Home made Subwoofer

Theory and measurements of speaker parameters

     To calculate the column you need to know at least the basic parameters of the speakers. These parameters shaft lie on the internet, but the trouble is that one and all, give implausible parameters for 75GDN-1-4 (for others I do not know). One of the sites I found a scanned passporton this speaker, so I believe these parameters more. It must have reasons, one of them - that such heads produced a whole bunch of plants and quite a long time and Soviet industry, so that over time may have changed settings. But the fact that they have changed as it turned out 2 times! And the parameters of the directories in the calculation of the program, the volume of the box gets 5 liters, it's me and guarded. By the way, almost all the articles I have read have been recommending measures its dynamics parameters. The decision to remove the parameters of its dynamics was made after a week of trouble with calculation programs saba (despite the terrible lazy to do it), which found that these parameters have a profound effect on the size of the box and on the AFC respectively. In order to understand what parameters are needed and what they eat, and how to get them makarom not having at hand intelligent instrumentation and will be discussed in this chapter. Will write briefly who wants to - can read in more detail in the accompanying literature from which the material was collected.

     It may seem strange, but the speaker is mainly characterized by three parameters proposed by Till and Small:

Fs - the resonance frequency in the open space;
Qts - total quality factor of the speaker;
Vas - equivalent volume.

Fs - is the resonant frequency of the speaker without any acoustic design. She measured and - speaker hung in the air as far away from other objects, so that now his response will depend only on its own characteristics - the mass of the moving system and suspension stiffness.

Qts - the ratio of the transfer function of the dynamics of frequency Fs to the transfer function at frequencies where the amplitude-frequency characteristic (AFC) of the horizontal dynamics, ie at frequencies above Fs. In other words, Qts-characterizes the efficiency of the speaker at the resonant frequency.

Vas - volume of air that has flexibility (the reciprocal of the elasticity) such as dynamic and flexible system. When placing the speaker in a closed box (Wl) flexibility of the air inside the box is added to the flexibility of the mobile system dynamics and its resonant frequency changes.There is the following pattern when placed in a box speaker volume Vas its resonance frequency and Q factor Qts Fs increased 1.4 times. Measure these parameters at the first look at the design gemoroyno pretty, but having done it once - all doubts disappear - everything is quite prosto.Dlya First you need to prepare:

   Download a program signal generator for sound card download Marchand Function Generator : 

- You can even download Oscilloscope 2.51 - oscilloscope sound card. Connect the output to    the input - you can see that the generator gets up 

- Find a 1k resistor
- To take the car stereo at least a power amplifier, because it is necessary to strengthen         the signal itself first, and then the measurement signal 

- Take a digital voltmeter desirable Schaub not recalculate and pereschelkivat ranges. I took the analog and digital, and compared the results for verification.

Next you need to collect the following scheme:

Next:
Take a pen and paper

Run the program generator, the volume on the computer set to the middle (otherwise squared off any turns), the rest of the amp will recover.

Connect the voltmeter to the points A and C (ie to the output amp), and set equal to the voltage of 10-20 V at 500-1000 Hz, adjust the volume on the amplifier.

Connect the voltmeter to points B and C (ie, the speaker).

Put on ~ 5Hz generator and retractable speakers away from all objects and walls (can be hung, if I can). Practice has shown that far from the objects lying on the floor and the speaker still gives different readings, but minor, but the purity of the experiment affected.

Changing the frequency of the oscillator look voltmeter reading - we are interested in the maximum and minimum voltage. Approximately near the resonant frequency voltage increases and decreases sharply after. Look at the maximum voltage frequency - this is Fs.

So we already have makarom Fs. Changing the frequency of upward relative to Fs, we find the frequency at which the voltmeter constant and much less than Us (with further increase in the frequency of the voltage starts to increase again). We write this value, Um. Better to repeat the procedure with decreasing frequency. 

According to these records get around this schedule:

Where, Fs - resonant frequency and its corresponding voltage Us. Um - the minimum voltage, U12 is calculated in Excel-e, after entering data in a cell. Once again we turn and look when frequency voltmeter readings coincide with the value of U12, memorize the frequency. Such values ​​should be two, as can be seen from the graph. It will be F1 and F2. Incorporate them into Excel. All - see value Qts. I made two calculation methods for checking the correctness of what I counted :) eventually agreed indications and small difference error calculations.Vas can not be considered, and take from the data, it looks like the truth, and this option does not affect the calculation of the box. In addition, to calculate it you will have to build a plywood box, hard enough and sealed, and then repeat the measurements. If you still have the desire to calculate Vas - read in the source code how to do it. I set myself the goal to create something like the outline of the material that I read, so discard unnecessary, and for a deep study of the question - after all links cite material. So we took the dynamics parameters. For 75GDN-1-4 that's what happened:

Toll dynamics	Reference parameters (from the Internet)
Fs	34	25
Qts	0.46	0.21
Vas	not measured	80

The most important parameter is the Qts. As opposed to you? TWICE! I must say that I downloaded about a dozen guides to the speakers, the other options they have fluctuated, but this was surprisingly stable. What trust then people ...

Go further. Now you have to choose the type of box where all this can be good shove. And then we just come in handy the exact parameters of the dynamics.

The most common types of acoustic design of all three:

Type box	ZakrytyyYaschik  (Wl)  Closed	Fazoinvertor  (FI)  Vented	Bandpass (BP)  4th and 6th order  Bandpass
The selection criterion	Qts <0.8-1.0, optimally 0.7  Fs / Qts = 50	Qts <0.6, optimum - 0.39  Fs / Qts = 85	Fs / Qts = 105
Distinguishing features	This is the most simple to manufacture the type of acoustic design speakers. With the simplicity of design has many advantages, but K.P.P smallest compared to any other type of acoustic design - as a consequence of the need and possible nehily power speaker output failure (from excessive effort :) To calculate the characteristics there is only one parameter - the volume of a box.	In its operating range bass reflex speaker creates quite hothouse conditions, and exactly at the frequency of the oscillation amplitude adjustment is minimal, and most of the sound emitted by a tunnel. Allowable maximum input power here, and the distortion introduced by the speaker - on the contrary, are minimal. Fazoinvertor much more capricious in selecting and adjusting as choice for a particular speaker, already subject to three parameters: the volume of a box, cross-section and length of the tunnel.	Champion efficiency. By selecting the appropriate volume and tuning frequency of the anterior chamber, we can construct a subwoofer with a high bandwidth, but limited impact, ie the bell will be low and wide as you can - with a very narrow bandwidth and high efficiencyin this band. Bandpass - capricious thing to calculate and most time-consuming to produce. But buried inside the speaker - less risk of damage to the speaker and virtually eliminates the need for a bandpass filter (although in practice it turned out - that still desirable)

Wl I immediately dismissed - why do I need a column with low efficiency? Speaker, and so not the most powerful - denomination 50W, 75W maximum. Moreover, the pressure inside the column nehily created, which requires special containment. And parameter Fs / Qts is not suitable for my speaker. However, it Wl and the smallest box of all - it can sometimes be important.

FI and I would not do for three reasons: 1) we need a super smooth hole and then do something to close it, otherwise my child immediately condemn speaker 2) should be put to filter cutoff frequencies above 200Hz, otherwise 75GDN further does not sound cool. 3) normal downturn bottoms was obtained by wild-box size 120-150l (another closet in the apartment), I would have kicked his wife with this subwoofer :) BUT! Fs / Qts = 74, i.e. speaker is most suitable for FI and bottoms gives great while, but that's SIZE 8 (. Here it should be noted that for single bandpassov fit virtually the same dynamics as for the phase inverter. Bandpass approached me and liked the most. Filters do not need to - he filter housing. Speaker hidden inside - not protknesh. And calculations programs showed the best results when opportune size ...

Calculation and design of the box

Calculations showed Bandpassa in relatively good size and decent downturn bottoms, but still decline strongly dependent on the volume and had to compromise slightly reducing box to 65l. Calculations I spent in three programs, so check the correctness of what I built. The results are almost converged. I used WinISD 0.44 , WinISD Pro Aplha and  JBL SpeakerShop or  BassBox (find 10 differences is called). I liked most the first prog, the second was terribly buggy (on it and Alpha), but in some ways useful, third - just confirmed my calculations (it is very awkward interface - bad to change parameters on the fly, choosing values ​​for the size of the cells and Bass reflex, and after each download you need to switch to the metric system). So what happened - looking graphics (take project files can be more):

Here we see compared BP and FI. When equal volumes box FI significantly inferior Bandpassu. And vice versa for large.So if you want to fuck a new wardrobe at home - you should look at the FI. About hump in the middle below. And almost the same in the JBL SS:

Here a little less than the amount in the FI, but still it is 2 times more reduced BP. BP is also scope for critical, can be made large, thus reducing the gap in the middle of recession and improving on bottoms.

Calculation reduces engine capacity in the selection of cameras and frequencies for which they are set, and of checking the AFC. I think the other three graphs you will not worry :).

Dip in the middle is not cleaned well, no - such a speaker :). To increase the volume, I did not want to go, I had to fit into the room, by the way rather big box anyway. But I think such a failure can be ignored - because failure to 3dB is very small (just schedule stretched in height), and given the frequency response of the speaker to 10dB, something about it just be forgotten. Besides, it is still an idealized frequency response, everything in life is much more complicated and confusing :) The failure can be made smaller if narrow frequency, but I wanted to hold on to 200Hz frequency response, which is not entirely successful, but certainly got 150Hz :). I will note here - the middle is still heard in sabe through thick hole so that the active filter does not hurt that I was then performed in amps.

Fazoinvertor payment reduces to the specification of the internal diameter of the pipe in meters and validating the value "Vent mach" to green when turn red - bad - too much air flow, ie it is not a subwoofer, and a musical instrument "pipe." Here it is necessary to look at the length of the phase inverter to it all together in a box, and it is desirable to finally up the middle. I could not be written off in size, for increasing the diameter, for proper airflow, long phase immediately became nemerjannyh. The length and diameter depend on the frequency to which the tuned phase - so the game can be more and frequency. When etoi will change the frequency response, consider this.

First, the sub was made with the same pipes, the benefit of the top, just the program shows less resistance to air, but after assembly at 100W amp and tapping it turned out - after ~ 50W began effect exhaust air (claps) just in the upper compartment (the smallest but ). Had all apart and cut out a big hole under tooolstuyu pipe 105mm inside than I butt almost fit into the camera height - 2cm left. If we consider that there should still vperet silencer - it's very small margin. I use sewer pipe plastic. Here I note that there is a 70mm tube, but they are not as shaft 50mm and 105mm. With the increase of the upper phase, immediately began to work better lower. 50mm really do not recommend to put - for such fools is very small. Bottom line is - for saba dimensions of the elements most important.

Downstairs in the windows lists all parameters dynamics, but it can take in the project files for both programs: for WinISD and JBL SpeakerShop .

Well, the volume defined, it is time to calculate the box and design. As I am doing my work on the debt 3D modeling, so I did - took the SolidWorks software, and built there a 3D model. If you notice - in the design of the site also enabled 3D graphics :) The program will calculate the volume of me. Similarly, do it yourself is difficult because all the connecting bars to eat a decent amount of design and construction had to invent and change on the fly. Another problem was the material and its thickness, and the program allows you to see immediately the size of each element, given as a time sheet thickness and dock with each other, ie I automatically get the size of each element.

Separately about the material. Well about MDF I did not even think, though of course the best option. The task was to find 20-22mmi particleboard, but it was almost impossible. The most common 18mm or 16mm chipboard laminated polish. 16mm small, and laminated on both sides, probably bad for sound, besides get expensive. Then I realized, as did the right thing, not buying laminated chipboard. And the usual 18mm could not be found about 1.5 weeks. Those that were in the shops - it is impossible doperet home for much large sheet. I have phoned all office and walked in all markets. Nerves began to take - I'm almost sorry that contacted the subwoofer, for all but the material box, it was already bought. Began to think about how to glue 16mm particleboard and plywood 4mm, but did not want to glue hard - it does take a special glue and press sickly. And then call from a friend asked me to help bring him home cement. So, while we ran looking for cement and its owners throughout the industrial area, we chanced upon the sale of office and sawing the very notorious Polish laminated chipboard. As the pads in shelves, they lay 22mm chipboard sheets. But the owner was not and had to wait ... Searching again hosts cement and not finding them :) We are back to the CPD. The owner has not given 22mm, arguing that it is difficult to pull out, and they have already caved type (short pull was too lazy) and offered me a 28mm. It's cool, I thought, and refused to looking at these a thick sheets. And then he offered me a 18mm prime polished chipboard - it is used as a transport for laminated sheets (top and bottom). So, all good, with sawing on imported machine, cost me $ 5 (cut $ 0.3 / m). Since he had never drunk - rovnenko, exactly the right size. Draw your own conclusions ...

Yeah, what was I saying? Oh yeah - on the box. Look what happened:

Downstairs in the windows lists all parameters dynamics, but it can take in the project files for both programs: for WinISD and JBL SpeakerShop .

Well, the volume defined, it is time to calculate the box and design. As I am doing my work on the debt 3D modeling, so I did - took the SolidWorks software, and built there a 3D model. If you notice - in the design of the site also enabled 3D graphics :) The program will calculate the volume of me. Similarly, do it yourself is difficult because all the connecting bars to eat a decent amount of design and construction had to invent and change on the fly. Another problem was the material and its thickness, and the program allows you to see immediately the size of each element, given as a time sheet thickness and dock with each other, ie I automatically get the size of each element.

Separately about the material. Well about MDF I did not even think, though of course the best option. The task was to find 20-22mmi particleboard, but it was almost impossible. The most common 18mm or 16mm chipboard laminated polish. 16mm small, and laminated on both sides, probably bad for sound, besides get expensive. Then I realized, as did the right thing, not buying laminated chipboard. And the usual 18mm could not be found about 1.5 weeks. Those that were in the shops - it is impossible doperet home for much large sheet. I have phoned all office and walked in all markets. Nerves began to take - I'm almost sorry that contacted the subwoofer, for all but the material box, it was already bought. Began to think about how to glue 16mm particleboard and plywood 4mm, but did not want to glue hard - it does take a special glue and press sickly. And then call from a friend asked me to help bring him home cement. So, while we ran looking for cement and its owners throughout the industrial area, we chanced upon the sale of office and sawing the very notorious Polish laminated chipboard. As the pads in shelves, they lay 22mm chipboard sheets. But the owner was not and had to wait ... Searching again hosts cement and not finding them :) We are back to the CPD. The owner has not given 22mm, arguing that it is difficult to pull out, and they have already caved type (short pull was too lazy) and offered me a 28mm. It's cool, I thought, and refused to looking at these a thick sheets. And then he offered me a 18mm prime polished chipboard - it is used as a transport for laminated sheets (top and bottom). So, all good, with sawing on imported machine, cost me $ 5 (cut $ 0.3 / m). Since he had never drunk - rovnenko, exactly the right size. Draw your own conclusions ...

Yeah, what was I saying? Oh yeah - on the box. Look what happened:

How do you know this 3D model. Length fazoinvertor seen in the screenshots above programs and is 19cm high and 25cm lower, inner diameters, respectively 105mm and 70mm. In a rear wall of the hole socket connectors. As the legs are manufactured in a factory and the spikes of hardened steel. I chose the size to your liking. In articles found people did spikes 2 times more, and I did not want to put much high subwoofer to the spikes were not visible, because the design and not so low. Spikes need 4, 3 stands I tried - terribly unstable design. It can be dispensed in three columns, as they are not deep, and the center of gravity in front of them. Height spikes dub plate, she needed only one, then checked with a known flat surface (the back wall).

Build box

Cutout holes in the chipboard I omit, we assume that they are cut. I cut out all the jigsaw. Without it, the pipe was probably used to me. Cant they too cool crop, one bad - it is incredibly buzzing.

Assembling the middle and bottom shelves. I pre-rounded top inner edge of the timber. Bruce 20x30mm 45mm long screws attached. For screws I always drillhole advance. The entire assembly was put on wooden nodes EXTRA PVA - PVA is thicker for a stronger bonding. All angles are greased with silicone sealant for wood - it is harder than usual for glasses, etc. But I think it is possible and in any.

Fasten the sides to the inner side of the bars with 35mm screws. Middle shelf should vymeryat quite smoothly. And do not forget the glue. To fasten the front side walls and timber. Obtained beams, front, around the perimeter of the large chamber.

Glue the front of the phase inverter. I pasted them with liquid metal - such multicomponent solid garbage, the smell as the epoxy. She kneaded her hands, turns like plasticine - soft and terribly lipnuschaya. Hardens almost instantly as soon cools, and if not cool anyway hardens after 5-10 minutes, so with it I shot it with a pumice stone and then with the skin :) I've been doing around the radius of the bulge at the base to secure the - they still stick away. In addition, the ideal is precisely cut hole jigsaw I personally did not work, and these are just wonderfully uneven Lead liquid metal. Tubes I wrapped windowed paper tape in layers 5-7, very tight all smoothing. Then paste over them with linoleum with insulation. Thickness mm commercials phase turned 7.

Fasten the front wall to the inner side of the bars 35mm screws.

At the top of it to fasten the corners for greater rigidity. I used plastic furniture with a lid.

Install the speaker on the sealant, bolted (can not remember the size). Coat with a good circle - on white photograph shows sealant. Fluster wires. In the rear wall I attached connector assembly. Now all this must okleit sound absorber. Many material options. I used foam insulation with linoleum, and then, in the end, thick sintepon ~ 2.5-4cm, which in addition to pasting, also fluffed up into a large chamber. To the top of the rear wall and also have to glue the linoleum. You can paste over several layers. Good stuff thick batting, but to get it I could not.

Put the top cover back on the sealant, no glue - then again they might have to shoot. I screwed them furniture coupler Allen - great gizmo attaches extremely durable, important not to disrupt, if tightening the drill as I am. Crepe-to-end, there must be careful with DSP - it can and flake. Ties for first drilled one long hole, and then bolted to the hole increases more drill. Two tie front top in the front wall, or rather the rest of the top cover and the rear just around the perimeter and in the center. As can be seen on ledges attracted sealant and attracted well.

More photos on the first option, with a small upper phase inverter and synthetic padding has not pasted. In the future, I papered with linoleum and bars too. It was, so to speak, the first test build, so listen - sounds or not. Rang. But yet I have only had power to 25W, it was fine until ....Then I threw the whole thing and began to gather on the 100W amp, it was necessary to check the design to the maximum possible. And not for nothing I feared more powerful amplifier - and here it was discovered that the maximum upper phase squelching, but the case has been sealed from the first time - no wonder I missed soundly all seams and corners with sealant. I had to disassemble the entire structure, to cut a big hole and put a thicker pipe. Here I have all "insulated" completely synthetic padding. Inside was like in the snow soft, white and warm.What can I say, it sounded quite good - smooth soft bass. At maximum volume amp speaker begins pereklinivaet already, but we must take into account that his denomination 50W. For apartments above the roof missing, the more that it should not be heard in the music, just when his hem - both the volume and the music added such a solid low sound. In short - the good work ...

The entire structure, I plan to paste the film under a tree. But first we must putty and sanded the edges - just as I do not cut on the machine, as I did not try to carefully collect all - all the same total error of 0.5-1mm still present. That's why still poorly laminated chipboard, because he must immediately glue on the end of the tape - in short, will be crooked.

And the feasibility of the construction itself, one could argue, looking at the costs ...

speaker 75GDN-1-4	$ 23.32
Particleboard	$ 5.18
mounting socket (jack)	$ 2.33
fasteners (screws)	$ 3.11
linoleum	$ 5.39
sintepon	$ 3.11
timber	$ 1.04
wire	$ 3.89
70mm pipe	$ 2.59
105mm pipe	$ 2.59
paint	$ 5.13
film	$ 7.77
putty for wood	$ 1.40
glue Moment	$ 1.24
PVA glue	$ 1.30
sandpaper	$ 3.89
spikes	$ 2.59
liquid metal	$ 3.11
trims	I dream about them until
Only	$ 78.96

It's certainly not $ 300. But I note - is inactive subwoofer. For comparison with the purchase, it is necessary to add here preamplifier with a tunable filter, amplifier and phase shifter things at 100W, and about the power supply must not forget (probably the most expensive in it).Since I decided to collect power to 6 channels at once, I did not bother inserting amplifier subwoofer. Still inside the body vibration is large, and the tightness of the box with ventilation amplifier combine hard. General power, it's a thankless job - terribly much trouble. I do this without showing heroism, maybe. 

Original sourse cxem.net