Hi guys!
I'm totatlly new in this field, I have done some research to learn the best cabinet I would build for my pair of chinese speaker 4" Aucharm DG-403 (1PCS 2019 New Aucharm 4'' DG 403 Full Range Speaker Driver Casting Aluminum Frame IIR Surrouding Mixed Paper Cone 4/8ohm 25W|Speaker Accessories| - AliExpress) which doesn't have many technical information on the store . So far, I found the TL more interesting, locazating the speaker 1/3 of the legth of the tunel to avoid the dip. So I used python to calculate dimensions given a design that I liked. According to the store the Fs is 72hz, so I end up with these two cabinets, the left tunned to 72hz and the right tunned to 50hz. I got a TDA7377 amplifier and a preamp with ES9018 decoder DAC with 12AU7 tube used probably for buffering. Every sigle piece of this system I found doing researches on youtube and forums, I'm completaly new in this field, therefore, I would appreciate so much any suggestion.
Cheers from Brazil.
I'm totatlly new in this field, I have done some research to learn the best cabinet I would build for my pair of chinese speaker 4" Aucharm DG-403 (1PCS 2019 New Aucharm 4'' DG 403 Full Range Speaker Driver Casting Aluminum Frame IIR Surrouding Mixed Paper Cone 4/8ohm 25W|Speaker Accessories| - AliExpress) which doesn't have many technical information on the store . So far, I found the TL more interesting, locazating the speaker 1/3 of the legth of the tunel to avoid the dip. So I used python to calculate dimensions given a design that I liked. According to the store the Fs is 72hz, so I end up with these two cabinets, the left tunned to 72hz and the right tunned to 50hz. I got a TDA7377 amplifier and a preamp with ES9018 decoder DAC with 12AU7 tube used probably for buffering. Every sigle piece of this system I found doing researches on youtube and forums, I'm completaly new in this field, therefore, I would appreciate so much any suggestion.
Cheers from Brazil.
Last edited by a moderator:
Greets!
Just a quick response for now: to tune a TL, it must be < a 1/4 WL of the desired frequency [Fp] due to its pipe end correction, so will be at least 0.613*effective radius.
Just a quick response for now: to tune a TL, it must be < a 1/4 WL of the desired frequency [Fp] due to its pipe end correction, so will be at least 0.613*effective radius.
Last edited:
Did you model the lines?
Let’s assume Fs = 72 Hz, these are given Qt = 0.59, Vas = 4.8 litre.
dave
Let’s assume Fs = 72 Hz, these are given Qt = 0.59, Vas = 4.8 litre.
dave
Yes I calculated 1/4 WL and used the Sd of 75,42cm2, assuming the value 97-98 is de diameter of the surface area of the cone. I didn't use any othe parameters, should I use?
Thank you everyone!
Thank you everyone!
Last edited:
Sd has little to nothing to do with designing a TL. It implies you ar eusing Classic Design Techniques to design your line, Rarely optimal, usually mediocre or worse.
dave
Ok, thanks.
So I can't model my TL with the parameters they give, so should I just follow their recommendations?
So I can't model my TL with the parameters they give, so should I just follow their recommendations?
You only need Fs, Qt and V as to model a TL. Some use BL but that can be derived.
You have all of those.
dave
You have all of those.
dave
Thanks! From dim memory remember using this, though also remembered it was closer to Qes based on ?? [don't remember, maybe Thorsten L.?], so did some 'close enough' calcs using the attached and HR impedance plots.
At these two extremes of 20 Hz/0.21 Qes/0.2 Qts & 49.6 Hz/0.78 Qes/0.71 Qts it comes closest to [Qes*Qts]^0.5, so in some cases will make the cabs a bit too small, though with the recommended 0.5 ohm min. series resistance [Rs] for wiring, etc., handles the difference in both cases.
I've never found a definitive for Qo, but from the Fostex site, it appears they use it as a ~equivalent for Qt, and from a (very) brief glance, it seems like it was in this, dated '86 also: https://pure.tue.nl/ws/files/3397667/240847.pdf
Quite possible it differed elsewhere though, but I suspect in most cases treating it like Qt will be 'good enough', especially with tolerances etc.
Quite possible it differed elsewhere though, but I suspect in most cases treating it like Qt will be 'good enough', especially with tolerances etc.
Falls in line with my understanding and calculating per the chart for a 0.403 max flat Qts, 0.43 Qes [ignoring [Rs] and with a ~1.01% error due to having to approximate F1, F2 in HR, Qo = 0.4133:
[0.43*0.403]^0.5 = 0.4163 mean or ~6.7% closer to Qes than Qts, so typically use it and 0.5 ohms to insure an adequately large net volume [Vb].
That said, early on, way before I was exposed to Qo, someone or something told me I should be using a 1/2 octave [0.7071x] on each side of Fo for F1, F2 [1.0 octave bandwidth [BW]], which equates to a simpler to calculate 0.4052, so in this scenario it is indeed Qo = Qts to all but the 'x' number of decimal points perfectionist.
Regardless, the Qo math is > Qts, though < Qes, so Qo = Qts is good enough overall if no measured impedance plot is available and use Qes if wanting a little extra 'cushion' of air in the box [Vb].
[0.43*0.403]^0.5 = 0.4163 mean or ~6.7% closer to Qes than Qts, so typically use it and 0.5 ohms to insure an adequately large net volume [Vb].
That said, early on, way before I was exposed to Qo, someone or something told me I should be using a 1/2 octave [0.7071x] on each side of Fo for F1, F2 [1.0 octave bandwidth [BW]], which equates to a simpler to calculate 0.4052, so in this scenario it is indeed Qo = Qts to all but the 'x' number of decimal points perfectionist.
Regardless, the Qo math is > Qts, though < Qes, so Qo = Qts is good enough overall if no measured impedance plot is available and use Qes if wanting a little extra 'cushion' of air in the box [Vb].
OK, now that we've 'hammered' Qo down to = Qts, back to our regularly scheduled program 😉
Converting your TL designs to HR is a bit too tedious for me ATM, so just going to remark that it looks like one size pipe from the closed end, then a stepped transition to a larger pipe [driver chamber], then back down to the original size to its terminus [vent], a design I don't recall ever seeing and while I'm pretty sure I can sim it in HR, haven't figured it out yet, so no clue if there's any audible advantage to it versus [Vs] just making it the same cross sectional area [CSA] its full length [basic TL].
As a general rule [ROT] then, the pipe needs to have a minimum [min.] volume [Vb] of ~ 4x Vas up to ~10x Vas depending on how small the Vas spec is and how much LF gain BW one wants, so CSA will normally be considerably > Bailey's [Sd] requirement for optimal response.
FWIW, for optimal driver TL loading a compression ratio of [Sd]/2 = 2:1 for optimal vented alignment driver specs.
WRT TL path-length, we need to account for the aforementioned pipe's end correction and the stuffing's impact on effective acoustical path-length, so another ROT is desired tuning [Fp]*0.81x for typical damping and 0.707x for max damping with the understanding that at some density [Fp] will begin moving back towards its higher empty pipe [Fp].
All that said, while any taper pipe can be used regardless of its Qts', for drivers with a < ~0.403 Qts' = constant taper [straight TL] and as it moves towards a lower Qts' it ideally needs to be inverse tapered [TQWT] and increasingly expanding [also technically a TQWT, but prefer to describe it as a horn], so with the DG-403's 0.57 + any added series resistance [Rs] it ideally 'wants' a bit of horn loading to offset its relatively 'weak' motor.
For the most accurate alignments based on full T/S specs, MJK's Classic Alignment Tables:
https://web.archive.org/web/20130328134100/http://www.quarter-wave.com/TLs/Alignment_Tables.pdf
Excel worksheet download: http://web.archive.org/web/20120304...om/TLs/Alignment_Tables_Calculator_3_3_09.xls
Qts': Qts + any added series resistance [Rs]: Calculate new Qts with Series Resistor
Converting your TL designs to HR is a bit too tedious for me ATM, so just going to remark that it looks like one size pipe from the closed end, then a stepped transition to a larger pipe [driver chamber], then back down to the original size to its terminus [vent], a design I don't recall ever seeing and while I'm pretty sure I can sim it in HR, haven't figured it out yet, so no clue if there's any audible advantage to it versus [Vs] just making it the same cross sectional area [CSA] its full length [basic TL].
As a general rule [ROT] then, the pipe needs to have a minimum [min.] volume [Vb] of ~ 4x Vas up to ~10x Vas depending on how small the Vas spec is and how much LF gain BW one wants, so CSA will normally be considerably > Bailey's [Sd] requirement for optimal response.
FWIW, for optimal driver TL loading a compression ratio of [Sd]/2 = 2:1 for optimal vented alignment driver specs.
WRT TL path-length, we need to account for the aforementioned pipe's end correction and the stuffing's impact on effective acoustical path-length, so another ROT is desired tuning [Fp]*0.81x for typical damping and 0.707x for max damping with the understanding that at some density [Fp] will begin moving back towards its higher empty pipe [Fp].
All that said, while any taper pipe can be used regardless of its Qts', for drivers with a < ~0.403 Qts' = constant taper [straight TL] and as it moves towards a lower Qts' it ideally needs to be inverse tapered [TQWT] and increasingly expanding [also technically a TQWT, but prefer to describe it as a horn], so with the DG-403's 0.57 + any added series resistance [Rs] it ideally 'wants' a bit of horn loading to offset its relatively 'weak' motor.
For the most accurate alignments based on full T/S specs, MJK's Classic Alignment Tables:
https://web.archive.org/web/20130328134100/http://www.quarter-wave.com/TLs/Alignment_Tables.pdf
Excel worksheet download: http://web.archive.org/web/20120304...om/TLs/Alignment_Tables_Calculator_3_3_09.xls
Qts': Qts + any added series resistance [Rs]: Calculate new Qts with Series Resistor
So I designed a TL Sl/So = 0.5, tunning to 50Hz. But I think the speaker is too close to the wood behind, there is any problem?
Last edited by a moderator: