Tech Forum Questions

I have quite a bit of experience building underwater LED lights for technical diving and well know the challenges of dissipating heat. You didn't mention what kind of material the enclosure is. Don't use plastic, plastic will just insulate and heat build up inside the box will probably destroy the Pi.

The best solution is use an aluminum box, I'd suggest some sort of rubber adhesive sheet to affix (and insulate electrically) the board to the box with a hole cut out for the processor which is in contact with the aluminum.

The thermal grease (use only a tiny bit) will wick heat to the box and the entire box will act as a heat sink.

First, I would check the voltage coming from the crank generator, if there's no voltage there then that's the problem. Usually mechanical devices die before electronics. Assuming the generator is working and you have more than 3.6 V (you need 3.72 for Li-ion batteries usually), the transistor is probably a three-pin regulator and the added other parts are to keep it to 3.72 volts. Any three-pin regulator will work OR you can buy a complete Li-ion charging circuit from China for under $5. I think I bought five of them for around $5 at one time.

I've just signed up to the tech forum and noticed the question #2163 Cranky Flashlight; two things stand out.

1, The zenner diode ZD1 should not be a 1N4148 and needs to be replaced by a zenner diode as close as possible to 4.9V, since the voltage at Q1's emitter will be about 0.6 to 0.7 V lower than ZD1's cathode.

2,  The transistor Q1 is not connected correctly to work as a voltage regulator in conjunction with ZD1 and the 47 ohm resistor. Note - the full part number is 2SC8050.

The diagram shows the corrected flash light circuit with the base and emitter transposed.

The zener diode/transistor combination are a voltage regulator to provide an upper limit to the battery's charge voltage.  (see schematic) 

Somewhere on the internet a few years back, perhaps on Youtube, someone suggested replacing the failed battery with a 1 Farad memory back-up capacitor. At the time Jameco carried a 1.5 Farad capacitor of the same physical size, and with some minor board modification I was able to mount it in place of the battery. It now seems to work better than with the original battery.

Use a DMM to see if the Li-ion battery is being charged when the crank is being turned. The battery should have around 4.1 VDC across it if the charging circuit is working correctly. A replacement for the 2SC8050 is the SS8050 (try www.mouser.com/Semiconductors/Discrete-Semiconductors/Transistors/Bipolar-Transistors-BJT/_/N-ax1sh?P=1z0z63xZ1z0z60l&). The Li-ion battery may need to be replaced because it cannot hold a charge. So if you have the correct voltage across the battery with the crank turning, you probably need another battery.

I've drawn the attached schematic to show what I believe the design intent for this flashlight was.  It is a little different than the presented board, but works to explain what is probably wrong.

Since the lights come on when cranked, the full wave rectifier section is working and the transistor is conducting. I believe the intent of the ZD1, R4 and Q1 is to maintain the voltage at the battery at a safe maximum level. This would occur if ZD1 was actually a zener diode instead of a 1N4148. The zener diode would have a breakdown voltage somewhere between 4.3 and possibly as high as 5.1 volts. The battery would then be held below the zener voltage by the BE drop of the transistor, for the range of values mentioned 3.6 and 4.4 volts. 4.4 volts is really too high, but 5.1 volt zeners are cheap and widely available so the manufacturer may have cheated a bit.

This is not a very sophisticated charge control circuit. The spec sheet for the battery allows no more than 50 mA of charging current. Depending on the generator this could easily be exceeded by this design. The most likely problem is a dead (not just uncharged) battery. The battery could be dead because of age or overcharging. The overcharging may not only be because of the unsophisticated circuit. A shorted transistor would make the problem worse, as would an open zener diode. 

If the transistor is shorted virtually any silicon NPN transistor will work as a replacement, as long as it can handle at least 50 mA. It will be easier to use if it has the same ECB arrangement of leads as the SC8050. A quick look through catalogs gives 2SC2655, TS13003 and APT27H as candidates.

Finally, there may not be anything wrong at all. I have similar flashlights, and when left unused for a long time it takes a lot of cranking to get the battery charged enough to light the lights. If the generator is specified to not exceed the battery limits too much it should take almost half an hour to bring a completely discharged battery to full charge, and probably at least half that to get it to where it will light the LEDs.

It looks like part of the PCB track connecting the cathodes of the 3 LED’s is missing from the diagram. The zenner diode can not be a 1N4148 (a small signal diode), you need a 4.7 or 5.1V zenner for the simple voltage regulator to limit the voltage applied to the battery to about 4 to 4.4V. Check the transistor connections. I expect the 2SC8050 collector should go to C1’s positive terminal, the base to the junction of the 47 ohm resistor and the zenner cathode, the emitter then goes to the battery positive terminal.

Several things seem amiss. Diode 3 should connect to diode 4, probably under R1,R2, and R3. This would complete the negative part of the circuit. Zener diode 1 is not a zener if it is a 1N4148. This is a small signal diode with a reverse break down of 75 to 100 Volts and with E-B junction of Q1 probably protects the generator overvoltageing of the circuit Speaking of Q1, the SC#### is in the form a Japanese transistor just a 2 first. Search the internet for a 2SC8050. The picture of the switch does not appear an ON-ON-OFF but an ON-OFF-ON. Now to what wrong, since it works when cranked everything except the battery is working. When not cranking the circuit fails, then battery is not charging. A bad solder connection to the board, a bad connection to the battery, or just a bad battery. I would suspect the battery A voltmeter across the battery would give you the answer.

I have worked with various microcontrollers over 30 years and the hard part was always the overhead getting started. The Arduino Uno has less overhead and can do interesting things. As for education, I learned a remarkable amount studying for a Ham Radio License, even if I never needed the license.

Get a copy of The ARRL Handbook for Radio Communications  (www.arrl.org/arrl-handbook-2015). It contains a lot of theory in an understandable format. It also contains full schematics and plans for projects. Go to www.arrl.org/arrl-publication-dealers to find a dealer near you.

You didn’t give specs, but, if you size a power supply, leave plenty of head room, tube amps are power hungry, transformer supplies have a soft drop off, and switching supplies simply drop out when pressed beyond their limit. Also I’ll bet that the transformer has two or three voltage outputs, that is a lot to spec and integrate, and expensive.

The thin aluminum oxide dielectric (energy-storing) layer in electrolytic capacitors is formed on the specially-treated anode metal, and the electrolyte contacts the outer can (on early parts) or the cathode foil (on later “dry” parts). When radios have been unpowered for a very long time, electrolytic capacitors tend to lose their dielectric layer and their voltage rating, but not usually much capacitance. They may destroy themselves and other parts when re-powered, unless slow-start techniques are used to renew their dielectric.

When restarting long-idle sets, it’s best to use a variable-voltage (Variac) transformer to increase voltage slowly over several days. Note that Variacs are usually not isolated. An alternative method is to power up with an incandescent lamp socket in series with the AC line, starting with a 60-watt bulb and increasing the wattage gradually. Turn the set off occasionally and check for excessive heating.

I have repaired many radios using higher voltage TV capacitors, and never known one to take more than a week, unless it was defective. If the TV parts aren’t ancient, they may still be sufficiently formed for your new (lower) voltage, but for safety, follow the above procedures.

There is a strong shock hazard posed by AC/DC sets. Many early versions of these sets had all B(-) connections grounded to the chassis, including one side of the switched line cord, no matter which way it’s plugged in. Missing or wrong size screws or knobs or rotting rubber mounting grommets could make outer metal cabinets lethal. Some otherwise-well-built later radios, such as the Hallicrafters SX-41, had a live chassis. Most later sets used an isolated internal ground system to minimize the hazard. It’s safest to work on AC-DC sets using an isolation transformer, especially if you’ll be connecting any AC-powered test gear. You can make your own, using back-to-back filament or power transformers of appropriate power. Good luck with your repair!

Let me start with a rule of thumb, electrolytic caps drop some value if formed to higher voltage (within working range). This is because of thickening of barrier layer, BUT this can vary, that is why most are rated plus or minus 20%. They generally only need to be big enough (power filter, or bypass). Don’t expect much value drift moving to lower voltage. Other concerns are, internal resistance (can screw with bypass performance in audio stages, and heats cap at high duty cycle).

Depending on your design layout and other components, I would suggest using a TVS (Transient Voltage Suppression) diode. While a conventional freewheel diode may be sufficient, not many clipping networks work in the picosecond range of a TVS. Placement of the TVS should be close to the source of the transient with as short a connection as possible.

Common TVS operating voltages range from 5V to 440V. They may be unidirectional (fine for relay coil flyback clipping) or bidirectional (AC transient suppression). If you can’t find one with the exact clipping level needed there are methods for adding to the clipping voltage by placing fast silicon diodes in opposing series (cathode to cathode) with the TVS. Adding a fast silicon diode in this way will increase the transient suppression clipping point by the silicon diode junction voltage drop (.7V).

Another option would be to use a Snubber Network. Again, use of these would depend upon your circuit layout and conductor lengths. These are simple RC networks. You can build your own by placing the correct capacitor and resistor in series across the relay coil or purchase a snubber as a unit. A TVS is usually less expensive and easier to add to the design.

The inductive reaction from opening a circuit which contains an inductor (relay coil) is due to the reverse voltage generated by the collapsing magnetic field around the coil when the current is interrupted by opening the contact. This high voltage can cause arcing across the contacts or spike on electronic components which leads to their demise.

For a DC powered circuit, you could add a snubber circuit (series resistor and capacitor) in parallel with both the diode and relay coil as shown in the Figure. Just make sure the resistor and capacitor can handle the power discharged from the diode coil when the contact opens. For an AC circuit you can use a properly sized Metal Oxide Varistor (MOV) to snub the spike.

When controlling inductive loads, I like to use an N-channel MOSFET such as a 2N7000, BS107 or equivalent to control the device in addition to the flyback diode. This provides further isolation and reduces the current requirement of the Arduino just in case you want to control more that a couple of relays.

There are a lot of solar to li-ion battery charger ICs. You might get one on a breakout board, like the following: www.ladyada.net/make/solarlipo

I made a small solar powered solar tracker that has a similar IC, and charges 2 li-ion batteries to about 8.0 volts. Works well, in operation over a year now.

Get an inexpensive 24 hour timer and adjust it so it is on for as much of the 24 hours as you can. If that provides a short enough off time, you are done. If not, you can do the following. Make a 555 monostable timer circuit to energize a relay for about 30 sec. Circuits to trigger a relay using a 555 can be found in many places. Power the 555 circuit using a (surplus) wall cube plugged into your 24 hr timer.

Power your equipment through the relay contacts (but not through the timer) so that when the relay is in the de-energized state, your equipment is on. When the 24 hour timer turns on, the 555 timer will go high for about 30s, energizing the relay which turns off your equipment for a short time.

You may not have resolution down to 30 seconds, but a simple light switch timer would do the job. Set it to switch power off in the middle of the night or at a time of minimal use.

You could use this timer designed for resetting a router on a timed schedule! It is a plug in timer! www.amazon.com/NetReset-NR-1000US-Automated-Cycler-Routers/dp/B00HUEU9H8

You can try this particular timer from Amazon www.amazon.com/Digital-Programmable-Socket-switch-Energy-Saving/dp/B00WHPNON6/ref=zg_bs_495340_1 Or, Google for household lamp timer, there are many available.