AVS › AVS Forum › Audio › DIY Speakers and Subs › Poll: How do YOU set your gain structure????
New Posts  All Forums:Forum Nav:

Poll: How do YOU set your gain structure????

Poll Results: Do you max your SUB amp's gain out to start or make it your last adjustment in gain structure?

 
  • 17% (5)
    I max my amp's gain then go from there
  • 44% (13)
    I start at the AVR and go down the chain
  • 34% (10)
    I haven't done proper gain structuring on my system yet
  • 3% (1)
    I don't know what gain structure means
29 Total Votes  
post #1 of 90
2/17/13 at 11:58am
Thread Starter 
I see a lot of questions rolling around about how to properly gain structure. With that said, many of us are not running the "standard" type systems so I encourage all that can, please post your answer to the poll and then explain your best approach to how YOU set your gain structure. There are many articles that I have read including Wayne's over at HTS, and they have proved invaluable in making the most of my system, however, each and everyone one of us I bet have some small tweaks we have done to maximize our systems to their fullest potential in our individual theaters. Let's hear what you all have to say!!!
Reply
Reply

AVS Top Picks

post #2 of 90
2/17/13 at 2:38pm
The thing I have noticed about gains structure is; there are a lot of different variables out there. I really don't think there is a one size fits all approach.

As I have talked with others on here when I setup mine; in my case it is physically impossible to achieve a gain structure due to overkill in some areas. I was almost at a point where I couldn't back down the gains and sub out any further to match speaker levels.

I do believe that everyone should know the limitations of there system however. Even though I didn't accomplish a gain structure, I did learn some valuable information about each piece in regards to output in relation to clipping. If nothing else, it gives me piece of mind knowing the "clean" capability of my system and how far I can push it before something bad happens. smile.gif
Reply
Reply
post #3 of 90
2/17/13 at 3:16pm
This is something that has greatly confused me since moving to pro amps for my subs.

System architect on my Crown amps are telling me that I am clipping the amps input at -20 volume, sub level -12.5.

To me that doesn't make any sense. Since i don't have access to an oscilloscope, it's hard for me to figure out what is going on. To make matters worse, I have read so many different threads and like those above me have stated, each person's setup is different.

Before I make my next DIY project, I think I need to really figure this out, because it is damn confusing to me. mad.gif
Reply
Reply
post #4 of 90
2/17/13 at 3:20pm
Quote:
Originally Posted by Donald2B View Post

This is something that has greatly confused me since moving to pro amps for my subs.

System architect on my Crown amps are telling me that I am clipping the amps input at -20 volume, sub level -12.5.

To me that doesn't make any sense. Since i don't have access to an oscilloscope, it's hard for me to figure out what is going on. To make matters worse, I have read so many different threads and like those above me have stated, each person's setup is different.

Before I make my next DIY project, I think I need to really figure this out, because it is damn confusing to me. mad.gif

That doesn't sound right. I am running a 3009 which has some of the hottest pre outs in the business, and I'm nowhere near clipping at that level. I would recommend putting in a 0dbs 60hz tone, turning the volume up to reference, and putting a DMM on your sub out to see how much voltage is being output.

I'm not familiar with the Crowns, but maybe they are very sensitive to input voltage. How far up do you have the gains on them set?
Reply
Reply
post #5 of 90
2/17/13 at 3:25pm
Quote:
Originally Posted by bass addict View Post

That doesn't sound right. I am running a 3009 which has some of the hottest pre outs in the business, and I'm nowhere near clipping at that level. I would recommend putting in a 0dbs 60hz tone, turning the volume up to reference, and putting a DMM on your sub out to see how much voltage is being output.

I'm not familiar with the Crowns, but maybe they are very sensitive to input voltage. How far up do you have the gains on them set?

I know! I have the crown dials turned up all the way. It doesn't really matter since it doesn't seem to affect the input side of the graph in System Architect. It's the input (Left) side that is clipping. When I turn the dials all the way down, the input side doesn't change. frown.gif

I'm running an Integra 9.8 XLR to Crown XTi 2002 amps. All the other channels are running RCA to a Fosgate Audionics 5 Channel amp.

How should I setup my meter to measure the voltage? I have one, just don't use it too often.
Reply
Reply
post #6 of 90
2/17/13 at 3:41pm
I like to allow for plenty of headroom in all components.
I typically set my AVR to 0 to +3 and then set the amp gains to half followed by the EQ gain to balance the levels to the desired curve.
Reply
Reply
post #7 of 90
2/17/13 at 3:46pm
Quote:
Originally Posted by jpmst3 View Post

I like to allow for plenty of headroom in all components.
I typically set my AVR to 0 to +3 and then set the amp gains to half followed by the EQ gain to balance the levels to the desired curve.

How do you know how much headroom you have though setting it up that way?
Reply
Reply
post #8 of 90
2/17/13 at 3:49pm
Quote:
Originally Posted by bass addict View Post

How do you know how much headroom you have though setting it up that way?

Well, I guess I don't know exactly.
But, I know that nothing in the chain is maxed out which can introduce noise, etc.. I run the EQ and the AVR as far away from maxing the gain as possible.
Edited by jpmst3 - 2/18/13 at 3:58pm
Reply
Reply
post #9 of 90
2/17/13 at 4:16pm
Thread Starter 
Quote:
Originally Posted by jpmst3 View Post

I like to allow for plenty of headroom in all components.
I typically set my AVR to 0 to +3 and then set the amp gains to half followed by the EQ gain to balance the levels to the desired curve.

Quote:
Originally Posted by jpmst3 View Post

Well, I guess I don't know exactly.
But, I know that nothing in the chain is maxed out which can introduce noise, etc.. I run the EQ and the AVR as far away from maxing the gain as possible.

See, in my experience, I have the DCX in the chain which has separate input/output level lights so I typically use those as a goto point to see what I am getting so it makes things a LITTLE easier. I know at ref if I am a good -10dB on both of the meters that I am not clipping either input nor output, so from that point, it is simply matching the AVR sub out's gain and the amp gain to match the rest of the system. I too typically stay between 0 and +3 on the individual sub outs on the AVR. Right now my FP14k gains are at about 1/8th to make a 5dB hot sub system work...No boosts or cuts to the input output signal going through the DCX. It took a while using "bumpboxes" to realize that I really didn't need them in the chain with the DCX as it can do up to 15dB boost in the input AND output stages if I truly needed it. For someone with a single driver sub setup, it might help, but not in my particular case, thus the reason I started this thread to see what everyone else is doing with their systems whether it be a single driver setup to a uber system such as many of have around here biggrin.gif
Reply
Reply
post #10 of 90
2/17/13 at 4:19pm
I have the DCX as well and that is what I use as the final adjustment. I try to stay less than +6 db or so on inputs and outputs.
It is a great device with loads of flexibility, especially in my 3 sub setup with multiple pro amps.

I set it all up that I could clip the amps if I turned the receiver the whole way up, but then I set the max volume on the receiver 5-10db below the clipping limits.
Even at the amp clipping level, it is far from the limits of the subs. wink.gif
Edited by jpmst3 - 2/18/13 at 4:01pm
Reply
Reply
post #11 of 90
2/17/13 at 6:45pm
I couldn't vote because none of the options are correct IMO.

So I'll just explain what I do...
After everything is equalized (that is critical because of boosting and cutting), I select a 0-db 16bit PCM sinewave in the middle of the bandpass (and with the speakers disconnected), I set the output level until I clip the inputs on my DCX (and amps that have input meters).
This is the max the pre-amp should be driven to.

(FYI: Black Hawk Down - F Irene or WOTW - Pod scene, and DTS intros works good as well for DD/DTS Master Audio signal checking.)

I then adjust the gain knobs on the amps until the output is just below clipping.
Again, this is the max the amp should be driven to.

I then turn the amps down and connect the speakers.
I drive the speakers to maximum excursion (usually 10-15hz ULF signals).
Again, this is the max the speakers can handle.

Never exceed those three values. Unless you like playing a hot & risky game of Russian Roulette.

[If the speakers can handle more than the amps can, they buy bigger amps or just live with it being under powered.]

I've never seen this NOT work; I mean, you'd have to be doing something pretty crazy to the signals/devices for this method to fail.

Obviously an o-scope/distortion detector would be even better still. Squarewave bad biggrin.gif

Now... whether you like how this will sound as a result, it's an entirely different ball of wax. Some like their signals really hot and compressed with dynamic EQ's and limiters. Nothing wrong with that per-se. How do you like YOUR goldilocks porridge?

The goal is to get all the speakers to max excursion/SPL at the same time, or all the electronics into clipping; whichever comes first.
This maximizes the drive signal to the speakers in a unmolested/pure way. Because nobody likes molested nuts, or roasted nuts for that matter. hehe smile.gif
Edited by BassThatHz - 2/17/13 at 7:11pm
Reply
Reply
post #12 of 90
2/17/13 at 7:15pm
Quote:
Originally Posted by BassThatHz View Post

I couldn't vote because none of the options are correct IMO.

Hmmm, seems like you picked option 2 to me. wink.gifsmile.gif
Reply
Reply
post #13 of 90
2/17/13 at 8:38pm
Quote:
Originally Posted by BassThatHz View Post

I couldn't vote because none of the options are correct IMO.

So I'll just explain what I do...
After everything is equalized (that is critical because of boosting and cutting), I select a 0-db 16bit PCM sinewave in the middle of the bandpass (and with the speakers disconnected), I set the output level until I clip the inputs on my DCX (and amps that have input meters).
This is the max the pre-amp should be driven to.

(FYI: Black Hawk Down - F Irene or WOTW - Pod scene, and DTS intros works good as well for DD/DTS Master Audio signal checking.)

I then adjust the gain knobs on the amps until the output is just below clipping.
Again, this is the max the amp should be driven to.

I then turn the amps down and connect the speakers.
I drive the speakers to maximum excursion (usually 10-15hz ULF signals).
Again, this is the max the speakers can handle.

Never exceed those three values. Unless you like playing a hot & risky game of Russian Roulette.

[If the speakers can handle more than the amps can, they buy bigger amps or just live with it being under powered.]

I've never seen this NOT work; I mean, you'd have to be doing something pretty crazy to the signals/devices for this method to fail.

Obviously an o-scope/distortion detector would be even better still. Squarewave bad biggrin.gif

Now... whether you like how this will sound as a result, it's an entirely different ball of wax. Some like their signals really hot and compressed with dynamic EQ's and limiters. Nothing wrong with that per-se. How do you like YOUR goldilocks porridge?

The goal is to get all the speakers to max excursion/SPL at the same time, or all the electronics into clipping; whichever comes first.
This maximizes the drive signal to the speakers in a unmolested/pure way. Because nobody likes molested nuts, or roasted nuts for that matter. hehe smile.gif
This seams like a nice approach. I like that it includes speaker excursion test, something I have not considered. How does one measure excursion?
Also, what MV/sub channel output level settings do you have on the AVR for the test? How do know it's not clipping or far below max clean output level?
Reply
Reply
post #14 of 90
2/17/13 at 8:50pm
Quote:
Originally Posted by zheka View Post

How do know it's not clipping or far below max clean output level?

The only sure way is with an O scope. You can use a DMM, check for max voltage output, and reduce by 30% to put you in the safety zone.

With EQ's it's really a crap shoot. Fortunately most of them have already been tested by other members, and you should be able to find clean voltage output before clipping occurs.

Sub excursion is all about the ears. You'll definitely know when you start to reach their max output.
Reply
Reply
post #15 of 90
2/17/13 at 10:43pm
You don't need a CRO to measure - use the SC in your computer. Use RMAA and do a loopback to get your baseline noise level. Then using test tones recorded onto a CD, REW etc, send a signal through your AVR (via pre out) and slowly increase until you can see the distortion spectra rise out of the noise, back off a little until they go again and then measure the signal level going into the SC. Also note the level it displays at on screen. This is the max leveel your SC can then take before it clips. Most will only take a couple of volts. If you have a cheapo non true RMS meter, do this at 200Hz and under and it should be accurate enough.

It is much easier to determine distortion by looking at the spectrum, rather than waveshape.

Some laptop SCs are too noisy but most desktop ones are OK. I used my M Audio192 for this until it started playing up recently.

Higher level signals can still be measured if you introduce a known voltage divider before the input of the SC. Single gang 24 stepped attenuators of 48 to 60dB range can be had from ebay cheaply enough ~$20, and 5k or 10k are the best.

Measuring a power amp output adjust the attenuation to be sure the SC isn't clipping via DVM (because you already know at what level your SC clips) until you see it clipping. Without adjusting anything, check the input to the power amp to make sure whatever is driving it is still ouputting clean signal. When you're sure that's the case, measure the input voltage to the power amp as this is the level it needs for full output. Make sure any DSPs internally are off and that front panel attenuators are at min attenuation.

Measure your AVR in a similar way. Play though a 0dBFS test tone and adjust the level until you see it clip from pre out. This is the max level it can put out. If this is less that the power amp needs to get it to clip, you need more gain. If it is enough, you don't. Set the offsel for the channel your testing to zero before this or note what it is.


To measure speaker Xmax make a small lightweight bridge that straddles the cone but is clear of the cone by it's Xmax at rest. When driven to Xmax by signal, the cardboard bridge will be tapping it, but because it's light and not fixed it will do no damage. The driver will need to be (temporarily) facing up to do this. If you watch the gap whilst inreasing level you can see it closing up until it touches easily. You can also play your reference scenes like Fn Irene a few times stepping up the level until it does touch to see if it will ever touch in practice with signal and/or at what level.
Reply
Reply
post #16 of 90
2/18/13 at 6:57am
Quote:
Originally Posted by A9X-308 View Post

You don't need a CRO to measure - use the SC in your computer. Use RMAA and do a loopback to get your baseline noise level. Then using test tones recorded onto a CD, REW etc, send a signal through your AVR (via pre out) and slowly increase until you can see the distortion spectra rise out of the noise, back off a little until they go again and then measure the signal level going into the SC. Also note the level it displays at on screen. This is the max leveel your SC can then take before it clips. Most will only take a couple of volts. If you have a cheapo non true RMS meter, do this at 200Hz and under and it should be accurate enough.

It is much easier to determine distortion by looking at the spectrum, rather than waveshape.

Some laptop SCs are too noisy but most desktop ones are OK. I used my M Audio192 for this until it started playing up recently.

Higher level signals can still be measured if you introduce a known voltage divider before the input of the SC. Single gang 24 stepped attenuators of 48 to 60dB range can be had from ebay cheaply enough ~$20, and 5k or 10k are the best.

Measuring a power amp output adjust the attenuation to be sure the SC isn't clipping via DVM (because you already know at what level your SC clips) until you see it clipping. Without adjusting anything, check the input to the power amp to make sure whatever is driving it is still ouputting clean signal. When you're sure that's the case, measure the input voltage to the power amp as this is the level it needs for full output. Make sure any DSPs internally are off and that front panel attenuators are at min attenuation.

Measure your AVR in a similar way. Play though a 0dBFS test tone and adjust the level until you see it clip from pre out. This is the max level it can put out. If this is less that the power amp needs to get it to clip, you need more gain. If it is enough, you don't. Set the offsel for the channel your testing to zero before this or note what it is.


To measure speaker Xmax make a small lightweight bridge that straddles the cone but is clear of the cone by it's Xmax at rest. When driven to Xmax by signal, the cardboard bridge will be tapping it, but because it's light and not fixed it will do no damage. The driver will need to be (temporarily) facing up to do this. If you watch the gap whilst inreasing level you can see it closing up until it touches easily. You can also play your reference scenes like Fn Irene a few times stepping up the level until it does touch to see if it will ever touch in practice with signal and/or at what level.

Thank you very much. Does RMAA stand for RightMark Audio Analyzer ( http://audio.rightmark.org )?
Reply
Reply
post #17 of 90
2/18/13 at 7:02am
Quote:
Originally Posted by zheka View Post

... How does one measure excursion?
... How do know it's not clipping or far below max clean output level?
One way is prior Klippel large signal analysis
Reply
Reply
post #18 of 90
2/18/13 at 7:04am
Quote:
Originally Posted by bass addict View Post

The only sure way is with an O scope. You can use a DMM, check for max voltage output, and reduce by 30% to put you in the safety zone.

With EQ's it's really a crap shoot. Fortunately most of them have already been tested by other members, and you should be able to find clean voltage output before clipping occurs.

Thank you.

Does anybody here know clean voltage output for sub channel in Denon 4311 ?
Reply
Reply
post #19 of 90
2/18/13 at 7:09am
Quote:
Originally Posted by WVSyd View Post

One way is prior Klippel large signal analysis

heh. that's something within the means of an average poster here tongue.gif
Reply
Reply
post #20 of 90
2/18/13 at 7:20am
Well Linkwitz details ( on his site ) a simple device for measuring mechanical excursion.
This doesn't really measure aspects of performance such as magnetic and suspension linearity and distortion.

FWIW: I use a similar methodology for setting gain as discussed prior: A calibrated reference signal is measured from the beginning of the signal chain - starting at the reference source and working through the gain blocks. Basically this is the same technique that has been used for a long time with a simple volt meter.
Reply
Reply
post #21 of 90
2/18/13 at 9:34am
Quote:
Originally Posted by WVSyd View Post


FWIW: I use a similar methodology for setting gain as discussed prior: A calibrated reference signal is measured from the beginning of the signal chain - starting at the reference source and working through the gain blocks. Basically this is the same technique that has been used for a long time with a simple volt meter.

+1

Start at the front end with 0db signal and go all the way to back end pushing each chain link nearly to full clipping.

IMHO, Problems really occur when you have extreme gain mismatches or lower sensitivity devices after higher, or vice versa which limits dynamic range in one way or another.
Reply
Reply
post #22 of 90
2/18/13 at 9:48am
so what if I cannot properly identify max clean output voltage of the AVR and end up with a more conservative value?
e.g. if the actual clean voltage is twice as high as what my SC can take, would the only problem be the 3dB loss of the system dynamic range or there are other consequences?
IOW, is it a big deal?

Edit: I am not at all sure my 3 db assumption above is correct. the more I think about this the more confused I get
Edited by zheka - 2/18/13 at 10:31am
Reply
Reply
post #23 of 90
2/18/13 at 10:09am
Quote:
Originally Posted by NicksHitachi View Post

IMHO, Problems really occur when you have extreme gain mismatches or lower sensitivity devices after higher, or vice versa which limits dynamic range in one way or another.

This is user dependent as well. I have pretty large gain mismatches and I don't foresee having a problem. For mains I'd see cause for concern, but for sub drivers, not so much IMO.
Reply
Reply
post #24 of 90
2/18/13 at 2:51pm
I added quite a bit of boost at some frequencies to flatten response, so I needed to lower the AVR sub level to keep from clipping - so ended up with power amp gain at max (a modest size power amp, QSC RMX 850).
Reply
Reply
post #25 of 90
2/18/13 at 9:32pm
@BassThatHz method seems to make the most sense to me. The idea is that you want to have more headroom and the most dynamic range, short of clipping an amplifier. Hence why its better to have amps that are more than capable of driving your drivers. Then Amp can reproduce the transients without clip. Obviously you also want to not have a voice coil bottom out so hence be careful with the remote.

My personal thoughts are that the higher the voltage from your pre-amp to your amp, the lower you can have your amp gain and still reach acceptable levels. higher voltage in the signal from one device to the other usually equals better signal to noise ratio. Balanced signal pathes (XLR / TRS) should mitigate that problem.
Reply
Reply
post #26 of 90
2/18/13 at 10:11pm
one question that arises is what to do when multiple channels of bass are re-routed into the subwoofer channel.

that can produce spl greater than -0dbfs because of summation. yes or no?

if it is assumed that a pre-amp is not clipping a -0dbfs signal from each channel, then calibrating the power amp to the pre-amp's output voltage that produces full power makes sense.

great.

but what if channels are summed in the pre-amp? other than putting a scope on it or measuring distortion, how does one know if the sum of the channels in the pre-amp output is clipping? or, do you just calibrate the system to -4dbfs or so instead of -0dbfs in order to guarantee that even after summation, the output signal from the pre-amp will not be clipping?

as for the power amp, if channels are summed and added to the lfe channel, a signal greater than -0db will be output and cause the amps that were calibrated to -0dbfs to clip.
Reply
Reply
post #27 of 90
2/19/13 at 6:42am
Quote:
Originally Posted by LTD02 View Post

one question that arises is what to do when multiple channels of bass are re-routed into the subwoofer channel.

that can produce spl greater than -0dbfs because of summation. yes or no?

if it is assumed that a pre-amp is not clipping a -0dbfs signal from each channel, then calibrating the power amp to the pre-amp's output voltage that produces full power makes sense.

great.

but what if channels are summed in the pre-amp? other than putting a scope on it or measuring distortion, how does one know if the sum of the channels in the pre-amp output is clipping? or, do you just calibrate the system to -4dbfs or so instead of -0dbfs in order to guarantee that even after summation, the output signal from the pre-amp will not be clipping?

as for the power amp, if channels are summed and added to the lfe channel, a signal greater than -0db will be output and cause the amps that were calibrated to -0dbfs to clip.

I thought the 0dbfs had most to do with the player output and was the max the player would/could output @that freq?

IOW is it possible for the player to output a higher signal than 0dbfs @ say 60Hz? IDK for sure.

If so what your saying sounds logically sound although im not sure how often 0dbfs is ever encountered simultaneously on 7 channels AND the LFE channel?
Reply
Reply
post #28 of 90
2/19/13 at 7:01am
http://www.sengpielaudio.com/calculator-db-volt.htm

http://www.jimprice.com/prosound/db.htm
http://en.wikipedia.org/wiki/DBFS
Edited by WVSyd - 2/19/13 at 7:19am
Reply
Reply
post #29 of 90
2/19/13 at 7:12am
How does Audyssey calibration/level matching play into all this? I suspect there is no guarantee that @ 0dbfs MVC settings the sub output would match the max clean output voltage used for the gain structure set up. Am I wrong about it? And if not, how do you address the problem? Do you adjust trim levels for all channels by equal amount to bring the sub output closer to the target ? I think NicksHitachi recommended doing something along these lines in another thread. If you have to bump up the levels, is there a danger of over driving the other channels?
Edited by zheka - 2/19/13 at 7:18am
Reply
Reply
post #30 of 90
2/19/13 at 7:17am
Quote:
Originally Posted by WVSyd View Post

http://www.sengpielaudio.com/calculator-db-volt.htm

http://www.jimprice.com/prosound/db.htm

Thank you. So doubling voltage or current is 6db and doubling power is 3db. got it.
Reply
Reply
New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: DIY Speakers and Subs

AVS Top Picks

AVS › AVS Forum › Audio › DIY Speakers and Subs › Poll: How do YOU set your gain structure????