Tech Forum Questions

There are two “inexpensive” ways to do a print screen from a TV screen:

1. Put a digital camera on a tripod and set it up in front of the screen — ensure the TV screen fills the viewfinder and ensure the flash is off. When you want to capture something, just take a picture of the screen.

After that, connect the camera to your PC via USB cable and download the screen capture. Be aware: the resolution of the photographed image will be limited by the screen resolution of the TV screen AND don’t be surprised if you get screen bars, etc. due to the timing between your camera’s shutter and the TV’s screen image refresh.

2. Install a stand-alone DVD recorder between your cable box and TV, or connect the recorder to your TV’s video out jack using a suitable interface cable. What you’ll do is record the program to the DVD (use DVD-R discs) and use the machine to finalize the recorded DVD (make it playable on other DVD drives) when the recording is complete.

Then, play that DVD on your computer’s DVD drive, use the computer’s playback program (i.e., Media Player for Windows) to get to the screen to capture, then PAUSE the playback. Use the image capture feature of the player to take a snapshot of the video image, then save it to your disc.  Be aware: the resolution of the DVD-recorded material will be no better than the video signal (Composite, S-video, HDMI) fed to the recorder.

After you’ve screen captured your images to your computer, use your Image Editing utility to clean up, etc. the images for your use.

Now then, there’s a more expensive way to do it: it involves installing a video capture card into your PC. Like the DVD recorder option, you’ll have to patch the video signal from the cable box or TV’s Video Out into your video capture card.

This way, you can watch TV using your PC and, using the Video Capture Application Software, capture a snapshot in (more or less) real time. Also, you can record the program, while you’re watching it, to your PC and edit/manipulate the recorded material at your leisure.

The big advantage is your captured images will (typically) be the screen resolution of your computer’s display (or at least much better than the inexpensive options above).

This would be most easily accomplished using an HDMI splitter and an HDMI-capture device used with your computer.

HDMI-capture devices can be external (to the computer) for PC or Apple machines, or can be implemented as a pluggable card for use with a PC desktop machine (assuming that the machine has an unassigned motherboard PCIe connector available).

Connect the splitter to the television program source – e.g., a cable box. Using an HDMI cable, connect one of the splitter outputs to the television set. Using a second HDMI cable, connect the remaining splitter output to the HDMI-capture device.

Record the program material using the software provided with the HDMI-capture device. See Newegg et al for available devices.

There is no such thing as a dimmable LED. The way we get around this is by using Pulse Width Modulation (PWM), which means the LED is turned on for shorter and shorter periods of time within a time frame. You can see that if the LED is turned on for 100% of the time, say 1/100th of a second, then it is on for 10,000 1/10,000ths of a second, the LED will be only 1/100th as "bright" if it is turned down to being on only 1/10,000th of a second for every 1/100th of a second.

I used 1/100th of a second as the time frame because the human eye can't see flickering when the flicker rate is around 1/50th of a second, this is why old analog TV's in the US used 1/60th of a second to build a picture frame, and in Europe they used 1/50th of a second for their TVs.

You can also think about this using a 1 second time frame if you want. In this case when the LED is on only 1/100th of a second for every second of time frame it would appear only 1/100th as bright. The problem here is that we would see the LED turning on and off, so we would lose the appearance of the LED dimming. We have to use a time frame that the human eye can't "see" such as a time frame of 0.01 (1/100th) seconds.
 

A LED is a beautiful dimmable device; lower the current, and the output brightness goes down. This can be done through simple resistors, or through PWM (Pulse Width Modulation). That is the good news.

In order to use a LED as a house incandescent lamp replacement, something needs to be done to make that possible. First of all, the voltage is WAY to high, it is AC (the LED prefers DC), and we need to feed it the correct current. So a power supply module is included to convert the 120 (or 220) Volt AC to some small DC current appropriate for the LED used.

Herein lies the problem. This PS module is designed to deliver x mA to the LED more or less regardless of the input. When you adjust the input, the PS module tries to compensate, to the point where it no longer can, often resulting in a flickering light. These are the now non-dimmable LED bulbs.

To make this “house” LED bulb dimmable, a radically different PS module must be built. It must sense that the input has changed (due to the dimmer setting), and adjust the LED current accordingly. Not easy to do! and of course, more expensive.

Several designs have been marketed, some better than others. In most cases, even with a “dimmable” LED bulb, a special dimmer is required. It is a mess, and most still don’t work perfectly.

I can’t answer specifically what is different about dimmable LED lamps, but will note that the type of dimmer is important. You need to use a dimmer specifically designed for LED lamps.

I have had good luck with Lutron CL line of dimmers that are made specifically for LED lamps. These are not X10, I have not seen any X10 dimmers for LED lamps. Also note that X10 light controls generally only work for incandescent lamps, they need a bit of current flowing through the filament to work correctly.

For LED lamps I have used the relay based X10 switch WS13A with good success, it also works with CFLs.

Different bulbs behave differently so you have try some options. There are some test results on the X10 Forum that you might find useful. X10 dimmers dim in 16 levels so some bulbs may not dim smoothly.

Texas instruments always encouraged its engineers to use a diode in the emmiter of thier transistors. Why, because the reverse temp curve, cancelled the xistor temp curve. So we built several little circuits to monitor the temp. So, a series circuit that lets a low current through your diode, monitored with a sensitive meter, will give a value for different temps.

Caution, some diodes are sensitive to light, so take that into account. The 1n914 and 4148 were a common device. Never messed with the gallium devices, just be cautious they are easy to let the smoke out. Luck in your endeavors.

I’m not certain what you mean by “crystal diodes” but most modern semiconductor diodes can be used as temperature sensors. A standard silicon diode will have a forward voltage drop, usually stated as “0.6V” or “0.7V.” However, it varies with temperature. Typically it will drop about 2 millivolts per degree C. By measuring that drop and calibrating at two extreme temperatures, the current temp can be determined.

The diode should have its cathode connected to ground and the anode should have a resistor to the positive supply. The voltage measured between the junction of the resistor /anode and ground will be the forward voltage to measure.

If by crystal diode you mean the old “cat’s whisker” crystals (e.g. galena) I suspect they will have much lower voltage drop. However, it will probably still show a temperature sensitivity, so may work. Silicon diodes work pretty well.

For an in depth explanation, with lots of math and theory, look here: https://www.eetimes.com/document.asp?doc_id=1279718.

Theoretically, the forward-biased voltage drop is about -2.2 mv/deg C at reasonable currents like a few milliamps. This works pretty well for silicon diodes like the 1N400x and the 1N914. It also works well for the base-emitter junction of a silicon planar transistor (almost anything in a TO-92), especially if you short out the base-collector junction. I have not tried it, but I think it does not work as well for point-contact germanium diodes like a 1N34 (which might include your "crystal" diodes).

I am not sure that time is the cause of loss of magnetism of permanent magnets. They normally lose their magnetism by three methods.

1. Heat.
2. Physical shock.
3. Exposure to opposing magnetic field.

I have two large toroidal magnets which sit on a tube with opposite poles facing each other. The top magnet ‘floats’ about 2cm above the bottom one. They have been like this on my workbench for at least 10 years and there is no discernible difference in the distance between them so I don’t believe that they have lost any of their magnetism.

The magnets weigh about 100g so the force required to hold the floating magnet against gravity is about 1 newton. I have no idea where the energy to do this for 10 years has come from and I have not found a physicist who can provide a satisfactory answer. I will keep asking!

Unless there’s something special about the old magnet it’s usually easier and cheaper to replace them. A case I’m familiar with where it’s worth “recharging” old magnets is in old motorcycle magnetos.

In this case we put a coil with a core inside the magnets and then pass a high current for a short period of time. The field from the coil has to be much stronger than the original magnet but doesn’t need to be maintained for very long. Discharging a capacitor into the coil is a good way to achieve this.

To little field and nothing happens, to high a field does no harm. Make sure the field you generate is in the same direction as the magnet or you’ll wipe out any remaining magnetism and possibly reverse it’s polarity.

There are a few ways to partially revive some types of magnets. However, the more powerful the magnet was originally, the more “brittle” is the arrangement of domains, and the harder it is to realign them.

For that reason, remagnetization is most useful for antiques, ornaments or teaching devices, made of materials such as carbon steel. Don’t expect a ferrite or neodymium magnet to gain back much strength.

First, determine the alignment of the original magnet. Rubberized magnetic backing, for example, often has stripes of opposite polarity on one side; these would be impossible to repair. A horseshoe magnet, often found on hand-cranked generators, or a bar magnet, is easier to fix.

Simplest is to put the old magnet in contact with a strong permanent magnet, e.g. a large rare-earth magnet, letting it align by greatest attraction (North touching South). Tap the old magnet a few times with a hammer to help rearrange domains. More effective is to wind wire around the magnet and pass a large pulse of DC current through the coil. Capacitive discharge magnetizers are simple to build, but involve high currents and voltages and present some safety hazards. See https://www.rcgroups.com/forums/showthread.php?307133-Inexpensive-magnetic-flux-(gauss)-meter/page3&highlight=capacitor%20discharge%20magnetizer for more details.

I would try making and using an electromagnet to re-magnetize your magnets. Works well for bar magnets, for other shapes it might be a bit more involved.

Since many “standard” type magnets are made from magnetic metal placed into a magnetic field to begin with I would say if this can be done for your magnets you stand a good chance of making them every bit as good as they were originally.

There are many low power radio modules in the market, which can be interfaced via the RPi’s serial port and/or GPIO pins, and which will allow point to point communications over short range (have a look at www.radiometrix.com; other suppliers are available). But, so far as communicating with existing IoT devices, you may have some trouble, since you will need a communication device compatible with whatever you are trying to communicate with.

Because you say the load is 35 lbs, you probably need a heavy duty drive with bearings to support it. While the unit listed below is expensive, $180 USD, it will do the job.

1. Speed can be slowed by reducing the voltage.
2. Reversing is done by changing the polarity.

https://www.surpluscenter.com/Electric-Motors/DC-Gearmotors/DC-Gearmotors/25-RPM-99-Volt-DC-Inline-Gearmotor-Ametek-119105-w-1-2-Dia-Rear-Shaft-5-1416-B.axd

SPECIFICATIONS

• RPM 25 RPM @ 99 VDC
• Voltage 99 VDC nominal
• Amperage 2.1 Amps @ 99 VDC, No load
• Gear Ratio 108:1
• Rotation Reversible
• Enclosure Totally Enclosed
• Duty Intermittent

This cheaper motor might also work:  https://www.surpluscenter.com/Brands/Parvalux/30-RPM-50-Volt-DC-Parvalux-Gearmotor-Single-Shaft-5-1549.axd

I have successfully plugged an old PC keyboard into that CD changer to enter data. However, what is it that you want to automate? Every entry i.e. artist, titles, etc.; are different?

Yes, you can emulate a PS-2 keyboard. An SPI-master port will do the job as long as it can handle 11 bits, or two 8-bit values in rapid succession. Not all microcontrollers can. (If you lack a good SPI port, emulate it in software.) Use the MOSI signal for the key-code transmissions and use the SCLK signal to emulate the keyboard’s clock output. The bits transmitted when you press a key look much like those put out by a standard UART: a start bit, eight data bits, an odd-parity bit, and a stop bit.

The keyboard’s clock output produces a logic-1 in its idle state and created a positive edge for each of the 11 bits. An SPI port configured as CPOL=1, CPHA=1, provides the proper timing. Note: The least-significant bit of an 8-bit key value gets sent first. My old PS-2 keyboard produces clock signals at 13.7 kHz. That’s a good frequency to start with for the SPI port.

A PS-2 keyboard assigns a unique non-ASCII code to every key. Find a list of the keys and assigned codes here: https://techdocs.altium.com/display/FPGA/PS2+Keyboard+Scan+Codes. When you press the “A” key, for example, the keyboard’s circuit produces the hex code 1C. When you release the “A” key, the keyboard transmits a key-release code F0, followed by the key’s code again.

The shift key works the same way. To send an uppercase “A” your microcontroller or processor board would send the SHIFT-key code, the A key code (then a short delay) followed by the key release code and the A key code again. Then it would send the key release code followed by the shift-key code.

Remember, your code must calculate an odd-parity bit and insert it in the 11-bit value to send. You can find parity-generator code on the Internet. Also, ensure you send the key and release values in bit-reversed form. Thus, the 1C code (0001 1100) for “A” must get transmitted as 0011 1000. You can simply set up your code so it uses the “reverse-bit” values to start with.

To provide a good answer your question, more information is needed. With the configuration of a LED and limiting resistor in series, with a capacitor in parallel, a time constant can be calculated. with a 2200μF capacitor and a 100Ω resistor, the time constant would be 0.22 sec. This would light the LED for about 1 second. I don’t know what your problem could be.

LEDs are pretty much digital (binary) devices, meaning they turn on and they turn off. They do not fade out when the voltage gets too low - they turn off.

The only way to make them fade out is to use an oscillating circuit to change/increase the time the LED is off vs the time it is on.

A joule thief voltage increasing circuit can be used to do this but so can a digital circuit.

The joule thief circuit could be made to do this if you use another circuit to control the voltage going to the joule thief circuit. Start with 1.0V for full brightness then decrease the voltage down slowly to ~ 0.20 V to dim & then turn it off.

The turn-off voltage will be higher (than 0.20V) if you use a simple joule thief where the number of windings/loops is the same for both sections. The efficient joule thief where the # of windings/loops is on a ratio of 2:5 and the 2-loop section goes to the base of the NPN transistor should have a lower shut-off voltage. I’ve had this circuit, with loops of 8:20 (8 to 20, 8 going to the base of the NPN transistor) still giving light down to 0.25 volts, so I don’t really know when it will turn off but you can simply turn off the voltage to the circuit when it gets to ~ 0.2V and that should look pretty good.

See my web site article on the Joule Thief Information Page at: http://cs.yrex.com/ke3fl/htm/JouleThief/JouleThief.htm - for all the experiments I did to find the best (efficient) joule thief circuit.