Converting 1948 radio to 12 volts

outbdnut

New member
My 1951 Special has a 1948 Sonomatic radio which is nearly the same as a 1951 radio. When I got the car, I had to cobble in a replacement power transformer that was not meant to be there, ( I took a transformer from a 240 volt to 5 volt Dc power supply and drove the transformer backwards, along with a full wave silicon diode bridge, and the voltage I got out of it, after also replacing the vibrator with a solid state one, was 110 volts DC; which is about half of what it should be, but the radio works. Now I'm converting it to 12 volts. I am putting in a 12 volt vibrator, and expect that the transformer output voltage should nearly double and that should put it right in the ballpark it should be. I'm going to rewire the tube filaments in series pairs, matching up pairs of 6 volt tubes with same filament current specs, so the tubes should run on 12 volts OK. Of course the dial light bulb I will change to a 12 volt bulb. I know the accepted way is to put a big resistor in series with the power wire instead of all this, or get an external power conversion box, but I'm a retired electrical engineer who also worked for a couple years in the late 1960s in an audio repair shop, so my conversion ideas are a fun challenge for me. I was wondering if anyone else has attempted a similar conversion method? Note I will also be putting a stereo in the glove box,but want the old AM radio to work too.
Dave
 
Buick radio 980797, 980798

The existing replacement power transformer should work OK as long as the increased primary voltage doesn't cause saturation of the iron core. If the iron saturates, the input current will increase sharply. Iron will be less likely to saturate as long as the vibrator operates at 115 Hz like the original. It's even better if the transformer was originally rated for use on 50/60 Hz power (meaning there is more iron present than on a 60 Hz unit).

Presuming the synchronous vibrator and the center-tap on the transformer secondary have gone away, might be a good idea to change to dual buffer capacitor/resistor setup over to a single buffer capacitor and resistor like on later sets using a non-synchronous vibrator.

The ideal solution of course would be to use a 12V vibrator power transformer from a parts radio.

A couple of other items to note:

Some schematics for the '48 sets indicate the speaker uses a 6V field coil. Easiest workaround here would be to install a permanent magnet speaker.

Also, there is a 6V clutch coil used on the tuner to de-couple the manual tuning knob mechanism when a pushbutton is used for station selection. This clutch coil was designed out of the '49 and later sets, so there is no 12V equivalent. The coil would need a series dropping resistor to maintain normal operation.
 
Last edited:

outbdnut

New member
Todd,

Thanks for your thoughts on my radio conversion to 12 volts. Since I did my first post on the subject, I rewired the tube filaments into series pairs. The output 6V6s now have their heaters in series, and the other 4 tubes all had a 300mA @6 volts spec, so I split them into series pairs for heater rewire. I did see the clutch on the tuner and measured the resistance at 2 ohms, so I added approx 2 ohms in series, wiring together two .56 ohm and a 1 ohm (it's what I had laying around), each with a 5 watt rating which should be enough, especially given the momentary, infrequent use when pushing tuner buttons. Someone had already put in a permanent magnet speaker. Then I wired in a 12 volt solid state vibrator from Antique Auto Radio.

When I put 12 volts to it and turned it on, the radio worked fine with the B+ supply going to 245 volts and then dropped down to 160 under load as the tubes heated up and started conducting (part of this drop, was when one side of the vibrator quit - read on). The radio worked OK for a short time, but the vibrator-induced square wave sinking the transformer to ground, viewed on a scope did have large spikes on the edges from the inductive collapse of the transformer at waveform switching transitions, and after a short time, first one output of the solid state vibrator died and then the other. I tore open the vibrator can and found the output FETs were in small, 8-pin DIP IC, and the plastic on one corner was fried. Whether the spikes killed it or the transformer I have was drawing too much current I don't know. The IC had no heat sink provision.

Running previously on 6 volts, before I converted the car to 12, the older solid state vibrator from the same company had worked fine with that transformer I cobbled in, and I used it for about 1500 miles of driving. Then I got thinking that since I had no further use for the 6 Volt unit, maybe the design was robust enough to withstand 12 volts, so I put that back in, applied 12 volts, and it promptly died. I opened up the failed 6 volt vibrator and being older (1998 date code) it's design was a circuit potted up inside 2/3 of the can with black potting compound so I couldn't see what was components were used. FYI - with this 6 volt setup, I was getting a B+ of 115 volts under load, so either the load was too much to sustain a higher voltage (which I doubt), or the transformer core was in partial saturation, or there were just some inefficiencies of driving the transformer backwards, as the now-primary low volt winding was not very close to the core, being wound over the high voltage winding.

So next. I plan on building my own solid state vibrator, with robust transistors, and will add some reverse diodes, to shunt the reverse inductive spikes. I can easily come up with a design, but also have done a websearch to see what circuits others have used and their comments. One suggested off times at the vibrator transition points of 10 % to 20%, which sounded like a good idea to clean up the switching transition. Since, by adding another wire, I can have access to the 12 volt battery to power a vibrator circuit, I have more circuit options than having to derive a supply from the transformer terminals I'm switching to ground. I'm thinking of using a 40106 CMOS hex Schmidt trigger as a building block. I have used Schmidt trigger stages in these 40106s in the past in medical product designs (before retiring) to very simply create oscillators and pulse width timing circuits with a minimum of components. With the one Schmidt trigger, I can set my base frequency and then build in the 10% delay windows using two other Schmidt triggers, and, of course, their are 6 Schmidt Triggers in the one IC. I'm thinking I will use 2N3055s for outputs to switch the transformer windings to ground, or, I am open to suggestions of FETS or other bipolar transistors. I'm thinking 2N3055s because I have some.

I'm putting more time into all this than most people would think it is worth, but this kind of stuff is fun for me, especially since I retired and no longer do this for a living.

Further thoughts and critiques are welcome! Also welcome are solid state vibrator designs you or others may have used.
Thanks!
Dave
 
Last edited:
Top