Nicad cell discharger

The life of Nicad cells can be extended by ensuring that they are deeply cycled from time to time, ie, discharged below 1V before recharge.

The ICL7665 is ideal for this job, as it incorporates two voltage comparators. Each comparator trips (changes state) when the voltage at its set pin (3 or 6) falls below Vref, nominally 1.3V. This circuit uses only one comparator and its output drives Q1 which connects a 68W resistor across the cells to discharge them.

While this circuit is designed to discharge two cells in series, the total-end-point voltage of 1.8V is not sufficient to reliably power the ICL7665, even though its minimum operating voltage is 1.8V. Therefore the circuit is powered from an external source via 5V regulator REG1. REG1’s output is effectively added to the 2-cell voltage, to give a total voltage ranging from around 7.75V with fully charged cells and a high-limit 7805 to below 6.7V with discharged cells and a low-limit 7805.

The resistive divider feeding pins 2 & 3 has been calculated using the measured output from REG1 (in the prototype it measured 4.99V) and the desired end-point voltage for the two Nicad cells (eg, 1.8V). The trip (change state) voltage (Vt) is therefore Vt = 4.99 + 1.8 = 6.79V

When Vt reaches the trigger point, the discharge load is removed and the cell voltage will rise again after a few seconds. Despite the presence of a hysteresis resistor (56kW), the rise may be enough to bring Vt back above the trigger point, turning the discharge transistor back on. Thus a low frequency oscillation will occur. Eventually the on period will become shorter and shorter, the net result being that the cells end up being discharged to the "no load" voltage of 1.8V.

Brian Critchley,
Elanora Heights, NSW. ($40)

CFL inverter has overload protection

This inverter was designed to run up to 10 compact fluorescent lights from a 12V battery bank at a remote location. This was because commercial inverters cause excessive RF interference which prevents shortwave radio listening. Standby pow-er was another issue.

This inverter has a low standby current drain of 40mA, very good regulation and no RF interference since the CFLs are supplied with filtered DC.

NAND gates IC1a & IC1b are connected as a square wave oscillator with a frequency of 400Hz. This is fed to flipflops IC2a & IC2b to produce complementary 100Hz pulse trains with are fed to gates IC1c & IC1d, to drive transistors Q1 & Q2 and Mosfets Q3 & Q4.

At the same time, a 200Hz pulse train from pin 1 of IC2a is fed to an RC network to produce a 200Hz sawtooth. This in turn is fed to the inverting input of comparator IC3 which acts as an error amplifier. It compares the 200Hz sawtooth waveform with a sample of the output voltage derived from zener diode string ZD1-ZD8.

IC3’s output is fed to the other inputs (pins 9 & 13) of NAND gates IC1c & IC1d via a 100nF capacitor and 820kW resistor. This provides pulse width modulation drive to the gates of the Mosfets (Q3 & Q4) at up to 50% duty cycle, corresponding to full power.

At above 50% duty cycle, the RC network at the output of IC3 effectively holds the gates of IC1c & IC1d at zero potential, depriving the Mosfets of gate drive. The circuit will shut down until the overload is removed and the reset button (S1) is pressed.

The transformer is based on a halogen lighting unit rated at 150-200W. The low-voltage windings were removed and rewound with 45 turns per side, using 1mm enamelled copper wire.

The Mosfets drive the low-voltage winding of the transformer which is used in step-up mode.

Dave Edwards,
Westland, NZ. ($60)

Spa heater control

This circuit can be used to replace old or faulty electronic thermostat units for spa/pool gas heaters.

Any spa/pool heater over about 10 years old is likely to use a simple continuous pilot gas valve. The pilot remains alight and a 24V solenoid valve, controlled by a relay and a simple analog electronic thermostat, is used to turn the main gas jets on.

This circuit is a big improvement over the original as it has a digital display and the temperature can be accurately set and controlled. It is based on a PICAXE-18X microcontroller and a Dallas DS18B20 temperature chip.

Because the PICAXE cannot easily drive multiple 7-segment displays, a 74C925 4-digit counter is used to do the job; it only requires three control lines from the micro to drive it. All the program does is reset the 74C925’s internal counters, pulse its clock line by the number that is to be displayed and then latch this count into the display register. Because the latch is pulsed at the end of the count cycle all the user sees is the new value being shown.

Safety features

The old controller had two safety features which are incorporated in this design. First, there is a pressure switch which is connected to the heat exchanger. When there is pressure in the pipe (ie, the pump is on), this switch is closed. The original controller had this in series with its power switch so whenever the pump was turned off the controller shut off.

In the new design, it was desirable to be able to read the temperature with the pump off, so the pressure switch is in series with the power feed to the relay. When the pump is turned off, no voltage is fed to the relay coil. This automatically shuts off the main gas supply but still leaves the controller with power.

There is also a sense line to the PICAXE so it knows whether the pump is on or off and immediately stops energising the relay control transistor (Q3) and extinguishes the pump-on LED (LED2) Even if the PICAXE fails to de-energise the relay due to a glitch, the relay can’t remain on as the power feed has been removed.

The second line of defence is the use of a high-temperature resettable fuse. This is mounted directly to the heat exchanger and will open if there is excessive temperature detected. As it is in series with the valve solenoid and power supply, if it opens there is no power to drive the gas control solenoid. While on the subject of the gas valve, be aware that the Honeywell valve solenoids have inbuilt diodes.

Brass tubing

The DS18B20 was mounted in a short piece of brass tubing with one end closed with a small piece of copper (soldered). Thermal paste was applied to the chip to ensure good heat transfer to the end copper plate. A shielded USB cable was used for the connection and the whole assembly was inserted into the heat exchanger in place of the old sensor.

It is important to try to isolate the sensor as much as possible from the surrounding heat exchanger metal; you want the sensor to read the water temperature, not the heat exchanger metal work. Even with this care, I still noticed a difference of about one to two degrees between the displayed value and the actual spa temperature.

The software (spaheater.bas, available for download from the SILICON CHIP website) is fairly straightforward and should be easily modified to suit other heaters.

The main program is just a series of subroutine calls and one of the general B registers is used as a status register to keep track of various heater conditions. Depending on the subroutine, it can either read the status of the relevant bit in the register and perform an action based on this or it can change the state of a bit.

For example, the temp subroutine executes a readtemp command on the DS18B20 and the value is stored in a free register. However, if the return value is zero, the program assumes there is a fault reading the chip and sets the fault bit on the status register.

When the gascall subroutine is called (operates the relay), the first thing it does is check this fault bit and if set, always de-energises the relay.

There are four changeable variable values in the program. The low and high point values are set to match the spa/pool heater minimum and maximum temperature values (eg, 20°C and 40°C).

The other two set-point values determine when the heating cycle stops and then restarts. Setting it 1°C above and 1°C below the set-point produced an on-off cycle duration of about 10 minutes, once the spa was up to temperature.

For example, if the set temp-erature was set to 37°C, the heater would stay on until it read 38°C and then not come back on until the temperature dropped back to 36°C. If you want to increase the on-off cycle duration, just change one or both of these offset values.

The unit is fully automatic in operation. There are only two push buttons and the on/off switch visible.

To set a desired temperature, simply push and hold one of the buttons until the "set temp" LED turns on. The display shows the last "set temp" value. As soon as the LED turns on, repeatedly pushing the Up button will increase the set temperature and likewise pushing the down button will lower the set temperature.

Once the preconfigured temperature limit is reached, further pushing will not change the display higher or lower than the preconfigured range.

After about 1.5 seconds of no pressing, the "set temp" LED goes out, the display resumes showing the actual temperature and the new value is stored in EEPROM. This value is recalled the next time the unit is turned on.

Clive Allan,
Glen Waverley, Vic.

Clive Allan is this month's winner of a Peak Atlas Test Instrument!

Contribute And Choose Your Prize As you can see, we pay good money for each of the "Circuit Notebook" items published in SILICON CHIP. But now there are four more reasons to send in your circuit idea. Each month, the best contribution published will entitle the author to choose the prize: an LCR40 LCR meter, a DCA55 Semiconductor Component Analyser, an ESR60 Equivalent Series Resistance Analyser or an SCR100 Thyristor & Triac Analyser, with the compliments of Peak Electronic Design Ltd www.peakelec.co.uk So now you have even more reasons to send that brilliant circuit in. Send it to SILICON CHIP and you could be a winner. You can either email your idea to silicon@siliconchip.com.au or post it to PO Box 139, Collaroy, NSW 2097.