Designing amp to withstand shorted output

[mr.duck]

I want to make a pre-amp / headphone amp using discrete components that must be able to cope with a shorted output. Not just as a precaution, but because 1/4" jack used by headphones causes a short when being plugged in. I don't want to use any overload protection schemes if it will compromise sound quality like putting a resistor on the output to limit current.

How about using many transistors in parallel in the output stage so that each transistor will not be overloaded, and all the transistors in parallel reduce the output impedance to something very low like <1 Ohm. I've attached image of what I mean. Not sure how many transistors are needed to reduce output impedance. The value of the resistor (if the transistors used are TO-92 with 500mA max current and 15v supply) would be 15/0.5 = 30 ohm. Is this correct and do you think it will work ok?

Also, is output impedance reduced when global negative feedback is used?

[Wavebourn]

You may use my output stage, it is not a secret anymore, I've published it several times already on the forum. However, if you age going to run production you can't use it without my permission.

It is made of pair of complementary operational amplifiers where one transistor is a difeerencial cascade is replaced by diode. Why diode? Because impedance of feedback loop when voltage gain is =1 is low, and because breakdown voltage of a diode is higher than b-e junction of transistors. As I said, both opamps have 100% voltage feedback (through diodes), so voltage gain is =1. But when output is shorted feedback by voltage stops working ("closed" diodes), so input transistors have 100% NFB by current, output transistors amplify it up to safe level (actually, a bit higher than possible peak current on nominal load).

[Sheldon]

Originally posted by Wavebourn
You may use my output stage, it is not a secret anymore, I've published it several times already on the forum. However, if you age going to run production you can't use it without my permission.

Anatoliy, unless you have already patented your stage, anyone can do anything they want with it. The sole significant exception is in the U.S., where you have up to one year after public disclosure to file. Even so, anyone can use the published method up until the patent issues. If a commercial vendor wants to give you a royalty when you have no legal right to demand one, they of course can. But they are not legally required to do so. If you want to protect your ideas, keep them to yourself, or execute an agreement with every user to keep them secret. BTW, making and selling a product that comprises the idea is the same as publishing it, unless you have a secrecy agreement with each buyer.

Sheldon

[Wavebourn]

Sheldon;
thank you for the patent law related lesson;
actually, I am for a freedom of flexible usage of electronic schematic solutions, like in case of Free Software Foundation. When you use some free software you have to follow some rules mentioned in the special file. Does not matter if it is protected by law, or just a jentlemen - like agreement, but I want the same with my electyronics ideas, so no crooks can patent what people are free to use as an air, or water, or whatever.

Originally posted by Sheldon


Anatoliy, unless you have already patented your stage, anyone can do anything they want with it.

...and patent it?

[Sheldon]

Originally posted by Wavebourn
Sheldon;
thank you for the patent law related lesson;
actually, I am for a freedom of flexible usage of electronic schematic solutions, like in case of Free Software Foundation. When you use some free software you have to follow some rules mentioned in the special file. Does not matter if it is protected by law, or just a jentlemen - like agreement, but I want the same with my electyronics ideas, so no crooks can patent what people are free to use as an air, or water, or whatever.

And I have no problem at all with that.

Originally posted by Wavebourn
...and patent it?

If you mean by this question, can someone else patent it? The short answer is no.


No one can take your published invention and file a claim on it. Even with the U.S. one year grace period, they would have to show that they invented first and made a filing within a reasonable period - not something they invented years ago and were prompted by your publication to file.

One technical addendum I should add to my prior comments on unissued patents: The publication of a provisional patent can allow for royalty claims from its date provided the patent issues with substantially the same invention claims.


Sheldon

[sam9]

For headphone amp (+/- 15 V rails), I have used the simple expedient of a 50-ohm or 75 ohm resister in series with the output. So far I have neither heard, seen or measured any adverse effect from this. Dead shorts appear to do no harm.

It's class-A so it's running full-out all the time anyway.

This may not be a suitable means in your case.

[mr.duck]

Originally posted by sam9
For headphone amp (+/- 15 V rails), I have used the simple expedient of a 50-ohm or 75 ohm resister in series with the output. So far I have neither heard, seen or measured any adverse effect from this. Dead shorts appear to do no harm.

Hi, I didn't want high output impedance for same reasons you don't want high output impedance from a speaker amp. Low output impedance for better damping factor, to give tighter bass. Some headphones may be designed to sound best from high output impedance (120 ohm is the official spec I believe). But most high end headphones work best from low output impedance like a loudspeaker does.

[sam9]

In principal, I agree that high output impedance can be problematic. In practice, it appears to be non-critical when there is sufficient current available. Such as an OpAamp driving a pair of medium sized BJT's such as mje340/350.

[mikeks]

Originally posted by mr.duck
I want to make a pre-amp / headphone amp using discrete components that must be able to cope with a shorted output.

Send me mail.

[GK]

For short circuit protection, just put a current limiting resistor in the collector leads of the output devices.