transistor substitutes

[Rich.Andrews]

For years I have struggled with the task of selecting transistor substitutes
with varying degrees of success. NTE/ECG is sometimes right but many times
they fail to perform anything like the original. What is a good way to
figure out which transistor *will* work without re-engineering the circuit?

r

 

[Ol' Duffer]

For years I have struggled with the task of selecting transistor substitutes
with varying degrees of success. NTE/ECG is sometimes right but many times
they fail to perform anything like the original. What is a good way to
figure out which transistor *will* work without re-engineering the circuit?

For me, the ECG numbers make a convenient filing system to
store similar parts in close proximity so that I know what
candidates I have on hand, but very few of the drawers
contain ECG or NTE parts. Instead, I stock a couple of
hundred common parts that I have seen used a lot in past
experience. And beyond my own stock, I am familiar with
a dozen or so distributors who can ship less common parts
in a couple of days.

More to the root of your question, the key is information.
You need to know the characteristics of the original device
and the replacement candidate and enough about the circuit
to know what information is important and what can be ignored.
Sorry if that sounds like a generalization, but your question
is rather general. A higher voltage rating will usually
work, all else being equal, but other factors such as gain,
frequency response, and power dissipation can be highly
dependent on the circuit. Some circuits barely care, while
others barely work even with an exact replacement. And don't
even get me started on all the different case styles...

I have shelves full of data books, but the internet has made
it a lot easier to find information and vendors of exact or
upgrade replacements. Most major semiconductor makers have
web sites where you can download data sheets. Or type your
part number into your favorite search engine. If you get more
hits than you know what to do with, start narrowing the search,
try adding ".pdf" to the part number, etc.

 

[Tom MacIntyre]

For years I have struggled with the task of selecting transistor substitutes
with varying degrees of success. NTE/ECG is sometimes right but many times
they fail to perform anything like the original. What is a good way to
figure out which transistor *will* work without re-engineering the circuit?

r

If you use an actual NTE/ECG match, it should work. People make the
mistake of thinking that all transistors with the same NTE/ECG match
are the same, and they aren't.

Tom

 

[Asimov]

"Rich.Andrews" bravely wrote to "All" (19 Jun 04 03:30:03)
--- on the heady topic of "transistor substitutes"

Ri> From: "Rich.Andrews" <[email protected]>

Ri> For years I have struggled with the task of selecting transistor
Ri> substitutes with varying degrees of success. NTE/ECG is sometimes
Ri> right but many times they fail to perform anything like the original.
Ri> What is a good way to figure out which transistor *will* work without
Ri> re-engineering the circuit?

The original transistor is often the best choice but sometimes it is
impossible to find or simply uneconomical. Substitution is sometimes
the only logical way to go and for this to be successful one has to
consider a number of factors.

The first thing to think about is that very often the original
transistor was either specially designed for the intended device or
sometimes the device was designed around an already available
transistor. Obviously the latter is the easier to substitute and the
first case is the more difficult.

- One important factor is to look for a transistor intended for a
specific task and a good example of this is a TV's Horizontal Output
Transistor. Thus one has to choose between a switching application or
a linear amplifier, perhaps an RF device, low noise preamp, etc.

- One has to match or exceed the original ratings for power, voltage,
current, gain, and speed.

- One has to decide if the new transistor will fit in the space occupied
by the original with the same pinout if possible.

Manufacters make it easy to select a sub by publishing tables of their
transistor specs sorted by their parameters. Motorola for example
offers programs which will try to match one of their transistors by
the parameters one enters as most important. Of course it tends to be
a more costly transistor but it gives a good idea what to look for.

A*s*i*m*o*v

.... When I was your age, we carved transistors out of wood.

 

[Rich.Andrews]

If you use an actual NTE/ECG match, it should work. People make the
mistake of thinking that all transistors with the same NTE/ECG match
are the same, and they aren't.

Tom

I have tried the NTE/ECG approach and have had the least amount of success
with that. Small signal IF or RF transistors is what usually gives me
fits. Either it works marginally or doesn't work at all. Base current
requirements threw things off the last time. I will just keep plugging at
things until I get them figured out. The sheer number of choices make it
even tougher.

Thanks all!

r