update on clipping detection.

[thylantyr 10+ year member]

@thylantyr, the circuit uses the amps output terminals. the circuit you describe doesn't seem to actually detect clipping. It sound more like an automated DMM method.

You don't need to detect actual clipping, this is a toy not a precision instrument like a scope.

You are really interested in detecting maximum voltage because if you exceed it,

that is when clipping starts.

Example, you have a sine wave with 50v peak. Drive the amp higher past 50v

and you still have 50v, but the clipping begins. If you overdrive your amp more, you

still have 50v so you can safely say that 50v is your high limit as that is the clipping

threshold. During operation, you drive the amplifier and when you hit 50v,

you are right at the performance envelope of the amplifier which is good enough

and your LED blinks.

This idea is frequency independent and there is no adverse loading effects on the amplifier.

 

[thch 10+ year member]

@thylantyr, the circuit uses the amps output terminals. the circuit you describe doesn't seem to actually detect clipping. It sound more like an automated DMM method.
You don't need to detect actual clipping, this is a toy not a precision instrument like a scope.

You are really interested in detecting maximum voltage because if you exceed it,

that is when clipping starts.

Example, you have a sine wave with 50v peak. Drive the amp higher past 50v

and you still have 50v, but the clipping begins. If you overdrive your amp more, you

still have 50v so you can safely say that 50v is your high limit as that is the clipping

threshold. During operation, you drive the amplifier and when you hit 50v,

you are right at the performance envelope of the amplifier which is good enough

and your LED blinks.

This idea is frequency independent and there is no adverse loading effects on the amplifier.

actually, in my testing thus far it has worked just as well as an oscope. likely better then a handheld scope. paired with a scope and you can get a good deal of detail (if you are a purist). your method should work, but isn't really an Oscope replacement. This circuit is not supposed to be a toy, but rather a cheap replacement for an OScope for this one application.

 

[thylantyr 10+ year member]

Why would you say your method is an O-scope replacement and mine isn't?

I don't consider any circuit an O-scope replacement if you can't see the waveform,

these ideas we throw around are just toys for boys who can't afford the real

thing {O-scope}. It's pretty cool to make these things but it would be cooler to

have a frequency independent design so you can use it on fully active systems

and not just optimzed for subwoofer applications.

 

[Raven 10+ year member]

Why would you say your method is an O-scope replacement and mine isn't?I don't consider any circuit an O-scope replacement if you can't see the waveform,

these ideas we throw around are just toys for boys who can't afford the real

thing {O-scope}. It's pretty cool to make these things but it would be cooler to

have a frequency independent design so you can use it on fully active systems

and not just optimzed for subwoofer applications.

Take a DMM and go out and test the voltage of your amp before and after clipping. In my experience, the DMM will bug out and give huge volt readings at high clipping. Like, 48 volts on a 600 watt amp with no load.

If it were possible, just build a relay that is totally insenitive to AC power and put a battery and LED between it. Once DC power hits, relay turns on, battery powers LED, clipping detected. Nowhere near as exact as thch's or your method, but workable. I've been thinking about it for a while, but my abilities with these things is limited.

 

[theCybe 5,000+ posts]

You should speak with werewolf @ elitecaraudio, he's very into reverse-engineering, and can probably give you some valuable pointers.

I'd buy a finished unit, for sure.

 

[thylantyr 10+ year member]

You should speak with werewolf @ elitecaraudio, he's very into reverse-engineering, and can probably give you some valuable pointers.
I'd buy a finished unit, for sure.

Werewolf is cool, but he didn't understand my design idea on how to make

a circuit that mimicked Rockford Fosgates new 15kw monster amplifier. I'm sure you heard of it.

They use hybrid caps to store energy and dump the energy

on burps. The amp is not really a 15kw amp, it's more like a 2kw rms amp

with tons of capacitance to deliver 15kw peak, patent pending. I spent 5 minutes one night thinking about

how to mimick this type of specification and I drew a block diagram on paper.

I posted my idea on the carsound thread where they were debating with the RF amplifier

design about the merits of the amplifier and claims...nobody had comment about my circuit,

perhaps it hit a nerve... //content.invisioncic.com/y282845/emoticons/smile.gif.1ebc41e1811405b213edfc4622c41e27.gif It's really simple, an existing class H power supply with a 'watch dog' circuit

to enable high rail on demand only when the big cap bank is fully charged. Perhaps I watch too much

Star Trek //content.invisioncic.com/y282845/emoticons/frown.gif.a3531fa0534503350665a1e957861287.gif

 

[Raven 10+ year member]

It's really simple, an existing class H power supply with a 'watch dog' circuit to enable high rail on demand only when the big cap bank is fully charged. Perhaps I watch too much Star Trek //content.invisioncic.com/y282845/emoticons/frown.gif.a3531fa0534503350665a1e957861287.gif

Yes, you watch too much Star Trek. //content.invisioncic.com/y282845/emoticons/wink.gif.608e3ea05f1a9f98611af0861652f8fb.gif Your idea sounds plausable, but in no way would I put "class h" and "simple" in the same sentence, unless I have my classes mixed up. =)

 

[theCybe 5,000+ posts]

Werewolf is cool, but he didn't understand my design idea on how to makea circuit that mimicked Rockford Fosgates new 15kw monster amplifier. I'm sure you heard of it.

Yes, and while I'm surprised it hasn't been done before, (to this degree) I'm not really in love with the idea.

I posted my idea on the carsound thread where they were debating with the RF amplifierdesign about the merits of the amplifier and claims...nobody had comment about my circuit,

perhaps it hit a nerve... //content.invisioncic.com/y282845/emoticons/smile.gif.1ebc41e1811405b213edfc4622c41e27.gif

Undoubtedly //content.invisioncic.com/y282845/emoticons/smile.gif.1ebc41e1811405b213edfc4622c41e27.gif

It's really simple, an existing class H power supply with a 'watch dog' circuitto enable high rail on demand only when the big cap bank is fully charged.

I will hunt your post down later; Currently at work. //content.invisioncic.com/y282845/emoticons/wink.gif.608e3ea05f1a9f98611af0861652f8fb.gif

 

[thylantyr 10+ year member]

Take a DMM and go out and test the voltage of your amp before and after clipping. In my experience, the DMM will bug out and give huge volt readings at high clipping. Like, 48 volts on a 600 watt amp with no load.
If it were possible, just build a relay that is totally insenitive to AC power and put a battery and LED between it. Once DC power hits, relay turns on, battery powers LED, clipping detected. Nowhere near as exact as thch's or your method, but workable. I've been thinking about it for a while, but my abilities with these things is limited.

I was under the assumption that the detector could be a permanent install //content.invisioncic.com/y282845/emoticons/smile.gif.1ebc41e1811405b213edfc4622c41e27.gif

LED near dash to monitor the amp.

The relay idea is actually more complex than the comparator idea. Comparators

are simple, look here;

Op Amps as Comparators section

http://www.bcae1.com/opamp.htm

Both resistors, 1000 ohm form a voltage divider at the + input. This is your

reference voltage, any voltage you want really. Assume 12v battery as shown in

the picture, 6v is the reference based on those two 1000 ohm resistors.

The capacitor connects to the ' - ' input, the 8200 ohm resistors is not

connected to the battery until you 'press that button' on screen, so the

voltage at the ' - ' input is zero. When the cap charges and exceeds

reference {6v}, the LED is lit up.

To make a peak detector for this application, you don't want a capacitor

because it will slow down the reaction time. You add a potentiometer

to the + input to preset a reference, you scale the voltage divider so the

opamp can accept high voltage from the amplifier so it won't blow up,

and to be clever, add a zener diode to make your circuit immune to

9v battery voltage fluctuations. A few more things to add.

 

[thylantyr 10+ year member]

OT: RF15K amp

Yes, and while I'm surprised it hasn't been done before, (to this degree) I'm not really in love with the idea.

There is alot of comedy behind that amplifier and how they displayed it

to impress the minions. I for one think the showing is more of a Jedi mind trick.

Here's my simple take on it. They had a booth with 60? drivers driven by that

amplifier. People were amazed like deer in headlights. Problem is, what was

impressive? was it the amp itself, the matrix array of drivers or both ?

An array of woofers is very impressive no matter what amp you use, the power

of array coupling does the work and all you need is some good power to drive it.

RF thread;

http://www.carsound.com/UBB/ultimatebb.php?ubb=get_topic;f=18;t=010998

Here is my idea for cloning the idea as I posted on that thread.

Lets clone the idea; Here is my 5 minute brain

storm session..

 

coil_1 -> diode_1 -> cap_1 -> diode_2 -> switch_1 -> rail

 

coil_2 -> diode_3 -> cap_2 -> diode_4 -> resistor_1 -> rail

 

Cap_1 & cap_2 are isolated from each other

using the diodes.

 

Coil_1 is the high voltage coil, coil_2 is the

low voltage coil.

 

Power is applied to both transformers {coil_1 and coil_2) charge the

two caps, lets use 200v and 50v for demonstration.

 

The resistor_1 is a current sensing resistor

connected to a 'watch dog' circuit monitoring

current draw from the 50v supply. When little

or no current flows, that means there is no

music playing or a pause in the music so the

watch dog connects the switch. {I imply an

electrical switch, not mechanical )

 

With switch_1 = on, cap_1 rail voltage is present

on the output stage, 200v.

 

When a music transient hits the voltage drops

as cap_1 discharges ultimately the voltage sags

down to ~ 50v as the low supply takes over

and the watch dog senses current draw and disconnects the switch, allowing coil_1 to

be charging cap_1 when the low voltage supply

is working ... when the watch dog senses very

little current draw {pause in music}, it turns on

the switch again to force high voltage on the

rail. /repeat.

 

You can make this more complex by adding

watch_dog_2 to monitor if cap_1 is fully charged

and it signals watch_dog_1 whether or not to

proceed.

 

 

You can refine this idea.

... maybe we should start a new thread and clean this one up, it's way

off topic //content.invisioncic.com/y282845/emoticons/crazy.gif.c13912c32de98515d3142759a824dae7.gif

 

[thch 10+ year member]

@thylantyr, your idea is a DMM replacement, you don't detect clipping at all, but rather a voltage output that you concider to be clipped. the reason behind an OScope is to determing exactly what voltage output this would be, but OScopes are expensive. Handheld scopes often offer low resolution screens, making the onset of clipping a bit vague. this is really only important for things like HU preouts where you want to garuntee there is no clipping.

@Raven, as it would turn out, clipping doens't increase DC power. its a common misunderstanding, but it actually increases AC power. this is also why clipping for arbitary waves can be difficult to do spectrally -- you have no idea if something is part of the music or distortion.

@Thylantyr, the RF amp may be a bit off topic. possibly start a thread? also, i'm a bit confused by the circuit depiction, are all elements in series?

 

[thylantyr 10+ year member]

@thylantyr, your idea is a DMM replacement, you don't detect clipping at all, but rather a voltage output that you concider to be clipped.

For sake of argument. Lets say the sine wave peak is 50v. But at 50.1v the

signal clips. If you calibrate the detector using 50v as the reference then it's

true that the detector may trigger and say it's clipped, but in reality you need

0.1 more volts to be actually clipping. Point is, who cares about that kind of precision ?

You are essentially calibrating the detector to the amplifier's performance envelope

which is pretty good, but every method has flaws. If you change

the load, clipping will occur at a different voltage. Depending on the amplifier

design, clipping can be different based on frequency too. Even clipping indicators

found internally in amplifiers {like proaudio} are not precision instruments,

hence I call them toys //content.invisioncic.com/y282845/emoticons/crazy.gif.c13912c32de98515d3142759a824dae7.gif

You need a circuit that operates at full bandwidth and is load independent too.

You can mod your circuit so it's opamp based to remove loading effects, but how do

you make it full bandwidth?

@Thylantyr, the RF amp may be a bit off topic. possibly start a thread? also, i'm a bit confused by the circuit depiction, are all elements in series?

I'm too lazy to draw a diagram on the computer or start a thread.

The elements are in series, there is two sets of elements. This is essentially an

ordinary class H power supply with dual rail but you have an electric switch and watch dog circuits to figure out when to enable/disable the switch.

It's a 5 minute brainstorming session concept, what happens if we give the idea deeper thought? /lol/

What is funny is that the idea can be tested on an existing class H amplifier

with PS modification so you don't have to design the amplifier from scratch. Just

make the smart power supply. This would make a nice science project for

interested parties.

*or.. wait until RF gets the patent and take a peek* /cheating

 

[AcidicDreams 10+ year member]

some repliesAcidDreams, i actually had planned to build active versions of the detector and have pricing in the $30 area. but at this time i'm still working on a full wave, low noise voltage controlled amplifier. I know of one way to do a VCA, but i'm not sure it'll work. the problem is that the circuit must work over a wide range of inputs. from 2V to 200V (Preouts of HU to some of the high powered amps).

That would be great even though I have no idea what your talking about //content.invisioncic.com/y282845/emoticons/biggrin.gif.d71a5d36fcbab170f2364c9f2e3946cb.gif Good work though.... if you made $30 clipping detectors I'd be interested even though I have no idea how they work internally //content.invisioncic.com/y282845/emoticons/smile.gif.1ebc41e1811405b213edfc4622c41e27.gif

 

[thch 10+ year member]

@thylantyr, your argument is that a DMM can detect clipping as effectively as an OScope (keyword is _effectively_). some people, like SPL competitors or SQ competitiors, or even engineers may want to know information about the exact onset of clipping. your method has one flaw in that it assumes things about the amplifier. this can be bad in that an unregulated amp may be capable of added output if the supply voltage is raised. an SPL competitor would want to know this. conversly, in a system with an overrated amp the peak detection would never trigger because the user assumes the output can be higher then is possible. Your circuit is a valid way of doing things, and i'd suggest that adding a timer or other short-time memory circuit on the output would improve performance as short time detections would result in human viewable output. unfortunantly, your circuit is not as easy to use or as reliable -- you must tune the circuit for each amp, and even then results aren't garunteed. My goal was to create a more generic way to detect clipping with performance similar to an OScope, but without the associated costs.

@thylantyr, i think i figured out the circuit. i assume the caps are shunt devices and the switch is some type of MOS device

@AcidDreams, basically to convert the highest input into something useable, i must attenuate by a factor of up to 100. this means my circuit could get small inputs that are then made even smaller. my hope is to detect an input of 200V, and adjust the gain of the amp automatically. or get a small input singal and get a higher gain automatically. that way any input to the amp works and the circuit overall just works. if i can do this, I can later add a micoprocessor to do an FFT analysis and determine the distortion components in the signal, which is something that some OScopes might not do.

 

[thylantyr 10+ year member]

@thylantyr, your argument is that a DMM can detect clipping as effectively as an OScope (keyword is _effectively_).

My idea is not a DMM or DMM replacement. It's an ordinary peak detector.

some people, like SPL competitors or SQ competitiors, or even engineers may want to know information about the exact onset of clipping.

Use a scope then //content.invisioncic.com/y282845/emoticons/smile.gif.1ebc41e1811405b213edfc4622c41e27.gif People that are critical don't use these toy circuits,

they use real tools.

your method has one flaw in that it assumes things about the amplifier. this can be bad in that an unregulated amp may be capable of added output if the supply voltage is raised.

You calibrate the unit. A peak detector is a simple cheap circuit, comparator, some resistors, pot, battery. It's very basic. I haven't placed more thought on how

to improve it as I only spent 5 minutes thinking about it. When a variables changes, you must recalibrate, but it's a toy circuit, I'd rather use a scope to see waveforms.

I have no interest in making this either, I'm just throwing out a bone.

My goal was to create a more generic way to detect clipping with performance similar to an OScope, but without the associated costs.

But you need to make it frequency independent to make it a great tool that you

can sell commericially //content.invisioncic.com/y282845/emoticons/smile.gif.1ebc41e1811405b213edfc4622c41e27.gif Why stop, continue the effort to make it more uber.

@thylantyr, i think i figured out the circuit.

Do you think it will work? lol