Phase settings on dissimilar subs

[grn1969c10]

I'm getting somewhat odd results when experimenting with the phase settings on my subs. I have a 12" JBL and a 15" CSX which I stacked so that speaker distance to the mic would be identical. The JBL has a phase switch marked 0 and -180. The CSX has a continuously variable knob from 0 degrees to 180 degrees.

With the JBL switch set to -180, I graphed the response while changing the CSX in 30 degree increments from 0 to 180. 120 degrees yielded the flattest response and 180 yielded the most SPL though very peaky.

Thats not so amazing I guess, but things changed when I set the JBL phase switch to 0 degrees. There was very nasty cancelling going on at every possible phase setting on the CSX. I verified that each subs output was normal and identical when operated alone at each phase setting. Yet together, there was no possible setting of the CSX variable phase control which let it work with the JBL set at 0.

I'm drawing sine waves on scrap paper and thinking if one signal led another by 90 degrees, delaying it from 0 to 180 (90 actually) would allow them to be matched. If one signal lagged the other by 90 degrees delaying it from 0 to 180 would never be able to match them. Even so, no combination of CSX settings with the JBL at 0 degrees were even remotely useable and nearly all CSX settings with the JBL set to -180 were useable.??

[grn1969c10]

Another unique observation of the two subs interacting. I picked what I thought was the flattest setting - The JBL phase set to -180 and the CSX set to about 120 degrees. While looking at the spectrum tab, I had the generator set to pink pn (I was trying to figure out the RTA funtion in REW without rereading the help file!) The displayed graph was nothing special. On a whim, I turned up the gain on the JBL and noticed an increase at 30hz and a decrease at 70. Turning down the JBL gain decreased 30hz but increased 70hz. Sort of like a seesaw pivoting around 50hz.

I was trying to view the computer screen from across the room as I turned the gain knob so I'm not sure about the precise numbers but the seesaw graph trend caught my attention. For upping the gain on one sub to decrease the overall output at a particular frequency indicates some phase cancelling going on right? Yet at the same time there's an additive effect in a different frequency range?

I feel like I'm beginning to babble like an idiot, so I'll go to bed. It would be cool if any of this actually made sense to someone and could be useful.

[brucek]

I have a 12" JBL and a 15" CSX

You do know that dissimilar subs is a no-no?

For upping the gain on one sub to decrease the overall output at a particular frequency indicates some phase cancelling going on right?

Yep, but you will never make sense of it with logic or drawing on a piece of paper.
Once those signals leave the driver and get into the bad world of room acoustics, it's a free for all. Use an RTA at the listening position to set phase - and be careful once you get the subs in phase harmony with each other that you then add in the mains....

I was trying to figure out the RTA funtion in REW without rereading the help file

See here for RTA and Spectrum Analyzer settings.


brucek

[tonyvdb]

brucek wrote:

You do know that dissimilar subs is a no-no?

brucek

This is true but it can work. In my case I also have two different subs and they work well together it just takes some time to set them up properly and the phase is a tough one.

[brucek]

This is true but it can work.

I don't see how it can possibly work if they have different responses, specifically if one has a lower extension than the other. Where they share equal output at a specific frequency range (i.e. 30Hz-60Hz), they will re-enforce and the SPL level will increase. Where the better sub extends to 15hz, yet the challenged sub, does not, the combined output reinforcement will be lower. This basically dumbs down the superior subs response to the challenged subs response. Why would we want that?

There's no problem with different models as long as they share a response that is the same, but for those who hold onto their old sub after they move up the food chain with a new low extension sub are doing the new sub a dis-service...

brucek

[grn1969c10]

Bruce, your argument for using one sub is obviously well thought out and compelling. Without REW, I would probably accept it and move on. Unfortunately, the results from REW in my listening environment really bother me. I mean really, really bother me. So much that I have been working on this dumb stereo weekly since probably September of last year.

I started out with a 10" velodyne (which I had been happy with for several years). It looked something like this.



Not very good. I bought a used 12" JBL and it looked like this.



Still not very good. I knew I shouldn't spend the money, but I couldn't pass up the 30% live.com discount, so I bought the 15" Cadence. It looked like this.



Hmmm. A little improvement. I know now that this data is flawed on several levels. I've learned a few things, such as my Onkyo's 12db/octave crossover at 100 Hz not REW's default 24db/octave at 80 Hz, etc. Still, none of my rookie mistakes have fully explained why three different subs have disturbingly similar (and bad) response curves.

I bought a BFD. It helped. So far, it may be the best bang for the buck investment I've made. I bought fiberglass and trapped the front two corners. So far I'm a little underwhelmed, but I want to test some more on that. I experimented with sub placement and found it definitely affects 55 Hz and up. I've tried several locations around the room. This graph shows how moving the subs in incremental steps closer to the corners produced gains in the 60 Hz to 120 Hz range, which I don't need. 50 Hz and lower seem to be somewhat resistant to position changes.



Which brings me to last night. After reading your post, I pulled the JBL out, stuffed it in the storage room and went to work with just the CSX-15. I ran a new baseline sweep, input new filters and it looked pretty decent.



When I ran test tones through the receiver, I had mains, center and surrounds at 75db but could only get the sub up to 72db - and that's with the plate amp trim maxed out. I played a movie, then went and got the JBL back out. I EQ'd the JBL by itself, then ran full range sweeps with both subs and mains and then modified one of the filters. With two subs there was enough output to let me get 75db and back off on the BFD input levels to get some badly needed headroom. I ended up with this sweep which is both subs and mains together.






The manufacturer's specs for the JBL is -6db at 23 Hz, if I remember correctly. Cadence is a little dodgy giving separate specs for the woofer and the amp; 20-500 Hz and 10-500 Hz respectively. I asked them outright in an email for the subs specs and they say they can't be more specific because they have no idea how the customers room will affect the response. (Smart guys over there.) I don't know if either spec has any relationship to the actual tuning of the sub. There is very little 3rd party info available on the unit. While the CSX does outperform the JBL, the actual in-room response of both is struggling at 30 Hz so I doubt the JBL is actually holding it back.

I sometimes wonder about the big boys like Hsu and SVS providing a detailed response chart for their products and then claiming in room response may even be 2 or 3 db lower. Wouldn't that be a fine problem to have?

[grn1969c10]

I can't explain why, but I had to set the JBL phase to -180 and the CSX phase to 0. I have tried different cables, so if anything is wired wrong it must be internal to the JBL plate amp.

[tonyvdb]

grn1969c10 wrote:

I can't explain why, but I had to set the JBL phase to -180 and the CSX phase to 0. I have tried different cables, so if anything is wired wrong it must be internal to the JBL plate amp.

I have my phase on my SVS PB13 Ultra set at 0 and my A/D/S sub set to 180 also and that made a huge difference. Totally different subs but they work really well together. And just like you it increases my lower end dramatically (20-40Hz).
I have the Low pass filter on my A/D/S set to 40Hz it has a frequency response of 19-100Hz.