While it is not compulsory to have indicators
and brake lights fitted to vintage cars in Australia, I thought it wise
that my Model T should have them. Few understand hand signals. So, I decided
to fit signal lights others would recognise.
I decided I'd take a different path to everyone else and design my own flasher unit. 6V flashers are available from all the vintage parts suppliers, but I wanted something more tailored to my requirements, and be immune to the limitations with thermal flashers.
The flasher circuit is built in a diecast box bolted onto the chassis cross member. The terminal strips on the sides go to the lights and power, while the DIN plug connects the indicator switch.
Advantages of electronic design
An electronic version would eliminate the heavy wiring to the indicator switch to and also provide a user adjustable flash rate. Some examples of 6v indicators I've seen seemed to suffer from an excessively long "off" period and dim lights.
As the voltage supply for the oscillator is regulated, flash rate is constant over the normal voltage range of the battery
Unlike thermal flashers, this design does not depend on correct lamp wattage to function properly. Thermal flashers operate by virtue of a heating wire and bimetallic switch. As the lamp current passes through the heating wire, it heats up and causes the bimetallic switch contacts to touch. This shorts out the heating wire and full voltage is applied to the lamps. The bimetallic switch cools and the process repeats. It can be seen that the time taken for the heating wire to actuate the bimetallic switch depends on the current flowing through it, and thus the lamp wattage. This often causes problems when LED's are used instead of incandescent bulbs. To overcome this problem, load resistors can be connected in parallel with the LED's to increase the current draw. However, this detracts from the efficiency aspect of using LED's.
LED indicator lights
While LED's and vintage cars may appear to be a contradictory combination, there is actually a lot of interest in using them. By default, this flasher is ideal for LED's because the performance is not affected by their lower current draw, and circuit completely breaks the supply to the lights in between flashes. It is possible to modify commercially made 12V or 24V LED signal lights to run on 6V as the individual LED's run on between 2 and 4V. To do the modification, you'll first need to power up the light on its intended voltage, and measure the current through each chain of LED's, (usually around 20-40mA) and see what the actual LED voltage is. The 12 and 24V lights have series/parallel wiring where groups of several LED's are in series with one current limiting resistor. The LED's need to be rewired so each has its own resistor. This will allow 6V operation. To calculate the resistor required, use the formula R(in ohms)=(6-Vf)/If, where Vf is the LED forward voltage, and If is the LED forward current. Quarter watt resistors will be adequate.
Circuit is based on two CMOS IC's and two relays. It eliminates all the troubles of thermal flashers and provides audible warning.
The flasher circuit is a schmitt trigger
oscillator built around a 4093 CMOS quad NAND gate, which then drives one
of two relays. The contacts of these relays switch the left and right side
The supply for the 4093 is stabilised at around 3.9 volts with the zener diode and 220 ohm resistor. While CMOS logic gates are happy with supply voltages from 3 to 15, the supply does need to be regulated so that the voltage of the car electrical system will not affect flash rate. I like to design for a 5 to 7 volt operating range. This allows for when the generator is charging and also if the car is stopped and other loads are on.
A second oscillator provides an audible tone from a miniature speaker so that the driver knows to cancel the signal when required. Additionally, to prevent having to listen to the beeping tone while waiting to turn when stopped, it is muted when the brake is applied.
The way the flasher oscillator works is very simple. The gates are simply wired as inverters with the output fed to the input. There is also a capacitive time constant at the input. Assume the circuit is first powered up. The 2.2uF capacitor will be discharged, meaning 0 volts at the input. Because the gate is an inverter, the output will be high at around 3.9V. Current flows via the 180K and 100K pot to the capacitor which starts charging. At a certain point, the input will see a high thus switching the output low. The 2.2uF now starts discharging through the resistors, and so it keeps oscillating.
Unlike commercially made flasher units, this one is adjustable. So, you can select your preferred flash rate simply by changing the RC time constant, hence the 100K trimpot.
As the CMOS gates can only supply about 20mA, a BD140 power transistor is used to drive the relays. The diode between emitter and collector bypasses back EMF when the relay coils turn off. I prefer this method to the usual one of wiring the diode across the coil as it means the transistor won't be damaged if the diode shorts, and the transistor is also protected against reverse polarity. The back EMF is absorbed by the power supply.
I did discover during the design that different types of 4093 will result in different flash rates. I'm using a CD4093BCN. If you use an MC14093BCP you'll need to increase the resistance...start with about 1.5M, or increase the capacitor value.
Note that the 220 ohm current limiting resistor is in the negative supply to the zener diode and 4093. This is done because a PNP transistor is being used to drive the relays, and therefore the 4093 positive rail must be at the same voltage as the transistor's emitter.
Audible warning circuit.
With my original design, a buzzer was connected in parallel with the pilot light on the end of the indicator switch. However, I found its operation to be unreliable because it was actually a 9V buzzer. Also, it was annoying while waiting at an intersection for several minutes having to listen to it beeping. The light and buzzer were switched with the extra relay contacts directly.
I wanted a reliable buzzer, but one that would cease operating when the brake was applied. A slight problem was to have the buzzer in the same location (under the indicator switch) while being able to switch it independently of the pilot light. It would require a separate wire run up to the switch. This was inconvenient.
So, how to run both the light and buzzer supply through the one existing wire but operate the light without the buzzer?
I designed an ingenious scheme where the buzzer would be a speaker running on AC, at about 1KHz. This would also pwer the light (6V 1.2W). When it was desired to silence the buzzer it was only necessary to change the supply to DC. This would be blocked from the speaker via a capacitor and thus have no effect.
The heart of this circuitry is another schmitt trigger oscillator, except running at a high frequency. There will be noticed a diode in the feedback circuit. This changes the duty cycle so that the "on" time is about 90% and the "off" time is the other 10%. The reason for this is so the pilot light runs at the highest power possible, thus maintaining its brightness, but still providing an AC component to drive the speaker.
Another BD140 switches the pilot light current. In series with the emitter is a 1R resistor. This is necessary only to act as a fuse and protect the transistor should there be a short circuit around the indicator switch wiring. The speaker is mounted across the pilot light, at the indicator switch. The 2.2uF blocks DC from the speaker voice coil.
When the brake is applied, voltage from the brake light circuit is applied to the BC548 via the usual base resistor circuit. This shorts out the .015uF and stops the 4093 oscillating.
Pin 4 stays low, turning on the BD140, sending 6V to the pilot light. It has no effect on the speaker which does not respond to the DC. In my Model T, the brake light is also the tail light running at reduced voltage. It can be seen that the indicator buzzer will also be silent with the lights on. However, the pilot light is quite visible in the dark, so no audible warning is required.
These are double pole units with 6v coils. The ones I used were actually 4PDT units with a coil current of about 200mA. The indicator switch determines which coil is activated simply by completing the earth return of the required relay coil. In the centre position neither coil is selected and the indicators are off. The second set of contacts simply switch the supply to the audible warning and indicator switch pilot loight circuit.
So that the oscillator isn't running all the time when the indicators aren't being used, there's a two diode OR gate connected to the indicator switch which completes the earth return for the 4093's supply. I used 1N914's in view of the low current but just about anything can be used.
I haven't shown it in the circuit diagram, but this design makes it convenient to add a hazard light switch. Simply add a small DPST switch in parallel with the indicator switch with the common contacts connected together and earthed. Of course a SPST switch could be used with the addition of two diodes instead.
Circuitry and relays fit in a small zinc diecast box.
The circuit was constructed on a small
piece of veroboard and placed along with the two relays in a zinc diecast
box. This was then mounted on the front chassis cross member, a convenient
hole already existing to mount the box. A bakelite screw terminal block
connects to the indicator lights and 6v supply, and a 5 pin DIN socket
used to connect to the indicator switch. Light stereo figure eight cable
was run to the switch. In the Canadian version of 1926 Model T exists a
tube running up the steering column for the horn switch wires. I used this
same tube to run the indicator switch wires...something you couldn't do
had the indicator switch been wired the conventional way carrying the full
The indicator switch used is a Hella 4208. This can be ordered in from your usual parts supplier. I paid about $74 from Scott's Auto 1, back in 2003. Repco wanted about $128! It is a far superior product to the metal chrome plated switch that all the repro parts suppliers sell, which is of very flimsy construction.
Have a look at the Hella 4208 switch here.
A tiny PC mount speaker was mounted on the underside of the switch, with the 2.2uF blocking capacitor mounted inside the switch body.
The actual indicator bulbs I use are 6V 10W as shown here. Of course 6v bulbs are available from other suppliers as well. I use 1929-31 Model A rear lights with the dual coloured lens; red for brake and parking and amber for indicator. The front lights are small motorcycle lamps I got at a swap meet.
The golden rule with 6 volt electrical systems is to run a separate earth wire from each light directly to the negative battery terminal. Don't rely on rusty chassis connections. If you have good connections and use wire with an appropriate current rating, your lights will work as well as those on a 12v car.
The circuit is adaptable quite easily to run on 12V. It's simply a matter of selecting appropriate relay coils. The other components are rated in excess of 12V so no modification is required. The 3.9V regulator needs no modification. However, given that the CMOS IC's are rated at 15V, it is wise to connect a 15V zener diode across the supply pins of the tone generator IC. That is, across the 47uF capacitor. This will prevent any high voltage spikes damaging the IC.