A type number is given to a transistor when it falls within certain maximum and minimum values called a production spread. This spread, for practical reasons, is large. Production techniques for maintaining rigid, uniform characteristics have not been developed. Often, the better transistors - with high current gain - are selected from a production run and given one type number while the lower gain units are given another. High-gain transistors are thus picked and marked for sale at a higher price.
Data sheets for lower-price transistors usually give only an average value of current gain without specifying any minimum. This is a "pig in a poke" situation where the buyer cannot be very sure about the quality of what he is getting.
Of two similar transistors purchased "off the shelf" one may operate well into the megacycle region; yet, the other may be good to only a few kilocycles. The purchaser of transistors must recognize these facts and realize that they exist even in brand new units that have never been connected to a circuit. The experimenter will find it highly advisable to measure the gain of his transistors even before they are put into service. By doing this he can then reasonably judge the capabilities of the transistor and thus use it in circuits where it will perform most efficiently.
Adding to these initial unit-to-unit variations are the changes caused by the transistor's electrical and physical environment, such as excess humidity, careless overheating, and exceeded ratings. Wrong socket insertion, the surge of a charging capacitor, and excess supply voltage are just a few of the things that can cause harmful changes or permanent transistor damage.
It is a good idea to keep a record of a transistor especially when it is used in different circuits. An occasional gain check can reveal any transistor damage and isolate the cause.
Current gain is the most useful single transistor measurement. This gain varies with the transistor's type of operation. For grounded - base operation, the current gain is termed alpha and is always less than 1 but approaches 1 for the junction transistor. Alpha is the ratio of collector current change to emitter current change for a fixed collector voltage. This is sometimes referred to as the short-circuit gain since it does not take impedance change into account. Manufacturers usually express the gain of their tran¬sistors in terms of alpha.
However there is another term that expresses current gain for grounded-emitter operation. This grounded-emitter gain is called beta and bears a simple mathematical relation to alpha. If you know one you can easily find the other.
While alpha is always a decimal fraction less than 1. beta is ex¬pressed in small numbers greater than 1. With the beta measure¬ment it is possible to glance at the gain and quickly know just how much better one transistor is than another. Alpha is not always so obvious. For example, how much better is a transistor with an alpha of 0.98 than a transistor with a 0.94 alpha?
The meaning and relationship between alpha and beta can be seen by observing the electron flow in the grounded-emitter and grounded-base circuits. These two circuits are shown in Fig. 201.
Suppose, as in Fig. 201-a, a signal change of 1 ma is applied to the emitter and this produces a 0.95 ma change in collector current. Alpha then must have a value of 0.95 because it is the change of collector current divided by the change of emitter current. Since the current flowing into a point must equal the current flowing out, the base current change is 0.05 ma.
If the same transistor is connected as a grounded emitter, a 0.05-ma signal applied to the base will create a 0.95-ma collector change. The gain in this case is the change in collector current divided by the change in base current. This is the beta gain, and it has a numerical value of 19 for this transistor (Fig. 201-b) . It is this gain that the checker measures.
The arrows in Fig. 201 indicate electron flow to the connections of a p-n-p junction transistor. Delta (a) means "a small change of" and the numerical relation of alpha to beta is:
input and output signals (Fig. 202) . The circuit is simple and its accuracy is limited mainly by the care with which the balance dial is calibrated and the accuracy of R1.
A 60-cycle signal is applied to the transistor base through R1, R4, R5, and C1. The emitter is grounded to point A, the neutral point for the checker. Capacitor CI prevents the dc bias on the base of the transistor from flowing into the measuring circuits. The input signal is provided by a small stepdown transformer T1.
The signal to the base is called "constant current" because the high values of R4 and R5 almost completely determine the signal current flowing between the base and emitter terminals. The base-to-emitter resistance can vary from zero to several hundred ohms
Since quantities are compared in a balance circuit, the absolute value of the signal current is not too important—the voltage from T1 may vary without affecting the accuracy—but the accuracy of the comparison standard Rl is important. Resistor Rl must be a 1 % precision unit.
A dc bias current, larger than the signal current, must be supplied to the base for class-A operation. The 4.5-volt collector battery supplies this current through R3.
The changing base input signal produces a larger collector-current output. This current flows through balance control R2, generating an output voltage proportional to the current. The voltage across R2 is opposite in phase to the voltage across R1. If they are equal in amplitude, they will cancel and no signal difference will exist between points B and C.
To measure transistor gain some null detector must be connected to the checker—an oscilloscope or audio amplifier will do—and R2 is adjusted until the 60-cycle output is zero or minimum. In this position the ratios of R1 and R2 are in the same proportion as the transistor input and output signals. The gain then can be read from the calibrated balance control.
Almost any reasonable layout is suitable for the checker. The original unit used breadboard type construction on sheet Bakelite
The parts will fit under any 5x7 inch, or larger, chassis base. The terminals and balance-control dial could be mounted on top. This probably would be a good-looking and simple way to assemble the parts. To keep the cost of the transistor checker down most of the components consist of standard parts.
Several transistor test sockets mounted on bakelite can be used "outboard" and connected to the checker by three short lengths of hookup wire. Otherwise, it might probably be more convenient to mount the socket on the checker.
Since the socket is used frequently, it is advisable to use the hearing-aid 5-pin socket. This type is considerably more rugged than some of the currently available sockets. The two unused pins can be pushed out and the holes plugged.
Step-down transformer T1 is an output transformer with half of the primary connected across the 117-volt ac line and with the voice-coil winding providing the test voltage. The transformer used matches 25,000 ohms plate-to-plate center-tapped, to a 3-4 ohm voice coil. Any brand of transformer with similar specifications should be satisfactory.
Filament transformers can be used for T1 by changing the total value of R4 and R5. The sum of R4 and R5 should be 100,000 ohms for each volt of the filament transformer. Thus, a 2.5 volt transformer would require 250,000 ohms.
Two components, R1 and R2, must be chosen with care. Either a deposited-carbon or metal-film resistor is satisfactory. These resistors are nominally rated at 1%, but stock resistors can be used if they are carefully checked.
Most wire-wound controls can be used. Even the dollar-or-less wire-wound controls are stable enough to be dial-calibrated for R2. The wire-wound laboratory potentiometers found on the surplus market are excellent for the checker.
Low checker impedances keep stray 60-cycle pickup problems to a minimum. The leads to the oscilloscope can be unshielded; and no shielding of any of the checker components is necessary. The ground terminal post from the oscilloscope is connected to terminal 1 and the input wire of the scope goes to terminal 2.
If the checker is built in a metal box or chassis, the chassis should be grounded to point A.
A television-type power connector is used to connect the 110-volt input. These TV connectors and cords make a very neat and inexpensive power-disconnect for experimental and commercial electronic equipment.
The collector battery can be any 4.5-volt battery. Three flashlight cells in series are also satisfactory.
Beta markings on the balance dial correspond to specific resistance settings of R2. These values are given in Table 1. The maximum resistance of 1,000 ohms is equal to a beta reading of 5 since this resistance is 1/5 of Rl. Minimum resistance of course indicates infinite gain.
One way to calibrate the dial is to connect an ohmmeter and rotate R2 to each of the resistance values in the table. At each of these points mark the beta value on a dial made from some firm material that can be conveniently marked on.
If the control is sufficiently linear, the dial shown in Fig. 204 may be used. Most experimenters will find this dial sufficiently accurate for their particular purpose.
Use of checker
Connect the oscilloscope, the transistor socket, and the 117-volt ac plug. Set the oscilloscope's vertical gain to maximum.
The horizontal deflection may be turned off. This gives straight-line vertical deflection, in which case the balance knob is rotated for minimum deflection when the transistor is checked.
Alternately, the horizontal deflection may be turned on and the scope may be synchronized with the line frequency. This procedure permits observation of the signal waveform in the transistor circuits. If the scope does not have a 60-cycle position on the sync switch, set the sync selector to "external" and run a jumper from the 60-cycle "test" terminal to the external sync terminal. Synchronizing the scope with the line, rather than the internal signal, makes the horizontal sync very stable and independent of both the phase and amplitude of the vertical deflection.
While S1 is open, the oscilloscope input floats and there is large 60-cycle pickup from strays (Fig. 205-a). With the transistor in place and S1 still open, the stray pickup will be clipped since the transistor functions as a simple rectifier.
When S1 is pushed, the scope display indicates how close to balance the checker is. The patterns either side of and at balance are shown in Figs. 205-b,-c,-d.
Because the signal from the checker is in the order of millivolts, the deflections with S1 closed will be small but large enough for use with even the lowest gain scopes. Do not expect large deflections.
If an oscilloscope is not available, the output from the checker can be fed into the microphone or variable-reluctance phono input of an audio amplifier. Rotate R2 until a minimum hum level is heard. Do not use an amplifier that does not have a transformer power supply.
To check n-p-n transistors, simply reverse the connections to the 4.5-volt battery. This is the only necessary change.
A total of 13 new transistors have been tested on the checker. The average beta should be 9 according to the data sheet. One transistor had a gain of 14, while three units measured gains dangerously close to 5. The remainder of the transistors had betas of about 7 or 8. Of course it is not implied that this was a representative group of transistors; but it does point up the earlier statement about variations that must be expected.
Testing transistors with an ohmmeter
Another technique (in addition to the one just described) that can be employed for testing transistors uses an ohmmeter. The conventional ohmmeter check measures the forward and back resistances between two of the transistor elements taken at a time, as if it were a double diode with a common base. Although this method reveals an open, short or other extreme kind of transistor failure, it furnishes no evidence that the transistor may be unsatisfactory when operating as a triode. For example, in testing whether a transistor had been damaged by soldering heat, the conventional diode-check method might easily pass the transistor because it may show a much higher back than forward resistance. Actually, it can be very poor in triode performance. Thus, it is important to be able to get some indication of triode action, if only a rough one. This test unit provides just such an indication. It registers a substantial change in meter reading, indicating a large change in output collector current resulting from a small change applied to the input base circuit of the transistor under test. In short, even though it is a simple check, it adds the highly desirable feature of checking the transistor as a triode.
Aside from the newness of the transistor art the main obstacle to a simple and straightforward test is the fact that the transistor is highly sensitive to the choice of its dc operating point. Even if we ignore the various other factors that transistor operation depends upon—temperature, impedance matching and the like—widely different results for current amplification are likely to be obtained for slightly different operating points. For a power transistor, for example, the operating point may be very different than for a general-purpose unit.
Despite these difficulties, it is still feasible to cut through many test qualifications by keeping firmly in mind that our object at the moment is to check a transistor's condition, and not make a comprehensive test of transistor characteristics. For our purposes, we can concentrate on the following two questions:
1. Is the test safe and simple enough to be practical?
2. Do the results reveal a defective transistor?
As to being practical, the test unit (see Fig. 206) requires only a few fixed resistors, a toggle switch and a terminal strip for mounting. When used with the proper multimeter scale, both meter and transistor are protected from damage by the current-limiting provisions built into the multi-range meter. Used under these conditions, the largest voltage that can be applied to the transistor is 1.5 (or, for some meters, up to 4.5 volts) and there is no necessity for switching ranges. The entire operation is no more complicated than taking two ohmmeter readings.
As to its ability to detect a faulty transistor, the test unit not only detects the more obvious resistance defects that would show up in a diode resistance check but also reveals faulty triode action by indicating relative current changes resulting from the amplifying ability of the transistor. Keep in mind that these readings are not in¬tended for comparing the merits of different makes or configurations of transistors, but rather as a practical means of comparing the action of a questionable transistor with that of a good one of the same type.
Transistor check circuit
The test unit (Fig. 207) operates with the transistor in a grounded-emitter hookup. This circuitry provides one of the most significant clues to the condition of a transistor - its ability to amplify small changes in its base current. To do this, the test takes advantage of the -circuitry in conventional service type multimeters on their resistance ranges (shown in dashed lines). This section already contains a 1.5-volt cell for the voltage supply, a sensitive meter and a series limiting resistor R8 which protects both the meter and the transistor under test. Using these, the test circuit boils down to providing a means of indicating changes in current as the transistor is checked.
The first meter reading is taken with the spdt toggle switch in it’s lo position, corresponding to zero base current bias. The toggle switch is then thrown to hi. This sends a small dc bias current through the base of the transistor, causing a much larger current to flow in the collector circuit. The increase in meter reading indicates the relative current-amplifying ability of the transistor under dc conditions. The difference between the high and low reading is then compared with the meter-swing increase caused by a good transistor of the same type. This provides a check on the condition of the transistor.
No attempt is made to measure ac signal performance since this is more properly a manufacturing measurement. It is more the function of a transistor check to detect those signs that point to possible deterioration or failure. Passing over the more obvious defects— opens or shorts—that show up immediately, there remain two major indications of possible deterioration: (1) a marked increase in the relative Ic, the amount of collector current that flows at zero base current, indicated by the meter reading in the lo position; (2) a substantial decline in the transistor's current-amplifying ability, the difference between the hi and lo meter readings.
Transistor check procedures
Since the transistor, like other semi-conductors, is a nonlinear device, its effective resistance may vary widely, depending on the voltage across it. Thus, different ohmmeters, or more important. different ranges of the same ohmmeter, will result in non-comparable readings. This need not present any serious difficulties as long as only one ohmmeter and one resistance range of that meter is used, but it is well to understand the limitations. The meter used here is a Precision model 120. On its X100 resistance range it uses a 1.5-volt cell and has a center scale of 2,000 ohms. Any ohmmeter, using not more than 4.5 Volts and having a center-scale reading of between
1,000 and 3,000 ohms is also suitable, with the proper allowance made for expected readings as compared with those of a good transistor.
The polarities shown in the circuit diagram are for p-n-p transistors.
The positive polarity does not necessarily coincide with the red lead of any particular ohmmeter. In the Precision 120, as in
many others, the red lead is connected to the negative terminal of the ohmmeter battery. It is a simple matter to check this polarity on a voltmeter. It can even be checked on a diode, known to be good, by observing which polarity gives the conventional easy current flow for forward resistance.
For the p-n-p transistor check, a negative polarity is applied to the collector. With the switch at lo there is zero base current and the meter indicates the so-called saturation current (Ic) flowing in the collector circuit (Fig. 208). When the switch is thrown to hi (Fig. 209), the base is connected to a sufficiently negative point on the voltage divider (Rl and R2) so that a small bias current (about 50 a) flows through the base circuit. This results in a greatly in creased collector current for the second reading. The difference in the two readings indicates the approximate relative effectiveness of the transistor for current amplification.
Strictly speaking, the second reading is made up of more than just the collector current; it is the sum of the collector current and the currents flowing in the base and the voltage-divider circuits. The situation does not introduce any serious inaccuracies, since, in the case of a typical transistor, the amounts of base current (around 50 a) and bleeder current (around 6 a) together form only a small part of the total reading (around 500 a). Thus, the hi reading is made up of collector current for the most part (around 90%) and, therefore, a large difference between the hi and lo readings gives a sufficiently valid indication of the current-amplifying ability of a transistor. The emitter resistor provides stabilization for the dc operating point used.
There is an incidental benefit derived from using the ohmmeter connection as shown. If an n-p-n transistor is to be tested, reverse the ohmmeter leads. The simple flip over automatically insures that the bias and collector circuits are connected to their proper polarities for the n-p-n transistor. The meter will still read in the forward direction and readings will be equally valid for this type of transistor.
Table 2 gives a summary of results obtained with good transistors (and two defective ones, as noted), using the test unit connected to a Precision 120 meter. In all, more than 30 transistors were tried, divided by types into three groups.
After the first column of the table, identifying the transistor group, typical meter readings are given in the next two columns for the lo and hi positions of the test switch. The difference between these two gives the net change, a measure of relative sensitivity.
If some other meter is used, the results of the table can easily be adapted, provided the meter's XI00 resistance scale is reasonably sensitive. The internal battery of the meter on this range should be between 1.5 and 4.5 volts and the center scale for the same range should read between 1,000 and 3,000 ohms.
The following example relates the table's meter readings to another meter having a center scale of 1,000 ohms and a 3-volt internal battery on its XI00 resistance range: The full-scale sensitivity of such a meter on this range can be calculated to be 3 volts divided by the center-scale reading of 1,000 ohms, giving a full-scale deflection (with the test leads shorted) of 3,000 a or 3 ma. This compares with 1.5 volts divided by 2,000 ohms or 750 a for the Precision 120 used in the table. Thus, the substitute meter has a sensitivity of. only one-fourth that of the meter used and it can therefore be expected that the readings obtained for the average transistor will be approximately one-fourth that of the readings shown in the table.
Before checking a transistor as a triode, check each B to E and B to G diode of the transistor for forward and back resistance. This eliminates transistors having obvious defects involving short, open or erratic (intermittent) readings.
Because of the rugged construction and the long operating life of transistors in general, be cautious about blaming the transistor for poor circuit operation. This is a vastly different approach from the method we have been accustomed to in tube circuits, where the tube turns out to be the culprit perhaps 90% of the time. In normal use, a transistor does not damage or deteriorate easily although it does not take much to damage a transistor through excessive voltages or excessive heat. Even polarity reversals need not harm the transistor as long as the maximum voltage and current ratings are not exceeded.