Electronics Forum

The Lounge • Best ebay purchases?

2012-08-11T16:54:43-05:00

W4GNS's post on Profit on Relays shows some of the amazing values you can find on ebay. I've mentioned Sure Electronics here a number of times and some of the great things they have.

Let's start a thread here of your best ebay finds. Let's limit this to new items that may be of interest to our community. Post pictures of the awesome stuff you buy from ebay or other suppliers and post a picture and a description of the item, along with the seller's name.

Gary mentioned the bread board power supply. When I saw what this was, I had to buy one too. It's a board designed to straddle you bread board and plug into the power rails on each side. It can be powered by USB or a wall wart and supplies 3.3v or 5v independently to each side, selected by a jumper. Pretty sweet for $3.25

bread board power supply.jpg

The seller is geiliable-store and the description is:

Breadboard Power Supply Module 3.3V 5V MB102 Solderless Bread Board DIY 2012 New

If you contribute, a link to the seller's store is probably better than a link to the auction which may not work after the auction expires.

What fantastic stuff have you found?

Statistics: Posted by Jon Chandler — Sat Aug 11, 2012 4:54 pm — Replies 2 — Views 23

The Lounge • The Nokia Asset Offload Continues

2012-08-11T09:56:28-05:00

http://techcrunch.com/2012/08/09/the-no ... id-for-it/

QT is a very nice piece of work. Personally I have too many coals in the fire now, to work with more new stuff.

Statistics: Posted by W4GNS — Sat Aug 11, 2012 9:56 am — Replies 0 — Views 11

The Lounge • Profit on relays

2012-08-11T06:57:02-05:00

I wonder how these folks are making money, here is an example with free shipping.
http://www.ebay.com/itm/New-1-Channel-5 ... 416667b4f1

Statistics: Posted by W4GNS — Sat Aug 11, 2012 6:57 am — Replies 8 — Views 55

Projects and Gadgets • A Ventillation Fan Controller - Simple Gets Complicated

2012-08-10T18:20:09-05:00

In the Seattle area, the climate is pretty temperate. It usually doesn't get hotter than 80°F (27°C) and it's a rare year that it that breaks 100°F (38°C). Usually the nights cool off to 65°F (18°C) or less. Air conditioning of residences has been a rare thing here until recent years.

We've found that most nights, a large box fan blowing out a sliding glass door at one end of the house with open windows in the bedroom makes for comfortable sleeping. The only problem is that some nights, the temperature drops so much that I wake up cold, reaching for the blankets.

This idea started off simply enough. Use one of my new sous vide boards to monitor the temperature and if it drops below some point, shut off the fan. The TI TMP513 chip has a local sensor that I usually use to measure the reference junction temperature of the thermocouple, so it will work well for this project.

I want to share my design process - figuring out what I'm going to do - before I start coding anything. I may not incorporate all of these ideas into the design but by planning ahead, they can be added without much trouble later on if I want to.

I want to shut the fan off if the temperature falls below some point. I'll use an SSR (solid state relay) to control the fan. I could use a fixed temperature but a better approach would be a reasonable default temperature with a means to adjust the temperature. The sous vide board supports an LCD display and a pair of push button switches, so I'll use the push buttons to set the temperature, which will be displayed on the LCD.

The above is the simplest implementation that will serve the need to avoid waking up shivering. But, since I have all the hardware, why don't I monitor the temperature, and turn the fan on when it starts to heat up again?

Well, if the sliding door is closed, I don't want the fan to run. A magnet and reed switch of the type used for alarms can detect if the door is closed, and come nicely packaged for the application. By planning ahead, the door detection is easy to incorporate. (This is one of the things I know RadioShack has...but not anymore!)

reed switch.jpg

So the fan will turn off at some low temperature. At some higher temperature, the fan will turn on. Now I don't have to remember to turn on the fan in the morning.

But here's another complication. If it's warmer outside than in the house, I don't want the fan to run. The TMP-513 has a provision to use diodes or transistors as external temperature sensor (based on the voltage vs. temperature dependency of a semiconductor junction. Unfortunately, when I laid out the sous vide board, I didn't consider that I might want to use these remote sensors, so the pins aren't brought out to pads or a connector. I could use a thermocouple to measure the outdoor temperature and the TMP513's internal sensor will measure the indoor temperature with is also the reference junction temperature but I don't have any spare thermocouples.

I think I can work around this pretty effectively. If I monitor the temperature after I start the fan for a period of time, it will fall if the outdoor temperature is low. If the temperature continues to rise, it's an indication that the fan in pulling hot air into the house. So if starting the fan results in an accelerated rate of temperature increase, the fan should be stopped.

Some further enhancements could be made if I incorporate a RTC (real time clock) chip - there's a provision for one on the board, or if I just measure elapsed time from shutdown and restart the fan sometime later with the idea that the outdoor temperature will fall in the evening.

So a simple project has escalated into something more complex. It's still pretty simple as long as the extra features were envisioned before coding started.

I explained what I was going to do to my partner Eric. Sometimes, that can be a mistake! Above, I explained that I was going to control the fan with an SSR. But he pointed out something I had purposed ignored. the box fan has a built-in IP remote control. Using this IR interface eliminates the need for an SSR and adds the ability to control the fan speed (slow/medium/fast). This makes sense and it ties into the need/desire I have to remotely trigger my Pentax SLR. I've got some research to do. Yet another camera trigger.

The use of the IR remote capabilities of the fan doesn't alter the program too much. I'll create a boolean called Running and use a subroutine to start/stop the fan based on its start.

I hope this look into my thought process was helpful. I'll post more details as the project develops.

Statistics: Posted by Jon Chandler — Fri Aug 10, 2012 6:20 pm — Replies 8 — Views 74

Swordfish Basic • Swordfish includes

2012-08-10T12:11:27-05:00

I would be very thankful for anyone could tell me %$@#%$#$^ the LCD.bas file is for Swordfish. This has been covered here before darned if I can find it, nor Windows for that matter using search.

Statistics: Posted by W4GNS — Fri Aug 10, 2012 12:11 pm — Replies 13 — Views 88

The Lounge • Wow. All I Can Say is Wow About a Hand-held CNC Router

2012-08-09T16:13:08-05:00

Forget about expensive complicated x-y motion for a CNC router. With this router you guide it by hand within half an inch of the path, and it adjusts the bit position to the exact desired shape!. You'll just have to check out the link yourself.

Position Correcting Router

Statistics: Posted by Jon Chandler — Thu Aug 09, 2012 4:13 pm — Replies 1 — Views 15

The Lounge • EEVBlog - Photocopier Teardown

2012-08-09T11:06:34-05:00

Dave has another great sequence of videos of tearing down a large copy machine that was in the trash at his office building. Anyone who's ever unjammed one of these things will have some idea of the complexity of one of these beasts but I think you'll be amazed when you get a good look inside.

The machine Dave was lucky enough to grab has a goldmine of stepper motors, dozens of optical-interrupter switches and lots of great parts for robotics.

The first link is a time-lapse view of the process condensed down into a couple minutes. I love the music he's used!

The second link is a detailed view of just the top lid sheet feeder assembly. Even this part is incredibly complex... and would fill your junk box with useful stuff.

The third link is the detailed tear-down of the main unit. It's a detailed look at the assembly. Warning - it's quite long.

EEVblog #305 - Photocopier Time Lapse Teardown

EEVblog #312 - Photocopier Teardown Follow-up

EEVblog #303 - Photocopier Extreme Teardown

EEVBLOG 305.jpg

Statistics: Posted by Jon Chandler — Thu Aug 09, 2012 11:06 am — Replies 1 — Views 18

Projects and Gadgets • Halloween 2012!

2012-08-08T18:23:22-05:00

Since I failed to get my gadgets outside to scare the kids into running away screaming and crying last year, I will strive to make that happen THIS YEAR. Thus, the continuation of the project on this thread with a more appropriate title.

More when I code it! I'm working a logarithmic scale issue now.

Statistics: Posted by hop — Wed Aug 08, 2012 6:23 pm — Replies 1 — Views 39

Swordfish Basic • Unable to open system file error [Solved]

2012-08-09T17:41:04-05:00

As the subject states in SF I cannot compile or build with the following 2 errors:
1 Unable to open system file
2 (13) unable to open system file

I'm running as administrator and have read the first to threads in the following Google hits. https://www.google.com/search?sourceid= ... g4s2&pbx=1

Any suggestions anyone? I am on a recently purchased computer, so I have not had SF running on the machine.

Statistics: Posted by W4GNS — Wed Aug 08, 2012 5:20 pm — Replies 12 — Views 70

Projects and Gadgets • Yet another camera trigger.

2012-08-08T12:58:36-05:00

Hello. Let me present myself first. English is not my first language,so I ask for your forgivenes if I make some mistake. Also, don´t hesitate to correct me when you think is necesary. I will try my best with the gramatics and orthografy.

Mechanical projectist by profesion, I like computers, wildlife, cats, photography, and anything DIY related.

My electronics knowledge is poor, but at least I know Ohm´s law. I have some experience programing in VB.net and C.

Now, to the project and questions:

I just want to made a break beam camera trigger, this part is easy and don't need a MC.

But I also want to have a flexible and modular system, so I can expand the capabilities in the future. Itervalometer, sound trigger, precise delays, etc... Anything imaginable. Here is where the mc is involved.

One of the things I want to try, is to not have dedicated ports for sensors, triggers, etc. All the sensors will have all the electronic in his own package, and only one wire (Analogic or digital, depends of the type of sensor) plus +5V and GND will reach the controller box. The same for the triggers.

photosensor.png

I can have the user define in the configuration what is conected were, but this is prone to errors. So I want to go one step furter. I want the control box to identify what is conected in every port. For this, I'm thinking about adding another wire to the conection. Betwen this wire and +5V will be a resistor, a diferent value for every kind of package. Inside the controller, another resistor will be conected betwen this wire and ground, to form a voltage divisor. This will be feed to http://www.electro-tech-online.com/custompdfs/2012/08/PCF8591.pdf (An I2C A/D converter). So reading the voltage, I can know what is contected to the port.

I'm using this I2C chip to not waste analog inputs in the MC, and I'm already using the I2C bus for a display and keyboard I have from the times I was playing with stamps. Speed is not a problem, since the I2C bus will only be involved in the setup phase.

What do you think about this arrangement? Is there a simpler way to identify the sensors?

Also, for the light sensor, do you think the LM311 is adecuate for the task? Will you sugest anoter comparator, or a totaly diferent setup?

By the way, I think I will use some PIC18F for the task... Waiting for the pickit2 to start blinking some leds.

I may sound ambitious with my first project, but I have plenty of time and patience to learn along the way.

Thank you in advance for your coments.

Statistics: Posted by Baldor — Wed Aug 08, 2012 12:58 pm — Replies 24 — Views 174

The Lounge • Transfer is Complete

2012-08-08T11:52:12-05:00

My house if very dusty, which is one of the reasons I cover my bench work with a towel when I am away from it. At least I do most of the time which is also why my bread board array got so trashed. Dust, dirt, AND the Stella Artois I accidentally knocked over onto it. Beer and bread boards do NOT mix.

So I bought a few new boards of the clear variety this last weekend and transferred the work onto it. With the exception of a cap I forgot to put back after I moved it out of the way to lay some jumper wires, everything transferred perfectly. The cap is a 1 μF electrolytic type tying a 18GF4550 PIN 18 (Vusb) to ground. Because I forgot to put it back, my project was an unknown alien to my Windows XP rig prompting quite a few minutes of driver diagnostics. I just love testing my troubleshooting skills.

So now I am back to where I left off but on a new board with some new technology. I have that little display to talk to and I also need to find a way to control both the TLC5940 and the display without adding another MCU.

prototype bread board images 003.jpg

prototype bread board images 004.jpg

Statistics: Posted by hop — Wed Aug 08, 2012 11:52 am — Replies 4 — Views 178

The Lounge • beer cooler 101

2012-08-08T11:37:15-05:00

You want cold beer it's not that hard make cooler so all sides and top bottom and door have
Same R value. Install coil on left side door on right side with temp control at door side.

Now set a temp cutoff on outdoor unit that cuts out the compresser. At 32dg

The dang thing will keep your cooler at a nice cold beer drinking level

Statistics: Posted by be80be — Wed Aug 08, 2012 11:37 am — Replies 3 — Views 117

The Lounge • DS30 Loader + PIC18f25K20 Woes

2012-08-08T08:56:34-05:00

Hi all,

I have been reading up on digital DIY's bootloader article and it has me quite intrigued indeed. I happen to have myself a pic18f25k20 (same as the one used in the amicus) and I also have an FTDI232 breakout board.

I decided to give this bootloader feature a go but have not had success unfortunately. Surely it must work because Graham's video shows him using an amicus board (which has the same FTDI chip that I am using)

Basically, I have loaded the bootloader onto the 18f25k20 using my pickit2 (I actually used Jon's pre-configured hex files from his site:
http://www.clever4hire.com/throwawaypic/tips-trick/a-bootloader-for-amicus) Thanks jon!

I then connected up my FTDI breakout board to the pics vcc and gnd and then RX and TX (these two lines are crossed between the pic and the breakout board)

I opened up the bootloader GUI and it had no problem finding my ftdi board (it showed there was now a com port) I set up the loader program with my specific pic details, opened my hex file and then tried to write. I reset the pic while the program was searching and it kept giving me time out errors.

Is there something special I am missing - perhaps a connection?

Any info would be greatly appreciated.

cheers,

-brad

Statistics: Posted by retrobrad — Wed Aug 08, 2012 8:56 am — Replies 2 — Views 54

General Electronics • Averaging Sensor Data

2012-08-07T19:34:54-05:00

I was thinking about MrDEB's plan to average temperature data (I'm not going to debate the validity of the idea - I don't think it will work well) and had an idea that I had to test.

Note to MrDEB: This is not the place to discuss your beer cooler software problems. Please confine that discussion to one of the many other threads you have going on the subject.

I think everybody is familiar with a conventional average: add up the values and divide by the total number of values. A plot which might be made from MrDEB's cooler is shown here.

Averaging - conventional.jpg

This doesn't look too bad. The temperature has changed about 4.5° (Fahrenheit) over time which is probably tolerable. Fat, dumb and happy, yes?

Well, maybe not. The average doesn't tell the whole story. In fact, it really doesn't even tell part of the story very well. The graph below shows the values of the three temperature sensors. One is increasing in temperature, one is staying steady and the other is decreasing. The temperatures have spanned a 22° range while the average showed only a 4.5° change!

Averaging - range of values.jpg

If you think about it, you've probably heard the story about the man with his feet in a bucket of ice water, his head stuck in the oven and he says "On the average, I'm just right."

Average data can be made more meaningful by computing the standard deviation. I'm not going into that here but it's an indication of the variance of the values in the average.

MrDEB has two issues to deal with. He hasn't been very successful in reading a number of sensors (DO NOT POST ABOUT THAT TOPIC IN THE THREAD) and the average looks like it will provide relatively meaningless information. One method of averaging data is called exponential averaging. It weighs the current measurement more than the previous measurements and is effectively a rolling average. If you're dealing with values that change rapidly, this technique can smooth out the changes. I use it in my sous vide routines to reduce the effect of small fluctuations in the data.

Truth be told, I didn't arrive at this idea starting from the idea of exponential averaging. I started from the idea of getting meaningful results easily. Instead of computing averages from each set of values, what happens if we average the average value with the new value? Let me restate this to clarify my thought. Take the computed average, add the new value and divide by two. Pretty simple math isn't it? If all of the (temperature in this case) values are similar, the rolling average computed this way doesn't change much. If the current value is much different from the average, it's going to make a big shift in the new average value.

I couldn't quite picture how this would work in my head so I plotted out some sample data. The graph below shows this.

Averaging - rolling average.jpg

The result is an envelope of values that provide an indication of the extremes in the data. Where the envelope is narrow, the variation between the values is small. Where the envelope is wider, the variation between values is large. This technique doesn't show the extremes like computing the standard deviation would, but it provides a strong indication of the man-in-ice-bucket-and-oven situation.

If a more realistic view of the data isn't enough of a benefit, consider how easy it is to calculate. You are always averaging the current sensor value with the previous average, so you don't even need to know the number of sensors being used; you always add current value and past value and divide by two. This works if you have one sensor, 5 sensors or 100 sensors - more sensors changes the sensitivity to a single sensor but the computation is the same nevertheless.

Please note, I was motivated by intellectual curiosity to do this little thought experiment. I could give a rat's ass about MrDEB's project at this point.

THIS IS STILL NOT A PLACE TO POST ABOUT MrDEB's SOFTWARE ISSUES. USE ONE OF THE MANY EXISTING THREADS FOR THAT.

In thinking about David Barker's example to read a bunch of One-Wire sensors, Manually Setting a OneWire (OW) Device ID with Swordfish, this method makes the process trivial. David's routine loops through all the sensors it finds and displays the temperature result. Since this rolling average method doesn't care where the number came from, the rolling average is simple to compute. Grab the new value, average it with the old value and Bob's your uncle (this completes the Australian content of the post).

Of course, acting on the results of this averaging technique will require a bit more than a yes/no decision but that's not my worry.

Not to confuse the issue, but the relative sensitivity to the new measurement and how fast the rolling average changes is easy to adjust. As discussed about, the past value and current value are equally weighted. To slow the change and reduce the effect of the new value, we need to increase the effect of the old value. Don't worry, it's not as complicated as it sounds. To reduce the rate of change a little take two parts of the old value and one part of the new...multiple the old value by 2, add the new value and divide by 3. To cut the rate even further, multiple the old value by 9, add the new and divide by 10.

This is a really good tool for data smoothing.

Statistics: Posted by Jon Chandler — Tue Aug 07, 2012 7:34 pm — Replies 6 — Views 149

The Lounge • Did not inhale

2012-08-06T15:01:39-05:00

This dude must hang out with Bill Clinton, he also did not inhale.
http://sports.yahoo.com/blogs/olympics- ... --oly.html

Statistics: Posted by W4GNS — Mon Aug 06, 2012 3:01 pm — Replies 2 — Views 42