Like the title suggests, i would like to know what happens to a typical complementary, dual rail, Class AB solid state amplifier when you short-circuit the output, under the following conditions...
1. Without input signal (and assuming a very low DC offset)
2. With input shorted (again low DC offset)
3. Does topology matter for the above conditions (i.e. CFP or EF)?
4. Without input signal, but with a capacitor coupled output
5. Does the input topology matter?
6. Assuming the no input signal scenario can survive an output short circuit, can that short trigger some kind of oscillation?
I couldn't find info on those questions with a search, and i am really curious. My intuition says that there shouldn't be a problem without signal, but then again, old diyers' experience > my intuition!
1. Without input signal (and assuming a very low DC offset)
2. With input shorted (again low DC offset)
3. Does topology matter for the above conditions (i.e. CFP or EF)?
4. Without input signal, but with a capacitor coupled output
5. Does the input topology matter?
6. Assuming the no input signal scenario can survive an output short circuit, can that short trigger some kind of oscillation?
I couldn't find info on those questions with a search, and i am really curious. My intuition says that there shouldn't be a problem without signal, but then again, old diyers' experience > my intuition!
Never had the misfortune to want to do that but assuming there is no DC offset and no signal, nothing. Like wise with a capacitor fed load where you can ignore any DC component as the capacitor will decouple it.
Give it a signal and toast!
Give it a signal and toast!
1. It should survive and be little different to normal quiescent conditions. This relies on there being some small series resistance between true amp output (amp output meaning the feedback node) and the short in order for the feedback to operate normally. Internal wiring, a small deliberate internal resistance, the speaker relay etc etc is all enough. Those same provisos should also prevent the amp oscillating.
2/ Same as above.
3/ Nope.
4/ Should be pretty much immune to any problem. So the same as 1 above.
5/ Nope.
6/ Yes 🙂 but its very amplifier specific. No hard and fast rules.
2/ Same as above.
3/ Nope.
4/ Should be pretty much immune to any problem. So the same as 1 above.
5/ Nope.
6/ Yes 🙂 but its very amplifier specific. No hard and fast rules.
Hi,
1 to 5 should survive in most cases. The other case is
6 and then it depends on what sort of protection is used.
Needless to say shorting an output is a very bad idea.
Generally relatively easy to protect against, whilst
the real difficulty is setting the SOA against too low
impedance loads.
Short protection is easy, what is hard / debatable
is setting SOA to be foolproof, against setting it
to allow more drive into say 2 to 4 ohms but
allowing the possibility of 1 to 2 ohms to fail.
rgds, sreten.
loads
1 to 5 should survive in most cases. The other case is
6 and then it depends on what sort of protection is used.
Needless to say shorting an output is a very bad idea.
Generally relatively easy to protect against, whilst
the real difficulty is setting the SOA against too low
impedance loads.
Short protection is easy, what is hard / debatable
is setting SOA to be foolproof, against setting it
to allow more drive into say 2 to 4 ohms but
allowing the possibility of 1 to 2 ohms to fail.
rgds, sreten.
loads
I have found shorting output usually blows up one or more output transistors in normal use. Apply zero volts is unlikely to blow it up unless there is a dc offset on the output.
In 1980 I built a Maplin 225WRMS amplifier for a mobile disco.
While out on a gig I tripped over the speaker lead pulling out the jack plug.
On its way out it shorted the output of the amplifier.
Luckily it just blew the amp output fuse and not the output transistors.
I replaced the fuse and carried on.
I rarely see fuses on amplifier outputs these days.
In 1980 I built a Maplin 225WRMS amplifier for a mobile disco.
While out on a gig I tripped over the speaker lead pulling out the jack plug.
On its way out it shorted the output of the amplifier.
Luckily it just blew the amp output fuse and not the output transistors.
I replaced the fuse and carried on.
I rarely see fuses on amplifier outputs these days.
This is actually an interesting point.
Pretty much depends on what we mean by "shorted".
A powerful amplifier will survive 0.01 Ohm wire with no signal.
Many amplifiers will survive 1 mOhm.
However, 1 uOhm will lead to significant increase of quiescent current in one of the push-pull shoulders - either protection will trigger, or current limiter will engage, or the fuse will open, or the output stage will burn.
NFB efficiency is the reason, as mentioned above. At some point feedback really looses control over DC balance.
No problem for unity gain power buffers with no global feedback loop though.
Pretty much depends on what we mean by "shorted".
A powerful amplifier will survive 0.01 Ohm wire with no signal.
Many amplifiers will survive 1 mOhm.
However, 1 uOhm will lead to significant increase of quiescent current in one of the push-pull shoulders - either protection will trigger, or current limiter will engage, or the fuse will open, or the output stage will burn.
NFB efficiency is the reason, as mentioned above. At some point feedback really looses control over DC balance.
No problem for unity gain power buffers with no global feedback loop though.
> What happens when you short a SS amp output?
Typical power transistors can pass FAR more current than we use for speakers.
If a circuit (any topology) can put good current (power) in a load, it will try to put FAR more current in a short.
Hardly matters how you "excite" the short. Signal, offset, etc. The output impedance is typically SO very low, that huge currents will try to flow.
So there are two cases:
1) UN-protected output stages try to flow "infinite" current. Something blows.
2) The Designer expected shorts to happen, and designed-in Protection.
Protection comes in several degrees. Simplest is plain current limiting. Problem is that in a short we have high current AND high voltage so HUGE power in the transistor. Next is fold-back where current falls along a line based on voltage. This perhaps most common. It is a narrow path between ample output in safe loads, and tolerable over-heat when shorted.
These amps will often survive momentary shorts.
Crown power amps have little computers monitoring transistor abuse and shutting down when needed. They also tend to have ample output capacity to handle "bad loads" without shut-down. These techniques are no longer Crown-only.
National's Overture power chips have internal thermal and V/I sensing fast enough to predict trouble and shut-down before failure. There isn't a lot of headroom in a chip-amp, and they have been criticized for abrupt waveform clipping in reactive loads. Also worst-case they shut-down and re-start constantly, which over a period of hours/days eventually cracks the seals.
Dynaco ST-80 transistor amp had no explicit Protection. The power device Base current was limited and you could sometimes short them for brief periods. The ST-120 had a single limiter in the power supply, was more tolerant, but could still be killed. (I used to earn a salary fixing these things.) Both of these are now vague vintage memories.
Cap-coupled does make DC abuse impossible. My untested impression is that several Musical Instrument amps which "should" have needed good protection survived decades in the wilds because it takes extreme AC abuse to really kill a cap-coupled output. I don't think that's an excuse for cap-coupled, or for failing to add protection.
Cordell's book has an extended section on Protection. I believe Self's does also. I recall Sloane's book showed protection circuits.
Typical power transistors can pass FAR more current than we use for speakers.
If a circuit (any topology) can put good current (power) in a load, it will try to put FAR more current in a short.
Hardly matters how you "excite" the short. Signal, offset, etc. The output impedance is typically SO very low, that huge currents will try to flow.
So there are two cases:
1) UN-protected output stages try to flow "infinite" current. Something blows.
2) The Designer expected shorts to happen, and designed-in Protection.
Protection comes in several degrees. Simplest is plain current limiting. Problem is that in a short we have high current AND high voltage so HUGE power in the transistor. Next is fold-back where current falls along a line based on voltage. This perhaps most common. It is a narrow path between ample output in safe loads, and tolerable over-heat when shorted.
These amps will often survive momentary shorts.
Crown power amps have little computers monitoring transistor abuse and shutting down when needed. They also tend to have ample output capacity to handle "bad loads" without shut-down. These techniques are no longer Crown-only.
National's Overture power chips have internal thermal and V/I sensing fast enough to predict trouble and shut-down before failure. There isn't a lot of headroom in a chip-amp, and they have been criticized for abrupt waveform clipping in reactive loads. Also worst-case they shut-down and re-start constantly, which over a period of hours/days eventually cracks the seals.
Dynaco ST-80 transistor amp had no explicit Protection. The power device Base current was limited and you could sometimes short them for brief periods. The ST-120 had a single limiter in the power supply, was more tolerant, but could still be killed. (I used to earn a salary fixing these things.) Both of these are now vague vintage memories.
Cap-coupled does make DC abuse impossible. My untested impression is that several Musical Instrument amps which "should" have needed good protection survived decades in the wilds because it takes extreme AC abuse to really kill a cap-coupled output. I don't think that's an excuse for cap-coupled, or for failing to add protection.
Cordell's book has an extended section on Protection. I believe Self's does also. I recall Sloane's book showed protection circuits.
Member
Joined 2009
Paid Member
I assume the feedback loop is the real culprit as it will provide whatever error signal to the output stage is needed to get the output back to the same dc-offset it had before it was shorted. Which means it'll apply close to rail voltage.
Last time I shorted the output of an amplifier it blew out the emitter resistor, a 5W 0R47 wirewound. I think it did some other damage too.
Last time I shorted the output of an amplifier it blew out the emitter resistor, a 5W 0R47 wirewound. I think it did some other damage too.
Once i had an amplifier (quasi complementary output stage with no short protection) connected to the speakers by long thin single core telephone wires.
I was turning the volume up till the wire begun to sing...yes the wires were shorted at the speaker end.
I was lucky nothing blew.
I was turning the volume up till the wire begun to sing...yes the wires were shorted at the speaker end.
I was lucky nothing blew.
- Status
- Not open for further replies.