Getting my feet wet with REW and the BFD

[Wayne A. Pflughaupt]

Probably shocks a lot of people that I’m so late getting into the REW and BFD game (especially the latter!). What is it they say about an old dog and new tricks? Thanks to Sonnie for dragging me into the 21st century! It took him personally coming to Texas and showing me how quick and easy REW is compared to the old method of manually plotting sine waves. It was so cool I just had to give it a try, desite my dread of learning new computer programs.

I thought I’d get my feet wet tweaking the system I have for my computer, which includes a pair of Swans multimedia speakers and a baby M&K sub with an 8” woofer. My wife and daughter would have had my head if I had tied up the main system downstairs trying to get the hang of all this (which took a while, because it really is hard to teach an old dog new tricks!).

Here is the baseline system response.

base measurement.jpg

Here is REW’s predicted response, using 8 filters:

predicted correction with 6 filters.jpg

And, here is actual response, using 8 filters.

actual response w-8 filters.jpg

Getting a feel for it, I went back and re-calibrated using less filters. This is actual response with 6 filters

actual response w-6 filters.jpg

As you can see, above 90 Hz this is even closer to REW’s predicted response, with less filters to boot!

The problem at the top end is because there is no high pass on the Swans speakers, so they’re “bleeding” into the graph. And obviously EQing didn’t make any change up there. I’ll probably go back and re-calibrate again, and this time not use any filters at the top end.

This is a problem with the Swans speakers, not being able to high-pass them (most computer speakers come with a sub and therefore built-in high- and low-passing). I think I’m going to look into an outboard electronic crossover to take care of this problem.

As you can imagine, the EQ made a noticeable improvement in the way the bass sounds. It’s tighter, and better extended. Looks like I need to eBay a second BFD for my office!

Still playing with the BFD and REW in the main system downstairs. Stay tuned.

In the meantime, it’s surprising the amount of time and effort it takes to tweak the filters on-screen in “real time” with REW. My hat’s off to all those courageous folks over the years who BFD’d their system by trial-and-error manual response plotting! Yikes! Maybe it wasn’t such a bad idea after all waiting ‘til the 13th hour!

Regards,
Wayne

[brucek]

Normally we like to post graphs with a vertical scale of 45Hz to 105Hz. It allows easier comparisons and evaluation when everyone uses the same scale.

I'm a bit confused - Is this the response of the sub and mains combined or the sub only? The usual chain of events is to equalize with the sub only and then add the mains once that's complete..

brucek

[Wayne A. Pflughaupt]

Hi brucek,

Normally we like to post graphs with a vertical scale of 45Hz to 105Hz. It allows easier comparisons and evaluation when everyone uses the same scale.

Yeah, I noticed that after I posted – Sorry. I’ll be sure and make the graphs to spec in the future.

I'm a bit confused - Is this the response of the sub and mains combined or the sub only? The usual chain of events is to equalize with the sub only and then add the mains once that's complete..

Sub and mains. Guess I should have elaborated. The Swans are more like studio monitors than traditional computer speakers – i.e., powered but with no sub. Since there is no bass management in the soundcard, the only way I was able to connect a sub that I could control from the Swan’s volume knob was to tie into the Swans’ woofers to get a signal to the sub – i.e., speaker-level. So, when it came time to graph with REW, there was really no way to fully separate the sub and mains. Sure, I could have split the soundcard’s output and ran one side to the sub, but that would have left me stuck using the soundcard’s volume control. I’d rather use the Swans’.

Not wanting to run the BFD full range, I later came up with an idea to solder 1/4” connectors to those speaker wires feeding to the sub, and instead feed them into the BFD’s inputs (the higher level shouldn’t be a problem – I can switch to the 4 dB setting if needed). That way I can dedicate the BFD to the sub. A rather inelegant solution, granted. That’s why I’ll probably be looking into an outboard crossover soon, which would connect first-in-line from the soundcard and give me separate sends for sub and mains.

Regards,
Wayne

[brucek]

I later came up with an idea to solder 1/4” connectors to those speaker wires feeding to the sub, and instead feed them into the BFD’s inputs (the higher level shouldn’t be a problem – I can switch to the 4 dB setting if needed).

I'm surprised the level wouldn't be a bit high for the BFD. Even the +4dBu setting assumes a maximum input of only ~4.9vRMS and nominal of ~1.24vRMS. Your woofers max out isn't more than 4.9volts? That would only be ~3 watts? I guess the BFD's input LED's will tell you that pretty quick though.

Even then, I would have thought the M&K sub would have wanted a consumer level rather than a level consistent with the +4dB setting. The -10dBV setting would assume a max of 1.25vRMS and a nominal of 315mvolts.

Are you sure you don't need to attenuate the level with a divider?

brucek

[Wayne A. Pflughaupt]

Hi brucek,

Well, the BFD’s meters are well below clipping, although I did have to turn the M&K down practically to the bottom. The Swans’ amplifiers are rated for 35 watts or so, and of course that’s divided between the two speakers via a passive crossover. Plus, I don’t listen at high levels, so it may well be that I’m not be driving more than a few watts to the woofer.

Are you sure you don't need to attenuate the level with a divider?

Obviously I need to do something... It occurred to me after dropping the last post that a crossover in front of the speakers won’t work as I want, as it will leave me controlling the volume from either the sound card or the crossover’s input control. I found a chart on Swans’ website that shows response dropping about 20 dB/octave below 100 Hz, so maybe I don’t need to high-pass them, just attenuate my speaker level outputs to the BFD, as you suggested. Any recommendations for that?

Regards,
Wayne

[brucek]

just attenuate my speaker level outputs to the BFD

I agree you probably don't need a crossover, given the spec you found.

It's certainly easy to attenuate the level, but didn't you feel that the signal level was OK since the meters weren't clipping?

brucek

[Wayne A. Pflughaupt]

It's certainly easy to attenuate the level, but didn't you feel that the signal level was OK since the meters weren't clipping?

It would mainly be for the benefit of the M&K. I used to keep its gain knob at around 12:00. Now it’s almost all the way down, so you get hair-triggered adjustment: barely moving it gets a lot of audible change. Man, I wish they had put a level control on the BFD!

Regards,
Wayne

[eddthompson]

you know, when you move the bfd to your main room. you can use rew to create an impulse responce derived from the filters which you can then plug into a convolver, you could then filter a bit higher, if your sound card is more than 2 channel, you could even create a crossover

edd

[brucek]

Man, I wish they had put a level control on the BFD!

A simple voltage divider with the appropriate input and output impedances would do the job of reducing that high level.

If I assume 50watts across a speaker then the max level at the speaker would be about 20vRMS.
If we want to reduce this to about 2vRMS at the BFD, then that's a factor of 10 (-20dB).

An output impedance of the attenuator of about 600ohms would be fine and an input impedance of ten times that would be acceptable, so a divider factor of 10 would result in a standard resistive value of 5.6Kohm.

The power across the attenuator would be small (E^2 / R =20^2/(5.6K+600) =~65mwatts), so 1/4 or 1/2 watt resistors would be fine.

If you wanted to make it variable from the -20dB down to no output you could replace the 600ohm resistor with a potentiometer of the same value and take the output off the middle leg.

If you need a higher level than -20dB, you could increase the 600ohms up to about 1K, but I wouldn't go higher than that. 1K would be about -16dB down....

level.jpg

Anyway, you could try this and see if it works OK. Play it loud and test the resistors for heat to be sure my napkin calculations are correct... hehehe

brucek

[Wayne A. Pflughaupt]

Cool! It shouldn’t be hard to wire all that up in a project box of some kind, maybe even install speaker connections on one side and RCA’s on the other. Being mathematically challenged, I would have had to “trial and error” it to get the correct resistor values, so thanks for doing all the heavy lifting - um, brainwork!

Just to be clear, 1/4 or 1/2-watt resistors will work for the 600 ohm as well as the 5.6k?

Regards,
Wayne

[brucek]

Just to be clear, 1/4 or 1/2-watt resistors will work for the 600 ohm as well as the 5.6k?

Yep...

The standard values for 600ohm will be 620 (blue,red,brown) or 680 (blue,grey,brown) and the larger value is 5600 (green,blue,red)................