Sunday 18 June 2017

Simple but effective Transistor Curve Tracer circuit.

This curve-tracer uses only 6 transistors and produces a beautiful curve display on an oscilloscope in X-Y mode. And it doesn't even matter which transistors you use to build it with!! I build it up on a bread-board first and it works very well! The component that is being tested does not get hot, unlike some Chinese Curve Tracers. I was impressed by the simplicity and effectiveness of this little circuit and since I didn't see it available on the internet I thought I'd share it with the world.
Here is the circuit:

This circuit can handle all sorts of components, NPN transistors (bi-polar junction transistors), diodes etc. However it does not work with FET transistors, like Jfet's, P or N channel mosfets etc. You can build a PNP version of this circuit. Instructions for that are at the bottom of this article. As you can see, very simple design. You could replace the square-wave generator formed by T1 and T2 for a NE555 square-wave generator if you wish (beware of the input voltage if you do, so you don't blow up the 555 and make sure the duty cycle of the squarewave is 50%). I've tried it with a 555 but on a digital scope the lines are a little bit thicker than with a transistor based square-wave generator. The chip seems to be noisier. (However you won't have this problem if you use an analog scope.) But the plus point is that you can easily make the frequency adjustable with a 555. This gives you a bit more control over how the traces look. Mind you, I only tested it on a bread-board so if you build a print for it, it might yield a better result. The traces are displayed from right to left which is the wrong way around for a normal graph display but you can change that by inverting the X channel on your scope. However most analog scopes won't allow you to invert the X-channel, usually only the Y-channel, so you'll just have to get used to that. That's what you get with such a simple design. (The graph for the PNP version requires you to invert the Y-channel (channel-2) on your scope, so that will look normal on almost any scope.)
TUT is the Transistor Under Test, TUN stands for Transistor Universal NPN, TUP is Transistor Universal PNP and DUG stands for Diode Universal Germanium. I used a AA119 for the diode, but you can also use a Schottky Diode if you don't have a Germanium one. I tried it with a 1N5819 and that works just as well. For the NPN's I used the 2N3904's (I also tried the BC547 and it works just as well) and for the TUP I used a BC559. I set it all up on a bread-board and it worked like a charm.
The circuit diagram advises a power supply voltage of 6 Volt but I use it with 10 Volt and I feel that works better. You can turn the voltage up to 15 Volt without any problems. I haven't tried higher voltages then 15V but I think the voltage is only limited by the transistors you use in the circuit and a higher supply voltage gives you more data about the transistor under test. As to power consumption: it hardly uses any power at all. At 6 Volt it draws a current of 4.8 mAmps and at 15 Volt 11.8 mAmps. That's equivalent to 0.177 Watts.

A reader of mine kindly sent in a link to a Falstad animation that shows how the step generator actually works which is very interesting to see. Here's the link so you can have a look for yourself:
- Click here for circuit animation -

The number of traces that appear on the screen is determined by the ratio between C5 and C4 and can be varied by changing the value of C4. Don't change C5 because that will change the length of the traces. Using a 2N3904 as test transistor, with a 68nF capacitor I got 5 traces on the screen and with an 82nF I got 6 traces. It also changes if you vary the power supply voltage. Depending on the type of transistor under test you can loose a trace if you turn the voltage up above a certain value. For instance with the AC187 under test, the display goes from 5 to 4 traces if the voltage goes above 11.4 Volt. Btw, should you ever short circuit this thing and it doesn't work any more, your best bet is to replace T4 and T5, They are the first to go. Believe me, I speak from experience ^____^

Here are some tips on how to set your digital scope (in my case the the Rigol DS1054Z) for best possible display of the curves:

1. Set your scope into X-Y Mode.
2. Put your probes on DC coupling so you can measure voltages with the cursors of your scope.
3. Set the bandwidth limit for both probes to 20MHz
4. Invert Channel 1 probe.
5. Set the horizontal timeline to 2.00mSec/Div
6. Set both probes to 1x multiplication (and don't forget to set it in the scope to 1x aswell)
    1x will give you sharper lines but if you want to do measurements it's better to put the probes at 10x to reduce the influence of the scopes impedance on the resistance in the circuit.
7. Set channel 1 (X) to 1 Volt/Division
8. Set channel 2 (Y) to 500 mVolt/Division
9. Go into the 'Acquire' menu and set the Memory Depth to 60K

This gives a good starting point from which you can fine-tune your scope depending on the type of transistor you are testing.

With these settings you should get a picture like this one:


This is what the graph from an NE555 based curve tracer looks like. You can see the lines are a tiny bit thicker although the difference is almost negligible.


You can also measure diodes with this circuit. Just put them between the emitter and collector of the TUT (Transistor Under Test) points and you'll get the characteristic diode curve. The anode must be connected to the collector and cathode to the emitter clip.

Here is a picture of the PCB layout including the parts list and a picture of how the curve appears on an analog scope:


I've posted a video about this circuit on my YouTube channel which you can watch below. The video description contains a link to the original article, written in Dutch, of which I made a PDF file which you can download for free. That link is also posted at the bottom of this article. I recently found an English language version of this same article on Archive.org and that link is also at the bottom of this article. If you plan on making this circuitboard then beware of the transistor polarity! The layout used in this article was designed for transistors like the BC547 with Collector on the left and Emitter on the right so if you plan on using the 2N3904 the Collector and Emitter will be reversed so you must put those in the other way around.


This circuit makes your oscilloscope an even more useful instrument than it already is because it allows you to easily match transistors together, which is sometimes necessary if you're building a high quality audio amplifier or a precision oscillator or for high accuracy current mirrors or if you (like me) are building your own synthesizer and need matched pairs of transistors for the filters (like the famous Moog Ladder Filter).
In the video I made the remark 'it's not that accurate of course because it's a very simple design' but I was thinking about that and that is actually a bit nonsensical because the oscilloscope doesn't lie. In fact this circuit is as accurate as your scope is and using the cursors you can make some very accurate measurements and calculate all sorts of parameters from these traces.

I've done some more experimenting and I've designed my own PCB with my own design layout. The new circuitboard has a lot of surface area for mounting clips to put the transistor under test in. I simply made some small coils out of copper wire and soldered them to the circuitboard. It works very well. I made the PCB so that it has a big ground plane, to reduce noise and it seems to have done the trick because I get nice thin/sharp lines when the curves are displayed on the scope. The breadboard versions always had thicker lines, at least on a digital scope. Recently I've also soldered on some short wires with alligator clips so I can easily measure big power transistors like the 2N3055 in TO-3 housing.
Here are some pictures of the new PCB. It doesn't look very professional because I simply draw the circuit layout straight onto the board with a permanent marker and then etch it. This is a double sided board:




Some screenshots showing the curves I get from this new print:
This is the curve from a 2N3904:


This is the curve from an AC187 Germanium Transistor. You can see that the back-traces are very prominent. That's also a drawback of the simple design. However it doesn't matter because all the information you can get from a graph like this is easily visible, plus you can dial it down a bit if you have a scope with an intensity graded display:


This is an AC176 Germanium transistor and at the top you can see the different wave-forms that make up the curves in the X-Y display. The yellow signal is from the X-axis (horizontal) and the blue from the Y-axis (vertical):


Below is the curve of a PNP transistor. The OC79 Germanium PNP transistor to be precise. So if you see a curve like this, you know your transistor is a PNP type and should be measured with the PNP version of this circuit. Notice how the wave-forms have changed. The blue Y-axis signal has changed from a square-wave to a Shark Fin wave.


This next curve is from a BC547.
Now we can do some calculations on this with the help of our cursors and determine the Collector current of the middle trace for instance. We select the 'Cursors' on our scope and set the first horizontal cursor on the middle trace and the second on the 0 Volt line. The readout of cursor AY says we are at 1.45 Volts. We know the collector resistor (R7 in the circuit) has a value of 330 Ohms so the current through that resistor and therefore also through the Collector of the transistor is 1.45/330=4.39 mA.


If we now want to calculate the Hfe or Beta or amplification factor of this transistor, at this value, we need to know the Base current. The Base current is biased through resistor R8 which is 270K. I soldered some copper-wire to each end of R8 so I could connect a probe to it. That's the purple waveform in the picture below. The probe we use to measure the voltage drop over R8 has an impedance of 10MOhm so the total resistance of R8 will drop to 262.9K.
The image below shows the voltage drop over R8 which is 3.64 Volt. If we divide that by 262900 Ohm we get 13.8 µAmpere. Hfe is then Ice/Ibe = 0.00439/0.0000138=318.1 for the middle trace.
From these curves you can also get a rough indication of the collector/emitter impedance; the flatter the horizontal bit of the trace is, the higher the impedance.



PNP version:
I recently build a PNP version of this design and it's very easy to do. Here's a video I made about that:


Like I mention in the video you need to exchange the NPN transistors for PNP's, and exchange the PNP transistor C5 for an NPN, reverse the polarity of the diode (DUG) and of capacitor C6 and don't forget to switch the power supply connections! Then, on the oscilloscope, you need to invert channel 2 (the Y-channel) and that's all. If you design your own print for this you could make both versions on one circuitboard for ease of use. Just let your imagination run wild and I'm sure you could build a tester that is better than many of the Chinese products advertised on eBay.

Here is the revised circuit schematics for the PNP version:


Here's a JPEG image of the original article in English from the Elektor magazine. I made a Photoshop compilation of the article and turned it into a single Jpeg image with high enough resolution to zoom in and easily read the text. Simply click on the image to enlarge it and then right click on it and choose 'Save As...' Then you can zoom in on it.



That concludes this article. Hope you enjoyed it.

You-tuber 'The Tube Roaster' made a cool little video about this circuit too. He's much better at explaining these things so I'll link to the video here, if you'd like to watch it:

https://www.youtube.com/watch?v=b2Qb6y-Ttkk

There's an other, more recent, video made by YouTuber "W4DXZ Signal Shack". He demonstrates this circuit on both an analog scope and a digital Siglent 1102CML+ oscilloscope. A very good demonstration which you can watch here: 
https://www.youtube.com/watch?v=DEtsD_TqJiU

Feel free to make your own video if you wish and you can use any picture/video from this website. No problem. And if you do, please send me the link and I will share it in this article.

DOWNLOAD THE ORIGINAL ARTICLE (PDF IN DUTCH) HERE:

Transistor curvetracer article (Dutch)

Here's a new link to the original article from Elektor Magazine from September 1980:

If you have questions about this circuit or see any mistakes in the text, please leave them in the comments. Leave a comment anyway please.
-- DON'T PUT ANY LINKS IN THE COMMENTS OR I WILL DELETE THEM.--
I see from my statistics that this is the most popular article on my website and gets, on average, more than a 100 visitors every day so this curve tracer must have been build by many people. So if you want to do me a favour in return, make a little video about it and show it here in the comments. It would be so cool to see your home made curve tracers.:)

24 comments:

  1. Hi Eddy. Thanks for the video and the schematic. I have put together a Chinese tester(first saw it on youtube), but it is getting the test transistor pretty hot. The reviewer had the same problem. Does your circuit suffer from the same? If not, I'll build it too.
    -osman

    ReplyDelete
    Replies
    1. No with this circuit the test transistor does not get hot at all. The current that goes through it is less than 10 milli Amperes so there's no danger of components under test getting hot. You can build it up on a breadboard first and do some testing. You will love this circuit, it works very well.

      Delete
  2. Great circuit. The retrace is the biggest problem and I saw that with my simple setup. One way to solve this is to generate a negative going pulse that coincides with the retrace. You can then use this pulse to modulate the brightness (and effectively blanking the retrace). You’d use XYZ scope mode vs XY.

    ReplyDelete
    Replies
    1. I'm afraid my knowledge of electronics doesn't go so far, but if you manage to eliminate the retrace I'd love to see a schematic of the circuit. :) Thanks for commenting and I'm glad you like the circuit.

      Delete
    2. The retrace can be easily fixed if using a true triangular waveform to feed T6 or the TUT collector. I eliminated all the astable oscillator and inserted two op-amps to generate a square and triangle waveform, it works wonderful, no retrace at all, since the up and down triangle ramp travels exactly the same path on scope. A little difference by TUT capacitances and turn-off delays, but much better than the original. The sharp corner of the square waveform supply the sharp and short pulse to generate the steps ramp. I also added an extra NPN to feed the base of the TUT, increasing the impedance at the bottom of C4, made the steps more square and base current more steady.

      Delete
    3. wlip that sounds amazing! Good job. I'd love to see a schematic of your circuit.

      Delete
  3. great article thank you for sharing

    ReplyDelete
    Replies
    1. Thank you very much! I'm glad you found it useful. :)

      Delete
  4. Hello Eddy
    This is a great circuit. I've done it on a breadboard and coupled to my analog scope. Working great
    Tested NPN transistors,few diodes and resistors.
    I will done it on a PCB during the weekend.
    Many thanks for posting this simple but effective circuit.
    73 de 4s6rav (Ravi from Sri Lanka)

    ReplyDelete
    Replies
    1. Hello Ravi, I'm very happy that you found this circuit so useful. I am actually using it myself too at the moment. I am trying to build my own synthesizer and for some of the filters I need matched pairs of transistors. This curve tracer is ideal for seeing the characteristics and matching transistors up. Thank you for commenting. Best wishes from Holland to Sri Lanka! ^___^

      Delete
  5. Hello Eddy,
    Many thanks for quick reply.
    As a HAM radio operator am into home brewing various test equipment to keep my old radios up and running.That's the reason of curve tracer ideal for me.I would appreciate it very much if you post your synthesizer once complete.I will try to build from this end based on your design.
    Let's keep in touch if you wish.
    Best regards and have a nice day Eddy.

    ReplyDelete
    Replies
    1. Thanks Ravi, as it happens I'm into HAM Radio aswell although I have not got my license yet but I will go for that in the near future. I've done some test exams and passed them all. :) But yeah, let's keep in touch. I will post about every module I make for the synth on this website because this is also a bit of an online diary for me, for my hobbies. So you can always catch me here by commenting, or join me on Facebook if you like. It's https://www.facebook.com/EddyBergman

      Delete
  6. Excellent write-up.
    Thanks for everything. Specially the English version of this article.
    I'll build mine with a 555 and pot, so I can play with different freq's.
    If it jitters like yours, I'll blame you.
    If it works flawlessly, I'll take all the credits. You understand...

    ReplyDelete
    Replies
    1. Hahaha, I take no responsibility of any jitter issues, LOL. I'm glad you liked the article. I aim to please ^____^

      Delete
    2. Oh and the jitter was not in my video but in The Tube Roaster's video. He used a breadboard version on an analogue scope that seemed to have some sort of ground-loop issues. I haven't heard from any other person that had that problem though. ;)

      Delete
  7. No worries. All is cool. Thanks again.

    ReplyDelete
  8. This is terrific. I had a pdf of the original article in Dutch from a while back but I don't have OCR to translate and... well, you know. Thanks for referring to an English version from my favorite Electronics magazine!

    ReplyDelete
    Replies
    1. You're welcome! :) Yes, I found it on archive.org and immediately put a link to it in this article. I'm glad you found it helpful. I can immagine Dutch is a bit problematic ^____^

      Delete
  9. Sir kindly mention how capacitors ans resistors value has been calculated in this circuit.

    ReplyDelete
    Replies
    1. I have no idea. I didn't design this circuit I only built it.

      Delete
  10. The values of component you have built is any theoritical calculation is there for fixing in that particular range.

    For ex: 270K, 330ohms, 10nf....

    ReplyDelete
    Replies
    1. I don't know what you mean but I didn't design this circuit. It's from Elektor from the 1980's

      Delete

Note: comments are moderated and do not appear straightaway. Your first comment is not allowed to contain any links.