6.5" OB "bookshelf" speakers

[BoomieMCT]

So, after reading the OB thread that 'Bob from St Louis' started I decided to make my own OB speakers. Since I also wanted to learn how to make a crossover myself I opted not to go with any kits or pre-made plans. So, striking out with Unibox, FRD Crossover designer and some forum advice (all free) I started my trip into the world of open baffles.

The woofers I settled on are the 6.5" Pyle Pro PA drivers. I picked these because they had a Qts of .89 (I had heard .9 was optimal), good bandwidth and they had a pretty decent xmax (I was scared of overdriving them with bass in no box). These drivers wouldn't dig much below 100Hz, but I always assumed I'd use a sub with them.

For tweeters I settled on the Hi-VI TN28. I'll be honest, I picked this mainly because it looked cool but also because the sensitivity was close to the woofers'. I also figured since they were on 'pods' it would be easy to mount them back in the 'Z' axis so the centers of the tweeters would match the center of the woofers. Give me a break, I was still learning.

The plan was to make small baffles (10" circles) with the tweeters mounted on top. The crossover was a second order / second order that I fiddled with a bid in FRD crossover designer.

The first version can be seen in the attachments. As you can see there is a big baffle step problem and a rising tweeter response problem. This was exacerbated by the fact that the 3 ohm padding resistor I needed was backordered.

[BoomieMCT]

Part 2 - The Auxiliary Speaker

So now I had a result but I also had some problems. First off, the baffle step was pretty bad. It could be reduced by making a larger baffle, but I didn't want to do that. I made a quick BSC circuit but it killed the sensitivity and reduced the openess and airiness I liked so much. So I opted to fix it with an auxiliary speaker. That, combined with more time spent adjusting the postion of the OBs, seems to work better for me.

When building the Aux speaker I used a very highly scientific criterion - I realized I never made a speaker with passive radiator and PE was having a sale on 8" PRs. So I got a Dayton 8" shielded DVC subwoofer (which plays to a much higher frequency then you'd think) and build a modified passive "Triska Sub" crossed over to match the baffle step problems I was having.

Attached is

1) A picture of the auxiliary speaker
2) A FR plot of just the auxiliary speaker (blue) (see how high it can go?) and the OB6 alone (red). It seems like the tweeter is decaying but in actuality I had my levels set too high and I did a poor job matching the impedence and it is clipping. More on that in the next post or two.

As an additional note, I added a stereo L-pad to the Auxiliary speaker so I can set the level of the original OB's. It works quite well. In my office there is very little cut. When these were in my living room (near the ceiling) I could cut down the higher freqencies to minimize the smear I was getting.

I'll write a part 3 about how I fixed the tweeter response when I'm done fixing it . I still have some things I want to try.

[BoomieMCT]

Part 3 - Fixing the Tweeter

First off, I'm going to have to write part of this and finish it at home as I don't have all my info here at work on my lunch break.

So, when I first tested my OB system there was a big problem with rising tweeter response. I attributed it to the fact that I didn't have a 3 ohm resistor in the tweeter's network (R3a in the diagram) to cut it down. When the resistor finally came in I installed it and the tweeter rise became a drop-off. What happened? I'm not sure but I have two guesses.

Theory 1: The tweeter still has rising response and it is making the tweeter clip in the freqency sweep causing it to be measured as decaying response. As I have mentioned before my measurement gear leaves a lot to be desired.

Theory 2: I should have padded the tweeter with an L-pad instead of just one resistor. When I first started I didn't understand the beauty of L-pads. I'm ordering new resistors and adding the R3b segment to try to see if that fixes anything.

When I get home I'll add the values for the componants in the diagram below. One thing I'll point out is that the box labeled "L-pad" is a variable stereo L-pad. I used this so I could adjust how much I cut the upper frequencies.

[Exocer]

This is a very interesting project...I'll be back VERY often to check up on your progress and hopefully learn a thing or two about crossover design in the process

Hopefully this turns out to be exactly what you wanted it to be.

[BoomieMCT]

Exocer wrote:

This is a very interesting project...I'll be back VERY often to check up on your progress and hopefully learn a thing or two about crossover design in the process

Hopefully this turns out to be exactly what you wanted it to be.

I hope so too. I won't be able to fiddle with it for a while though because s/h on small shipments through PE is kinda expensive - I have to wait until I need a lot of stuff. Luckily one of my friends needs an amp and enclosure built for his TC-3000 subwoofer driver so that might be sooner then I think!

[BoomieMCT]

Well, I've been fiddling around trying to see if I can get just the upper sections (no aux speaker) to flatten out a bit. According to this site I used a .8 mH inductor and a 20 ohm resistor to make a BSC to flatten out the high end. The results were less then spectacular.

Oh well. Back to trying to tame the tweeters. Actually - that's not a fair statement. The tweeters sound fine, they just don't seem to measure well. I'm trying to verify that the problem is in the measurement equipment (where I think it is).

[BoomieMCT]

So I figured the falling tweeter response could be from rising impedene. So, I tried to do impedence matching by adding and changing resistor R3a. Here is a plot with different R3b values:

Red: R3b = infinate (no resistor)
Yellow: R3b = 20 ohms
Pink: R3b = 8 ohms.

As you can see this is not the problem I'm having as there is little difference between these values. So, I'm left with two theories;

1) The TN28 doesn't play as high as advertised. I think this drop-off is too severe to be a small error in specs.

2) The soundcard on my fiancee's Mac G4 sucks. Given my experience with this so far this is my current leading theory - especially since the tweeters *sound* good.

[BoomieMCT]

Had a friend bring over his test gear (which is much better then mine) and it confirmed what I suspected - that Mac G4 soundcard sucks. The plots match up pretty well until about 9kHz where the Mac card starts to take a dive. By 11kHz the Mac won't pick up much of anything.

Or, these speakers sounded fine because they were fine.

I have one more experiment left with these. The pro-sound Pyle drivers I use as the OB woofers start their roll-off much sooner then I expected. I'm going to order a set of Audax 6.5" drivers which model well and see how they work. Hopefully I'll be able to eliminate / reduce the need for the auxiliary speaker.

[Exocer]

Yeah, thats some severe premature roll-off on the tweeter. You're right in asssuming if both tweeters behave the same way or similarly that they must not be living up to tall spec claims. Fortunately they sound "good" to you, or else i'd switch those out as well.

Let us know what happens with the Audax drivers

[BoomieMCT]

Exocer wrote:

Yeah, thats some severe premature roll-off on the tweeter. You're right in asssuming if both tweeters behave the same way or similarly that they must not be living up to tall spec claims. Fortunately they sound "good" to you, or else i'd switch those out as well.

Let us know what happens with the Audax drivers

The spec is fine, my measurment gear was not. As I said, I had a friend with better gear come by and verify what my ears were already telling me - the tweets are fine. It is the stupid Mac soundcard that needs help!

The Audax drivers arrived today. Sadly one had a big knife wound in its cone. PE is sending a replacement postehaste.

Enclosed are two pics. The first is my super-technical test stand. Yes, I just hung a driver from a vacum cleaner. It is breaking in right now. Already I can tell the cone moves a lot more then the Pyle one I was using. I think the bass response is better, but once it is broken in some test will tell me what is going on. I have one of the OBs on the other channel (w/out the augmentation speaker) so I can switch back and forth to compare.

The second pic is a possible baffle idea I was thinking of trying. Since I'm trying to get the most baffle in the smallest place I figured a bowl might simulate wings in three dimension. Target had "medium" salad bowls for $9 each so I picked up two. A 6.5" driver will *just* fit on the flat part on the bottom. Even if I don't use these for this project they may become an OB full-range set at some point.

As a side note playing both channels actually sounds really cool right now.

[hdspeakerman]

Hi! I stumbled onto your project and I am quite interested. Your last post was in May. Anything new to report? Are you still experimenting?
Howard

[BoomieMCT]

hdspeakerman wrote:

Hi! I stumbled onto your project and I am quite interested. Your last post was in May. Anything new to report? Are you still experimenting?
Howard

Nope. I felt I got all I could out of evolving this design and moved on to a new OB project.

These are now in my parent's living room - they like them very much. I'd be happy to answer any questions about them that you might have.