I came into amplifiers through the door of tweaking and an understanding the underlying physics. I never went the route of hardcore electronics design.
I probably could design my own and then fiddle fiddle fiddle, until the cows come home..and manage to get there. I have read One of Doug's editions front to back about 4-5 times. I could muddle my way through it. however, that would be a terrific amount of wasted time and effort, when I could get someone who has designed amplifiers before, to involve themselves with a given design.
Then I look it over and make changes, according to my own discovery work. That would be at least 10x less complicated, for my end.
I'm too busy doing other stuff, to get involved designing an amplifier from scratch. Too much room for error, too many missteps, too much time, etc.
regarding the final result: I stand on my life that it would be one of the best amplifiers anyone has ever heard. Unconditionally.
This is very true. This matter can be very layout-dependent.
It is also true that, if one must use base stoppers, it is nice to have an inductor in parallel to eliminate the effect of the resistance at low frequencies, especially at DC in regard to thermal stability. But if one is using one stopper network per transistor, the approach of using an inductor in parallel with a resistor becomes expensive and bulky.
2.2 ohms on each base is usually enough. 4.7 ohms is common. 10 ohms is way too much, and leads to significant thermal instability, especially when RE is small (e.g., less than 0.22 ohm).
Cheers,
Bob
also kbk, those of us with the electrical background to understand the circuit theory can analyze and correlate.
mlloyd1
mlloyd1
I use this approach , 1000 hobbyist's are happy with 4.7R stoppers and .22Re at the outputs. No blown stages (so far).
We even will use 1206 smd right at the to-3p base pin for the stopper.
There was another HK design that did use a L/R for each output. I guess
expensive and bulky was planned.
OS
Careful, 1206 SMD resistors can take very little peak current before they blow.
KBK, what kind of amplifier would you prefer to work on? Do you have a set of requirements for an amplifier you think would be worth your tricks?
This sounds like a good conservative approach. The space is not really an issue. I imagine the base current could get quite high under some fault conditions or conditions where there might be serious beta droop in the output transistors at high current.
Cheers,
Bob
1r0 250mW can pass 500mA continuously and pass a 1Apk transient for a very short time.
Any transistor that demands more than 1A of base current while connected to a load/short circuit is already toasted. A 2pair output stage with a drooped gain of 20 for hFE would be passing 40Apk into that short circuit and not blow those 250mW base stoppers..
Any transistor that demands more than 1A of base current while connected to a load/short circuit is already toasted. A 2pair output stage with a drooped gain of 20 for hFE would be passing 40Apk into that short circuit and not blow those 250mW base stoppers..
You don't get it, it's not about power dissipation. A fuse doesn't blow because of exceeding its power dissipation limit.
Yes it does.
It overheats to such an extent due to power dissipation that it melts.
In the case of a base stopper one wants it to survive undamaged in the event of some amplifier abuse.
If the output devices did catastophically fail, then a blown base stopper is not the end of the world.
It is about power dissipation !
It overheats to such an extent due to power dissipation that it melts.
In the case of a base stopper one wants it to survive undamaged in the event of some amplifier abuse.
If the output devices did catastophically fail, then a blown base stopper is not the end of the world.
It is about power dissipation !
Hi Andrew,
You are exactly right about both fuses and resistors, perhaps with the exception of a few extreme cases for resistors.
Many people have only a vague understanding of how fuses work. They must have some resistance and must heat up in order to work. It is about power dissipation, although there are some significant nonlinearities in the way it functions. As it heats up, its resistance increases and that in turn increases the amount of power dissipated in it. Fuses take quite a long time to blow unless they are subjected to current much larger than their rating. Just go to the LittleFuse web site and look at the blowing time vs overload curves.
However, fuses can "wear out" if repeatedly subjected to significant overloads.
Many resistor types can be remarkably robust to brief overloads of easily 10 times their rated dissipation. Metal Oxide Film (MOF) resistors are very robust this way. I usually use 3-Watt MOF's for output emitter resistors rather than non-inductive wirewound resistors. I don't know about the level of robustness of SMT resistors to brief large overloads.
Cheers,
Bob
Yamaha used fusible resistor base stoppers in one of there 1980s designs. The base stoppers blew before the output transistors. It was pain having to change them all. The drivers, pre drivers however didn't survive.
I don't remember the model.
I don't remember the model.
SMT resistors are good fuses. I think this is what Waly is referring to as well. A 1/4W 1206 won't take anywhere close to the abuse a through hole one will.
I think a fuse blowing has more to do with it's tempco than it's static resistance. The tempco needs to be the dominant mechanism, otherwise the fuse could not be conductive enough in normal operation. Maybe this is Waly's point.
What about MELF resistors? There are high surge versions as well.
What about MELF resistors? There are high surge versions as well.
an oddity of the tech - untrimmed smt flat chip resistors have usefully higher surge ratings than the typical L cut trimmed ones - the cut concentrates power and V stress
some advertise a "smooth cut" trim on higher surge rated smt parts
some advertise a "smooth cut" trim on higher surge rated smt parts
It is true that in some cases of amplifier failure, the drivers go first and then take the output transistors with them. Driver SOA matters too.
Cheers,
Bob