Amp Clipping Issue

[panteragstk]

Decreasing the gain on the amp while increasing the output level on the AVR did nothing. Still clips/peak limits at the exact same spot. I think lowering the gain on the amp while increasing the output level is canceling out what I'm trying to do. I still end up at "zero".

I'm going to try an analog signal straight from the computer into the amp and see what happens. That will tell me if it is on the preamp side or not.

 

[panteragstk]

It's the amp. I changed the output from the AVR to the internal PC soundcard and get the exact same result after level matching. The drivers don't seem to be stressed. I'm still getting 105db+ peaks from the scene I'm using to test. The one where the thing is coming out of the ground near the beginning. No other movie has been able to make this happen, so I guess I'll just have to live with it.

...I wonder what turning off the limiter will do

 

[PENG]

Decreasing the gain on the amp while increasing the output level on the AVR did nothing. Still clips/peak limits at the exact same spot. I think lowering the gain on the amp while increasing the output level is canceling out what I'm trying to do. I still end up at "zero".

I'm going to try an analog signal straight from the computer into the amp and see what happens. That will tell me if it is on the preamp side or not.

I was hoping you would do the opposite by decreasing the preout of the AVR and increasing the gain of the Crown to get the same output but eliminate the possibility that the AVR was clipping. In your last post somehow you seemed to have concluded that the Crown power amp is the culprit?

 

[panteragstk]

I was hoping you would do the opposite by decreasing the preout of the AVR and increasing the gain of the Crown to get the same output but eliminate the possibility that the AVR was clipping. In your last post somehow you seemed to have concluded that the Crown power amp is the culprit?

When I had the gain on the amp all the way up it clip/peak limited much easier than it does now. I arrived at my current setting after lowering the gain and increasing the output level of the AVR after the first time I had the issue. After totally eliminating the AVR from the equation the amp still went into protect mode. Not as quick, but it still happened. I turned the gain on the amp as low as I could to make sure it was level matched (LFE output on the soundcard was all the way up) and it peaked in the exact same spot. Gain all the way up on the amp had the same result as well.

I don't think there is anything necessarily wrong with the amp, but it doesn't like the test I'm throwing at it.

EDIT: Since the input impedance is higher with the balanced input, do you think using that may help? Would it make a difference at all?

 

[Irvrobinson]

How do you know this? I'm not saying you're wrong, but it's counter-intuitive based on how clipping indicators usually work. I'm not even sure how you could differentiate a clipped waveform from an arbitrary one. Most clipping indicators are based on either a fixed voltage threshold or a difference between an input and output waveform.

Please clue me in how this could happen.

I've wondered about this myself, and I found it by experimentation. I found that the front-end of my Velodyne sub is subject to easy overload on the balanced inputs, because Velodyne stupidly included level controls only on the single-ended inputs. At one time I was using the high-pass filters in the Velodyne to feed an ATI AT3000, and at high listening levels on one really well-recorded rock and roll CD I noticed the clipping lights being activated on the amp, which are worth about 450W/ch into my speaker's load. There was no way I was clipping the amp, so I thought, and I was right, because the addition of -10db attenuators in the balanced lines feeding the Velodyne cured the problem.

Without the attenuators I was also able to clip a loaner QSC GX5, which is even more powerful, and again the problem was cured with the attenuators in the circuit. Two amps behaving the same way? My guess is that when some of these line-level ICs clip in some products, like the Velodyne, they're losing control of the output voltage in some way, and they're inserting spikes or DC or who knows what in their output, and this is tripping protection circuits, which I'm guessing are also on the input driver stages. This is just a SWAG, but I can't explain the outcome of my experiments any other way. Can you?

 

[PENG]

EDIT: Since the input impedance is higher with the balanced input, do you think using that may help? Would it make a difference at all?

To use the balanced input you need balanced output from the preamp. Regardless, it won't help because your Crown clips when used beyond its limit, as simple as that..

 

[avengineer]

To use the balanced input you need balanced output from the preamp. Regardless, it won't help because your Crown clips when used beyond its limit, as simple as that..

Actually, to use the balanced input AND get the full advantage of the additional noise immunity, you need to feed it from a balanced output with a balanced cable. But you can use a balanced input with an unbalanced output, it still will pass audio just fine. The disadvantage may be lower sensitivity, but that depends on the input design. So in this situation, it would probably be of no advantage.

 

[avengineer]

When I had the gain on the amp all the way up it clip/peak limited much easier than it does now. I arrived at my current setting after lowering the gain and increasing the output level of the AVR after the first time I had the issue. After totally eliminating the AVR from the equation the amp still went into protect mode. Not as quick, but it still happened. I turned the gain on the amp as low as I could to make sure it was level matched (LFE output on the soundcard was all the way up) and it peaked in the exact same spot. Gain all the way up on the amp had the same result as well.

I don't think there is anything necessarily wrong with the amp, but it doesn't like the test I'm throwing at it.

Just to be clear, do you detect clipping/limiting by the front panel indicator, or some audible effect?

Is there any play level where you don't hear/see the clipping effect? Even very low?

EDIT: Since the input impedance is higher with the balanced input, do you think using that may help? Would it make a difference at all?

No, probably not.

 

[avengineer]

One more Q: What exactly are the subs?

 

[avengineer]

I've wondered about this myself, and I found it by experimentation. I found that the front-end of my Velodyne sub is subject to easy overload on the balanced inputs, because Velodyne stupidly included level controls only on the single-ended inputs. At one time I was using the high-pass filters in the Velodyne to feed an ATI AT3000, and at high listening levels on one really well-recorded rock and roll CD I noticed the clipping lights being activated on the amp, which are worth about 450W/ch into my speaker's load. There was no way I was clipping the amp, so I thought, and I was right, because the addition of -10db attenuators in the balanced lines feeding the Velodyne cured the problem.

Without the attenuators I was also able to clip a loaner QSC GX5, which is even more powerful, and again the problem was cured with the attenuators in the circuit. Two amps behaving the same way? My guess is that when some of these line-level ICs clip in some products, like the Velodyne, they're losing control of the output voltage in some way, and they're inserting spikes or DC or who knows what in their output, and this is tripping protection circuits, which I'm guessing are also on the input driver stages. This is just a SWAG, but I can't explain the outcome of my experiments any other way. Can you?

Very interesting. I'd have to look at the output of the Velodyne when it clips with a scope to know what's up. If it did something really nasty, like break into oscillation when it clips, that would possibly trigger a clip light on a power amp following it, even if it didn't actually clip itself. Several IC Opamps do bad things when clipped, others clip very cleanly. I'd look at that first.

Many clip detectors are looking for a difference between input and output wave form. Extreme high frequency content might not make it to the amp output, but if present on the input might just trip the light because it's different. But I would point at the Velodyne doing something strange, because generally the clip detectors in amps have no way of knowing if the input signal is pre-clipped or not. The situation is fairly abnormal, but interesting.

 

[PENG]

Actually, to use the balanced input AND get the full advantage of the additional noise immunity, you need to feed it from a balanced output with a balanced cable. But you can use a balanced input with an unbalanced output, it still will pass audio just fine. The disadvantage may be lower sensitivity, but that depends on the input design. So in this situation, it would probably be of no advantage.

That's why I said "regardless", instead of getting into technicalities. It would be like playing with the gains between the pre and the power amp, that he had already tried so I wasn't going to waste his time.

 

[panteragstk]

Just to be clear, do you detect clipping/limiting by the front panel indicator, or some audible effect?

Is there any play level where you don't hear/see the clipping effect? Even very low?

Yes, the front panel indicator is where I'm seeing it. At lower levels it never happens. It honestly never happens with anything other than WOTW so it's not a big problem, but I'd still like to solve it.

I just wonder if a new amp is the solution. The next thing I plan to try is wiring the subs to 8 ohms so the demand on the amp is lower. It may just be that the amp isn't quite strong enough for the 2 ohm load.

To use the balanced input you need balanced output from the preamp. Regardless, it won't help because your Crown clips when used beyond its limit, as simple as that..

I knew that, but I was clinging to some false hope.

One more Q: What exactly are the subs?

Infinity 1262w drivers in a custom sealed box.

 

[Irvrobinson]

Very interesting. I'd have to look at the output of the Velodyne when it clips with a scope to know what's up. If it did something really nasty, like break into oscillation when it clips, that would possibly trigger a clip light on a power amp following it, even if it didn't actually clip itself. Several IC Opamps do bad things when clipped, others clip very cleanly. I'd look at that first.

Many clip detectors are looking for a difference between input and output wave form. Extreme high frequency content might not make it to the amp output, but if present on the input might just trip the light because it's different. But I would point at the Velodyne doing something strange, because generally the clip detectors in amps have no way of knowing if the input signal is pre-clipped or not. The situation is fairly abnormal, but interesting.

I've thought about getting a scope now and then, but for the once a year or less I'd use it I haven't been able to push myself into acquisition mode.

I suspect poorly-designed circuits with very high gain ICs in them. The Velodyne is obviously one, I also wonder about those AVRs that clip at 1V, what sort of junk signals are they sending down the line at clipping?

 

[avengineer]

Infinity 1262w drivers in a custom sealed box.

Ah. Well, that explains a bit. A new amp that raises the clipping threshold would help, but only a little. You're at 775/ch now, if you went to 1500/ch (also into 2 ohms) you'd raise the clipping point by 3dB, which is just audible, but not exactly night/day. And at that point you'd be smacking the back of the voicecoils into the magnet as you push them past XMax. In fact, you could be doing that now.

Assuming you've run Audyssey on the Onkyo, and in your room it's tuning out a null or three, Audyssey can boost as much as 9dB. If your room/sub combo has a bit of weakness at, say 25Hz, and pushes you put 6 or 7 dB at that frequency, you could be clipping or over driving the sub, but only with material that has a concentration at that frequency.

You're pretty much at the practical limit of the sub design for LF SPL. I don't have quite enough data to run your design through simulation, but just using the published T/S parameters of the drivers in a 4cu ft sealed box, I get a -3dB frequency of about 38Hz. So at 20Hz you're already well below the cutoff, and will pay a huge penalty in maximum power handling and efficiency down there. If Audyssey tries to fix it, and it would, in combination with a room that doesn't want to do 20Hz either, you'll definitely be above the limit with some material.

Sorry, but the fix is to replace your sub and amp with two powered subs. The discussion of why more than one sub is sort of outside this thread, but look around, it's not just about SPL.

 

[markw]

Remember, just because a speaker is specced to be ale to survive X number of watts, it doesn't guarante it'll sound good while doing it.

 

[avengineer]

Remember, just because a speaker is specced to be ale to survive X number of watts, it doesn't guarante it'll sound good while doing it.

+1.

Subs need to be selected based on required acoustic performance, not watts alone. 105dB SPL @20Hz is not easy to do.

 

[jinjuku]

My initial impression is that the .1 track on the WOW is probably mastered pretty hot given what I have read. I have two challenging tracks for my dual opposed 1262w. WOW and The Incredibles. My fix is to add a 2nd sub and come off the signal by 3dB. That will give me the head room I need without running into the limits of a $140 sub array. While it does a hell of a job for the money, boy does it ever, it will still have it's limits.

 

[markw]

I've said many times that, in this hobby, if one looks hard enough for imprefections, one will never be disappointed.

 

[jinjuku]

The one thing I would like people to keep in mind when recommended the 1262w as a very affordable driver for DIY sub use is that they are best served in building up an array of them.

It's one thing to spend ~$140-50 on two drivers and some ply-wood. It's when you spend $1000 on drivers and ply-wood that things get REALLY interesting with this driver

 

[avengineer]

The one thing I would like people to keep in mind when recommended the 1262w as a very affordable driver for DIY sub use is that they are best served in building up an array of them.

It's one thing to spend ~$140-50 on two drivers and some ply-wood. It's when you spend $1000 on drivers and ply-wood that things get REALLY interesting with this driver

Sure, but with the -3dB point at 38Hz, you really need a lot of them and a ton of power to do 105@20Hz. There are other designs for that kind of cash that can do it with two subs, though.