TV signal
transfer problem

Are you aware of a device that can transmit TV RF signals room to room without coax cables or a VCR; ie, incoming RF from aerial to remote TV, not using A/V transmitters? I have an application where it would make life a lot easier. Perhaps you could advise? (I. M., via email).

One possible approach may be to adapt the portable masthead antenna described in the August 1996 issue. It will handle UHF and VHF signals. You would need to connect its output to a dipole antenna.

Data logger connection for magnetic field meter

A few years ago I bought a Magnetic Field Meter, as featured in the October 1991 issue. It has an LCD panel meter and a 3-way rotary range switch and runs from a 9V battery. I now want to connect it to my PC logging multimeter and want to know which pin I can get an analog voltage equivalent to EMF reading? (G. J., via email).

There is no convenient place to connect a data logger since the DC signal is a high impedance that is suitable for the ICL7106 but not a data logger.

To enable a Data logger connection, you would need to add a 10kW resistor to the cathode side of diode D1 (this diode connects to the pin 6 output of IC2, an OP77). Then connect a 10mF capacitor to the end of this resistor to ground (plus side to the resistor). The data logger can connect to the capacitor terminals.

10-Level Tank Gauge Wanted I recently wanted a water tank gauge and searched high and low for an economical commercial unit. My search unearthed the "Water Level Indicator" in the April 2002 issue of SILICON CHIP. Its only shortcoming was the 20% increment spacing. In my tank’s case, this represents about 7400 litres and I would prefer more accurate readings. It was noted that the LM3914 had 10 comparator outputs and I toyed with the idea of obtaining this kit and modifying it to accommodate 10 LEDs. However, I recognised the "Multi-Voltage Monitor" in the May 2006 issue as a similar project and I think I have my answer. If I construct this monitor, is there any reason why I couldn’t add the power supply, the "critical level" indication (if required) and a sensor array to convert this kit to the level indicator I desire? I see my main hurdle as the component selection for the sensors, where I now will have 11 sensor inputs as opposed to the original six. How do I calculate the resistor values please? As well as providing me with this answer perhaps you will consider a "Mark II" version of the Tank Level Indicator as I’m sure there are a lot of punters out there that would appreciate this luxury, especially now that there has been a drought-induced tank population explosion in Australia. (P. W., via email). The calculations for the resistors are quite tedious, particularly for 11 values. An easier method is to start out with trimpots and adjust these so the LEDs show correctly for each tank level.

Light dimmer has reduced brightness

I recently assembled one of your projects titled "Touch/Infrared Light Dimmer" back in January & February 2002. It passed the low-voltage test using the small bulb as suggested in the article. And yes, I did change the .001mF capacitor back to 0.1mF but once I installed it, the unit can only reach ~25% brightness level.

I suspect maybe the phase is too late, as shown on one of the scope shots published in the January issue. What should I do to troubleshoot this?

By the way, I changed the 4MHz PIC16F84 to a 10MHz device and re-programmed it with the hex file from your site. Did I miss something here? (H. H., via email).

The correct phase control is dependent on the correct detection of the mains waveform zero crossing. So the 680kW resistors and the 10nF (.01mF) capacitor at pin 6 of IC1 must be correct. Check them out.

The 10MHz version of the PIC should be used and in fact is what we specified in the parts list. Use either the 10MHz or 20MHz versions. A 4MHz PIC cannot be used as we are running the PIC at 10MHz using crystal X1.

Curious non-linearity in amplifier module

I recently built two audio amplifiers using Altronics kits but replaced the LM3876T with a LM3886TF and dropped the rails to ±28V as suggested by you for a 4-ohm load.

Both outputs are linear up to about ±9.5V peak-to-peak on sinewave but at this point the tops of the sinewaves clip. With increased drive, the bottoms of the sinewaves clip at about -22V which, for a symmetrical waveform, would correspond to 60 watts output. The linear gain was measured at 19 as expected.

Have you any idea what may be wrong? Presumably, something must have happened to each LM3886. (R. E., via email).

This is an interesting problem which could cause anyone to tear their hair out. The likely cause is that you do not have both the supply pins (1 & 5) properly connected to the +28V rail.

Modifying the
SLA float charger

I have purchased several 12V SLA float chargers as described in the March 2003 issue. Rather than use an AC supply, I would like to be able to use a DC supply, as I have a large regulated supply that will at the same time be powering Nicad/NiMH battery chargers.

What would be the best way of modifying the circuit? Can I complete the kit as normal and just feed the DC in on the correct rails or omit the rectifier part of the circuit? (G. I., via email).

You can build the circuit as it is shown and power the charger from DC. The supply voltage needs to be at least 15V. Alternatively remove diodes D4, D6, D7 & D8 and place a link in the D6 position. You can then use a minimum of 14.5V DC for the input supply.

Queries On 3-Phase Motor Controller Our electric cooperative serves a large sewer-lift station that utilises a solid state GE soft start on a 300 HP 3-phase 480V motor. We supply the station from a 480/277V transformer. They began experiencing problems while starting the motor from day one. I used a Metrosonics PA-9 PQ Analyser to capture the motor start sequence, including voltage and current waveforms both at the transformer and at the output of the soft starter. This particular soft starter achiev-es a reduced voltage level for motor starting by utilising solid-state switching devices to turn the voltage applied to the motor off and back on at various points along the voltage waveform. After looking at the voltage waveforms coming from the soft starter (measuring the phase-to-ground potential), I noticed that the waveform showed the voltage dropping below the zero level during the "off" time for both the positive half cycle and negative half cycle. Keeping in mind that the motor was connected in a delta configuration (phase-to-phase) to the soft starter and my measurements were phase-to-ground, do those zero crossings really mean anything? I have heard that a solid-state switching device exhibiting zero crossings on the output when connected to a sinusoidal source is a good indication that it is bad. Any information you can supply would be greatly appreciated. R. C., via email). We are not at all familiar with the system you describe but we can make a number of comments. First, if the controller only varies the voltage and not the frequency to achieve soft starting then the only way it can do it is by phase-controlled switching, probably using Triacs or SCRs. Second, while Triacs or SCRs turn off when the current reduces to zero or goes slightly negative, the voltage can still swing quite a lot negative as the motor is commutated off. If the commutation from cycle to cycle is erratic, it could mean that the snubber networks across the SCRs/Triacs are not correct for the load. Of course, if the controller varies the frequency as well as the voltage, then none of the above applies because the electronics will use IGBTs. In any case, it sounds as though you need an experienced technician to check it out.

Bigger transformer
for preamplifier

I’m completing my third pair of SC480 50W amplifier modules. They will mate up with the Studio Series Preamplifier and the Studio Series Remote Control. I have a toroidal transformer on hand that I would like to put into service for the power supply but it is 18V + 18V (at 50VA) and not 15V + 15V as specified.

Although the difference seems slight, would it cause any problems, such as overheating regulators? (J. H., Falmouth, Tas).

Unfortunately, an 18V transformer is unsuitable for two reasons. First, the higher voltage will indeed cause excessive power dissipation in the voltage regulators, particularly when the headphone amplifier is installed. In addition, the rectified output voltage will exceed the voltage rating of the 2200mF 25V filter capacitors. Higher voltage (35V) capacitors will fix this problem but there may be problems fitting some physically larger variants on the board.

Problem with
telephone dialler

I have a problem with the Telephone Dialler (April 2003) which I cannot figure out. When setting up and testing, it works fine and dials my mobile perfectly. However when I disconnect the dialler from my PC and turn off power and then hook up to modem and turn power on again, the LED flashes slowly (trying to get communication with Modem). It doesn’t matter what I do, I cannot get the dialler to go into scan mode (LED flashing faster).

To get the dialler to work, I have to disconnect the modem and hook up dialler back to the PC again. I then type (0) to simulate that the dialler has found the modem. Then WITHOUT disconnecting power to dialler, I just pull out serial cable plug it into modem then it works fine UNTIL you turn the power off which leads me back to square one. Why is it doing this every time I turn the phone dialler off? My guess was the configuration of modem but all the commands in article work fine. The modem understands everything.

My Modem is a Dynalink 1456VQE. I am using the serial cable the modem came with for hooking up to the dialler. One more interesting thing I don’t understand is when the dialler is hooked up to the modem through its own serial cable, the SD and RD lights on the modem flash every time the LED on the dialler flashes while it’s looking for the modem. However when I make my own serial cable (not XOVER): pins 2-2, 3-3, 5-5, then the SD and RD lights on the modem do not flash. Is this a Dialler to Modem Communication problem?

The alarm dialler uses the AT command set in terse mode. That is, the response from the modem must be in single digit numbers not in words. For example OK is a 0 (zero). When the alarm dialler is first turned on it sends out AT and waits for a zero response. As you have noted this can be simulated by typing zero on the PC when you are directly connected. Only when the alarm dialler gets the zero will the LED flash quickly and go into scan mode.

As you have discovered, you can fool the dialler by sending it a zero from the PC and if you keep it powered you can then plug it back into the modem and it will work as expected. So the problem you are experiencing is that the alarm dialler is not seeing a zero when it sends an AT request. There are probably only two causes:

(1) the modem is not programmed to respond in terse mode. Try connecting your PC directly to the modem and entering AT. If you get OK, then the modem is not configured correctly.

(2) there may be a cable problem. The alarm dialler uses a 9-pin connector, whereas most modems use a 25-pin connector. The important thing to note is that the send and receive pins are reversed between 9 and 25-pin connectors. So you can’t simply connect pins 2 to 2 and 3 to 3. You need to have them reversed. If the cable is not correct then you will not be able to communicate.

When a device is talking to a modem the SD (send data) and RD (receive data) LEDs normally flash. This is because the characters sent into the modem are echoed back to the device. If you don’t see any LEDs flash then the cable is wired incorrectly.

High-voltage
meter wanted

Do you have a circuit for a high-voltage voltmeter that can measure up to 5kV? I need to know the voltage from a voltage multiplier that I am working on. The maximum voltage that I would like is about 3kV, so a full-scale meter readout of 3kV would be good. (R. B., via email).

We have not produced a high-voltage meter. However, in May 1999, we published three electric fence testers which could do the job. Alternatively, you could use the resistive divider string in one of them as a voltage divider to be used with a conventional digital multimeter.

4-Digit Tacho Has Microcontroller Fault I hope you can assist me with the 4-Digit Tacho described in the April 2000 issue. I made it up and it appeared to be working OK but in trying to set up the calibrations I have run into trouble. When it was first turned on, I set it for a 4-cylinder motor and it seemed to be working. I then tried a couple of the other adjustments and now I cannot seem to get back to basics again; eg, I now want to set it for 8-cylinder operation. How can I reset the micro to the defaults and start from scratch? For instance, now when I turn it on I get --00 on the display and all LEDS lit and after a while it reverts to 000 with no LEDs. Any help would be appreciated. (G. B., via email). If you cannot change the settings with the up or down switches, then perhaps the PIC needs to be reprogrammed. You should be able to change the "0" to any number between 1 and 12. If not, return the PIC to your kit supplier for reprogramming.

Mesmeriser clock
has faulty LEDs

I have built the Mesmeriser Clock featured in your June 2005 issue and I have a problem. The "seconds" circuit is working great, doing everything it should. My problem is in the four digits. There are quite a few LEDs that don’t come on at all. In fact, 26 LEDs don’t come on.

I have checked my work over and over and even replaced U6 which is responsible for the digits. I would appreciate it if you could point me in the right direction on what I should be checking. (J. T., Renmark, SA).

The fact that the circle LEDs all work indicates that the column drivers (Q1-Q5) are working OK. We must therefore assume that the problem is related to the row drivers (U6). How you now proceed depends on what test equipment that you have available.

First, disconnect power and measure the resistance between each output of U6 (pins 10-16) and ground. It’s possible that one output is shorted to ground somewhere on the PC board and the chip is being destroyed. Next, use your meter to do a continuity check between pin 8 of U6 and ground. Perhaps the IC socket is faulty and the ground pin is open circuit.

If you have an oscilloscope or logic probe, power up and trace the signal lines back from the non-operational LEDs. You should immediately be able to see where the problem lies. We’d also suggest removing the PIC from its socket and checking for bent pins or damaged contacts in the socket.

Using microwave transformers in a welder

I am looking to buy an inverter "stick" welder but they are quite expensive compared to a traditional transformer type. I have come across an article on the web at www.dansworkshop.com/Homebuilt%20arc%20welder.shtml and I wonder whether you could comment on its design, safety, etc.

Is this a possible project for the "Salvage It" column as it could save several microwave ovens from being entirely scrapped. (R. C., via email).

At first sight, building a welder from recycled microwave oven transformers is an attractive idea – it saves all those transformers. However, as pointed out in the link you sent, the transformer secondaries must be removed and rewound and then quite a lot of other stuff added to get a workable unit. By the time you buy all the necessary bits, you could easily end up paying more than $100 to get a unit which is pretty much a Heath Robinson affair.

By contrast, you can pick up welders very cheaply on eBay and elsewhere and they would be safe and work well.

Even apart from the dubious value, we would be reluctant to advocate people rewinding microwave transformers. Inevitably, someone would energise the unmodified transformer and that is an exceedingly dangerous act.

Transistor equivalents for PN100 and PN200

Could you possibly tell me some equivalents for the PN100 and PN200 transistors used in the Slave Flash Trigger, in the July 2003 issue? I live in the UK and can’t find PN100 and PN200 here. (E. J., via email).

Both the PN100 (NPN) and PN200 (PNP) are low-cost general-purpose bipolar transistors in the TO-92 plastic package, sold here as "multi-replacement" types. They both have a Vcbo rating of 45V, a maximum Ic rating of 500mA, and an hfe (beta current gain) range of 80-600 at 10mA collector current.

Here are some of the common devices listed as electrical equivalents but note that some of them may not be suitable for all applications. Some also have different lead connections, so you’ll need to check this as well:

PN100 - 2N2222, PN2222, PN3565, 2N3704, 2N3904, BC337, BC547.

PN200 - PN2907, 2N2907, 2N3906, BC327, BC557, 2N3702.

For reference, the pin connections for the PN100 and PN200 are both C-B-E when viewing the transistors from below, with the ‘flat’ side downwards and reading from left to right.