The Smoothflo AC79/6 Inverter is a compact unit built around the transformer.

This inverter was obtained from the Ballarat swap meet around 2001 or 2002. Being 6V and not yet owning the Model T, I didn't actually have a use for it, but as it was a vibrator inverter I had to have it.
Not much is known about "Smoothflo" which is a brand name of Industria Manufacturing Engineering, who once existed in Sydney, at least into the mid 1960's. I have seen one other inverter identical to the one described here; I have a larger 12V Smoothflo inverter, and I have seen one of their 240V isolation transformers.
The inverter described here is their model AC79/6, which produces 240V 100c/s at 30W from 6V DC.

This ad from 1950 paints a rather optimistic picture of why one needs a Smoothflo inverter.

Inverters such as these had a brief popularity in the 1950's and then suddenly disappeared. As can be seen from the ad shown, "going on holidays" was a valid reason as to why one might need an inverter.
Back then, "holidays" meant staying in a beach shack, country home, or caravan, and such places often had no reticulated electricity supply. If there was no lighting plant (usually 6, 12, or 32V) on site, the battery could be borrowed from the car, or a separate battery taken for the trip. Interestingly, the ad shows an Astor radio, which if it's the one I think it is, is probably a poor choice to use in a remote area being a simple one stage TRF design. Added to that, it turns out the Smoothflo inverter has very poor RF interference suppression. Not a good combination for country listening!
By the 1960's the power mains had spread further afield, and with transistor radios starting to become common, the idea of lugging a car battery, inverter, and full size mantel radio along would seem rather unappealing for areas still without mains.

Back of inverter lists details. Power cable was replaced.

The design of the inverter contains nothing unusual and is the usual vibrator and transformer circuit. The top half of the transformer is fully exposed, and the bottom half is enclosed with a deeper cover that contains the vibrator and forms the base. This type of construction was common with other transformer based devices around the time. A bakelite HPM three pin socket provides the 240V output.

The vibrator is an MSP/Oak V6606 dual interrupter type, and thus operates at 100c/s, as is typical with radio type vibrators. Rudimentary RF filtering is provided by a .47uF paper condenser across the 6V supply. When this inverter was first examined, the buffer condenser was another .47uF paper type. This turned out to be completely unsuitable. Ideally, with dual interrupter vibrators the transformer should have two identical primary windings so that current sharing between the contacts is more even. The simpler approach of simply paralleling the contacts is taken here, which limits the output power to about 30W. This is because it is impossible to ensure each pair of contacts opens and closes at exactly the same time.

Underneath the unit shows the simplicity of construction.

Opening the base is difficult because it's a tight fit and is hard to lever it out without damaging the sides of the case. The original power cable was shielded but only a short length was still attached. There was no fuse anywhere to be seen. If there had been one, it must have been an inline type in the missing length of power cable. So, I replaced the cable and terminated it with battery clips. I installed an 8A inline fuse. The .47uF filter condenser on the 6V supply was left in situ as any leakage is harmless.
The vibrator was checked and not surprisingly found to be in excellent condition, as the Oak type usually is. As it turned out, I doubt it had much use anyway.
When it came to replacing the buffer condenser, I simply replaced like for like. Upon powering up the inverter, there was arcing at the vibrator contacts and when I felt the power cable connection to the power supply it was rather hot. That "7.5A" rating stamped on the back was the no load current draw! Once a load was added, the supply current only worsened of course, and there is no way a V6606 vibrator would last long operated like that. Obviously, the operating conditions were totally wrong. That the vibrator contacts looked hardly used suggested the owner found the high current drain unacceptable so only used the inverter once.
All the time I had thought that .47uF was higher than normal for a buffer condenser in this kind of circuit, with a 100c/s vibrator, so I tried lower values. As it was, the output waveform was rather rounded with .47uF which threw suspicion on this value.

Waveforms of the output. Left is loaded with a 28W light bulb. Right is unloaded.

And that was it. With .047uF, the no load current at 6V was about 600mA and no arcing on the vibrator contacts. The waveforms were now correct. Fully loaded, there was no arcing either. I used a polypropylene 2kV replacement (KP series from WES Components). I have a suspicion that when the inverter was assembled, the person doing so picked out one of the .47uF's meant for the 6V supply instead of a .047uF for the buffer. Easy to do if one isn't taking notice of the decimal point. Although, the physical difference in size should draw attention to the mistake.
Alternatively, the inverter really had been designed with .47uF with no one actually measuring the current draw or checking for contact arcing. The qualifications of the designer are not known of course, and it's quite possible he did not know how to properly design a vibrator power supply. Given the construction of this, and my other Smoothflo, there is an air of backyard or low quantity production construction quality, about their products. The earth pin of the power socket had not been connected to anything, so I added a wire to the chassis, which is also connected to one of the 6V input leads. It is true that mantel radios were connected with only twin flex so it would have been unimportant at the time. Of course, there are no polarity sensitive components so the inverter may be run with the battery connections either way round - except when used in a car, in which case the inverter case should be the same polarity as the car body so as not to cause a short circuit.

Here the inverter is powering a 28W light bulb.

In terms of 6V to 240V conversion, the inverter now works properly. With a 28W incandescent light bulb, output voltage is 237V rms with an input of 6.2V DC. Input current is 5.5A. Vibrator noise is minimal, and the rubber grommet mounts for the vibrator socket help minimise this.
When used with a radio the RF interference is unacceptable. There is just inadequate filtering. I did find that earthing the case of the inverter directly helped considerably and with a strong station the interference was tolerable. When I tried earthing through the black wire which connects the chassis, the interference was not subdued. Obviously, this is because of the current flow in the supply leads.
Even on VHF, the interference was terrible. It was this inverter I was using to do tests on the use of a super regenerative receiver for the Model T Ford. I took my mains operated 12AT7 receiver out for a drive in the Model T, running it off this inverter. It confirmed the idea would work, in terms of receiver sensitivity, but the interference from the inverter was pretty bad.
On the contrary, the Ferris is an example of an inverter that has been properly designed and is suitable for radios.

I would prefer the Smoothflo to have had a power switch. Connecting or disconnecting an operating inverter (or any other load) from a gassing lead acid battery runs the risk of explosion as the spark can ignite the hydrogen gas that is released.

It is interesting to speculate as to what kind of loads this inverter would be used with in 1950. Apart from mantel radios, there wouldn't be a lot. 30W output is too low for soldering irons of the day.
100 c/s output frequency kills off any idea for things like table fans with induction motors, or fluorescent lamps. You could run a 25W light bulb, although I would expect a 12V counterpart to be more likely used. An AC/DC shaver is one suitable candidate, but I can't think of anything else. In the modern day there are many more things of a low power solid state nature which could be used. For example, some mains operated test instruments, portable sound equipment like CD players and cassette recorders, solid state monochrome TV sets, etc.

This advertisement from 1951 shows what appears to be the same inverter in kit form.