MDF dust is evil (and a crossover question)

[ptwining]

MDF dust is evil.

The table-saw is bad but the router is worse.

My God, that stuff is as fine as talcum powder, sifts through crummy masks like a gas and coats yer eyeballs with gluey scum.

I encourage all and sundry to wear a good OSHA quality mask and use your shop vac while you use the tool or you'll be hacking spooge for a week!

Take care of yourselves!

Now, for the XO question.

I am in the process of building a published design but (because I'm impatient while waiting for parts) I used a canned Dayton XO instead of the published spec and some close-ish drivers.

The drivers I used (temporarily) are close to the right ones and the XO points are reasonable for the published response graphs.

Why is this XO solution so bad if the XO points are nice with the driver curves? I know they don't match the original design (and I fully intend to build the right ones when the parts come) but they curves of the temp drivers overlap nicely at the XO points of the canned XO.

Why should the temporary result suck if these things have been taken into account?

Should this temp design disappoint and be shunned like a bad witch?

Am I ignoring something magic that an XO does except distributing power and shunting freq's around?

No intention to stay with this temp solution but I am curious...

Regards,
Tom

[poobah]

It should work... keep in mind the sensitivities of the drivers should be equal as well.

[jackinnj]

MDF dust, wax and newspaper will make great fireplace logs.

Real world MDF processing:



i am showing this off in 2 weeks at IWF2006

[sdclc126]

I'm sure some more learned members will contribute here, but I have read many times in this forum how people have experimented with different crossover designs for given projects, and have described the various audible improvements (or lack thereof) they have achieved by optimizing their designs. So it's not a myth that crossover topology is critical.

Crossover order (db/octave), frequency response, driver break-up, impedence fluctuations, distortion, rolloff, etc. are all affected by crossover design, and this will affect what you hear from the finished product. Much of design is to taste also - some will aim for a "warm" midrange, an enhanced bass response, or an attenuation of very high frequencies to dampen "brightness" or sharpness in the treble.

However, if you don't have measuring apparatus (or design software at the minimum) it's virtually impossible to do ANY kind of true design work in audio. So the safest way to go is with a proven design; "slapping" something together has often been described here as throwing darts in the dark and, though whatever you've put together can and may sound fantastic, you have nothing to compare it to until you've done it the "scientific" way.

Having said all that, IMHO the true scientific method is seldom ever used in audio design, and for good reason - it's very difficult to do properly. On the one hand there is no arguing with how crossover, enclosure etc. design will change loudspeaker performance when you have the recorded measuring data to prove it. But what is exceedingly rare would be to perform a double blind listening test only, similar to how pharmaceutical companies test drug candidates in laboratory and clinical trials --

Essentially you would randomly select a bunch of people (probably give them all a hearing test to make sure you're comparing apples to apples) and have them do A/B tests with two loudspeaker sets sharing the same source system, in each case differing in structure by only one parameter - say crossover frequency for starters (drivers, cabinets etc. are identical). They listen to the two speaker sets "blinded" - they can't see them and don't know anything about them; the same recording is played on a single source, switching back and forth from one set to the other. They are then asked to describe as best they can what, if any, differences they hear between the two speakers. BTW - the subject will NOT be told that there is actually a difference between the speakers nor what it is, and indeed in some cases there IS no difference (placebo). Also, the people conducting the test will be blinded to the same things, so as not to set up any possibility of prejudicing the test (this is why it's called double blind - both the test-ors and test-ees are blind as to whethter what's being tested is the real thing or a placebo.

When all the data is collected at the end of the test it is then interpreted to see if there is any statistical difference between the placebo - no difference between the loudspeakers - and the real thing - the speakers where one set has a single design difference.

You can see the difficulty of doing this sort of thing, but it's not impossible. I would also like to see a comparison between a group of "laypeople" and us audiophiles, to see if we are any better at detecting subtle differences in speaker design.

People always have their preferences, making this a very subjective endeavor, but if you want to assess something scientifically you need to level the playing field. I have often wondered what the results would be if a designer were blinded to his/her own designs and asked to assess differences with a test as described above - I'm sure a lot of eyes would be opened at the outcome, with possibly a storm of controversy and debate arising afterward.

You see, if you're the one listening to the results of your own work, you may be hearing a difference because you want to or you're expecting to, or because someone else told you there would be a difference - but this is hardly scientific and the test is indeed "rigged." Very difficult to be impartial here! AND, measured differences don't always translate into AUDIBLE ones!!

Well you are probably surprised to read all this philosophy in response to a very fundamental (and important) question (have I helped at all or did I digress too much?), and I know there may be a few sharp responses to what I've said, but this is something I've thought about for some time, and your post gave me a "forum" in which to voice my thoughts. (If anybody does wish to take a different perspective, please make sure you've read what I've said thoroughly and understand my intent, and don't read anything into it that isn't there!)

At the end of the day I think you'll find the answers to your questions about crossovers once you've finished your published design project and can do a comparison. Now, what happens if you don't hear a difference when you've done this? Well, for one thing it doesn't mean there wouldn't ever be any differences with other designs and how much you've digressed from them. On the other, I have an unofficial "truism" about audio, summing up my long-wnided dissertation above: If you don't hear a difference, there IS no difference.

[Michael Chua]

My God, that stuff is as fine as talcum powder, sifts through crummy masks like a gas and coats yer eyeballs with gluey scum.

I encourage all and sundry to wear a good OSHA quality mask and use your shop vac while you use the tool or you'll be hacking spooge for a week!

Take care of yourselves!

Hello Tom

You should have seen me after I sanded my plaster ceiling. I think the dust from the vinyl filler is even finer.

Regards
Mike

[Scottmoose]

Just take care of yourselves guys. MDF dust is a known carcinogen. A filter mask and eye / ear protection are mandatory for safety.

[ShinOBIWAN]

Originally posted by Scottmoose
Just take care of yourselves guys. MDF dust is a known carcinogen. A filter mask and eye / ear protection are mandatory for safety.

Yeah, not to mention the formaldehyde content thats release when cutting.

Didn't they ban it in certain parts of America? Everything is a potential liable case over there so they tend to be more vigilant regarding legislation.

[Puggie]

Mix MDF dust with polyeurothane resin, makes quite good filler

[Scottmoose]

Of graves?

[andy2]

That's why my next project will use plywood. Much less dust and stronger, livelier sound.