One of the very first practical types of transistors but made with a
technology that would be obsolete within a year or so.  To the left is
a later silicon PNP transistor similar to the one I used.

A 1955 advertisement for the first practical
commercial transistors.  Notice that the Ad
includes a small picture of the 7TP radio.

      After I repaired the transformer and after I replaced all the many electrolytic capacitors, but before I replaced the paper/wax capacitors, I attempted to get some sound out of the radio, but heard nothing.  For some reason this discouraged me and I got it into my head that this radio would take a lot of troubleshooting, so I'm sorry to admit it, but I put the radio in the back corner of my workbench and let several months go by before doing anything further on it.  However, I did make an inventory of all the capacitors that would need replacing and I did make up a kit.

     There is one thing I didn't have and that was a proper schematic.  I couldn't locate one and I didn't want to trace all the circuitry and try to make up my own schematic, so I just let the radio set until I could figure out how to get a schematic.  My number one rule is to never start a project unless you have a good schematic, but here I violated my own rule and it came back to bite me.

     All my usual sources of schematics failed me so until just a month ago this is how matters stood.  The radio was only half restored and I no schematic and that's the way it sat "on the back burner" for months and months.  A couple of weeks ago Lisa asked me about the radio and that brought it to my attention once again and since I had no other projects pending, I finally got busy and started looking for a schematic.  The little paper tag glued to the case indicated that the radio was a Raytheon 8TP2 and so that's what I started looking for.  I joined the Radio Museum as a dues paying member and looked up the 8TP2 radio.  There was a picture that looked exactly like my radio, so I downloaded the schematic of the 8TP and went to work.  


     I thought I'd trace the signal path from the antenna through to the second detector stage with my oscilloscope to see where I was loosing signal.   I couldn't make sense out of what I was seeing in the chassis using the schematic I downloaded.  Nothing made sense at all and then I thought of something, since this radio had only 7 transistors, maybe this wasn't an EIGHT TP (assuming TP stood for 'transistor portable') but was instead a SEVEN TP.  Well, I looked up the museum's 7TP and by golly, there it was.  Now all the circuits made sense and I had something I could really work with.

     Before I could trace the signals through the radio, I had to be sure all the old paper/wax and electrolytic capacitors were replaced or I'd just be wasting my time trying to trace signals.


     After all these many months I finally replaced all the old capacitors (except one) from my kit of parts I bought a year or more earlier.  When I had them all in, I applied power to the radio making sure that the chassis was the positive side of 6 volts.  Again I heard nothing and on tracing the signals, I discovered that the audio output (push pull) stage wasn't working.  I got no signal out of the output transformer (T7), but I did notice that one of the old CK721 output transistors was very hot to the touch indicating it was shorted.  My heart kind of sank when I realized that this transistor was destroyed (probably through moisture getting in there over the decades) and causing it to go into "thermal runaway."  I did a quick search to see if the CK721 transistor was still available, but at around $50 apiece, I didn't think it was worth it especially if there was something else wrong that would destroy it too.




     The original CK 721 transistors in the output stage were gone so I figured that I had nothing to loose by removing them and trying two PNP silicon transistors that had the power dissipation that I'd need.  Well, the first attempt was a disaster as they went into thermal runaway and were destroyed.  I changed the bias resistor to match the greater greater junction barrier voltage of a silicon transistor and discovered that the 100 ohm resistor was bad.  The bias resistor's function is to bias the transistors so that there is no or almost no current flowing through the transistors when there is no audio signal present.  This establishes a "crossover" point where one transistor conducts when the audio signal just starts going "positive" and the other conducts when the audio signal just starts going "negative" and if the crossover point isn't carefully adjusted, you get the dreaded "crossover distortion" which I'll discuss at length later.

     I thought I had the problem licked, but the radio still sounded bad and then I noticed that there was a resistor in there that fed back some of the output of the speaker's transformer to the input of the driver stage.  I had never seen anything like this done before and I could only make guesses why it was in there.  Obviously it was causing distortion and an oscillation that caused so much current to flow in one of the silicon transistors that it burned it out --- again.  I was very reluctant to mess with the biasing of the ancient V5 (CK721) transistor for fear of hurting it, but that resistor just had to go.  By taking measurements of the biasing of the V5 transistor, I discovered that removing the resistor would have no effect on biasing, however I was still a long way from having a restored and operating radio.

     These new high gain NPN silicon transistors I put in there had a huge amount of crossover distortion and the radio, although now making sound, sounded just horrible.  I just couldn't get the right balance of bias voltage on the transistors.  Just when I thought I had the bias right, one or the other transistor would start running away and I destroyed an additional transistor.

     I tried using an ordinary silicon diode to establish the bias because it would have the same junction barrier voltage as the transistors, sort of.  In other words, simply by replacing the bias resistor with a silicon diode, the push-pull stage should be biased close to Class AB2 operation and the transistors could not go into thermal runaway.  By golly, it seemed to work, but only later did I realize that at low volume, the crossover distortion was bad and this diode bias was no solution after all.  I'll discuss the troubleshooting I did and how I finally got the radio working properly, but first I want to change the subject and say something about how this radio works.  

    It seems to me that this radio was designed by people who were more used to designing for tubes because the design principles are so similar.  Even the identification of the transistors (V1 through V7) uses the 'V' for valve designition that was (and is) common with tube radios, but the 'V' was soon to be dropped in favor of identifying transistors by 'Q' numbers.  

     The 7TP is a conventional Armstrong superheterodyne consisting of a tunable local oscillator, a mixer to produce the intermediate frequency (IF), IF transformers for bandpass filtering, an IF amplifier, a diode detector for recovering AM audio and three stages of audio amplification to boost the audio signal up to speaker level.  

     Transistor V1 is a mixer who's base is connected to a coil that picks up the radio station from the ferrite loop antenna.  This pickup coil also has on it the oscillator signal from the V2, the so-called local oscillator (LO).  The station frequency and the oscillator signal mix in V1 to produce a 455 KHz IF signal which is filtered by T3 and then amplified by the IF transistor, V3.  The amplified IF signal is further filtered by L1 and L2 and then the audio is demodulated by V4, a germanium diode detector.  Again, the diode is designated with a 'V' number as if it were a valve (tube), but soon diodes would be have a 'D' name.

     The audio thus detected goes to audio transformer T4 where it is coupled to the volume control potentiometer.  There is a network on the V4 side of the T4 transformer that forms an automatic volume control circuit (AVC).  A strong carrier produces a negative DC voltage and through a network of resistors and capacitors, that AVC voltage, going up through coils in T1 and T2 changes the bias on the mixer, V1 to make weak stations louder and cut back on strong stations.  See what I mean about everything being so interactive and thus hard to tell what might be wrong with a circuit?

     After detection and recovery of the sound, the low level sound is amplified by transistor V5 where it is coupled over to transistor V6 through audio transformer T5.  From V6's collector, the signal is connected to transformer T6.  The output of T6 (the secondary) is phased split and feeds the bases of the push-pull audio output transistors V7 and V8.   A push-pull (class AB2) circuit is used in radios like this because there is no power wasted  and the power output transistors run a whole lot cooler than a single transistor running at class A.  By the way, for proper sound reproduction, the amplifiers must run at class A or if the transistor is biased for the more efficient class B, there must be two of them working in push-pull so that the result is a class A amplifier.  Here's how this radio's class AB2 amplifier works:

     When the audio waveform at the primary of T6 is going -- say -- "positive", one leg of the secondary of T6 is going positive too and other side is going negative (opposite phases input circuit).  The side going "positive" biases the transistor it is attached to even further into cut off and no current flows through this transistor.  When that transistor is cut off -- let's call it  V7 --  it no longer conducts current through its collector to the output transformer T7.  On the other hand, the "negative" going wave that is produced at this exact moment at the secondary of T5 biases V8 into an "on" state and a proportional (linear) current flows through its side of T7's primary windings resulting in a speaker voltage at the secondary of T7.  When the waveform reverses, it is now V7 that conducts and V8 is cut off.  It is that critical region where the waveform goes through zero going from positive to negative (or vice versa) that is is what's known as the crossover and the biasing on the base of the transistors must be adjusted for a smooth crossover to prevent "crossover distortion" and bad sounding audio.  In other words, when the waveform is at zero, both transistors must be biased off, but just barely biased off so that when a tiny voltage starts building up in the audio waveform, one or the other transistors starts to conduct in a linear fashion.  


         The AB2 amplifier is perfect for battery operation because they draw a lot less current than a single transistor class A1 circuit of equivalent power output and the AB2 amplifier is as linear and sounds just as good.  The reason for this is because the A1 amplifier's transistor is biased so that it operates at a 100% duty cycle whether the radio is playing loudly of softly.  The duty cycle of an AB2 transistor is never more than 50% and only reaches that at the highest volume levels.  Normally the duty cycle of an AB2 is waaaay below 50% and it only draws current from the battery as it needs it.

     Finally I should mention that the push-pull (double phase) currents from the collectors of V7 and V8 are at relatively high impedance (~6 volts and ~0.2 amps) and can't drive a loudspeaker directly.  T7 combines the two (double phase) collector currents from V7 and V8 and transforms their relatively high impedance down to 8 ohms which drives the (single phase) loud speaker (~3 volts and ~0.4 amps).

     The two original CK721 transistors in the push-pull circuit were biased for a class AB2 amplifier crossover configuration by resistors R26 and R27.  Of course, when I put in the two silicon PNP transistors in place of V7 and V8, the crossover bias was all wrong and way too low (silicon junction voltage is double that of germanium).

      After experimenting with biasing resistors and burning up a couple of transistors, I tried using a silicon diode to establish a safe crossover bias because, being made of silicon, the diode would have an junction drop of 0.6 volts that would be close to the static crossover bias necessary for those new transistors.  In other words, if you put 0.6 volts on the base of a silicon transistor (-0.6 for a PNP), the transistor is cut off (but just a little too much) and is close to a class B amplifier configuration.  Add another transistor identically biased but fed 180 degrees out of phase and you have a class AB2 amplifier --- sort of.

     To my mind at the time, the amplifier stage biased this way should have worked and for a while that's the way I left it.  The more I tinkered with the radio, the more I realized that I just wasn't satisfied.  After extensive listening and tuning around, I realized that the radio sounded pretty good if the volume was turned way up, but pretty bad when it was turned down to normal listening level.  This just wasn't satisfactory and so I tried something exotic.

     I had earlier bought one of those tiny, tiny amplifier boards from Adafruit that has a class 'D' amplifier chip on it (PAM8302 amplifier).  When I first got the board ($ 4 for the board and $ 4 for the shipping), I was absolutely amazed at the level of sound you could get out of the tiny little thing.  Another thing that amazed me was how cool everything on the board stayed, even at full power output.  In addition, the marvelous little unit has a differential, reasonably high impedance audio input and a screwdriver adjustable gain level.  I was sure that the board would be perfect for this application so I wired it up and hot-glued it in as shown in the photo below.  I proceeded to congratulate myself for finding a quick and easy solution to this vexing audio problem even if I was modifying the radio in a very radical sort of way.  Ah, but there were problems and eventually I had to go back to the drawing board.


     Even before putting in the Adafruit board, I had noticed that the two original geranium audio transistors (V5 and V6) were extremely noisy and even if they had a clean signals going into their bases, the signal coming out the collectors had a huge amount of "white noise" on them.  This was especially true of V6, the output driver transistor, it was extremely noisy.  I don't know if the noise was because of the very, very early fabrication techniques of these transistors or if the geranium crystal got contaminated by moisture over the decades.  

     Something had to be done about the noise, so I looked at the schematic and I realized that the original designers had made a bad mistake.  They had put the volume control potentiometer in the wrong place and that it really should have been connected to the input to the driver transistor (V6), not to the input to the audio preamplifier (V5).  The solution was to "move" the potentiometer over to the driver transistor.  Actually it was pretty easy to rewire the potentiometer by simply moving a few wires and capacitors and so I did.

     Changing the function of the volume control really helped a lot, but the radio still put out a lot of noise with no signal present and a lot of noise would accompany weak or quiet signals.  This accompanying noise made for a very poorly performing radio.  I finally decided that transistor V6 was just too noisy and it would have to go.  With great reluctance I replaced the noisy CK721 with a low gain, low noise Motorola BC416 silicon PNP transistor.  Wow, what a difference.   The static noise level with the volume turned down went from annoying to nothing and at first it was so quiet, I thought I had done something wrong.  

     With this solution to the noise problem and the Adafruit board installed, I thought I had the radio ready to give back to Lisa, but I had to give it a more complete trial before certifying it as ready to go.  During my tuning around on the AM band, I discovered some fatal (and I mean fatal) flaws in the Adafruit amplifier board that doomed its use.

     To my great surprise, I discovered that the class 'D' amplifier on the Adafruit board puts out a huge amount of RF interference at several locations on the broadcast band.  This interference is strong, it sounds terrible and makes using this little module impossible.  This little board is in clear violation of FCC Part 15 and perhaps this is why I never heard of using a class 'D' amplifier before.  No, the Adafruit board, cute as it is, would have to go.  Fortunately, a little hot air from my heatgun carefully applied to the board softened the hot-glue and I was able to remove the Adafruit board and pull out all its wiring without damaging anything.  So, there I was, back at square one trying to get a couple of modern silicon PNP transistors to work in the original Class AB2 amplifier.  

     After removing the class D amplifier board I again played with the bias on two push-pull transistors and again I burned up two more transistors, but just as all hope was lost, I had this "brilliant" (for me) insight.   Dense as I am, I finally remembered and realized that I wasn't working with tubes and bias voltages, I needed to think 'transistor' and bias currents.  Instead of trying to use a voltage divider and set a voltage for these high gain transistors, I should use a large value resistor in the base circuit and let the transistors draw a small current.  Since these transistors are (as mentioned) high gain, all I would need is a tiny current from a single high resistance at the center of T6's secondary and that would insure that my transistors wouldn't overheat and burn out.  

     Wow, wow, wow, did that work well.  I put in a single 75 K ohm resistor from the negative side of the battery to the center of the input transformer's secondary and that caused these high gain transistors each to draw a small current that established a good crossover point for class AB2 operation.  Oh joy, the radio was now low noise, the audio was very sensitive and easy to listen to with no distortion and because of this, I could hear weak stations that I couldn't hear before.  Best of all, there was NO RF interference and no overheated transistors.  Boy, if I would have only thought of this in the beginning, I would have saved myself hours of frustration.

     Finally, the radio is restored and is working just great.  Except for a few minor changes, the radio is in a configuration very close to its original design.   No, it is not completely original with original parts and yes, I had redesigned the audio section slightly, but now the radio performs as well and sounds as good as it did when it was new back in 1955 and as well or better than a modern AM portable radio.

     In paleontology we have several examples of so-called "missing links" although paleontologists just hate that expression and prefer "transition fossils."   Throughout the geologic record there are numerous fossils that show transition features that point to the evolutionary paths that all animal and plant life on earth has taken.  For example, there are very old, very primitive amphibian fossils that clearly show features of the fish they evolved from.  Seymouriamorphs were amphibians that were on their way to becoming the first reptiles.   Likewise, there are certain reptiles from the Permian that show mammalian features back when there were no mammals and then later, early creatures that were clearly mammals, but showing reptilian features.  These so-called "missing links" don't "prove" that we, as mammals had ancestors that were fish, and later ancestors that were amphibians and still later ancestors that were reptiles, but the evidence is there and it is very strong and it doesn't rely on magic or myths to explain why we are related to all life on earth.

     The evolution of electronics from spark gaps to personal computers is also full of transitional fossils too.  One such transitional device that shows characteristics of its earlier tube ancestors and yet is a fully functional transistor radio is the 7TP.  All the original capacitors were directly from tube technology with working voltages in the hundreds of volts.  Of course, when I replaced those capacitors I used much smaller capacitors with only 50 working volts, a voltage more suited to transistor equipment.  Even the way the transistors were laid out reminded me of triode tubes and what had to be done to "neutralize" them so they wouldn't self oscillate.  What really indicated to me that this was a radio designed by guys who were used to designing tube radios is how all the components were soldered to terminal strips and interconnected with wires and all installed by hand.

 

     Is it important to know the history and operation of this radio?  In a word, no.  For the vast majority of people who simply consume electronic products for the marvelous kinds of entertainments that are now available, the invention and evolution of early electronic technology is beyond their understanding and beyond their least desire to know.  It is only for the tiny few of us who like to delve into this early technology simply for the fun of it and certainly not for any kind of career or profit.  In the end, it's all a question of what makes you happy and what interests you and, in some cases, what satisfies your unique expression of that human trait called curiosity that we all have to a more or less degree and that can take so many different forms.

     The next question is, now that this wonderful old 7TP radio is working once again, is it good for anything?  Yes, it can receive AM broadcasts as well as any portable AM radio (and better), but is it worth tuning in on AM radio?  For me personally, no.  The only AM radio I listen to on a regular basis is a rebroadcast of my favorite FM stations that I do through my own little AM transmitter.