My collection of Ferris inverters. The unit on the right is used on my home lighting plant and is fitted with a two pin polarised plug.

Ferris are most well known for a series of portable car radios; radios which could operate off the car electrical system, but then be taken out of the car and used at home on the mains supply. These radios had brackets to mount them in the car, and had a battery input connection, as well as a connection for 240V AC. The first, and best known of these was the post-war model 74. However, a much later set was actually a conventional mains/ dry battery portable, which was also designed to be used in a car by virtue of the extensive shielding and a socket to connect the set to a car aerial. This was the model 106 which is described here.  So that the dry batteries in the radio did not have to be used in the car, a companion 240V AC inverter was available. The M106 would be plugged into this just the way it would be at home.

The Inverter.
It was available for 6,12, or 32V supplies. 32V was of course a common home lighting plant voltage, while 6 and 12V models were available for car use. The inverter design is completely conventional, using a 6 pin MSP/Oak non synchronous vibrator. Output power rating is 240V 30W. Because of the vibrator type used, the output frequency is 100c/s. This increase in frequency is acceptable for the small valve radios intended to be used, because if anything, their power transformer will operate more efficiently, as will the B+ filtering circuit. Of course, resistive loads such as soldering irons and incandescent lamps do not care about supply frequency.

Front panel details.

The inverter is constructed in a steel box with a conventional flush mounted power point on the front. The power point has been modified so that the switch controls the DC input. For all practical purposes, using the inverter is just like plugging in an appliance at home. While it was introduced to power the 106, the inverter is advertised as being suitable for certain other appliances up to 30W.

The battery cable is screened and includes a shielded fuse holder. C1 and C2 provide input filtering, along with L1 and C3.
R1, R2, C4, and C5 provide RFI suppression for the vibrator contacts. C4 and C5 are 0.1uF. The 20R series resistors limit the capacitor discharge current when the vibrator contacts make, otherwise the capacitors would simply be shorted out having just been fully charged. A sudden discharge current would flow causing RFI.
C6 and C7 , each 0.25uF, make up the buffer capacitance, being 0.125uF in total. The output is balanced, with the common connection of the two buffer capacitors earthed. The 106 radio has a three core power cable, so that when plugged into the inverter, the radio is earthed to the car body.
The circuit for the three different voltages is essentially the same, with the main difference being the transformer primary voltage, and the vibrator driving coil voltage being altered to suit.
However, for the 32V model, some of the buffer capacitance is connected across the primary winding. In this model, there is also a 1uF primary buffer, and the components C4, C5, R1 and R2 are omitted. With the higher input voltage, it becomes practical to include buffer capacitance across the primary, which is actually the most desirable location because imperfect primary to secondary coupling does not interfere with correct buffer operation.

Here we can see the transformer, vibrator, and DC input filtering components.

This inverter has been restored and is in regular use. 0.22uF 2000V polyester capacitors were used to replace C6 & C7.

This one has also been restored, but is is missing the fuse holder and battery cable.

Interestingly, this inverter has been copper plated while the others are not. As can be seen, the buffer capacitors are original. They will have to be replaced before use.

Restoration.
I found two of these inverters at car swap meets, and the third came from the Wyong amateur radio swap. I paid between $5 and $20 for them. All three are 12V models and use a V4012 vibrator.
Restoration is very easy. Basically all that needs doing is to replace the 0.25uF buffer capacitors and cleaning the vibrator contacts if necessary. See the vibrator notes here. The primary capacitors do not need replacement as any leakage cannot cause harm. Depending on how the inverter is to be used, a suitable power lead and plug may need to be attached. For the unit I have in use, a I fitted it with period cloth covered twisted flex and a bakelite two pin plug for the house 12V supply. It is most important to see that the supply is fused. The value of this fuse is not specified, but 5A is a suitable choice for the 12V model.

Performance.
Provided the 12V supply is earthed, RFI is a non issue. For the last 14 years, I have been using one of these inverters to power a Breville portable valve radio with a clip-on mains power supply. This power supply was designed for the radio and uses a 6X5 rectifier for the 90V B+ and a selenium recitifier for the 1.5V A supply.

No load. The second waveform shows vibrator dead time.

Waveforms were taken from across the entire transformer primary. This is the standard method. Of course, the 12V supply to the inverter must be floating because the CRO has one side of its input earthed. The voltage at the pins of the 12V plug was kept at 12.6V for the tests.
The no load waveform is of the correct shape, with sufficient but not excessive buffer capacitance. Note that the dead time of the vibrator is about 1.72ms. It should be closer to 1ms. However, given the inverter has been operating a radio over the last 14 years, this is acceptable wear and would only require a slight adjustment.

From left to right: 15W load, 30W load, 11W CFL.

Under load, the waveforms are also as expected. A resistive load of incandescent lamps were used. The waveform shapes for both 15W and 30W loading are identical, although the amplitude is of course higher at the lower loading. It can be seen that for AC inverters, a resistive load discharges the buffer capacitance very quickly during the vibrator dead time.
The result is the buffer capacitance actually has little effect. Under these conditions, it is the loading of the inverter that prevent the destructive high voltage overshoots, which would otherwise occur with no, or insufficient, buffer capacitance. So, either way there is no contact sparking. However, a problem is that the buffer capacitor now is no longer able to reverse the transformer primary voltage at the correct point, meaning the vibrator contacts now carry the full current when they open and close. With a properly operating buffer circuit, the primary current flows just after the contacts close, not when they're starting to close. This is obviously kinder to the contacts. From this, it can be imagined that an AC inverter will have a shorter life from the vibrator compared to a DC-DC converter. That may well be, but as this particular inverter has been regularly operating a valve radio for 14 years, it serves as an interesting life test. I think we can say all is well, and in practice it isn't something to worry too much about!

The third waveform is with the inverter powering a Philips 11W compact fluorescent lamp. Here, the mains input is rectified inside the lamp to create a DC supply of the mains voltage peak. As a result of this, the inverter is actually operating along the lines of a DC-DC converter, as is the usual situation with a vibrator supply in a radio. The buffer capacitor works properly therefore. However, notice the slope is actually steeper and for a shorter time than the unloaded waveform. This is indicitave off too much buffer capacitance. It could well be there is a mains filter capacitor across the AC input of the lamp, adding to the buffer capacitance.

It should be obvious from all these waveforms that the type of load has a considerable effect on the operation of a vibrator inverter, and some loads will in theory cause a shorter vibrator life than others. It is not sufficient just to state, for example, "30W maximum loading" and to assume any appliance of less than 30W is suitable.

Furthermore, as the output is a square wave and at 100c/s, not all appliances are suitable, even if they are 30W or less. Resistive loads such as incandescent lamps or soldering irons are ideal, as are switch mode power supplies, such as USB and phone chargers. These rectify the mains input so only draw current at the peaks of the voltage. Compact fluorescent lamps of the electronic ballast type are also usually suitable loads as these rectify the incoming mains.
Some LED lamps or other devices that incorporate a capacitive dropper should not be used because they present a low, leading power factor, which adds to the buffer capacitance of the inverter output. This is harmful to the vibrator. Also, the reactance of the dropper capacitor will be much less on 100c/s, causing a much greater current to flow than at 50c/s.
Likewise, inductive loads are unsuitable because connecting these effectively subtracts from the ideal buffer capacitance. Also, such loads like induction motors and fluorescent lamps with iron cored chokes are frequency dependent, and will not function correctly at the higher frequency.
In short, the load should be kept to incandescent lamps, soldering irons, small valve radios, and solid state electronic devices.
 

Load	Input Current 12.6V	Output Voltage
No load	420mA	273V
15W incandescent	1.7A	238V
30W incandescent	2.85A	207V
11W CFL	1.3A	246V

Table of performance. The output voltage would be slightly higher if the vibrator was re-adjusted to specifications.
For the purpose it is still very satisfactory.

The above table shows the various output voltages with different loads. Regulation is not perfect of course, and no doubt the connecting leads, fuse holder, and fuse would all contribute to resistance in the supply line. Nevertheless, results are very good for this type of inverter. Note the excellent efficiency.

This ad appeared in RTV&H for November 1957. It appears Ferris was getting rid of excess stock, which would not be surprising given that transistor portables were starting to appear, making valve portables obsolete.