Seriously, what is the value of a $30 thing that is unreliable?

[watchmaker]

I get that many of these things are great for initial ID of components.  I have a couple (the TC3 is my go to) and even use a $25 two channel USB scope for the display on my Thaikit Curve Tracer.

But what is the point of trying to select components with a measurement system that yields unreliable results.  Not even talking about calibration, just consistency and documented approaches to the calculations?

When is "good enough" a path to failure and frustration?

Because of this forum, I have learned the importance of verifying every component. I am just an amateur doing things to amuse myself, so time is not an issue.

Capacitors in particular are more complex than they first appear (at least to me) in their behavior both in and out of circuit.  It would seem to me that using these $30 things for their selection can lead to much frustration.  Long gone are the days when the choice was ceramic or big electrolytics.

OTOH, I get the challenge of working to improve these devices.  That provides many with a great deal of fun as evidenced by 300 plus page threads. 

[coromonadalix]

mmm i would say    follow the huge threads and see remarks and feedback's and fw evolution's

for sure an 20-30$ gimmick has some usefulness, and i would say precise 80-90% of the time,  unless you have some difficult parts like scr's and some mosfets, they help to basically diagnose stuffs

and sure  in some cases  if you try to test on pcb's,   you'll be greatly limitted

same for users who tests caps with an dmm, it's time constant based not a real cap tests with esr dsr z q  etc ...

you would need lcr meters with serial or parallel functions dc / bias etc...

always a compromise between real equipment and gizmos and learning their limits

my 2 cents

[watchmaker]

mmm i would say    follow the huge threads and see remarks and feedback's and fw evolution's

for sure an 20-30$ gimmick has some usefulness, and i would say precise 80-90% of the time,  unless you have some difficult parts like scr's and some mosfets, they help to basically diagnose stuffs

and sure  in some cases  if you try to test on pcb's,   you'll be greatly limitted

same for users who tests caps with an dmm, it's time constant based not a real cap tests with esr dsr z q  etc ...

you would need lcr meters with serial or parallel functions dc / bias etc...

always a compromise between real equipment and gizmos and learning their limits

my 2 cents

 

[Kean]

always a compromise between real equipment and gizmos and learning their limits

This also applies between two different models/brands of "real equipment" - there are always compromises!

It is important to understand the limits of any test equipment, even if you don't fully understand how it works.

[BeBuLamar]

What $30 thing? I don't understand.

[Kean]

What $30 thing? I don't understand.

The cheap LCR/transistor/component testers.

[KungFuJosh]

The only value is in what they're good at, and they're usually good at transistor testing. And usually repeatable in that specific case.

I would never use them to test basic components that a DMM or real LCR is better suited for.

[madires]

The Transistortester can measure R, C and L within some limits, but it's not an LCR meter. Unfortunately, many sellers tell nonsense. And the manufacturers of the tester clones cause another issue by poorly porting and modifying the OSHW firmwares to alternative (cheaper) MCUs. The OSHW firmwares are optimized for ATmegas and porting them to other MCUs will take a huge effort. My advise is to buy only Transistortester clones with a genuine ATmega. This way you can also upgrade your tester to one of the two OSHW firmwares.

[Phil1977]

Cited very very freely from Werner Heisenberg: There is no measurement without compromise. Know yourself, know your equipment and know the limits.

Knowing the limits I often can also live with cheap equipment. If I have a €7-Transistortester and a bunch of caps that are expected to be equal I can test them and if they all show the same value then it´s an indication that all parts are okay.

If I have a €2000 LCR meter then I can test them for absolute values and directly send them back to the supplier if the test fails.

But neither of both devices can tell me if a cap is leaky at 400V. For this test I need specialized equipment that I can quite easily build on demand.

From my point of view we are living in amazing times because even super cheap test equipment can give you so much information. But no instrument can release you from cleverly interpreting its results. Not even the most expensive one.

[temperance]

I wonder why you need to verify component with curve tracers and transistor testers?

[madires]

Because the n-ch JFETs you buy via eBay or AliExpress often turn out to be jellybean NPNs. Or you wonder why the replacement MOSFETs for the SMPSU you're repairing blow up the second time (they are counterfeits).

[axantas]

I wonder why you need to verify component with curve tracers and transistor testers?

Because we can and we have a collection of these "things", which is to be used 

[AVGresponding]

I wonder why you need to verify component with curve tracers and transistor testers?

You don't always, it's entirely dependent on your use case.

If you want a couple of transistors to turn on some relays, all you need to know is if they can handle the voltage and current in general terms.
If you want a couple of transistors for matched input stages (for anything that might need such things) you're probably going to want to check very carefully that they match in a number of characteristics, in which case a curve trace becomes useful.

The point about caps being tested for leakage current at 400V is again, use case oriented. You only really need to do this if you're going to be using them at that voltage.

[temperance]

Because the n-ch JFETs you buy via eBay or AliExpress often turn out to be jellybean NPNs. Or you wonder why the replacement MOSFETs for the SMPSU you're repairing blow up the second time (they are counterfeits).

Interesting. So you spend $30 on some test equipment to verify to cheap junk you obtain from whichever place.

[KungFuJosh]

Interesting. So you spend $30 on some test equipment to verify to cheap junk you obtain from whichever place.

As opposed to what? The exceptionally expensive test equipment we also use to very the expensive components we also purchase?

I use the transistor testers to make sure the parts I get from Mouser are functional, and sometimes I check to see if values are close for matching. If the test results are repeatable, that's usually useful.

If you know of a significantly better transistor tester that doesn't cost $50,000 and isn't the size of a 1980s computer, I'd love to check it out. 🤷

[temperance]

A company where I sometimes work places between 300 and 500 K components / hour. After assembly all components are tested with a flying probe tester. 0.0000001 or even less of all rejects are actually defective components. Alter all this, all assemblies are tested by means of an in circuit tester measuring voltages and currents by means of test points on each PCB trace. Here the same applies. Almost each failure is usually caused by assembly problems and not defective components.

Of course we do have defective components. But defective components failures are usually failures of a different kind. Those components fail after a short period of use. And even then, some of those defects were engineering mistakes like operating components outside their maximum ratings under startup conditions for example.

In my own experience I never received a bad component from a reputable distributor. (On average 400 K components on the shelf for small production runs) If something didn't work it was usually my own error.

-Matching BJT's or MOSFET's: with a bread board and a DMM. (But matched pairs are not that expensive anymore)
-You will need an RLC meter for special cases and for very special cases you probably have to invent your test equipment. (measuring the saturation behavior of inductors for example up 20 A and higher)
-Finding broken semiconductors is easily done with a DMM. If leakage due to aging is a problem I measure some voltages and gently apply some heat to see how things drift around. (I think it is easier than putting every suspect candidate in an dedicated transistor tester.) By doing so you also know if there is anything else wrong in the circuit. (Calculate on the fly which voltages you are supposed to find in there. A rough calculation is good enough to start)
-How to put SMD components into such a test device? DMM or scope probes can probe almost any SMD component or a nearby available solder joint.

I'm just someone failing to see the value of those test devices and prefer poking around in the circuit itself.

[shabaz]

There is value for hobbyists. I have a cheap $20 or so LCR meter, I used it when I didn't have anything better, for seeing if my inductors were in the right ballpark, despite knowing that the test is not applying a known frequency signal; that's not the way those low-cost meters work. I still have it, and still occasionally use it, despite having access to better instruments. I sometimes (very often!) do not want a precise measurement, nor might I care that there's no specified accuracy for a $30 instrument. Often with uncertainty, I can live with it, it's part of normal life.

You can learn a lot by observing the change in value as you compress the inductor, or applying different materials close to it. You also can visually observe drift as the instrument warms up, learn your requirements for a better instrument, and notice the need for calibration, or zeroing, and observing the amount of change that occurs when the test wires are moved.

I don't think many engineers would use it to 'select' a component as the OP asked, and there is often no need either, since many circuits can be 'tuned' through voltages or other adjustments. But a $30 meter for observing some quite interesting behaviour for a beginner, and providing results that you may indeed need to verify in other ways (e.g. in-circuit running the end application and seeing if it works) for non-critical work, sounds like a very acceptable thing to do.

[thm_w]

I use the transistor testers to make sure the parts I get from Mouser are functional

There is no need to do this though.

[KungFuJosh]

I use the transistor testers to make sure the parts I get from Mouser are functional

There is no need to do this though.

You're joking? I build some complex enough products and I don't want to guess if there's an issue. Every single component gets tested before install. Every single time. And yes, I have found failures that saved me a lot of frustration. Sometimes thanks to a DMM or LCR, and sometimes thanks to these cheap transistor testers.

[temperance]

You're joking? I build some complex enough products and I don't want to guess if there's an issue. Every single component gets tested before install.

I find it hard to believe you will be using a $30 transistor tester powered with two AA batteries to verify component performance. Because that's what this topic is about.

I do know that some companies building industrial kW power supplies tested all power transistors before installing them. The same for a company building kW audio amplifiers in de USA. Unfortunately the switching power supplies they developed later on to replace kW transformers were not of the reliable kind and the brand evaporated over night.

Edit:
I do verify what I receive from suppliers after being burned a few time. Getting 27 mOhm resistors instead of 100 mOhm for example. Instant smoke... But I don't test all components. That would be about 90 K components this month.

[KungFuJosh]

You're joking? I build some complex enough products and I don't want to guess if there's an issue. Every single component gets tested before install.

I find it hard to believe you will be using a $30 transistor tester powered with two AA batteries to verify component performance. Because that's what this topic is about.

Believe whatever you like, but I only check transistors with the devices this thread is about. I don't go through 90k parts a month, so it's not a big deal for me. What I do is on a smaller scale, so failures are a larger problem.

[thm_w]

You're joking? I build some complex enough products and I don't want to guess if there's an issue. Every single component gets tested before install. Every single time. And yes, I have found failures that saved me a lot of frustration. Sometimes thanks to a DMM or LCR, and sometimes thanks to these cheap transistor testers.

So you've had multiple failed transistors from a major distributor? Not something I've ever heard of before.

[KungFuJosh]

So you've had multiple failed transistors from a major distributor? Not something I've ever heard of before.

I've been in business for 30 years. Yes, I've had multiple failed transistors over the years. I'm surprised anybody is surprised by this. I'm in the guitar business, even the bigger distributors aren't always that big, and who knows where they get their inventory from. Probably more relevant is that some devices call for vintage transistors, and that should speak for itself when it comes to testing before using.

[porter]

Markus Frejek, the inventor (I think) of these types of devices introduced the tester this way:

Every hobbyist knows well the problem: one builds a transistor made from a single device, or extracting one
from the junk box. If the type designation can be seen (and known one is), everything is fine. Often, however,
this is not so, and then the usual times goes off again: type and pinout of the transistor from the Internet or a
book out looking. It's annoying in the long run of course pretty. Also transistors in the same housing do not
always have the same pin assignment. For example a 2N5551 has a different assignment as a BC547,
although both have a TO92 housing. If in a transistor, the term is no longer visible (or not even Google
knows what to do so), it is with conventional methods really complicated to figure out the types of transistors:
it could be to NPN, PNP, N or P­channel Mosfet acting etc.. Possibly the component also has a protection
diode internally. This makes the identification more difficult. By trial and error by hand is thus a rather time­
consuming task.
Why not let do it by microcontroller that? So this automatic transistor tester has arisen.

Is this an entirely new type of instrument?

[Fungus]

Their original purpose of these was a component identifier.

The fact that they have to do a few basic tests on a component to figure out what it is was secondary.

Now people expect a $30 gadget to do the job of an LCR meter and curve tracer?