Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Nathan Riddle]

I know we seem to talk about limiters every few months. But I don't believe we've spoken about this.

I am NOT talking about RMS voltage to peak conversions (1.414 multiplier). [Though I imagine it could be this easy for some 'under the DSP's hood calculations]

Specifically, I have an amp (QSC PLD 4.5) that has two limiters:

RMS
3-100v
attack: 1ms - 9.951s
release: 10ms - 59.01s

Peak
5-155
attack: 0.1ms - 20ms
release: 1ms - 1000ms

---

The RMS is used as a thermal limiter (~1/4 power) 40v 3s attack & release.
The peak is wasted as a peak limiter as the amp can't send more voltage than the woofer can handle (given a properly set HPF) it can't physically pop the cone. Thus it would be best used as a RMS limiter of some sort (if it's needed at all).

[David Sturzenbecher]

I know we seem to talk about limiters every few months. But I don't believe we've spoken about this.

I am NOT talking about RMS voltage to peak conversions (1.414 multiplier). [Though I imagine it could be this easy for some 'under the DSP's hood calculations]

Specifically, I have an amp (QSC PLD 4.5) that has two limiters:

RMS
3-100v
attack: 1ms - 9.951s
release: 10ms - 59.01s

Peak
5-155
attack: 0.1ms - 20ms
release: 1ms - 1000ms

---

The RMS is used as a thermal limiter (~1/4 power) 40v 3s attack & release.
The peak is wasted as a peak limiter as the amp can't send more voltage than the woofer can handle (given a properly set HPF) it can't physically pop the cone. Thus it would be best used as a RMS limiter of some sort (if it's needed at all).

Is there a question?

[Nathan Riddle]

Is there a question?

In the title?

Obviously, there are assumptions we would take on the DSP's part. [is the peak limiter measuring peak or RMS, is the RMS (thermal) limiter measuring RMS or some other value, eg. Vavg (under the curve)?]

Can we come up with a method of determining what type of limiting algorithm it is using for peak or RMS?

What is typical?

[David Sturzenbecher]

In the title?

Obviously, there are assumptions we would take on the DSP's part. [is the peak limiter measuring peak or RMS, is the RMS (thermal) limiter measuring RMS or some other value, eg. Vavg (under the curve)?]

Can we come up with a method of determining what type of limiting algorithm it is using for peak or RMS?

What is typical?

I have done extensive measurements of amplifier limiters in order to copy presets from on amp to another.  ITechHD, Lab Gruppen D series, Powersoft X, and QSC CXD-Q are the ones I have tested. There is by far no hard and fast rules for how any of them behave. My method involved hours, upon hours of time in front of an oscilloscope in “roll” mode in order to capture the actual peak voltage values and time constants.  Any ole DMM should allow you to set the RMS values, but peak takes some work.

The QSC CXD-Q doesn’t allow you to enter values at all, so nothing I measured would carry over to your amps.   The CXD-Q settings want voice coil size, power rating and if it’s a compression driver or not. I was thoroughly disappointed in the results of the CXD-Q though.  There were no provisions to let through any transient peaks. Everything was brickwalled to the RMS calculated voltage... with a very long term limiter coming in at 10-30 seconds.  Also, don’t even get me started in how wrong the limiter settings are when you put those amps in bridge mode. I had to tell the amp to limit to a “100watt driver” to get a 700W equivalent limitation.   Had I blindly typed in values and not measured, it would have been a service nightmare.


Sent from my iPhone using Tapatalk Pro

[Ivan Beaver]

I have done extensive measurements of amplifier limiters in order to copy presets from on amp to another.  ITechHD, Lab Gruppen D series, Powersoft X, and QSC CXD-Q are the ones I have tested. There is by far no hard and fast rules for how any of them behave. My method involved hours, upon hours of time in front of an oscilloscope in “roll” mode in order to capture the actual peak voltage values and time constants.  Any ole DMM should allow you to set the RMS values, but peak takes some work.

The QSC CXD-Q doesn’t allow you to enter values at all, so nothing I measured would carry over to your amps.   The CXD-Q settings want voice coil size, power rating and if it’s a compression driver or not. I was thoroughly disappointed in the results of the CXD-Q though.  There were no provisions to let through any transient peaks. Everything was brickwalled to the RMS calculated voltage... with a very long term limiter coming in at 10-30 seconds.  Also, don’t even get me started in how wrong the limiter settings are when you put those amps in bridge mode. I had to tell the amp to limit to a “100watt driver” to get a 700W equivalent limitation.   Had I blindly typed in values and not measured, it would have been a service nightmare.


Sent from my iPhone using Tapatalk Pro

This is yet another reason to MEASURE and DO NOT assume the "simple numbers" are what they think they are.

There is one MAJOR manufacturer whos "thermal limiter" does not make any sense to me.

To me, a limiter should start to act whenever the threshold is exceeded.  Not hard limit to that value quickly, but rather "bring the level" down to that level over the time period specified.  If you just barely exceed the level, then it is a very slow attenuation.  Exceed it a whole lot and it is a faster attenuation.

In this case, the thermal limiter does NOTHING, until the attack time has passed, and THEN is starts to reduce, at the time period specified.

What is really strange, is that the manufacturer recommends (in several places) for subs to be around 3 seconds of thermal, but yet HF drivers be 10 Seconds of thermal.

That is AN ETERNITY for HF drivers.  That provides NO protection what so ever to the HF driver.

As much as we "want" the digital numbers to all be the same, very often they are VERY far off, not even close, unless you count a ratio of 2:1 to be "close".  Not me.

It amazes me the "blind assumptions" people take as gospel these days.

Just because somebody says something, many take it as the truth, yet do nothing to try to figure out if it is actually fact or not.

-3dB, -27dB-  Hey close enough to "believe" -right?

I better stop now--------------

[Nathan Riddle]

Thanks, guys.

David, seems you're the one to talk to about learning to test DSP (amidst others, Ivan/Langston)

I knew everything in DSP can be/is different from previous posts from Ivan/others. But I was hoping there might be some uniformity in the general application of a peak limiter vs RMS limiter. Seems as though there isn't :/

Next on the chopping block would be walking me through testing one. I know how to find thermal limiters well enough via DMM's. But I'd like to learn more advanced techniques of testing such as the thorough reports by Langston on the interfaces or from others on amps/dsp.

https://forums.prosoundweb.com/index.php/topic,154105.msg1553151.html#msg1553151
http://www.bennettprescott.com/downloads/DSP_Differences.pdf
https://www.poweraudio.ro/diy/albums/userpics/10001/BENCH_COMPARISON_TEST.pdf

---

I think I'd need some tools/test equipment,
I have a BK Precision 1474, but it's too old to be of any real use.
I'm looking into budget o'scopes (Siglent Technologies SDS1202X-E) to help facilitate this.

[Peter Morris]

I know we seem to talk about limiters every few months. But I don't believe we've spoken about this.

I am NOT talking about RMS voltage to peak conversions (1.414 multiplier). [Though I imagine it could be this easy for some 'under the DSP's hood calculations]

Specifically, I have an amp (QSC PLD 4.5) that has two limiters:

RMS
3-100v
attack: 1ms - 9.951s
release: 10ms - 59.01s

Peak
5-155
attack: 0.1ms - 20ms
release: 1ms - 1000ms

---

The RMS is used as a thermal limiter (~1/4 power) 40v 3s attack & release.
The peak is wasted as a peak limiter as the amp can't send more voltage than the woofer can handle (given a properly set HPF) it can't physically pop the cone. Thus it would be best used as a RMS limiter of some sort (if it's needed at all).

There are 3 things that you need to protect - Xmax, thermal and mechanical stress limits.

The RMS limiter is usually used to stop the voice coil being overheated. The peak limiter is usually set to protect a mechanical limit. i.e. from a dangerous transient that could rip a cone, shatter a diaphragm, damage the voice coil, or result in early fatigue failures. You can also set it just to stop the amp clipping.

If you get these correct and subject to the speaker design they can also protect Xmax, but there are other more sophisticated ways to model and protect Xmax ... and thermal limits for that matter.

How these are set will depend on the characteristics of the device you are trying to protect, what its peak/transient limit is compared to long term thermal limit. 

[Nathan Riddle]

There are 3 things that you need to protect - Xmax, thermal and mechanical stress limits.

The RMS limiter is usually used to stop the voice coil being overheated. The peak limiter is usually set to protect a mechanical limit. i.e. from a dangerous transient that could rip a cone, shatter a diaphragm, damage the voice coil, or result in early fatigue failures. You can also set it just to stop the amp clipping.

If you get these correct and subject to the speaker design they can also protect Xmax, but there are other more sophisticated ways to model and protect Xmax ... and thermal limits for that matter.

How these are set will depend on the characteristics of the device you are trying to protect, what its peak/transient limit is compared to long term thermal limit.

I think the terminology gets butchered a bit between manufactureres, publications, forums, etc. I'm trying to see through the veil 

I was trying to state that with the HPF set according to manufacture recommendation over excursion would be taken care of.

Taken from Nick here:
https://soundforums.net/community/threads/help-me-understand-system-limiting-setting-it.1139/page-2

There's also talk of a difference between a true RMS limiter and a thermal limiter.

Perhaps there is no need for an additional rms/peak limiter when using a thermal limiter and the amp can't send peak voltage?

[Mark Wilkinson]

Specifically, I have an amp (QSC PLD 4.5) that has two limiters:

RMS
3-100v
attack: 1ms - 9.951s
release: 10ms - 59.01s

Peak
5-155
attack: 0.1ms - 20ms
release: 1ms - 1000ms

---

Nathan,  I've looked at the PLD4.5's limiters on a scope, and AFAICT both the rms and peak work exactly the same, other than the different ranges for attack and release.  They seem to simply compress the signal, in hardwall fashion base on attack and release.

But I should add that I'm not all that experienced with a scope....
The only thing I know how to do, is look and see if the waveform is clipped or compressed, after the limiter is fully engaged.
I do not yet know how to look at the time it takes to engage or release a limiter, and whether there is any voltage ramp to on or off.

David, what do you mean by oscope "roll mode" ?  Any tips on how to look at attack and release?

Peter, can you describe the more sophisticated ways to model and protect xmax, ..and thermal ?
I know how to use hornResp to model displacement with HFP in place, but that's about it.....

I gotta keep learning here....despite it all, I still managed to lunch a sub driver last week during testing 

[Nathan Riddle]

Nathan,  I've looked at the PLD4.5's limiters on a scope, and AFAICT both the rms and peak work exactly the same, other than the different ranges for attack and release.  They seem to simply compress the signal, in hardwall fashion base on attack and release.

But I should add that I'm not all that experienced with a scope....
The only thing I know how to do, is look and see if the waveform is clipped or compressed, after the limiter is fully engaged.
I do not yet know how to look at the time it takes to engage or release a limiter, and whether there is any voltage ramp to on or off.

David, what do you mean by oscope "roll mode" ?  Any tips on how to look at attack and release?

Peter, can you describe the more sophisticated ways to model and protect xmax, ..and thermal ?
I know how to use hornResp to model displacement with HFP in place, but that's about it.....

Do you recall if the peak works on peak or rms voltage? [I know, I know... measure; I don't have one to test and I'm creating presets for a friend]

I think he means single where it waits for a signal to cross the trigger threshold and measures from that point forward?

I gotta keep learning here....despite it all, I still managed to lunch a sub driver last week during testing 

While you might be bummed (sorry), I busted out laughing and thoroughly enjoy the candor 

[ProSoundWeb Community]