Tapered TL satellite prototype questions for MJK spreadsheets

[skov]

Hi there. I have some newbie questions about a prototype I just built for a satellite speaker using the Tang Band 1364S bamboo 3" driver.

First, I am completely new at speaker design so please bear with me if (or I should say when) I am doing things wrong. Before posting, I have designed and built a prototype for the speaker design, and now have some questions about the observed results.

The design goal for this speaker is a relatively compact full-range satellite (front/center/surround) speaker for use in a 5:1 system with an HT receiver having a crossover at 90Hz. The Tang Band 1364S available at PE seems to have many desirable characteristics for this design objective, especially given that high SPLs are not required. Using Martin J. King's wonderful (and infinitely amusing) spreadsheets, I chose a conventional single-folded tapered TL (taper approx 5:1) to fit into a conventional rectangular cabinet design. My target TL length was 26.5", which should yield a cutoff well below 90Hz given the 1364S manufacturer specs. Due to the obvious newbie error when translating the TL length to the exterior dimensions, the prototype's realized length based on design assumptions is really around 24.25".

Even at this length, and allowing for an actual Fs of 115Hz instead of the manufacturer's stated 105Hz (I have no way to measure Fs), modeled response should be flat close to 90Hz, whereas the eardrum-observed (I have no SPL meter either) in-room frequency response using warble tones at 90Hz, 95Hz, 100Hz is something like -5db (noticeably down), flat, flat. Using the spreadsheet, I have to shorten the model's TL length to 22.0" to get this frequency response. So, I some questions about how to reconcile a short single-fold TL model with the observed response.

One possibility is that the signal path aournd the fold should not be considered in the TL length. MJK's application notes say that the path around the fold should be counted. OTOH, Bjorn Johannesen's MJK for Dummies seems to say that the path around the fold should not be counted.

So, I have some questions. I'm partly asking out of curiousity, and partly asking because my plan is to now build the 5 satellites, and then move on to a powered subwoofer based on a double-folded tapered conventional TL using the Tang Band W6-1139SG, and so I'd like to know how to correct the design.

Is it simply that constructed cabinet response is simply never as good as the model, especially for incompetent carpenters working with hand tools in NYC apartments?

Is it that the path around the fold should not be considered when calculating the TL legnth?

Is there some other factor I have not considered?

Thanks very much for taking the time to read my post.

- Skov

[sreten]

Hi,

you are not considering baffle step compensation.

http://www.zaphaudio.com/audio-speaker18.html

for the B3S

http://www.zaphaudio.com/archives.html

for the TB and Aura 3" drivers

are essential reading.

/sreten.

[MJK]

Using Martin J. King's wonderful (and infinitely amusing) spreadsheets

Infinitely amusing! That is a first ..... but probably the right way to look at the worksheets.

Is it simply that constructed cabinet response is simply never as good as the model, especially for incompetent carpenters working with hand tools in NYC apartments?

I have always found the measured respons to line up well with the worksheets if done nearfield. So in my opinion the worksheets do a good job ofredicting the acoustics in the enclosure. When you move out into a real listening room then other factors influence the response which are not accounted for in the current worksheets. I hope to address some of these with the new versions of the worksheets. One good way to check the accuracy of the model is to measure the electrical impedance and see if the peaks, dips, and wiggles line up with the worksheet predictions. If you have not measured the T/S parameters of your drivers then this is probably the biggest source of error.

Is it that the path around the fold should not be considered when calculating the TL legnth?

I always model the fold in the worksheet as shown in my application note. I have verified this by correlating the MathCad plots witth finite element models and measured results.

Is there some other factor I have not considered?

I think sreten is right, you need to account for the baffle step. One other influence will be the floor bounce cancellation but I would expect that in the several hundred Hz range if the speakers are ar ear level.

[skov]

Thanks very much sreten and MJK for your replies.

I did consider baffle step response as a factor in the frequency response and I definitely agree that these speakers are candidates for BSC based on how they sound (tilted a bit to midrange and treble). It's not clear to me how baffle step would create (or modify) what appears to be a bass cutoff, as opposed to a gentler rising response through the upper midrange and treble, but I definitely don't have the test instruments to speak authoritatively about the frequency response of my prototype speaker. I'll continue to research baffle step and try to learn more about it. To be perfectly honest, though, since this is my first DIY audio design and implementation, I will probably be too enthusiastic about getting something (anything) built to implement BSC in my first project.

Not having anything like proper test equipment is a bit of a hindrance. I lean towards MJK's suggestion that there are additional T-S parameters beyond Fs that need to be properly measured and modeled. However, I don't even know how to observe the electrical impedance as MJK suggests, let alone measure the T-S parameters of this particular driver.

Perhaps the lesson here is that novices should design their TL enclosures to conservative standards (e.g. cutoff >5Hz below target x-over) and be extra careful about measurements when translating their designs into reality. With my prototype I was successful on the former but fell down on the latter. Nonetheless, the prototype has performed really well as a center channel for front projection video and given a lot of encouragement for completing the project promptly with 5 properly sized satellites and a powered TL sub. AFAICT there are no commercial full range sat + sub designs for HT and music that present a high WAF, so cracking open MJK's spreadsheets is an excellent way to go. I am definitely looking forward to finishing this project and many others beyond.

Thanks again,

- Skov