Restoring a Canadian Philco Model 3650 Gerry O'Hara VE7GUH Jun 2011

	Cabinet awaiting completion.

This mid-1930's Philco radio was brought into SPARC in February, 2011. The set was reportedly not working, and the owner also reported that the speaker voicecoil had been rubbing.

The Philco Model 3650 is the Canadian-made floor model version on the US-manufactured Philco Model 650, however, the chassis is essentially the same on both models, comprising a 8 tube superhet with an RF stage for additional selectivity/sensitivity, plus a fairly beefy push-pull output stage.

This article is also available in PDF form.

Assessment

Given that the set was reported as being not working and the power lead had been spliced and insulated with some sticky-tape, I decided not to risk switching the set on and instead took a look under the chassis. Nothing obviously wrong, though I noticed that a few capacitors had been replaced at some time in the (dim distant) past, including one of the filter capacitors and two coupling capacitors. I also noticed that a toggle switch and phono socket had been installed on the side of the chassis, wired between the detector stage and the volume control, allowing a gramophone pickup to be connected to the audio stages of the receiver. Apart from that, the remaining components looked all-original with the exception of a 10 Watt wirewound resistor. The first job was to remove the thick wads of fluff from under and on top of the chassis, together with some waxy tar where capacitors had deposited their coating over the years. Next, a continuity check of the power transformer was undertaken - replacing this adds considerably to the cost of any repair, so I wanted to check it out right at the start - it was fine, as was the filter choke and audio driver transformer.

Restoration

	Underneath chassis after restoration.

Philco sets from the early-mid 1930's period were fitted with a mixture of Bakelite-bodied tar-encapsulated paper capacitors and waxed paper capacitors. The Bakelite-bodied capacitors have solder lugs on their upper surface that are used as anchor points for wires and other components, making replacement somewhat problematic. These units often contain more than one component (several capacitors and, sometimes, a resistor). I tested a couple of these capacitors and found them to be leaky, and so decided to replace all them all with modern plastic film types. There are two ways of replacing the Bakelite-bodied capacitors:

• Remove all wires/components from the solder lugs (or at least enough to turn the capacitor upside-down such that the tar can be removed with a screwdriver blade or similar) once the fixing screw has been removed, removing the old capacitors and tar, and then 're-stuff' them by installing replacement capacitors, pushing their wires through the holes in the solder lugs and soldering in place, then re-attaching all wires and components. There is no need to re-fill with tar as the (open) base of the Bakelite case is against the chassis, so the repair is invisible; or
• Snip the (thin) wires that run from the internal capacitors where they are fixed to the solder lugs, then use a 1/16” drill to remove the remnants of the wires, tamp the tar into the hole and fill with a spot of epoxy glue. The new capacitor(s) can then be fixed to the solder lugs external to the Bakelite case.

All electrolytics were replaced (under the chassis), with the old capacitor cans left above the chassis for cosmetic reasons. During capacitor replacement, I took the opportunity to test resistors. I found that the replacement 10Watt wirewound resistor was open circuit - on close inspection this was also found to have been an incorrect value resistor. The failure of this component was likely the reason the set was not working - it was replaced with the equivalent of a higher-wattage resistor (20Watt) as a precaution (even though the dissipation was calculated to be only 7.5Watts). The underside of the restored chassis is shown below.

Some under-chassis wiring was replaced where insulation had degraded, as was some of the top cap connections on the tubes. Also, an in-line fuse was fitted to the power transformer primary.

Replacing the Tuning Dial

	Fitting the new dial.		Nice new dial.

The original dial was cracked in two places. A replacement (reproduction) was obtained from Radio Daze and fitted. However, it was found that the replacement was more opaque than the original, which resulted in the dial light 'pointer' (which shines through a mask, the position of the aperture of which corresponds to the band selected from the rear of the dial) being almost invisible, even with a new dial light bulb fitted. The fairly simple fix for this was to use a high-output LED fixed into a #46 (screw) bulb base. A 100ohm resistor was used to limit the LED current to 18mA and a yellow LED was selected to provide a more suitable colour tone for the 'pointer' (I find white LEDs tend to have a bluish tinge). This assembly was wrapped in heat-shrink tubing and installed in the dial light socket - it can therefore be easily replaced with a standard bulb if the owner wishes to do so, however, an advantage of the LED solution is that it will likely never need replacing due to burn-out. A regular new bulb was installed in the 'Shadowgraph' tuning indicator which was bright enough (as this was still shining through the original translucent window above the new dial).

Finishing up

	Chassis before restoration ...		... and after.

The re-furbished chassis was then tested and re-aligned. It is quite a performer on both Broadcast and Shortwave bands - that RF stage really does make a difference. It is also remarkably stable - I left it on soak test on WWV (10MHz) for over ten hours and it remained 'on the nose' throughout, and the audio is rich and full.

At the time of writing this article, the cabinet is still very much a work-in-progress. The veneer on the top of the cabinet was in very poor condition, in particular, the centre section need to be replaced.

The article will be updated once the restoration work on the cabinet is completed and the radio is ready for return to the customer.

Cutting out old caps.

Removing remnant leads.

Fluff found under chassis.

Refurbished tone control.

Cabinet before restoration.