Old EA frequency reference kit wanted

I’m after a pre-loved EA kit and that’s the Low-Cost TV-Derived Frequency Reference, published in Electronics Australia in the October & November 1993 issues, by Jim Rowe. The kit’s PC board number is 93TVF8.

Ideally, I’d prefer a working unit in its case but a "dead-un" would be considered. It’s part of my excursion into high-precision and digital HF and VHF transmission techniques. It’s a strange world looking at millihertz stabilities of TV carriers and E-layer ionospheric Doppler shifts.

Could any readers please let me know the pickup suburb/state and asking price on 0411 232 734. (C. S., Beacon Hill, NSW).

Modification to PIR sensor lights

I have a surveillance system at home. One of the features of this system is reliant on a sensor light and a relay that starts my video recorder. This works well until we get a momentary power failure. Most sensor lights have a feature that allows you to momentarily switch the power off and on so as to leave the light on. To return to sensor light mode, you have to switch off for about 20 seconds, then switch back on.

Is there a chance that one of your boffins could devise a circuit to allow the sensor light to return to sensor light status after a momentary power interruption? I would envisage an override switch if you wish to have the sensor light work normally. In addition to my needs, this would save people coming back from holidays and finding that the day they left on their 6-month journey, the power glitched and they’ve paid a bigger electricity bill than otherwise and contributed to global warming.

It’s the simple things like this that drive me to buy and read your magazine and I hope that this could "spark" an article for me to build. (N. O., Moruya, NSW).

Class-D amplifiers get hot

Recently I was in the market for a modest home theatre amplifier to replace my Panasonic SA-HT70 which is rather large and clunky, has an inbuilt and unhackable DVD player, and lacks any auxiliary input for the internal surround decoder circuitry.

I was interested in a couple of very low-profile and fully-featured JVC models and searched the web for reviews. Apart from some negative comments about their automatic set-up features, the most common thread was the large amount of heat they produce, even when idling! That put me right off the products.

While browsing in K-Mart one day I came upon a Sansui SAN0310 amplifier complete with five surround speakers and a subwoofer for the princely sum of $129.00! "Can’t go wrong at that price", I thought, and made my purchase.

I left the original Panasonic speakers in place and connected the new amplifier, taking immediate advantage of the coaxial and optical inputs. The Sansui amplifier gives a good account of itself as far as sound level goes but the lack of tone controls and the most bizarre "DSP" reverb effects I have ever heard (why would anyone listen to a movie in a mineshaft full of cotton-wool or in an empty cathedral?) make me sad when I think back to the excellent Sansui hifi gear of the 1970s and 1980s.

Anyway, you do get what you pay for, so I’m not complaining about that. What I do find objectionable is the large amount of heat the Sansui amplifier (and apparently the JVC class-D units) produce all the time. Even when ‘Off’ it gets quite hot, too hot for any gear sitting on top of it.

When the unit is turned off by the IR remote, the power LED blinks slowly – just what you need in front of your face when you’re not using it. There is a power switch on the front panel which disconnects the power but unfortunately all your settings for delay, etc are lost when you power off this way. This Sansui amplifier will not have a long career in my set-up but will suffice for the time being.

I have not studied class-D much but I assume that the heat is being produced by the switchmode power supply, the class-D switching transistors and the low-pass filters in the six power amplifiers. How can it be claimed that class-D is 80% or so efficient when it wastes power so freely? This is an outrage in these times when we should be conserving energy.

I’d be most interested to hear your views on this topic. (J. R., via email).

However, there does not seem to be any good reason why the unit should still get hot when it is in standby mode. It may be that the power supply it quite simple and does not have a low-power rail to run the micro when it is "asleep".

Speed controller for a Go-Kart

I have built an electric Go-Kart for my two young boys. It is powered by two 24V 300W motors, one driving each rear wheel via a 2-stage chain reduction with the overall ratio chosen to limit top speed to about 12km/h.

In use, the motors easily reach their rated RPM even with an adult aboard, do not blow the 40A supply fuse and are not overly warm even after 30 minutes or so of use. So I am satisfied that the motors are working well within their ratings.

The speed controller used was originally constructed many years ago using a combination of ideas from various "Electronics Australia" and SILICON CHIP train controllers. The controller was originally used with a much smaller motor, with only three FETs, and was actually only used a few times.

When the controller was modified to suit the current application, the output stage was revised to include a 4093 and eight FETs. The FETs are mounted as two groups of four on two separate heatsinks but drive both motors connected in parallel. Reverse is provided via a heavy-duty relay. The idea of using the 4093 and eight FETs was based on information gained from reading some recent articles in SILICON CHIP.

The Go-Kart has been used many times over the last three or four months and is a "hit" with the boys but on two recent occasions the speed controller has failed. On both occasions, it is the output(s) from the 4093 driving the FET gates that has failed. The FETs test OK and no other components appear to have suffered.

As the same failure has occurred twice, I suspect that I have a design problem. The latest failure occurred on a fairly hot day so the problem may be temperature related. Can you suggest a fix? (A. C, Boronia Vic).

The second and the most likely cause of the damage is that the switching transients at the FET drains can be coupled back to the gates (capacitively) to cause 4093 output damage. Thirdly, the gates of each FET are not isolated using 10Ω resistors (one resistor should be in each gate); the FETs tend to oscillate at switch-on without them.

As a minimum, we would recommend 10Ω gate resistors for each FET and a 12V zener diode between each gate and ground to clamp voltages above 12V. Further reliability would be gained by using the 4093 outputs to drive a complementary transistor buffer comprising a BC337 and BC327. The NPN BC337 transistor would have its collector to +12V and the PNP BC327 transistor would have its collector to 0V (ground).

Car radio noise cure

I noticed the letter asking for help with AM car radio noise in "Ask SILICON CHIP", January 2007. The usual cause for this is due to poor earthing to the car body at the antenna. ( D. A., via email).

Serial I/O Controller & Analog Sampler

I want to use the Serial I/O Con-troller & Analog Sampler featured in November 2005 to monitor and control the charge/discharge of some large batteries, a total of 48VDC at 700A.

I want to monitor both temperature and voltage and switch the load and the charger during daylight and at night. To do this, I need to be able to locate the temperature sensor and LDR remotely. Is this OK or will long leads (up to five metres) cause a problem?

Also, what do I need to do to monitor up to 52V DC at the analog input port (normally 25V DC max)? Do I use a voltage divider? I assume software changes would be required in the Windows program as well but what about the PIC code? (K. K., via email).

The second analog input can be modified to accept 52V as follows:

(1).replace the 330k&Omega resistor with 110kΩ.

(2).replace the 100kΩ trimpot (VR1) with a 2kΩ unit.

(3).Break the track between VR1 & LED4 (the ground connection) and insert an 11kΩ resistor in series with the pot.

The above changes will give an input range of about 49V-59V.

You will need to modify the Windows software to get the correct reading. The source code for the software is written in VB5. The PIC firmware should not require modification.

Regulated Power Supply For Amplifiers Wanted

I would like to suggest a regulated split rail power supply project for audio amplifiers. The design should have an adjustable voltage of as wide a range as possible but should be able to go up to ±37V and at least 3A of current.

There are many audio amplifier designs which I would imagine could benefit from the addition of a quality regulated supply, especially those with a high continuous current draw (such as class-A designs). The main reason I want a regulated supply is to get that extra bit of quality from the amplifier – to lower distortion. I would have thought that it would be impossible for any amplifier to have zero distortion (theoretically) unless the supply had a constant unvarying voltage.

The design should be able to be used for amplifiers in the small to mid-power range (say up to 60W) and perhaps the design could allow for a variable number of pass transistors to allow the current capacity to scale with the intended use. (P. T., via email).

As published, it delivered ±20V but it could be made adjustable, depending on the transformer input voltage. However, we do not regard such a power supply as providing any advantage for most class-B amplifier designs.

For a start, typical class B amplifiers have a PSRR (power supply rejection ratio) of 100dB or more, over a wide frequency range. This means that the supply rails can have quite large fluctuations with negligible effect on the amplifier’s waveform and distortion performance. There is also no advantage in terms of residual hum and noise from the amplifier itself. Provided the supply voltage does not droop so low that the amplifier is pushed into clipping, the fluctuations will have negligible effect. Amplifier designers have been depending on this factor for decades.

In fact, there is a disadvantage in terms of power output. Any normal class B amplifier will have considerably higher music power output than its continuous power output (the ratio is referred to as "dynamic headroom") if it is used with an unregulated power supply. That is why all commercial amplifiers do not have regulated supplies.

Making the supply adjustable to cover a range of output voltages also presents a difficulty because if the supply is to have a big difference between the input and output DC voltages, it will need big heatsinks to dissipate the waste heat.

Connecting PIR Sensors To An Alarm

I am a year-12 student at Ulladulla High School doing electronics and I am about to start my major project. It is going to be the PC-Controlled Burglar Alarm from the February 2006 issue of SILICON CHIP.

I have a problem with the PIRs for the sensors. They have a positive, negative and two alarm terminals. By contrast, the PC board in your magazine has positive, negative and one terminal for the alarm signal.

Does this mean that I put the two leads out of the PIR into the one on the board or do I need different PIRs? (N. P., Ulladulla, NSW).

The second two terminals provide "normally open" (NO) and "normally closed" (NC) outputs. Only one of these must be connected and it’s not important which one you choose. As described in the article, the NO or NC output of the sensor connects to one of the "ZONE" inputs on the alarm board.

When running the alarm software, be sure to select "N/O" or "N/C" in the "Configuration" panel to match the sensor wiring.

Converting A TV To An Oscilloscope

Have you ever designed a project that converts a miniature or portable B&W TV into an oscilloscope? It doesn’t have to be fancy but just has to have some accuracy. Would this be possible? (K. S., Adelaide, SA).

These days, a much more practical approach would be to use our sound card interface for PCs together with software available over the internet to give a much better instrument. Two articles on this subject were published in the August 2002 issue of SILICON CHIP, together with a follow-up Circuit Notebook modification in May 2003.

By the way, Dick Smith Electronics has a 10MHz single-channel oscilloscope (Cat. Q1803) available for just $128.00