DIY Audio Projects Forum

Have you been fustrated becasue fake transistors blew up many of your expensive power transistors and caused you to waste a month on fixing your amp?
Running away from aliexpress and other cheap supplier of transistors becasue your afrarid of being fake?

No need for that any more.
Build your self a transistor tester. (You can get full refund if you have sufficent evidence of fraud when buying on aliexpress)

You should stress transistors to 70% of maximum rated voltage and 70% of maximum rated power.
For to-92 transistors, a test at about 400mW is sufficent enough to prove that its real. A stress of 5-10minutes should be done.
For to-220 and bigger transistors, reasonabllly large cooling should be used, heat sink should be directly bolted on to transistor with heat sink grease. The transistor should be tested to about 70-80% max power and voltage. If the transistor withstands this power for 5 seconds or more, then its 100% geniune.
Fake trasnsitors tend to fail at high voltages and fails within 5 seconds with less than 65% of max rated power.

For saftety reasons, you should always have the test transistor facing away from you, in case of a fake transistor, non short proof supply and fuse fault, the transistor will blow up.
An approiate fuse suited to test current should be used. A short circuit proof power supply is always recomend.
For voltages of 36v or less the lm317,337 schematic is recomended as it is blow out proof and overpower proof.
For voltages of 36v or higher the discrete transistor veresion should be used.
Always mark clearly which way the pot should go for minimum or maximum current, always power up with current adjusted to minimum.
A single turn pot should be used, because you should have a piece of mind knowing your adjusted to minimum test current before powering up.

In case of fake transistor being tested, the discrete version may have its test transistor driver (to-220) damaged, this to-220 transistor along with its to-92 tranistor should have maximum ce voltage rating of at least 60% above test voltage. (the to-92 is genrally short proof and will not be damaged because of 8.2k current limiting resistor).

The emitter resistor of the test transistor should have a millivolt to volt meter connected. By monitoring the voltage and doing v=ir. you will find out the current through transistor.
Then by doing p=VI you can estimate the required test current for a 70-80% power load. The voltage of transistor can be estimated as vcc since the drop through emitter resistance is small, the votlage the transistor sees is about the supply voltage.

For example, you are using a 0.25ohm resistor and have monitored 1v accross it. I=V/R. 1/0.25=4Amps accross transistor. Your supply is 24v, this means about 96W is being disipated by test transistor (4*24v) (4*(24-1)to be 100% accurate)

Example 2 you are using a 220ohm reistor for a to-92 transistor. you monitored 0.8v accross 220ohms.
(I=V/R) 0.8/220=0.00363636363A
0.00363636363*1000=3.63636mA
You are using a supply voltage of 82v and P=IV.
The transistor is disipating about 3.636mA*82v=298mW
to be 100% accurate, it is disipating （82-0.8)*3.636mA

Feel free on asking me to design a transistor tester at 100v+ or a tester optimised for a certian operaing voltage and test load.
Should you always test transistors, you should have sockets for quick installtion of transistor and a proper case everything neatly installed.
( you can do this your self by adding collector resistances with approprate power rating to limit power to drivers. chaging zener resistor for higher volts so appropriate current flows through zener etc.)

The theory of operation for 317, is that the transistor current is dependant on its emitter voltage and resistor, the emitter voltage is dependant on the base voltage, the lm317 supplys a constant adjustable votlage to the base of transistor. A zener is in series because the minimum output voltage is 1.25V, with a zener in series the transistor can be completly turned off with voltage adjusted to minimum compared to a ~0.6v at emitter when adjusted to minimum voltage.

Theory of the discrete is virtually the same. it relies on a zener refrence, you then take a voltage divider from zener refrence with two pots. this controls the voltage into the base of test transitor.

Um, um, um:

Why not purchase your transistors from a legitimate source?
If one uses transistors in any kind of volume, maybe one of these?

https://cdn.shopify.com/s/files/1/0915/ ... 1447743059

I will purchase transistors from Mouser, Newark (Element 14) or DigiKey. I will not purchase from Ali-Express, on e-Bay or any other no-name source. I have never had a bad one. Certainly never one that caused down-line damage.

Nice idea, however, for those who simply do not have reliable sources. Were I in that position, I would be all over this device, so, thank you!

Great idea, KochiyaYamato



As a pro repair tech, sometimes it isn't possible. Sometimes you have to go across the pond and deal wiith somebody sitting on a pile of 2SC-unicorn, especially for odd amplifiers dependent on original type, like a Marantz 300DC that will not work with any sort of substitute.

Cheers!

_________________
-= Gregg =-
* Ratings are for transistors - tubes have guidelines*
Home: GeeK ZonE
Work: Classic Valve Design

Those in picture test transitor for gain, possiblly capacitance etc.
Quite meaningless test, fake transitor can have smalle die same gain capacitance.
However fail under large load.

Only way to be sure if a transistor is real is to give it large stress 70-80% maximum rated power on large cooling. withstand 5-10sec, also printed model name does not rubb off with pure alchool ok geniune.

I do not buy from reputatable store because its too expensive. Espcally small signal transisor, buy some test for quality, good quality, geniune and sound good then buy more.
Especally with aliexpress, if you make good evidence like video showing the whole testing process from working to fake transistor damaged, which is what I do you feel confident about a full refund while keeping the fake goods.

I'd say testing one or two transistors from a large batch is usually enough to prove a transistor is real of not.

I tend to perfer testing transistors near the maximum ce voltage and near max power. Some fake transistors withstand 80% of maximum power but has lower ce voltage ratings.

I'm paying 22usd for 10 matched pairs of tip35c, shipping included.

On usa website they are asking 2.8usd per transistor. Total of 56usd excluding shipping. Also transistors not matched.
https://www.alliedelec.com/nte-electron ... /70726223/

Thats a massive diffrence in price, you also don't get matched pairs if you cannot be responsible for testing the transistrs you've brought.

I forgott to point out one important thing on aliexpress, One shoulod never go for overcheap transitors for power ones.
All 0.5usd per transitors I got were fake.
However when going to arround 1usd per transitor I get real deal.

With small signal ones,generally all the ones I brought were real deal, it matches up with max voltage and power ratings.
I also payed a small price for them. For fet 2sk30a I also never had a plorbem.

Dear KochiyaYamato,
Thanks for posting this information. I built the NPN version of your tester and it works quite well indeed. I can use it to verify that the transistors I have can operate at Vce of 56 volt while passing at least 120ma DC. I think that if one measures the base current int the DUT it is also possible to get a nice measurement of the hfe parameter. I have lot of driver transistors that are candidates to go into a Sansui AU-D907 where they are at Vce of 56 volts and need to pass enough current to drive a pair of 2sc6145A or 2sa2223A. But before soldering the ones i select i will make sure they do not blow up.
I just wanted to say that it is well worth building the testers you posted.

Cheers!!

Jose

Nice to see it helped

Measruing base current with a series resistor or multimeter can also give you the gain.
However if too much voltage is droped through this measruing resistor or multimeter, the maximum test current avalible is reduced.
Most transistors have a gain of 100-300 this can give you a good idea on how large your base resistor should be.

Keep in mind if your amp is runing on duall 50v, then all transistors used should have a ce voltage of at least 100v. typically 30% head room is sufficent.
At max swing the transistors will see 100v.
If your amp runs on duall 50v. Then I recomend testing the transistors at 100v. Withstand 70-80% of max rated power for 5 seconds.

I recomend testing transistors at higher voltages than currents, normally fake ones have low voltage ratings and break dow.
70-80% max Power testing will hopefully indicate a fake one cause normally not going to do rated power.
However theres still a chance where its not rated for the working voltage even if it has the rated power.

I got my tip36c from Aoweziic Raw material spot Store, the tip36c pairs are pefrectly matched, model no dosen't rub off with acetone and geniune high quality craftmanship was observed. Further more has the rated 125w dispation.
https://www.aliexpress.com/store/238305 ... 3bacOQjoFW

I will buy transistors from moser digikey allicdata, these are some reliable sources.
https://www.mouser.com
https://www.digikey.com
https://www.allicdata.com