The truth about subwoofers

[kevintomb]

That's a good point and very possible for my JVC example because the amp and speakers in "all-in-one" systems can have built-in EQ to compensate for what are always bass-challenged mains, but it's not the case for any of my other systems using speaker-level -- those are analog 2-channel only with no active EQ.

Again, I would never recommend speaker-wire for theater use, but I prefer speaker-level for analog 2-channel in my direct experience. I only recommend that people try both (for 2-channel only) and determine what works for them.

For those who believe that all amps sound the same, its easy to see why they also believe speaker-level can't have any benefit. For those that don't believe all amps sound the same, they don't really have a leg to stand on to say the using speaker-level can't be beneficial (<- that doesn't mean it "will" be beneficial, it only means the amp can make a difference in how the sub sounds when connected by speaker wire). There are no "absolutes" about which is better -- that will depend on individual room acoustics, the specific system components used, and the types of music played (e.g. techno with ULF content vs jazz). The best approach is to try both for 2-channel, and stick with the one you prefer (and report back so we all benefit from your experience! ).

I have done it both ways, and honestly think more in theory it is a lesser signal by using the high level inputs.

The difference was not what I am talking about, but more the potential for there to be slightly more degradation, if that makes sense.

As far as amps, I do not recall anyone saying all amps sound the "Same", but I do for sure feel that many are similar enough to not alter the signal enough to make any significant difference.

 

[Audiofreak71]

That's an excellent point. I would think *any* rumble filter would exactly counter the effort to reproduce frequencies handled by subwoofers. And that the use of subwoofers at high volumes would preclude a turntable?
For those of you spinning vinyl, how do you resolve rumble with subwoofers? What type of rumble filtering/arm damping/turntable suspension works for you? Have you found volume limits before rumble feedback occurs?

I don’t use a rumble filter with my tables never needed to. I have my sub isolated very well , as well as my whole room. Another thing is having the proper cartridge compliance to arm ratio, if you have that correct then there usually isn’t a problem except in extreme situations in which that case good isolation or a rumble filter should work.

Audiofreak71

 

[kfa888]

six weeks to come up with a response lol?

I didn't think I had to... You said you were done and not responding like 20 times.

 

[Tube Radio]

However, I don't buy that the transfer function of the upstream amp for the mains matters one bit, if it's competently designed for high fidelity. Sure, some archaic capacitor-coupled or (ahem, archaic) tube amps *might* have poor low bass response. The answer there is to not use them where they aren't capable, and to split off the sub signal *before* the incapable amp (i.e., don't use speaker-level). But for the vast majority of amps, any high pass filtering should have minimal response drop at 20Hz, and therefore minimal phase shift

Glad to see someone else mention phase shift.

It's near impossible to get any amp that uses a capacitor to couple the output transistors to a speaker to where it has zero phase shift at 20Hz as the capacitor needed for that would be around 60,000uF for 8 ohm speakers and even more for 4 ohm speakers and the resultant thump upon turn on could damage the amp or speakers.

I have two of those types of amp. Both are the same model and I used 15,000uF output coupling caps which provides zero phase shift to 70Hz for 8 ohm speakers. One is used with an active crossover set to 70Hz which goes to a powered sub.

The other I added a couple caps from the + end of the main speaker capacitors and those caps go to the remote speaker jacks which feeds the speaker level inputs of a powered sub. That way I don't have to use a 60,000uF cap and there's no phase shift in the signal going to the sub amp.

High pass filters especially those in plate amps unless properly designed do cause phase shift.

I've had to replace the two caps forming a 20Hz high pass filter in a plate amp before with values that were much larger so there was zero phase shift at 20Hz.

LFE filters can also cause phase shift.

Had to eliminate one from a plate amp as there was phase shift at 70Hz caused by the filter.

I've even gone so far as to replace coupling caps in two plate amps with larger values just to get zero phase shift to 20Hz.

In my main system I had two subs at the time and the subs weren't integrating well.

Removed the output coupling caps in my DIY 70Hz 24 dB/octave active crossover (not needed as it uses +/- supplies) and that helped. Got the main amp to zero phase shift at 70Hz and that helped more. Got the phase shift of the plate amp to 0 at 20Hz along with removing the LFE filter and the subs simply disappeared. Added two more and I have proper bass without being able to pinpoint where the subs are located.

Some people have responded that's why they like the variable phase knob on their plate amp, but the variable phase knob is not for compensating for a phase shift in the plate amp or main amp. it's for adjusting the sub to the room correcting for any phase issues caused by the room or where the sub is located.

 

[twiiii]

In a stereo system its hard to always get the arrival times to be the same and finding a Stereo processor to do everything that is needed in an integrated unit is almost impossible. HT systems thats all taken care of. So that means an add on external processor, an integrated unit with the ability to insert the processor and calibration equipment and the knowedge to set up everything correctly. Save the money and head aches and buy two good full range speakers.

 

[Tube Radio]

How many speakers out there are flat to 20Hz that dont cost a small fortune?

 

[fill35u]

It's near impossible to get any amp that uses a capacitor to couple the output transistors to a speaker to where it has zero phase shift at 20Hz as the capacitor needed for that would be around 60,000uF for 8 ohm speakers and even more for 4 ohm speakers and the resultant thump upon turn on could damage the amp or speakers.

I have two of those types of amp.

What make and model are your capacitor-coupled amps? The inherent issues of distortion and frequency response worsening at lower frequencies are well-known, as is the cure of large(expensive) output caps. Why choose such an uncommon, archaic design?

I've even gone so far as to replace coupling caps

If your plate amps don't have DC servos, the shunt feedback cap is the biggest offender. Coupling caps obviously should be large enough, but "enough" depends on the associated ground-return resistor (or input impedance of the next stage), and how low in frequency you want to go. If you have DC servos, you might consider removing/bypassing the coupling caps entirely.

The phase shift goes hand-in-hand with frequency response. You can't have a high or low pass filter without it. The natural *acoustic* rolloff of drivers/speakers themselves introduce phase shifts. Using any kind of equalization will introduce phase shifts. It's a matter of best matching the overall phase shifts (or more precisely, *timing*), because there will always be some phase shifts on both sides of the sub XO. FIR filtering can evade that, but even then you need a lot of horsepower to get it right.

The phase shift issues shouldn't matter, because you should have enough subs and freedom of location to fill in all room mode nulls at the listening positions, enough parametric equalization to clamp down peaks, and enough delay adjustment in all channels to time align everything.

There's also little research on the subject, but the audibility of group delay drops dramatically at lower frequencies. The extrapolated rule of thumb is to keep it below 1ms per Hz . But that's through the entire chain, after all integration has been applied.

 

[fill35u]

Save the money and head aches and buy two good full range speakers.

Which will be almost impossible to locate simultaneously for both best stereo imaging and room mode mitigation.

Excitation of room modes is by far the biggest, and least predictable, offender of the entire endeavor. If you don't have the freedom to locate the sub(s) to best deal with room modes, the end result will be a crap shoot at best.

You *can* "cheat" by using near-field subs, whose direct output swamps the room modes by sheer proximity. Of course, that also requires a certain freedom of location.

You could construct a Double Bass Array to negate room modes, but the location (and number of drivers) requirement is even more strict.

 

[Tube Radio]

My amps are Lafayette LA-375.

Those capacitor couple the input to the amp section (for my main stereo i have the signal going straight to that capacitor through a resistor voltage divider as I didnt need the volume and tone controls given its driven by the high pass filter of a crossover which is driven by a preamp) and capacitor couple the speaker plus there's a capacitor in the feedback loop. All three had to be increased in value.

However I love those amps due to their simplicity and excellent sound quality.

Whats cap coupled in the plate amps is the crossover board which has the input on it as well. The amp board itself has one cap on the input along with the cap in the feedback circuit.

Agreed about matching the phase shifts.

If the amps have zero phase shift it makes matching the phase shifts easier.

When I check for phase shift I use a dual channel scope so I can see the input and output waveform with both traces at the same screen position and amplitude. Any phase shift shows up as one of the waveforms beginning to lead or lag. I simply check each stage before and after a coupling cap and will select one twice its value and test again. Once a stage has zero phase shift at 20Hz I move to the next cap coupled stage until the whole amp has zero phase shift at 20Hz.

The most complicated was a mono full range plate amp/mixer with three balanced inputs. Many capacitors to increase the value of.

I do know that sometimes when using an amp with full range speakers and no sub eliminating phase shift can sometimes make the bass sound better in the room.

However I always use at least one powered sub. When space allows ill do two subs or four as I find a distributed bass setup sounds best.

The reason for always using at least one powered sub is it gives me much more options when setting my main speakers as I no longer need speakers that can go to 20Hz.

Take my main system for instance.

The mains (Electrovoice FR-200) only go to 80Hz, but they sound simply amazing over the range they do reproduce and that -3dB point is close enough to 70Hz which is what my active crossover is set at.

Its especially beneficial when building speakers as well given I dont have to design them with a bass driver capable of going to 20Hz.

 

[theophile]

I haven't read more than a couple of pages. I hope Earl Geddes name has been raised.​

 

[imral3]

I haven't read more than a couple of pages. I hope Earl Geddes name has been raised.​

Yes, by you just now! But, yes it was. As a matter of fact, I just started my very own three sub setup based (loosely) upon Dr. Geddes directions.

 

[Grenadeslio]

I didn't think I had to... You said you were done and not responding like 20 times.

oh, you didn't have too, but apparently YOU feel you had to

But SIX weeks lol? Now that's weak

 

[Grenadeslio]

Thanks for the info! Some more questions, if you don't mind:

What record/song were you playing when you took that dB reading?

Have you found any rolloff at the lowest frequencies, through the overall turntable/preamp chain? Have you measured for it?

What do you use as a test record? Does it have tones down to 20Hz or below? Are they cut at reference level?

The Doors

Tables used;
Mitsubishi DP-EC8
AR XA
Thorens TD166MKII
Kenwood KD500/Grace G707 arm
Kenwood KD770D
Well Tempered Lab Turntable

The table used when that db reading was taken? The Kenwood KD500, the AR XA being suspended, and concrete floors never had any issues with acoustic feedback in the same location. Damped the table, put it on Vibrapods and the issue was solved. Moving the table would have been easier but problematic aesthetically.

The Well Tempered Lab by it's very nature is immune to most feedback issues.

The little AR was my main table for many yrs, with concrete floors it's almost impossible to disrupt. You can bang pretty hard on the plinth while playing without any issues. The Thorens also but not to the degree the AR does.

One day cranking Ray Parker's "Ghost Busters" for the kids the neighbor kid came running over saying "what was that?", their windows were rattling lol. That system had two 10" subs from Madisound built into the huge entertainment wall unit I built. They were powered by a 1000watt amp, and the little AR never missed a beat.

I should add, the neighbor kid was about 19 at the time, with one of those car stereos that rattle your house windows when they drive by, and HE was startled at just how loud the system was.

No spl measurements were taken at the time so......, but it was a good bit louder than the time I took that reading in a different house, different system, 8yrs later.

 

[Grenadeslio]

I have done it both ways, and honestly think more in theory it is a lesser signal by using the high level inputs.

The difference was not what I am talking about, but more the potential for there to be slightly more degradation, if that makes sense.

As far as amps, I do not recall anyone saying all amps sound the "Same", but I do for sure feel that many are similar enough to not alter the signal enough to make any significant difference.

Just as I've been saying. Members here seem to ignore the fact I've stated several times I'm currently using both methods with good results. I'm just not deluding myself into thinking the high level is the cleaner signal. And definitely not the better method.

 

[Tube Radio]

I too have used both high and low level inputs on a sub with good results.

I always prefer low level as I can use an active crossover as otherwise I have to either select a frequency below what my mains -3dB point is or add a high pass crossover to my mains which can affect the sound. The expense is when using higher order passive crossovers. However with a 6dB/octave crossover all one needs is a cap in series with the woofer.

 

[fill35u]

The Doors

Tables used;

Thank you for the detailed reply! And The Doors are a good choice for blasting.

However, I was hoping for some measurements more meaningful to the current discussion than an instantaneous SPL reading of full-range playback of music. Especially since I doubt The Doors on vinyl had really loud bass in the bottom octave?

I was hoping you had played at least some test tones or sweeps, to glean what the overall capabilities of a vinyl system might be. Can vinyl encode a sustained 20Hz tone *at reference level*? Can most turntables read it? If not, what are typical limits?
Those of you not using rumble filters, have you played back 20-40Hz steady tones *at (measured) reference volume* without feedback?
If so, what test records, with which tones, recorded at what level?

My point being: if the capabilities of the media are better understood, it's easier to make decisions about the required capabilities of any downstream subwoofers.

 

[kevintomb]

So exactly how are high level inputs on a sub worse?

I assume you are asking seriously?

You are taking a low level signal, boosting it to a high level, then taking THAT signal, and consequently reducing it to low level again, then boosting that signal to a high level a 2nd time.


With the sub out, you eliminate 2 steps. I am not sure it makes a BIG difference, but why would one subject a signal to 2 levels of amplification when not truly necessary is my thinking.

Mind you I have done it both ways, and do not see or hear a huge issue, but never thought it was the best idea.


Question: Do you know how the high level signal is turned back into a low level signal before being boosted again to a high level sound?

 

[Grenadeslio]

Thank you for the detailed reply! And The Doors are a good choice for blasting.

However, I was hoping for some measurements more meaningful to the current discussion than an instantaneous SPL reading of full-range playback of music. Especially since I doubt The Doors on vinyl had really loud bass in the bottom octave?

I was hoping you had played at least some test tones or sweeps, to glean what the overall capabilities of a vinyl system might be. Can vinyl encode a sustained 20Hz tone *at reference level*? Can most turntables read it? If not, what are typical limits?
Those of you not using rumble filters, have you played back 20-40Hz steady tones *at (measured) reference volume* without feedback?
If so, what test records, with which tones, recorded at what level?

My point being: if the capabilities of the media are better understood, it's easier to make decisions about the required capabilities of any downstream subwoofers.

I have no test records, only music. But here's the thing, I've blasted many records to such a degree but didn't take any readings and I'm sure there were a few that had more bass content like DSOTM.

If you're that interested why not do some testing and get back to us with the results?

 

[fill35u]

I'm sure there were a few that had more bass content like DSOTM.

The heartbeat on "Speak to Me" is around 30Hz.

If you're that interested why not do some testing and get back to us with the results?

Because response of the vinyl playback part of the chain depends (at least) on the particular pressing, the particular cartridge, and the particular preamp. A report doesn't do the discussion any benefit unless it helps define the performance envelope of the medium. And since I don't spin vinyl, I can't do that testing.

A test disk with low bass recorded at the same nominal level as other frequencies would help by showing linear dropoff in the (particular)system. A reference level recording would show the ultimate limitations of the system.

Sure, vinyl might be able to record 10Hz in the groove... at -60dB from reference level. That doesn't mean it's *practical* for vinyl to reproduce it. Because of its inherent variability, is it impossible to define vinyl's low bass performance? If so, anecdotes are even more meaningless without comprehensive measurements.

 

[Tube Radio]

You are taking a low level signal, boosting it to a high level, then taking THAT signal, and consequently reducing it to low level again, then boosting that signal to a high level a 2nd time.

Nothing wrong with that for a sub.

With the sub out, you eliminate 2 steps. I am not sure it makes a BIG difference, but why would one subject a signal to 2 levels of amplification when not truly necessary is my thinking.

Most don't unless necessary.

A lot of times doing it right if the receiver has a pre out jack and main in jack involves either building a buffer to drive a commercially available two way active crossover or building a two way active crossover with a higher input impedance.

Either way it involves building something which a lot of HI-FI people don't know how to do.

So for them using a speaker out on a receiver makes more sense.

Some receivers have no option of a pre out so they have no choice.

Mind you I have done it both ways, and do not see or hear a huge issue, but never thought it was the best idea.

I've done both and have not noticed any real quality difference when done right which means adding a passive high pass crossover to the mains or setting the sub's crossover to the -3dB point of the mains.

Question: Do you know how the high level signal is turned back into a low level signal before being boosted again to a high level sound?

Using resistors. At least that's how I've seen it done in the two plate amps I've dealt with.