More advanced Limiting.... questions

[The Chinese]

Ok, so I want to go to the next level in my theory and use of limiting...

 

There are many who believe in overpowering a driver in order to get maximum performance, and there are those who believe in powering a driver conservatively in order to prevent damage. Seems to me that controlling BOTH the peak power AND the RMS power going to a driver is key in getting the most out of one's system.

 

So I am curious to know how people do that kind of control, and how they set their limiters in their systems- Do you let the Amps peak limiters control peaks, and set your system limiters to control the RMS values (obviously by slowing attack & release times, Ratio, etc.) How do you measure your crest factor?

 

Thanks in advance,

 

Todd A.

[Unalaska]

In the 2 amp racks that have a DRPA the limiter is set around where the amp limiters kick in. The amps are a little small for what the speakers can take and I leave the amp limiters on (QSC PLX/CX and Yamaha P7000's). Redundant I'm sure and there prolly is a better way, which someone here will help explain.

[dennis a]

I have a calculator that is a .xls file. It calculates several things and one of them is limiter settings. By entering a little information about your amp, you can calculate a limiter setting that would, let's say, limit the output of a 1000W/Ch amp to 850W/Ch. I can't attach a .xls file on this forum...if your interested, I would need an email address.

 

Dennis

[agedhorse]

There are several aspects to limiting that are important, mpost importantly are those of the speakers being used. Limiting will be approached differently on different frequency bands too.

 

For subwoofer drivers, the damage is most often mechanical in nature unless the program material is highly compresed or limited. That means that PROVIDED YOU HAVE ACCEPTABLE BACKUP LIMITING AND HPF, for short periods of time a typical sub driver can safely handle 2x the RMS rating. It doesn't necessicarily mean that you will get 3dB more SPL than powering at the RMS level (maybe 1.5dB more) but it won't destruct immediately. Now, at you look into the amount of time that you are driving this speaker, if the system is being pushed so that you are hitting this mechanical limit regularly, it's prudent to begin to limit this level back to closer to the RMS rating of the speaker. Depending on the driver, I like to start this limiting after about 80-120mSec and have the limit rate (attack time) be relatively long (say 50-100mSec) with a relatively long release time (200-400mSec) which results in a rather benign effect. The idea is to limit the magnitude of the pounding when things get pushed to what is no longer realistic for the driver to handle in the long term. As limiting occurs, we want to back down to close to the RMS level because you are now altering the peak to average ration profile of the program material and shifting mechanical energy to the thermal limits.

 

For lows and low-mids, rarely is the problem due to mechanical issues, so we must pay closer attention to the thermal limits and again the 2x RMS is fine but the limiter parameters will need to change a little bit... the holdoff will be in the neighborhood of 25-75mSec, the attack will be around 30-75mSec and the release time will be around 100-150 msec. HPF is provided by the crossover of course. I will limit back to the RMS rating or maybe even a little less.

 

For the high frequency driver, the damage is almost purely thermal provided a proper xover frequency is selected and the driver is suitably matched to the horn. The horn factors into the xover frequency since it provides mechanical loading for the driver. I will typically use between 2x and 3x the RMs rating while using approx. 5-10mSec of holdoff, 10-20mSec of attack time and between 30 and 75 mSec of release time. I will typically limit back to roughly 50-70% of the RMS rating of the driver. Note that for many drivers, this is only 35-50 watts.

 

I always verify that what I program is what comes out at the end. It's very easy to make an error and if somebody is going to drive their speakers hard, this is the one place where you need to get it right.

 

This is the kind of stuff that powered speaker designers put into their speaker processing, and it's not uncommon for the subwoofer HPF to slide up a little bit as limiting occurs also.

[The Chinese]

Interesting! Ok, So at what level processor do I start seeing holdoff time? My Ashly stuff only has attack and release, and almost all the other processing gear I have owned doesn't have that feature. Can you give me some examples of gear that has that feature - This is one of the missing links that I've interested in learning more about.

 

@dennis,

 

I have that particular calculator, thanks for offering though. Appreciate it.

[dboomer]

 

There are many who believe in overpowering a driver in order to get maximum performance ...

 

 

NO, no, no ... you can't overpower a driver and expect it to function. That's the point - delivering the power.

 

Now you can use an amp that is rated for more power than a driver can handle but you can't actually deliver more than rated power without problems.

 

Why are these two things different? It's because using music as a source you will never drive the amp to full power (assuming you can actually supply wall power to do this - which is a separate discussion.

 

The first thing about limiters is that they do not guarantee you won't blow a driver, they just give you a real good chance that you won't (assuming correct setup). That's the next problem ... there is not one and only one way to set them up. it's a bunch of tradeoffs that you are willing to accept.

 

So using both peak and average protection is better that either alone. Usually peaks do not present a real problem (depending how high the peaks can be) so getting a limiter setting that you are happy with to control average power is step one.

 

Measuring crest factor is just measuring the instantaneous peaks compared to the average level and expressing it in dB. Easy with a sine wave ... but real live music is all over the place.

[agedhorse]

DSP based processing is kind of the go to device for these applications, matched to their amplifiers.

 

Even the DriveRacks have a holdoff feature, I think they call it overshoot or something like that. Sometimes it's disguised in the attack window,and sometimes you just need to calculate a compromise attack time that's about 1.5-1.75x the holdoff time.

 

Also, note that there are 2 limit points, the fast mechanical limiter that I generally choose to be the limiter in the amp because that looks for loss of (internal) feedback of the amp which is a prime indicator of clipping independant of power supply voltage. Then there's the thermal limiter that is secondary. Of course you must choose your amp size based on your desired mechanical limits of between 1.5x and 2x the RMS rating of the speaker. Before some of you guys jump up and down saying "I told you so", this is not a problem provided the backup limiting is precicely set. The variables really do depend on the drivers being used, I gave some rule of thumb generalizations.

[agedhorse]

I agree with Don BTW, that this is all about managing compromises and tradeoffs, and peak to average ratios are all over the place with different music types and styles.

[The Chinese]

 

NO, no, no ... you can't overpower a driver and expect it to function. That's the point - delivering the power. Now you can use an amp that is rated for more power than a driver can handle but you can't actually deliver more than rated power without problems. Why are these two things different? It's because using music as a source you will never drive the amp to full power (assuming you can actually supply wall power to do this - which is a separate discussion. The first thing about limiters is that they do not guarantee you won't blow a driver, they just give you a real good chance that you won't (assuming correct setup). That's the next problem ... there is not one and only one way to set them up. it's a bunch of tradeoffs that you are willing to accept. So using both peak and average protection is better that either alone. Usually peaks do not present a real problem (depending how high the peaks can be) so getting a limiter setting that you are happy with to control average power is step one. Measuring crest factor is just measuring the instantaneous peaks compared to the average level and expressing it in dB. Easy with a sine wave ... but real live music is all over the place.

 

 

 

Well, Of course not! But that's the cruz of the discussion: How to properly protect one's gear while maximizing performance, from a scientific and practical standpoint. I also recognize that measuring Crest factor is a very difficult thing to do- so IMO, the generalizations I read/Hear about using an Amp 2x the RMS rating don't really fly with me; evidence is that I power my Rig At RMS level. This is a theoretical thread inthe sense that I'm not looking to replace my Amps. I really just want to have a killer grasp of what is going on in the world of limiting.

 

T

[agedhorse]

And note my disclaimers and words of caution as you go beyond RMS.

 

I say this as I plow through a bunch of recones that I am working on. Every one of them due to too much power. I'm getting too old for this {censored}.

[The Chinese]

 

DSP based processing is kind of the go to device for these applications, matched to their amplifiers. Even the DriveRacks have a holdoff feature, I think they call it overshoot or something like that. Sometimes it's disguised in the attack window,and sometimes you just need to calculate a compromise attack time that's about 1.5-1.75x the holdoff time. Also, note that there are 2 limit points, the fast mechanical limiter that I generally choose to be the limiter in the amp because that looks for loss of (internal) feedback of the amp which is a prime indicator of clipping independant of power supply voltage. Then there's the thermal limiter that is secondary. Of course you must choose your amp size based on your desired mechanical limits of between 1.5x and 2x the RMS rating of the speaker. Before some of you guys jump up and down saying "I told you so", this is not a problem provided the backup limiting is precicely set. The variables really do depend on the drivers being used, I gave some rule of thumb generalizations.

 

 

Are you using the terms "Secondary limiting" and "backup limiting" interchangeably? I use QSC PLX, RMX, and PL Amps. All have the peak limiters engaged. All my Amps are matched to the RMS ratings of my Speakers; so I Generally don't worry about Mechanical Limits. i Do worry about thermal Issues; the EAW LA325's have some history of having crossover failure from overpowering.

 

So in my system, one method you might propose is to set the Limiter attack time to 1.5-1.75x the standoff time. (ashly 3.24 does not have standoff time).

 

Now, what you're telling me is that If I'm hitting the "Clip" lights on the QSC's, I'm actually hitting the Limiters in the Amp. At what point am I still good on that? Let's assume I have my Ashly limiters set correctly w/ proper gain structure. Just let 'em hit the red, so long as they don't stay lit? Or am I basically just wasting my time unless I have amp that are more powerful?

 

T

[Unalaska]

So Andy, how much is it to recone a 2258 these days?

[agedhorse]

Todd... if you are powering at the RMS level and using the limiters in the amps, you are good to go. You are potentially sacraficing maybe 2dB on maximum SPL but your worries are less.

 

The terms that I used are perhaps not a great description... maybe primary limiting and secondary limiting would be better, with the primary limiting beig done by the amp at whatever level above the RMS rating that you choose to apply. The secondary limiting is the thermal protection limiting that limits back to the RMS rating of the speaker.

 

Now with the high frequency sections, you will probably need to use external limiters for both functions since it's likely that the amp will be many times the RMS rating of the driver, so processing with 2 limiting stages will be necessary. The DriveRacks will allow this as will many others.

[agedhorse]

 

So Andy, how much is it to recone a 2258 these days?

 

 

Middle-high $200's.

[malcatron]

I use a PV1500 (500w ch) to power Peavey UL115 subs (500w RMS). I engage all protection on the amp (limiters & 40Hz hpf) - is it OK for the amps clip/limiter lights to come on occasionally at the biggest peaks or should I aim to keep them from coming on at all ?

[The Chinese]

 

Todd... if you are powering at the RMS level and using the limiters in the amps, you are good to go. You are potentially sacraficing maybe 2dB on maximum SPL but your worries are less. The terms that I used are perhaps not a great description... maybe primary limiting and secondary limiting would be better, with the primary limiting beig done by the amp at whatever level above the RMS rating that you choose to apply. The secondary limiting is the thermal protection limiting that limits back to the RMS rating of the speaker. Now with the high frequency sections, you will probably need to use external limiters for both functions since it's likely that the amp will be many times the RMS rating of the driver, so processing with 2 limiting stages will be necessary. The DriveRacks will allow this as will many others.

 

 

Yes, this is the gap in my knowledge i seek - The Secondary, thermal limiting.

 

Although I gather I'm very safe since I power at RMS, I've been asked to think about doing some larger festivals in the future (I got some really good feedback from the NYE party I did sound for). So There's potential for either new gear or rental gear to expand my current arsenal.

 

Hmm 2 limiting stage in a Driverack? Hmmm. Well certianly not in an Ashly 3.24c, whcih is what I am using. However, I do Have 8 Ch worth of Ashly 4.24G, which is a 28 Band GEQ w/ HPF,LPF, Delay & Limiting.

 

So using my gear, typically I setup my master Feed into 2 Ch of the 4.24, for FOH EQ. I then run that into the 3.24c For Crossover, Limiter, Delay. So in theory I actually have 3 limiting stage - 4.24G, 3.24C, and Amp. So I could set the 4.24 to be a thermal or secondary limiter, with the 3.24 as the primary peak limiter, and have the amp as backup. Or would you change that order?

 

I also Y-Cable the master feed to 2 more channels of the 4.24 for my satellite mains, so I can EQ and time them into sync with my 325's.

Hmmm...

 

I am getting warm and fuzzy as all this is startign to coalesce in my mind...

 

 

T

[The Chinese]

 

I use a PV1500 (500w ch) to power Peavey UL115 subs (500w RMS). I engage all protection on the amp (limiters & 40Hz hpf) - is it OK for the amps clip/limiter lights to come on occasionally at the biggest peaks or should I aim to keep them from coming on at all ?

 

 

At 500RMS power to a 500RMS driver, you shoudl be just fine with occasional clip lights on.

 

T

[dboomer]

 

At 500RMS power to a 500RMS driver, you shoudl be just fine with occasional clip lights on. T

 

 

Of course you should ... since you are likely only DELIVERING about 50W to the driver. Even with highly compressed bass you are likely to only be delivering 150W of average power.

 

Here's where we are not communicating. A 500W continuous rated power amp never delivers 500W continuous to a speaker in normal use ... during catastrophic disaster maybe.

[agedhorse]

 

Of course you should ... since you are likely only DELIVERING about 50W to the driver. Even with highly compressed bass you are likely to only be delivering 150W of average power.

 

Conversely, by the time you size an amp to approach the speaker's thermal limit, you will typically be well beyond the mechanical limits. Generally for LF drivers, the mechanical limits dominate.

[dboomer]

 

the mechanical limits dominate.

 

 

Totally in agreement. Proper HPF setting on a sub will save you from 10x the problems that "overpowering" will cause.

[malcatron]

Thanks guys, sorry for lowering the level of the conversation

[The Chinese]

 

Here's where we are not communicating. A 500W continuous rated power amp never delivers 500W continuous to a speaker in normal use ... during catastrophic disaster maybe.

 

 

Don, I understand this- why are you jumping on me? I didn't mean to upset you or anything, my apologies if I did...

 

T

[agedhorse]

I would also suggest that the peak to average ratio of much modern music on subs can be aslow as 6dB, or 1/4-power, and as temperature increases, mechanical power handling capability does decrease as the VC winding adhesives begin to weaken.

[dboomer]

 

- why are you jumping on me?

 

 

My intention was not to jump on you but only to point out the technical reason. Your statement, while correct, I thought painted the wrong picture. So while your statement doesn't rule it out it is perfectly acceptable to use 1000W - 2000W amp with that same speaker.

 

I thin most readers will think ... OK 500W amp with a 500W speaker, gee 500=500, I'm ok. But the real reason they are OK is because they will only ever deliver 10 to 30% of the amp's power to the speaker.

 

So again, if you know what you are doing and have a sense for the dynamic range of your material and system that 1000-2000W amp will still very likely deliver power to the speaker that is well below the speaker's rating.

 

Are you more likely to suffer damage with a larger amp than a smaller amp ... YES, duh! But the damage while delivered by the amp is likely not initiated by the amp, but more likely by something stupid happening in front of the amp. So don't shoot the messenger:thu:

 

Amplifier's power ratings have NOTHING to do with music while speaker power rating DO (although there is still plenty of debate about it). Looking at the numbers together is not an apples to apples situation.

 

Thank you ... I feel sooo much better now!

[The Chinese]

Ok, Cool man. It's hard to get the gist of meaning/ intent on the internet, I dig what you're sayin'. and I agree.

 

Peace, My Brother in Sound!

 

T