DIY way to check working voltage of audio power transistors

[rajeev luthra]

DIY way to check working voltage of audio power transistors

Is there a simple way by which we can select higher working voltage power transistors from a lot .

If we make a variable unregulated supply by using a variac in the primary of the power supply transformer , check the transistor using a current limiting resistor in series and measure the breakdown voltage using a digital voltmeter.
will this work ?
can anyone give some more simple ideas

[UrSv]

Why not read the datasheet?

[consort_ee_um]

If you have a 100V power supply or supplies you could wire in series it would be easier. I would use a relatively high value resistor say 47k as there is the possibility of damaging the device.
In the old days the maker would do the selection i.e a TIP31A was a 60V device a TIP31B was an 80V device and a TIP31C was a 100V device.
I suspect you are wanting to push the device beyond the data sheet limits. There is usually some margin but reliability may be compromised

[consort_ee_um]

Gosh, I forgot to say you need the correct polarity DC supply. Your proposal would apply unisolated AC which is not a good idea and is possibly dangerous.

[rope]

Hi Rajeev, you must read about fundamental electronics and semiconductors The Vce max depends on temperature,shematics(Rbe) and many others so you can not rely on one measuring with some shematics-dont push your luck!

[rajeev luthra]

You are right Rope,
I knew that the devices may get destructed on breakdown testing, but I still wanted to try and see if there is a method,

Consort says that in old days they used to select Tip series in this way, Is that really possible ??

In another thread DJK had said that his friend had made the leach amp with selected transistors of higher working voltage 20 years ago which is still working ,

My problem is that here in India it is very difficult to get devices , specially for DIY purposes ,some replacement devices are avilible for professional amps but there rates are sky high , hence we have to use various alternative nos for our application , hence I got tempted to find out if there is a simple way to select devices and I want select devices which can withstand only 20 % more than the specified voltages

[Tony]

i damaged several power trannies this way, i used very high series resistors, 1meg, yet i still damaged them...so i never tried again...they were unmarked to3 power trannies, i was able to get the breakdown voltage alright, but then after that it had leaks...

[alfsch]

To test breakdown, use setup below. i tested many tranistors this way, no damage...important: no capacity at collector, this would make a avalanche peak current and destoy the dut. when testing mosfet, connect always (!!) gate to source before test.
to see temperature effects, watch voltage while heating up the dut with a solder iron.
with base open, voltage is VCEo.

[ashok]

To ensure that you do not bust the transistor , charge up a very low value capacitor ( maybe 0.1uF) and apply that through a switch to the device under test. Obviously the charging source for the capacitor must be disconnected while testing. A DPDT switch can be wired to achieve this. One contact opens while the other closes.
That way when the transistor breaks down at VCEO , there isn't much charge left to blow it. I've done this without blowing transistors . The DMM must be set to 'HOLD'and of course the series resistor must be there.
Use a low value resistor ( 1K?) between the dc supply and the 0.1uF cap to avoid a spark.
It was long ago and I can't remember what cap value I used. It was low however.
Cheers.

[jneutron]

Originally posted by rajeev luthra
You are right Rope,
I knew that the devices may get destructed on breakdown testing, but I still wanted to try and see if there is a method,

The old type 575 and 576 curve tracers use a variac powering a rectifier. no capacitor..

There is a selectable series resistor, the selector knob has on it maximum dissipation that the combination of voltage and resistance can deliver to the device under test.

Keep the dissipation to less than half a watt to a 1/10 th of a watt, and you most likely will not destroy the power device. Keep the current under 10 uA, and you will be safe, 100uA for power devices.

Bipolar, always tie base to emitter, mosfet, gate to source.

For bipolar devices, there is an additional issue...

The base width of the device is always a function of the device speed, and sat gain specs. The higher gain, faster devices will have thinner base diffusions..

You can certainly operate a device beyond it's designed values, as long as you remain below the breakdown levels...but you will not be able to easily test the device's ability to resist punch through, a rarely discussed failure mechanism, as most designers don't exceed design numbers..nevermind secondary breakdown, which gets nasty in the higher voltage regime.

Punch through causes a very high gain area under one of the emitter fingers, and causes thermal runaway in that spot, this runaway will not be at large currents, typically I've seen this kill a 250 mil square device with 10 milliamps current. For normal devices like 2n2907, I've seen it happen as a result of a defect in the primary mask from wafer fab, causing every fourth device on the wafer to fail. For power devices, it can happen directly under one of the bonding wires, as there is a concentration of stress under that that can push the depletion zone down. For RF and switching devices, it's nature of the beast, as the base width is minimized for speed.

Detection of punch through can be done by comparison of the BVCES vs BVCBO...the breakdown between the collector and base will have one value...if the breakdown collector-base with the emitter tied to the base is a lower voltage, the device is punching through...do this test at 10uA, and it will be non destructive.

Most manufacturers do not know where their devices will do this, or even if they do it at all..and they will destroy the devices during datalogging if they perform the normal BVceo test..so they typically do not do that, but simply a pass/fail at the maximum voltage of the device.

Keep below 10 uA, and you'll not destroy any..

Cheers, John

PS...well, you asked...