Project:OPTIMUM SHORT CIRCUIT PROTECTION IN MOSFET AMPS

Status
Not open for further replies.
Hi everybody,
Recently I came across bassbins fitted 18" Peavey Lowrider speakers along with Passive 24dB/octave linkwitz - Riley passive crossovers.
Our further experiments shows that there is lot of Reactive energy [due to passive crossover]fedback to amplifier, thus turning the overcurrent protection ON without any coniderable power output to load.The normal trip level of protection sensing was 28A peak, but the measurement shows 40A peak of destructive current was present without the protection circuit involved.
This event triggers us to think a new way of activation of shortcircuit protection involved than using traditional approach of sensing overcurrent through a series resistor.

We want to propose a new idea of protection which includes :

1 Sensing the gate to source voltage overdrive during overcurrent conditions, then limit the gate drive or if necessity arises even mutes/limits the input which also avoid harsh clip conditions.

2. A comparator scheme which involves:Sensing the input voltage at the input of amp, sensing the corresponding voltage at output of amp. If the input voltage is present and output voltage is not present, which clearly indicates the conditions of short-circuited output, the the comparator latches or activates the mute circuitry, thus protects the amp.

Kindly , give your valuable suggestions so that we can further continue the discussions involved and a new project will be started soon involving various parameters to be catered along.

Regards,
ampman
 
Proposed Idea

Hi Everyone

Here are the simple image illustrations involved.


Regards
ampman
 

Attachments

  • combo.jpg
    combo.jpg
    32.4 KB · Views: 1,610
Limiting MOSFET GATE :
I WOULD suggest a low current,high voltage mosfet as the sense element fed thru a zener, which will conduct as soon as the drive level exceeds a pre set level,The Mosfet should switch an optocoupler for isolation,the output of Optocoupler should turn on a small signal FET connected across the input of amplifier.Of course the amplifier input should have a resister in series with the signal sourse.This resistor and FET will form an attenuator or even mute.
 
Hi ampman, your contemplation is uncorrect : every amp have phase shift, which bring to the comparator circuit big fault. Also is uncorrect to evaluate input vs output voltage - how will be function of this circuit, when will be any signal on input ? In mosfet amp must be current limiting circuits quite the same as in bipolar amp, but must have still " zener " protection for overvoltage on gates, that's all ;)
 
For "better" current limit protection, i'd probably go with some sort of multistage approach:

- keep the existing super-fast current protection, which limits instantaneous peaks. Make it less sensitive than before - keep it as an "oh @#$^" circuit that will perform instantaneous protection from dead shorts across the output.

- add a second, similar circuit which has a lower current-limit threshold. But low-pass filter the current sample going into it, so that it doesn't act as fast. This protects from low-resistance loads that can damage the amp from sustained high current, but shouldn't clamp down on a quick current spike coming from a funny crossover reactance.

As for the implementation, that's an exercise left to the reader ;)

I've designed RF amplifier protection circuits that work a similar fashion, and they work well.
 
hallo amp_man1
as Upupa Epops said amp has a phase shift, but there are also other problems with simple compairing: amplitude distortion('filtering'), overdriving (even slight), you also should know EXACTLY the voltage amplification factor.

for me it is much simpler (maybe I am missing something):
protection shuold turn on when
1.current via source resistor is to big
AND (even ttl or cmos and)
2. output voltage is zero (null, ground, gnd...);)
 
This reactive overcurrent effect happens when passive crossovers :

- Are not designed properly
- Have capacitors failing in an intermittent short due to excessive voltage
- Have iron core coils saturating due to excessive current
- Have any component with bad solder joints losing contact intermintently
- Have any of their outputs not loaded with a suitable impedance or not loaded at all
- Include impedance compensation for a particular loudspeaker but this driver is not the one used
- Etc..

So I would rather not waste my time trying to make my amplifiers capable of driving a defective passive crossover. I see it pointless. When passive crossovers are working properly there are no reactive impedance drops

Also, loudspeakers only are reactive at frequencies where they show high impedances, so if the problem persists without the passive crossover then the loudspeaker is defective [probably shorted turns on the voice coil]

Your protection is already optimum enough. It trips just because the load is defective, so fix the load instead of the amplifier :D

Passive crossovers for subs are very critical, that's why almost nobody uses them
 
Reply

Hi everybody,

ThanX all of you for your valuable suggestions.

I myself feel that the comparator scheme is not good as there is some amount of phase shift still present at output.
Hafler also uses this comparator scheme but he senses the drive signal and output instead of input signal.

Now I will work on following aspects:

1 Gate overdrive voltage,
Overdrive on gate will be dealt with appropriate design.


Sivanand , your suggestion regarding optocoupler is good but it is not as fast because optocouplers are slow.

2 Traditional method of sensing current through resistor is still good but there must be some tradeoff value for peak current requirements for limiting.


EVA
your suggestion is absolutely right, but some customers will never want to put blame on their speakers rather they blame the amps for not driving them properly.:D


Any more suggestions are most welcomed

regards
ampman
 
amp_man_1 !!

Why u need high speed !! Opto's are usable upto 50Khz.(Your not in the RF arena).I 've tried a current limit scheme in a UPS system having 25KVA it was a PWM sinewave output. With Mosfets, I noticed the system would trip whenever a small surge appears the output of UPS,then i deliberately slowed down the response.The Mosfets can and will sustain short duration overloads w/o damage.
GOODLUCK!!
 
Reply

hi Sivanand ,
Do know about the delay characteristics of optocouplers.
In linear amps mosfets have very little time , of order of few micro seconds to survive the dead short circuit.
Therefore it is necessary to activate the protection instantly.
A combination of Pre-Parallel delay would be the most viable solution.

regards,
kanwar
 
AX tech editor
Joined 2002
Paid Member
Re: Reply

amp_man_1 said:
[snip]In linear amps mosfets have very little time , of order of few micro seconds to survive the dead short circuit.
Therefore it is necessary to activate the protection instantly.
A combination of Pre-Parallel delay would be the most viable solution.

regards,
kanwar

Hi amp_man,

Why do you say that? I think because power fets have no secondary breakdown mechanism and can sustain HUGE currents for some time, they can only "break" because of thermal limits. (Assuming they are operated inside their voltage limit of course). Building up thermal load takes some time also. Therefore, they should be able to survive longer than bjt's.

Anyway, instant protection is impossible because you need time to detect a short that is not in reality a strong lf signal.

Jan Didden
 
Reply

Hi janneman,
Mosfets have no second breakdown voltage, but that doesn't mean that they are able to handle large short circuit current for long time. Their is certainly a time limit for this sustainance which has to be overcomed by good protection activation circuitry.

large current simply damages the N-Epitaxial layer of mosfets without considerable heating because of excessive carrier concentration on the die.

One thing more ,Do you know the fact that "Jannemann" is also known as "Life of Heart" in Indian language.

Regards,
kanwar
 
In linear amps mosfets have very little time , of order of few micro seconds to survive the dead short circuit.
you can have multiple steps of short circuit protection.
First level: Very high current at the output(dead short) circuit responce should be very fast, cut off should latch until reset by user.
Responses can tailored for different levels of output current.

At first switch ON of amplifier in live circumstances, a small high power resister can be connected in series with load.the drop can measured and compared,If within resonable limits then only the added power resistor, should be bypassed.Can be also used to protect against oped load also.
An ideal protection scheme will be .....will post later..after some thought.
 
AX tech editor
Joined 2002
Paid Member
Re: Reply

amp_man_1 said:
Hi janneman,
Mosfets have no second breakdown voltage, but that doesn't mean that they are able to handle large short circuit current for long time. Their is certainly a time limit for this sustainance which has to be overcomed by good protection activation circuitry.

large current simply damages the N-Epitaxial layer of mosfets without considerable heating because of excessive carrier concentration on the die.

One thing more ,Do you know the fact that "Jannemann" is also known as "Life of Heart" in Indian language.

Regards,
kanwar

Sure, but you talked about "microsecond" destruction speed. That's not realistic in audio, even with full short. It takes time even to build up current just by the inductance in the wires!
And even you cannot built protection circuits that work faster than sveral tens of milliseconds - try it, and if you think you can, post the circuit. Or is that a secret?

"life of heart"? interesting. I visited Rajasthan this spring, enjoyed it very much. The drive, the dynamic and ambition, India is really moving forward! But still a lot of poverty also of course, and I guess it will take time to disappear.

Jan Didden
 
amp_man_1 said:
The normal trip level of protection sensing was 28A peak, but the measurement shows 40A peak of destructive current was present without the protection circuit involved....
In linear amps mosfets have very little time , of order of few micro seconds to survive the dead short circuit. Therefore it is necessary to activate the protection instantly.....
Didn't you say somewhere that you amps could take normal PA environment... anyway something things confuses me.

You build extremely fast amps (200-500 V/us?), made on single sided boards so they will be easy to repair... because they got broken... and total breakdown in micro seconds... I can't really put these things together. This is not criticism, I just don't understand.

Let's say you have 4-10 pairs => 80-200 A CONTINUOUSLY and short circuit current of up to 1 kA and not in microsecond perspective. OK, the current value is lower in a AB-stage but 28-40 A you mention seems low. Datasheets mentions milliseconds when it comes to current at the same time as voltage across the transistor.

Maybe you trouble with your products has something to do with your extremely high slew rates and difficult loads. How does your current limitation look like now? You mentioned that your amps had full short circuit protection which I'll gather protects you to 100% if you have a short circuit but with this thread title in mind they might work properly after all.
 
Reply

hi Sivanand,

Your approach is good about multistage short circuit protection but I think it will not do much of good when it comes to dead short circuits.

Hi Jan-Didden,
Surely you are right about this way, but we have never tested the short circuit protection by shorting the outputs with long wires, rather we have used short [4 to 6 inches]and heavy wires to make the short circuit happen and measured the time taken to provide protection and it came out to be 75mS flat.
We simply use a series resistor and sense the current and then activates the muting of input signal for about 10 Seconds repititive, and nothing else.

Nice to hear that you have visited Rajasthan in spring season. Its a Great Phenomena.
###############
Now here comes the Moderator Senor
P E R A N D E R S

Hi mod,

Yes we had said earlier that our amps can take any PA environment.

We build amps with 200V/us [bridged]slewrate only.

Our protection circuits are reliable against all types of faults.

Please read the first post on this thread , which clearly illustrates that what type of protection is necessary if some one puts a mismatched load with passive crossover, rather than simply tripping because of current limit exceeding. In that situation it simply mutes the input and any other amp in this world will also do the same thing with the help of a relay or by any other means

Kindly eliminate your senseless confusions.

regards
kanwar
 
AX tech editor
Joined 2002
Paid Member
Re: Reply

amp_man_1 said:
[snip]Hi Jan-Didden,
Surely you are right about this way, but we have never tested the short circuit protection by shorting the outputs with long wires, rather we have used short [4 to 6 inches]and heavy wires to make the short circuit happen and measured the time taken to provide protection and it came out to be 75mS flat.
We simply use a series resistor and sense the current and then activates the muting of input signal for about 10 Seconds repititive, and nothing else.[snip]


Series resistor? Where?
What do you mean "measure the time taken to provide protection". What did you measure?

Jan Didden
 
Re: Reply

amp_man_1 said:
Our protection circuits are reliable against all types of faults.

Please read the first post on this thread , which clearly illustrates that what type of protection is necessary if some one puts a mismatched load with passive crossover, rather than simply tripping because of current limit exceeding.
... so all types of loads but not all....?

You don't use foldback current limitation which protects better to reactive loads.

What is a mismatched load according to you?
 
Status
Not open for further replies.