[Amps] Tubes, transistors, and 'abuse'

[Manfred Mornhinweg]

It's fun to see you quarreling, once again, over the "are tubes or
transistors better" question! I thought this had been about settled, 
back around 1960 so or...

>> When you handle them, do you have an anti-static mat and wear a
>> ground wrist strap?

And in this regard I must have missed something... Because I don't have 
an anti-static mat, nor do I have a grounding wrist strap. At the job 
some people used those straps, but I refused: Grounding one of my wrists 
means that if I put my other hand into something that has enough 
voltage, it's _me_ who needs to be replaced, rather than a MOSFET!

I started using MOSFETs in 1980, after doing my first steps in 
electronics (detector radio) in 1977. I'm still using them. I have 
"handled" lots of them. And I can't remember having lost any of them due 
to static while handling it. In fact the only case I can remember when I 
did lose something to static was a case of sheer stupidity: At work, I 
walked over to the office of a colleague, an unprotected card in my 
hand, on a dry day, and when he extended his arm to grab it, a nice 
spark jumped from his fingertip to the card edge and gave all three of 
us a jolt: My colleague, me, and the card. After that the card didn't 
work, and I had to replace a CMOS chip on it.

But that was the only case of static-induced damage, in all of my career 
so far!

I have read enough terror stories about static discharges partially 
damaging devices, which will then fail later in the field - but so far I 
have never witnessed any case like that, so I think that the stories, 
like all good terror stories, are considerably exaggerated.

I do apply basic precautions, but I do NOT ground myself, nor do I work 
on a grounded surface. The precautions include, for example, to avoid 
touching an ungrounded soldering iron to the high impedance circuit of a 
MOSFET gate, while that circuit is connected to anything else (for 
example, grounded). Instead I lay the circuit board on the insulating 
work desk, and then happily go along soldering. Almost all MOSFETs 
include little zener diodes to protect the gates. These zeners can 
perfectly well take the current caused by a poorly insulated soldering 
iron, and coupled to ground through the few picofarads of capacitance 
the card might have to the rest of the world. But they might not be able 
to take the current put in by that same ungrounded and poorly insulated 
soldering iron, when the card is grounded!

In the same line, when grabbing a MOSFET from a metal surface, grab it 
by its body, or the source or drain leads, but not by the gate lead. 
Instead when grabbing it from an insulating surface, grab it any way you 
want, you won't damage it.

With just basic precautions like these, I have fared well. Despite 
having used thousands of times more MOSFETs and CMOS devices than tubes, 
I have lost more tubes due to handling mistakes than solid state parts: 
  One time I dropped a tube on the floor, and it broke; three times or 
so I cracked tubes while trying to remove them from their sockets (all 
three were small receiver tubes); and I lost a 50GY7 tube simply by 
opening the door next to the TV that used it, and the cold air stream 
was enough to crack the tube from thermal stress! Talk about being fragile!

Of course, if that card I handed to my colleague had been built with 
tubes instead of CMOS chips, it would have survived the static hit just 
fine. But then I don't see how I could have built that realtime 
telescope tracking control card, using tubes...

Cathy, that "good" little amp that eats MOSFETs for breakfast, lunch and 
dinner cannot really be "good". It must have a serious design flaw, to 
behave like that. So it's a BAD amplifier, not a good one. The IRF510 is 
a small, rather low power MOSFET, with a pretty good gate voltage range, 
and it's no rocket science at all to make an amp with it that works 
without blowing up such MOSFETs. I would almost bet that the amp is 
cooking those poor IRF510, due to insufficient heatsinking, and very 
possible combined with a thermal runaway effect. At any drain current 
below roughly 4A, the current will increase when the IRF510 heats up, 
leading to thermal runaway if not properly controlled by the circuit. In 
a class A amplifier it's easy to stabilize the current, but I have seen 
some design that don't do it! On the other hand, class A involves high 
dissipation, and thus is prone to frying transistors. What power level 
does that little amp run? With very good heatsinking, each IRF510 might 
deliver up to about 7W in class A, before frying it. With less good 
heatsinking, less power. Instead in class AB the power could be 
significantly higher.

The designer's task gets significantly harder when you are trying to 
milk a high power LDMOSFET for the last watt it can deliver. Not only do 
these LDMOSFETs have a much smaller gate voltage survival range, and a 
high gain and wide frequency range that can lead to self-oscillation, 
but also it's really a beastly task to keep them cool enough! Yet it can 
be done, and has been done.

Okay, let me end this post now, and return to the other side of the 
shack, where I have to install some MOSFETs in a circuit I'm building. 
Without antistatic mat nor grounding straps. And I don't expect to kill 
them.

No tubes today for me. Nor yesterday. Nor tomorrow, in all likelihood.

Manfred

========================
Visit my hobby homepage!
http://ludens.cl
========================