higher than the filter capacitor voltage. This is probably a very low duty cycle so the instantaneous current is much higher than the average current.

The crew I work with has the highest safety rating, but possibly the highest turnover rate as well. Five quit this week, one wasn't even on for two full days. I guess that's one drawback to also having the highest productivity of all shifts. The wage pays the bills. The medical coverage is the best I've ever seen- $250 per year deductible (including chiropractor and dental) and paid entirely by the company. The product du jour varies, as to the tasks for each person (with a few exceptions, the breakdown operator doesn't like doing anything else and nobody else wants his job).

I'm afraid that for an intelligent sequencer you'll have to use a microcontroller. Since you only want one LED lit at a time and the kill switches only need to be read prior to their renk in the sequence, you can probably do it with 9 I/O pins. Is this for a game-show device?

Am I the only one who disagrees with the layout of this H-bridge?

I'd use a microcontroller with a TRIAC. I'd have it periodically check the line frequency so that it could work in different regions (with different frequencies) and in areas with poor frequency regulation (or on a portable/emergency power generator).

I've found a source for 3.875MHz, which is close enough for my purposes. Thanks for reading my post.

Do you already have a 39 cell pack? I wouldn't advise making a charger for it as the voltage dip for the first cell completing it's charge may be difficult to pick out of 50+ volts.

With that requirement, you're not likely to find many chips that fit the bill. You'll probably have to use an FPGA.

Yes, I have. She just might- she already overtook me in math 3 years ago. Did I mention the holograms she made last quarter?

Do you have local friends who use microcontrollers? if so, consider using what they use- they're a help resource. What's easiest for you to get? Choose one on price, convenience, support- they're about equally capable. Both companies offer samples (last I checked Microchip had more devices to choose from though).

I've bored her to tears some time ago, but her major requires that she take some classes in electronics.

The distributors I've checked with don't keep these in stock. I'm hoping to find someone who'll sell ten or twenty of them.

Ouch! Those meters suffer from a design shortcut. Neither input pin is common to either supply pin, and it may be difficult to stabilize if there is an external current path. They're great for a battery powered panel meter, but inconvenient in some other applications. The current is a bit high to try capacitive isolation, unless you build a driver for higher frequencies (low kHz)- but that adds a bit to circuit complexity. I'm sorry I couldn't contribute anything useful. I'll be watching to see what comes of this as I am curious.

Have you seen office chairs that roll around? The wheels can turn to roll any direction. Those are casters. By dedicated device I mean something that will take your commands for what the motor is to do, then it will make the motor do it. To find which motor is suitable, you balance your priorities. For me, I need them common and cheap so I go to the scrap recycling place and get stepper motors out of 5.25" floppy disk drives. Those are all made by Teac and have similar (if not same) specifications. If I break one, it cost me nothing. If I need a better motor later, my software is already developed. Sorry, too many other projects come first.

Sumit, Most of the hobby robots I've seen are based on two drive wheels. Some have three wheels (one caster), some have four wheels (two casters), some have tank treads (still same concept). My first robot (I still haven't made one yet) will be this way too. It is true that you need to learn to control a motor. If you make a dedicated device for two motors, then you will have half the required brains for a robot. The other half decides where to go. Whether you choose a DC motor, a stepper motor, an AC motor, or a modified servomotor (or whatever), you must use a device that can control it. If your CPU has PWM, that can be used with most but is not necessary. I have yet to see a microprocessor that is entirely unusable for any of the motor types, though all of them will require some sort of interfacing. As such I suggest you use whatever CPU is available in your area at a good price.

I'm not getting anywhere fast with my postings elsewhere, So I'll paste it here. Some of the things I'd like to make will need I2C communication. I figured I'd have to make a terminal first, but that leaves the same problem- how to talk to it while debugging. I went looking for some downloadable software, but the stuff I found that claimed I2C (or SPI) without just mentioning serial EEPROMS required me to shell out some zorkmids (and I think they had some hardware requirements too). Is there a minimum bells-and-whistles software that will do it with no more a few passive support components on one of the ports? I have a notebook dedicated to my addiction (toshiba satellite 115cs, win95), so all the ports are available (com1, parallel, ps/2, there is no USB). At this point I don't even care if I can get 100kbps.

Does the transformer have standard E-I laminations? If so then with the dimensions of one of the outer legs (cross-section) and the power frequency I can give you a guess as t o the total VA rating. If the outputs are the wire they're wound with (not terminated to a terminal strip or spliced internally to more durable wire) then you can divvy up the output current according to the wire cross-sections.

Good afternoon, Read-modify-write means that the port is read, the bit selected is set or cleared, and the result is written to the port. The disadvantage of this is that the output that is driven high by the output may be overloaded and read as low (or vice-versa) as the output latch is not read only the actual pin. This is why sometimes it is good to use 'shadow' registers to modify, then transfer their contents to the port. Tri-state means that the output driver is disabled- writing a 0 or 1 to the latch will not affect the external circuitry.

What you want is a hobbyist with carpentry, mechanical, and electronic skills. Lo and behold! Here you are. You can probably get more informed suggestions if you can provide some pictures. If you have a web site of your own to post them on it would be easiest. Posting too many pictures here might be irritating, but one each of the inside and the outside (.jpg and not too high resolution) may give an idea of what needs to be worked with.

An appliance repair shop may be willing to sell you an "infinite range control" for an electric stove. This is a self heating thermal switch which is rated for 240V and at least 10A, usually more. Run the elements in parallel if they are identical. What are you planning to use for a crucible? What kind of flux? Casting in sand?

The 16F59 is a fairly new product in the 12 bit core family. I would recommend getting the 16F877 for several reasons. Any issues are likely to be known and documented, more people are familiar with them, "projects' are designed around them, and they have features you will want that the '59 lacks (I2C, USART, data EEPROM, analog to digital converter, more memory, more registers). If you qualify for and are requesting samples, you might consider getting the 16LF877, the 'L' indicates they are designed to run on a lower voltage if necessary. There is the '877A as well, this is a silicon revision (basically the same thing, but a few issues got dealt with). I'm sure that Dallas makes good microcontrollers, but I don't know anybody that uses them. You may have trouble finding people to talk to. The DS1307 is common enough. You should be able to get some comment from fellow hobbyists on that one. I have two, but I haven't started comunicating with them yet (need to write the routines). Maxim has several varieties of RTC. Look into the features, you may find something you like but remember you'll have to write a communication routine (repetitive use of another's routine is good) and you'll have to mount it (surface mount requires a PCB, just get DIP). As for the keypad, keep it simple. My next project that uses one will have five buttons (up, down, left, right, commit). They will be independantly wired. That's one of the benefits you'll enjoy with so many extra pins- you won't have to multiplex your buttons. You needn't be in a hurry to order the optoisolators and TRIACs just yet. Sure you can get familiar with them, but for the most part they'll have to wait for the microcontroller. I hope you have a solderless breadboard, DMM, DC power supply, fair array of components to play with, soldering iron, the list goes on.

The fewer communication protocols you have to deal with, the less code you will need to write. I would choose I2C because it allows several different devices to share the same communication bus- the first byte transmitted is an address header. The CO2 sensor does not have this, so it will have it's own dedicated pins and won't have to share. I recommended getting DIP because it is the largest. The pin spacing will be easier to work with than surface mount. It will work with solderless breadboard for experimenting, and common protoboard types for construction- no need to make a PCB. Plastic is cheapest, and the only material likely to be available in samples. The clock chip "ISL1208IB8" from Intersil is a surface mount package. Maxim has some in DIP. As for the microcontroller, don't get anything with a "C" in the middle of the name. Be sure there is the letter "F" after the first two digits. The "16x74" is good, but why not ask for the best available. The '777 has twice the program space, six more analog inputs and master I2C built in. The '877 also has twice the program space, the same number of analog inputs- but ten bits of resolution, master I2C, and has 256 bytes of EEPROM (non-volatile). If available to you, Microchip will allow three of any device and up to five devices per sample request. The optoisolators need not come from Farichild specifically, they're common enough that lots of companies make them. You'll need one for every device you want to control power to (fan, light, CO2 valve, heater, window flaps, water pump, stereo). You'll want to be sure the output can handle more than double the voltage that powers the devices- look for 400V ratings. The current will only be good to 100mA or so, don't worry- these will be controlling beefier TRIACs (which you'll have to select as well). Another option is to get SSRs (solid state relays), which will be more spendy but take care of some issues. MPlab is available at http://www.microchip.com/stellent/idcplg?IdcService=SS_GET_PAGE&nodeId=64 . It will let you run simulations of your programs and debug your code among other things. ICprog is available at http://www.ic-prog.com/index1.htm , but may be of little use to you if you don't use a programming module it was written for. I looked on Ebay, and the programmer I was talking about isn't there anymore. I guess the guy ran out of them. A google search for "jdm programmer" should get the plans for one. They're simple enough, and don't need to have a PCB. Ugly is okay for these- mine sure is. Programming ability comes with practice. I figure you'll have to spend some five or ten hours reading the data sheets and looking at other people's code, then you'll have enough to write a program to blink an LED. Once that much is done, you'll start writing more stuff. You'll test it, debug it, pull out some hair. This is a big project and you'll probably take several months to do it because you're starting at ground-zero, you're making the tools you'll need to work with.

The part you are missing is a microcontroller. The CO2 sensor has two types of serial communication available, either of which can be used by just about any uC. The LCD module (you didn't say it was just a DVM) will need to have te information spoon-fed. You will probably want to have a real-time-clock with it's own battery backup. Possibly an external EEPROM to keep any changes in schedule intact. Of course a keypad, maybe five buttons. External I/O for the CO2 valve, water pump, lamp. Assuming none of these items can share a line with anything else, that'll be 21 I/O pins needed. One of the mid-range microcontrollers from Microchip in a 28 pin package has 22 I/O lines. If space is not a concern, a 40 pin chip will give you pins to spare. Microchip also has serial EEPROMs. Depending on your location, you may be able to request free samples from them. Be prepared to tackle a few programming projects before the greenhouse controller though, you'll need to be familiar with the hardware. Maxim has a few real-time-clock chips. They offer engineering samples too. Be very careful that you pick one with a DIP package and a communications protocol that you want. Some of these have some extra memory on them, so they could be the external EEPROM as well to save on the parts count. For the external controls, you'll almost certainly want to use optically isolated TRIACs to control more powerful TRIACs or relays. Fairchild has MOC301xM and '302x' (3010, 3011, 3012, 3020, 3021,3022, 3023). I haven't received any samples from them, but the devices aren't too spendy. Once the unit is hooked up, the code can be written and modified without picking up the soldering iron again- this is a definite advantage. Oh yes, a google search for a "JDM type programmer" will give a schematic if you feel like building it- or a trip to Ebay will find one for $15 or less (shipping may even be part of that). The software for writing the uC programs and sending it the chip are available for free. I use MPlab for development, but ICprog for actual programming. Note: if you want to use basic or C, I don't know who offers free compilers- I use assembly.