Back To The Top
The following image shows the pin configuration for several types of transistors. It also gives part numbers for several common devices in each case style.
Click
HERE to open this graphic in a new tab. Right-click to zoom. Left-click to drag.
Information About the Transistors:
On the face of most semiconductors you'll find the part number but you'll also find other information. In THIS photo, you can see that there is the part number, the logo, a production code and a date code. You can generally find the key to these codes in the datasheet.
If you find a group of parts (output transistors, power supply transistors...) and none have the same exact set of numbers but you find a number common to other similar parts, the number that's common to multiple parts is VERY likely the part number (or part of the part number). The part number is 'generally' at the top but
not always (J108). Also, the entire part number
isn't always on one line (another J108 JFET). The following parts are the same component but have different production/date codes. If you were repairing an amp that had transistors with a set of numbers like those on the following parts and you didn't know which number was the part number, you could assume with virtually 100% certainty that the IRF9640 was the part number. There's 'virtually' no chance of having identical date codes on multiple parts that aren't the same exact type of component. Since the top numbers match and you know that the part number is generally at the top, there's little about which is the part number. The following parts are both IRF9640s but they are from different batches.
This is mentioned elsewhere but it's important so I'll touch on it here also. When replacing parallel parts, you need to use matched parts so that each part carries an equal share of the total load. You don't have to make any measurements to get parts that match precisely. If you use parts from the same batch, you'll generally get parts that are very closely matched. To know if they're from the same batch, you can look at the production/date codes. This is generally foolproof but there are instances where it doesn't work. The following transistors were pulled from an amp and checked. When checking the forward voltage of the intrinsic diode (covered later on this page), I found that the voltage was SIGNIFICANTLY different. One had a forward voltage of ~0.45 (a common value). The other had a forward voltage of more than 0.90. These would obviously not be a good match if they were operating in parallel. If the style of the logo and lettering differ you should not consider the parts as closely matched even if they have the same date code.
From above...
If you use parts from the same batch, you'll generally get parts that are very closely matched. <<< This only applies if you're buying genuine parts from reputable distributors. Buying counterfeits (common on eBay/Amazon), you don't know what you're getting. For an example of the difference you can expect between distributors go to TT10, here.
Back To The Top
The following diagrams will show you how to check an FET to see if it is defective. This is only for N channel enhancement mode FETs (the type used in most amplifiers' power supplies). Since I use Fluke meters almost exclusively, all measurements will show what you would see with a Fluke meter. Other digital meters should give very similar readings. Until you become extremely familiar with FETs, you should check the FET in this order from top to bottom. If the order is changed, you may not get accurate test results. To test P-channel FETs, reverse the meter leads where they plug into your meter and perform the tests as shown below.
Click
HERE to open this graphic in a new tab. Right-click to zoom. Left-click to drag.
Method 2:
This test (below) is checking to see if the gate is leaking or shorted. In this type of FET (the type used as audio output and power supply transistors), the gate should be COMPLETELY isolated from the other two terminals when checking the transistor with a multimeter set to ohms or diode-check. The meter should read the same as if the probes are open (not touching anything). During this first test, you are inadvertently charging the gate and turning the transistor on. If you were to place the red probe on leg 2 and the black probe on leg 3, you may see some leakage/conduction (more on this later).
Click
HERE to open this graphic in a new tab. Right-click to zoom. Left-click to drag.
The following are some of the FETs that can be checked with this procedure.
The following are some of the FETs that can be checked with this procedure but you'll need to reverse the probes. These are P-Channel FETs.
In this image, you are discharging the gate and making sure that the transistor is turned off for the next test. The meter should again read the same as if the probes are open (not touching anything). If it gives any other reading, the transistor is defective.
Click
HERE to open this graphic in a new tab. Right-click to zoom. Left-click to drag.
In this image, you're checking to see if there is any leakage between the drain and the source. The meter should again show no continuity between the drain and the source. If your meter probes slip and short between any of the terminals, you need to go back to the step where you discharged the gate. When doing this sort of test, you need to be absolutely sure of the test conditions. If your meter probe slips, you may have charged the gate which may give false readings.
Click
HERE to open this graphic in a new tab. Right-click to zoom. Left-click to drag.
If you get any reading other than 'open' on any of the previous tests for an (N-channel enhancement-mode FET), the transistor is defective.
In this image, your meter should read approximately 0.4 to 0.5 volts on diode check (the same meter setting as the previous tests). Here the meter is showing the forward voltage of the 'intrinsic' diode of the FET. The intrinsic diode is the Zener shown on the schematic symbol on the datasheet for the FET.
Click
HERE to open this graphic in a new tab. Right-click to zoom. Left-click to drag.
If the readings on the previous test is below approximately 0.4 volts, the transistor is likely defective (assuming that you properly discharged the gate). If it reads around 0.3 volts or lower, the FET is definitely defective and would likely have shown the defect in one of the previous tests.
Note: In FETs that have relatively high RDS(on), the forward voltage of the intrinsic diode may be higher than 0.5 volts but you're unlikely to see any power FETs that read higher than 1 volt.
The transistor is defective in both of these examples. It shows that the drain and source are shorted together. The transistors don't necessarily read 0 ohms when defective. Sometimes, there is simply leakage and it will show up as a low resistance reading.
Click
HERE to open this graphic in a new tab. Right-click to zoom. Left-click to drag.
In the image above, the meter shows 0.1 ohms. For most meters, the lowest reading you'll get is 0.1-0.2 ohms with the probes shorted together. If you place the probes across a shorted transistor, the meter will read no lower than what it would read with the probes shorted together.
Back To The Top
This (below) is a close approximation of the junctions in both NPN and PNP bipolar transistor. I included it to help you to remember how/why the following tests are done.
This is the pin out configuration of a
TO-3 transistor. The test procedure is the same as the bipolar transistor in the next diagram. Take note of the pins labeled on the diagram for the TO-220 case transistor.
This shows how the collector is directly connected to the tab. You can see the (now destroyed) silicon die. This is in the area of the TO-3 transistor where the labels are located above.
This may help some remember the configuration of the TO-3 transistor when checking them. They are the same as most all large transistors.
The following is a small sample of the transistors that can be checked with this procedure.
These can also be checked with this procedure but you'll need to reverse the probes/leads.
The two images below show the junctions for both NPN and PNP transistors. If you understand that the red meter probe has the more positive voltage (using the black probe as the reference), you can see why you get the ~0.6v reading. It's the same as if you touched the meter probes to a standard rectifier diode. When the meter is set to diode check and the probes are not touching anything, the voltage across the probes is ~3.5v. When you touch it across the 'diode' junction, the voltage across the probes is limited to the forward breakdown voltage of the junction. If this is not clearly understood, read the Diodes page of the tutorial.
Click
HERE to open this graphic in a new tab. Right-click to zoom. Left-click to drag.
Click
HERE to open this graphic in a new tab. Right-click to zoom. Left-click to drag.
In this diagram, the volt meter applies a small voltage to the transistor's junctions. With the probes in this position, the junctions are forward biased. The reading should be between 0.5 and 0.7 volts. Readings outside this range are likely defective.
Click
HERE to open this graphic in a new tab. Right-click to zoom. Left-click to drag.
In this diagram, the junctions are being checked with reverse bias. No current should flow through the transistor's junctions with the probes in this position. The meter reading should be the same as it would read with open probes.
Click
HERE to open this graphic in a new tab. Right-click to zoom. Left-click to drag.
Both of these connections should also read as open. any other readings indicate defective transistors. The only exception will be with Darlington transistors.
Click
HERE to open this graphic in a new tab. Right-click to zoom. Left-click to drag.
Click
HERE to open this graphic in a new tab. Right-click to zoom. Left-click to drag.
Click
HERE to open this graphic in a new tab. Right-click to zoom. Left-click to drag.
Back To The Top
Darlington transistors are two bipolar junction transistors in a single case. Some Darlingtons have internal resistors and possibly even a diode. This makes them a bit more difficult to check. If you're unsure of the readings of the Darlington you're using, use a known good transistor as a guide. The values below are for a TIP102. Due to differences in individual transistors and multimeters, your readings will be somewhat different. No matter the probe positions, you should not read anything near zero ohms. If you do, the transistor is defective.
Click
HERE to open this graphic in a new tab. Right-click to zoom. Left-click to drag.
Back To The Top
If you made a short list of perfectly good components that were replaced for no reason (other than ignorance), JFETs would be near the top of that list.† With regards to their basic use/function/operation, they are one of the least understood components in an amplifier. People test them expecting them to read like other transistors and that's simply not what they should/will read, when checked with a multimeter. JFETs, when working perfectly, will read as a leaky or shorted transistor. <<< When working perfectly. This section certainly won't tell you all there is to know about JFETs but should clear up most questions that you have for their use as muting transistors.
The pin configuration is another aspect that can cause confusion. The TO-92 package has the pin configuration of Japanese transistors. If you have an amplifier (older Orions) that's populated extensively with MPSA06/56 type transistors and you see the J108, you may expect it to be EBC as well. The SMD parts are problematic as well. Virtually no one expects the first pin to be anything than the base/gate. When they read low resistance between what they believe to be the base and emitter, they immediately believe that it's defective. The gate at the pin 3 location is new to most people who encounter the SMD JFET for the first time.
†Along with PS driver ICs, op-amps, and perfectly good capacitors. The PS driver ICs have a very specific list of requirements that will allow them to produce pulses. Op-amps are similar in that they are very predictable. For more on op-amps, click HERE.
To help understand JFETs, let's compare them to a double-throw (DT) relay (a simple, well understood, mechanical 'equivalent'). Knowing relays, would you diagnose a relay as being defective if you read 0 ohms between terminals A and B in
THIS image? Of course not.
For those who don't know relays, in a
DT relay, there is one pair of normally closed (NC)
† contacts and one pair of normally open (NO) contacts. The NC contacts will pass current when the relay coil is NOT energized. If there is no coil voltage or
insufficient coil voltage, current will flow through the NC contacts. To get current to pass through the NO contacts, you have to energize the relay's coil (with sufficient
operate or pull-in coil voltage to toggle the contacts).
†Normally closed as in closed
circuit and able to pass current with no power/voltage applied to the relay's coil.
The standard
enhancement mode FETs that are used in the
switching power supplies and as output FETs are the equivalent of the relay's
NO contacts. You must apply voltage across the gate and source to allow the FETs to conduct between their drain and source terminals.
Depletion mode JFETs are the
NC contacts. To
prevent current from flowing through the JFET, you have to apply voltage to the gate. It's no more complicated than that. And the same (sort of) as the relay, if you apply no gate voltage or
insufficient gate turn-off voltage to the JFET, it
will continue to conduct (just a bit less efficiently).
Enhancement mode FETs have a
threshold voltage (
VGS(th)) that, when reached, allows the FET to begin conducting. JFETs have a
VGS(cutoff) that specifies the voltage required to stop them from conducting. A JFET is
on until voltage is applied to its gate terminal to drive it
off.
The FETs have a
linear operating mode that prevents them from reaching 100% of their R
DS(on) ratings.
† Operating in linear mode means that they're not fully on or fully off. The example above described the action of the JFET and relay as being imperfect when the control voltage was insufficient to switch them fully on or fully off.
To get back closer to the relay analogy, imagine the example as having sufficient drive to switch the standard FET fully on (about 10v) or the JFET fully off (varies but let's use 15v). If the control voltage snaps from 0v to 10/15v (depending on the type of FET), the relay analogy is again valid (or as valid as it can be in relation to the simple relay).
†Resistance from
Drain to
Source with the transistor fully
on
JFETs are used as muting transistors in many amplifiers like Kicker, JL Audio, Zapco, PPI, Orion, etc... The most common JFET is the J108. This JFET will conduct between the drain and source (it will be 'on', as will all depletion mode FETs) with no voltage applied to the gate. In the photo below, the meter reads 6.7 ohms when connected across the drain and source. The RDS(on) will vary depending on the JFET (as it does for any FET). According the J108 datasheet, that JFET should read no more than 8 ohms when fully on. 6.7 is perfectly OK. Others may read 100 ohms. Consult the datasheet to determine the RDS(on) of the JFET you're testing.
To determine if the gate is intact, set your meter to ohms and measure the resistance from the gate to the other two legs. Then reverse the probes. One way, you should read an open circuit. The other direction, you should read ~4M ohms. You should also perform the test on diode check. On diode check, you should read open in one direction and ~0.6v in the other direction (like a diode).
Generally, the muting JFETs that are most commonly used in car amps are turned off (mute defeated) by negative voltage. Negative 15v to 20v is commonly used as the drive voltage to switch them off. When on, they clamp the audio signal (short it to ground). When off, they allow the audio to pass. The audio doesn't pass through the transistor. The transistor is connected between the audio line and ground.
It's important that you not forget that JFETs pass current when no gate voltage is applied (like the
NC contacts of a relay that has no coil voltage).
Remember this
before you (needlessly) pull transistors marked
J108 (PPI),
J111 (various amps) or
2N5639 (
Orion). This also applies to the SOT-23 surface mount transistor used in many of the MTX class D amplifiers. The transistor is connected across the optocoupler diodes. These transistors are typically marked
C1 for the
SST111 or
6P for the
MMBFJ111. These last 2 can be seen in the
MTX 7801 and the
JL 500/1.
If you're concerned about the JFET readings in the board and there are others in the amp (there generally are others, except possibly for mono amplifiers), compare their readings while they're still in the board. It's rare that multiple JFETs fail and there aren't many (any
?) amps that connect them in direct parallel.
† If you thought that one was leaking
† and all the rest (performing the same duty) read the same, they are all likely OK.
For the SMD JFETs, remember that you can simply
lift them which makes testing and reinstalling them much easier.
†The drain and source are what are going to be of concern and those are
very unlikely to be in direct parallel with any other JFET. It is, however, possible that you will find that the gate and source are in direct parallel with other JFETs. This is done in the JL 300/4 and similar amplifiers. In other amplifiers, there are generally gate resistors (Orion generally uses
100k ohm) for each individual JFET. For amps like the JL, there are no individual gate resistors (all JFETs driven by a single mute output) and all of the source terminals are directly connected to ground.
†This is the most common mistake when checking JFETs. People believe that they're shorted/leaking (electrically) because they read a low resistance or low voltage on diode-check.
This is just a bit more to try to help those struggling with these transistors. I'll admit that I had a problem remembering whether a P-channel or an N-channel conducted with positive/no voltage. The image below helps me remember that the N-channel JFET is like the NPN bipolar transistor. It conducts with positive or no voltage and stops conducting when the gate/base is driven with negative voltage. The 2N5486 is an N-channel JFET. The MPSA06 is an NPN Bipolar transistor.
Above, you can see the N-channel 2N5486 conducting (pulling the voltage down on the 100 ohm resistor) with either positive bias or no gate bias. The NPN transistor is also conducting when the base is forward biased by positive voltage. It's also clear to see that neither passes current (pulls down the resistor voltage) when the bias voltage is negative. There isn't a perfect correlation for the two but this is close enough for me to remember that the N-channel JFET is the one that functions a bit like the NPN transistor.
One more similarity is the diode action from G-S and from B-E. When forward biased, the internal diode conducts and limits the gate and base voltages but when the polarity is reversed (+10v to -10v), the diodes do not conduct and don't limit the voltage .The gate/base simply follows the drive voltage.
I try not to do this but this is so far off the beaten path that there isn't much use for the information, the Wikipedia link:
^^^ Copy and paste into your preferred online browser and... Enjoy.
JFETs have some specs that are much like vacuum tubes, specifically transconductance. With that, you may see the terms
siemens and
mhos. A mho is simply the reciprocal of ohms. 3 ohms would be 1/3 mhos or 0.33 mhos. The symbol for the mho is the upside-down ohm symbol.
The siemen and the mho are the same thing. That 0.33 mhos would be 0.33 siemens or 333.33 millisiemens. The mho is from the 'olden' times when the concept originated. The siemen is the unit used in the SI system (International System of Units
<< the
bleeping metric system).
The older datasheets for JFETs like the 2N5486 and the newer ones like the J108 (both linked above) vary significantly. The siemens are still used in the newer J108 but specs like R
DS(on) are more prominent. This is especially true for most of the JFETs that the circuit designers chose for muting (clamping audio). The 5486 uses mainly siemens and can be confusing (even more confusing than the J108, which can be a bit difficult to wrap your head around). If you have to select a replacement muting JFET for one that has gone obsolete, choose one with the same polarity and a similar (or lower) R
DS(on).
Back To The Top
This is essentially the same as checking standard rectifiers but here there are 2 diodes per package. A rectifier diode is used to pass current in one direction and prevent the flow of current in the other direction. When you have your meter set to ohms or diode check, the meter produces a low voltage across the probes. When placed across the diode one way, the diode will allow current to flow from one diode terminal to the other. With the probes reversed, the diode will prevent current flow and the meter will read as it does when the probes are open. There's more on the meter function at the end of this section.
These first 2 diagram show how to check the dual positive rectifiers (diodes) and the associated meter readings. The windings of the transformer would go to the two outside legs and the center terminal would go to the positive rail capacitor.
Click
HERE to open this graphic in a new tab. Right-click to zoom. Left-click to drag.
Click
HERE to open this graphic in a new tab. Right-click to zoom. Left-click to drag.
The next 2 diagram show how to check the dual negative rectifiers (diodes) and the associated meter readings. The windings of the transformer would go to the two outside legs and the center terminal would go to the negative rail capacitor.
Click
HERE to open this graphic in a new tab. Right-click to zoom. Left-click to drag.
Click
HERE to open this graphic in a new tab. Right-click to zoom. Left-click to drag.
If any combinations of meter probes to component terminals on any of the aforementioned devices reads less than 0.2 volts (the JFET excluded), the device is defective. Remember to test the FETs in the proper order. All of the devices must be removed from the circuit to obtain reliable test results. Some of the devices can be checked in circuit to see if they have any internal short circuits. There should be no connections on any of the devices that will cause the meter to read 0.00 while on diode check. The only exception will be the rectifiers. The outside legs of the dual rectifiers will be directly connected through the transformer windings and will appear to be a dead short.
Back To The Top
To better understand the tests above, you should understand the way the meter operates when set to ohms or diode-check. In either setting, the meter applies a small voltage to the meter probes. It does this through an internal series resistance. When you have two resistors in series, the supply voltage (provided by the meter in this case) will be divided between the resistors (if this is not clear in your mind, read the Resistors page now - go to the 'Using Springs to Demonstrate Voltage Division' section). Since the internal resistance of the meter is a known quantity, the meter can determine the value of the external resistance by the voltage drop across the internal/external resistance. Typically, the internal voltage source is ~3 volts but will vary depending on the meter's range setting (most meters are auto-ranging). The diode check function works in a similar fashion but instead of calculating the resistance from the voltage drop across the external resistor, it reads and displays the voltage across the meter probes.
If the meter is set to diode check and connected across a device like a diode and the black meter probe is on the striped end of the diode, the meter will be able to forward bias the diode. When the diode is forward biased and the applied voltage is equal to or greater then the breakdown voltage of the silicon junction, the diode will conduct. The voltage displayed on the meter will be the forward breakdown voltage of the diode/junction. When testing a diode (or similar semiconductor), you are checking the condition of the junction. The junction (the actual piece of silicon inside the package) should pass current in one direction but not in the other. If you have the meter set to diode-check and place the meter probes on the diode one way and get no reading (just as you get with open probes - probes not touching anything), and then get a reading of ~0.6 volts when the probes are reversed, the diode is OK. If you get the same reading no matter the polarity of the probes, you know the diode is defective.
If the meter is set to resistance/ohms and you check a diode, the reading will depend on the breakdown voltage of the device. Since the meter is assuming that there is a resistor between the meter probes, it will display the resistance that's equivalent to a resistor that would create the same voltage drop as the forward biased junction. This reading is not actually the resistance of the device. For example, a diode may read 3.4 mega ohms when forward biased when the meter is set to read resistance. If this were actually the resistance, the diode would be useless for anything requiring more than a few thousandths of an amp. Although you can test semiconductors on resistance or diode-check, it's generally best to test them on diode-check. The only exception is when you're testing for reverse leakage (passing current when reverse biased). That test may require more sensitivity from the meter and the resistance setting provides that.
Back To The Top
For those who don't have a diode-check function on their meter, you can use the following circuit to make a crude test jig. The circuit applies battery voltage (~3 volts) to the meter probes through the resistor. When connected across a junction with forward biasing, the meter will read the breakdown voltage of the junction. When reverse biased, the meter will read the same as when the probes are not connected to anything. You will have the meter set to DC volts when using this test jig. When not in use, you must remove it from the meter to prevent draining the batteries.
When checking high-power LEDs or diodes with more than about 3v of forward voltage, you can use this method to find the forward voltage if the two batteries are replaced with batteries (or a power supply) with greater voltage. This can be a pack of batteries like the three 18650 Li-ion batteries or virtually any supply that has more voltage than the forward voltage of the LED. While the purpose is different, the probe setup can be seen HERE. The T-taps allow you to disconnect the clip-leads when not needed. You can use a cheap pair of spare leads to do this.
Click
HERE to open this graphic in a new tab. Right-click to zoom. Left-click to drag.
As a side note... The battery holders seen in the previous paragraph are available for virtually all types of batteries and in up to about 8 cell packs. The pack above is 12v (batteries connected in series) and is fairly capable (even with the cheap batteries) of powering small automotive LED running lights (cheap ones found on eBay) in 2x2 or 3x3 configuration. Bear in mind that those running lights can be run at 12-24v so larger packs can be used. At higher voltage, the running lights will draw less current.
For good batteries, don't buy from eBay. Try Battery Junction. The Nitecore (expensive but good), Samsung, Panasonic and LG batteries are typically good choices.