 |
|
|
|
HASIB'S ARTICLES AND AUDIO PROJECT REVIEWS
3-WAY MAIN SPEAKER SYSTEM (FOR TRI-AMP SEPAKER SYSTEM) Introduction I spend many time building loudspeaker systems using passive crossover but never satisfied by their performance. Then I found an article about bi-amping technique on Rane Corporation. I spend many years “investigate” what is bi-amping means accurately, the procedures, techniques, theories, circuits needed etc. I found ESP page and I think the author didn’t left any points about beauty of bi-amping/tri-amping technique. Here is the project intended for loudspeaker system on tri-amped speaker so there is no passive crossover details. The active crossover (3-way 24dB/octave Linkwitz-Riley) that I currently used also available from ESP page. I choose premium drivers and do some measurements for their frequency response to know what exactly the optimum frequency that need to be cut on active crossover. Also using my only measurement software (Sound Forge) I’m able to determine the drivers sensitivity, not accurate but the idea is there. I measured Small-Thiele parameters for bass and midrange driver. Tweeter driver not measured. The Drivers I choose very premium 10” woofer that has an impedance of 8-ohm/200-watt (?) power handling. Here are the woofer driver parameters that I measured myself:
For box simulation I use WinISD beta Version 0.44 (by Juha Hartikainen). This is freeware software and recommended to other who decides to do box simulations. I’m very lucky (the speaker retailer don’t have the actual specifications) because after measuring the T-S parameters, the EBP (efficiency bandwidth product) is 41.22 (Fs/Qes) and this is suitable for closed box system. Here is the frequency response based on the box volume I choose:
There is a slight hump around 60 Hz to 110 Hz, I can live with that since it will make the midbass sounds a little louder. Clean internal volume is 80 liters. This is equal to big box. I can tolerate with big box because my listening room is quiet big. I have done frequency response measurement on this driver (not using special equipment, just Winamp and Sound Forge Program). I use sine wave frequency that sweeped from 20-Hz to 280-Hz.
Here is what I have measured (free air frequency response): As can be seen the driver quiet happily produced frequency from 40-Hz to 230-Hz. From the graph shown above, the maximum cut-off frequency for low channel on active crossover must be equal or lower to 220-Hz (it is 100-Hz actually if select one octave down) if flat frequency response desired. The response is not quiet good after 230-Hz. After determining this I have to cut the cut-off frequency on active crossover to 220-Hz (217.5-Hz actually) I use very “conventional” midrange driver from Sony. It is not really a midrange but a midbass driver. I have measured it parameters and here it is:
Here is the frequency response based on my selected box:
Here is the graph when I measured it actual frequency on free air.
The frequency was sweeped from 200 Hz to 4000 Hz and it looks a bit nasty. Using my own method I have determined that 2200-Hz is the optimum frequency (flat response) because the output (220-Hz to 2200-Hz) in linear region. So on active crossover, the selected frequency for midrange driver is from 220-Hz to 2200-Hz. This is the range for “brained frequency”. So tweeter driver must handle frequency starting from 2200-Hz and up. Tweeter driver using 1” soft dome tweeter that has an impedance of 8-ohm, 20-watt power rating and 95dB/w/m sensitivity. Using frequency generator and DVM I’m able to measure what exactly it resonant frequency – about 900-Hz. This is useless parameter for tri-amped system since this parameter is not needed unless for designing passive crossover network. Here is the frequency response that sweeped from approximately 1-kHz to 20-kHz.
The range from 1-kHz is quiet good up to 10-kHz. Output from 10-kHz to 15-kHz is subdued and there is no output at all after 15-kHz. Crossing over the tweeter at 2200-Hz (2175-Hz actually on active crossover) is safe for this tweeter. The Box The box is made from plywood, sorry, no MDF (medium density fiberboard) available in my area. The plywood has a thickness of 11-mm but doubling together to give 22-mm of thickness. Using thin plywood for big box is not an option because the box will not hold together. On doubling the two sheets of plywood I use wood glue and screws. I did not use nails because the joint will not as strong as screws. I use lots of ¾” screws for this doubling process. The box measured externally 112-cm height, 40-cm depth and 32-cm width. This is reasonably big box but cheer up because the assembling is not too hard if you use 22-mm thickness because there will be no combining process. To assemble the joints between the sheets (front, rear, side etc panels) I use long nails, long screws and wood glue. These joints need to be strong. The box designed as mirror image for good imaging and tweeter/midrange driver placed closely to minimize diffraction effect. Tweeter and midrange driver has it own sealed chamber about 12.65 liters in volume excluding the fiberglass, braces and the drivers itself. The tweeter/midrange chamber made from 11-mm plywood sheets because the box is quiet small. Woofer has clean volume about 80 liters after subtracting the volume displaced by braces, the driver and “port”. I know this is confusing since I designed this box as sealed box but there is a 4” port mounted on the front baffle. No it just a convenience to me if I want to “transform” the woofer box to vented configuration by taking out the vent “stopper”. Ok just forget the port. There are two brace panels installed in the box, at ¼ and ¾ quarter by total height. The braces were made from 22-mm plywood and I use wood glue and wood screws to secure them. I don’t want the box rattle if I did not mount the braces properly. I spend about two weeks just for assembling basic box. The combining process takes full one week. My philosophy on building box is measured twice-cut once to avoid the cutted panels goes to scrap. For attaching the front and back panels to main frame I use lots of glue and wood screws. This is very important since large airforce will boxing these two panels. For felt recess panel I separate them in to two sections. Section one is half of bass driver and section two for half of bass driver/tweeter/midrange. The felt recess has a thickness of 11-mm. So bass driver can be flushed mount. It is not really true flush mount technique but close. On front panel I use 11-mm spin-wood attached at every corner to reduce diffraction loss. The rounded side mounted closely at side panels. Then I mount the 11-mm felt recess panels using wood glue and some small nails. I can’t use wood screws here because it needs many works and I must cover the screws with wood filler to keep them invisible. Final Works Just simple solution. The joints that have rough insertion covered with wood filler and then use medium grit of sandpaper finished it up. I use three coats of black wood paint for finishes. Only the front panel sprayed with clear spray. Tweeter/midrange chamber covered with ½” fiberglass on every internal panels excluding for front panel where the drivers were mounted. Bass chamber covered with 1” fiberglass. I use stapler and thumb tacks to hold it. Before mounting the drivers I use double side foam tape behind the driver’s mounting flange to eliminate air from coming out. I don’t use speaker gasket. Holes for appropriate drivers drilled before mount with screws. If you don’t want to use foam you can use clay or something. I use hot glue before and it is a real paint to remove them, if I want access inside the box. The hot glue also will “pull out” the finishes pretty bad. On back panel there are three binding post speaker terminals. I use regular type because my retailer doesn’t have three-way type available. Tweeter terminal on top followed by midrange and woofer. The speaker terminals also need foam tape to avoid air leak. Tweeter/midrange chamber needs a hole for wiring speaker wire. I drill a hole about 10-mm in diameter, run the wires and secure the holes using hot glue to keep the box airtight. This is important because the woofer air pressure will not leak to tweeter/midrange chamber. Speaker spikes? No I don’t use regular speaker spikes but four sheets of thick rubber mounted below box and the result is quiet good. Test Result I'm quiet happy with the result. The sound really forgiving. I have tested the box air tight or not and yes it passed. I don't hear any whooshing air coming out from panel joints, driver sealant and speaker terminals. About my own method to test the box airtight or not - I run about 40Hz sine wave (from frequency generator) signaling a power amplifier and powered the woofer. I put my hand/ear closely at all joints, speaker terminals, woofer baffle and able to detect there is an air leak or not. Unfamiliar procedure isn't it? But it works. The sounds really show up after three weeks - after full break in. The drivers have good combination between them. I don't hear any peaks and dips introduced. Crossing over the active crossover at 220-Hz and 2200-Hz is good crossover frequency with the drivers I used here on this project. Cost for this project is around RM600 and I can say it is a real bargain. The final sound really good and beyond my expectations. That's all folk. I hope I have given you some ideas on this speaker building project. And yes no fancy building materials/techniques available but the idea is there. Completed Project Pictures
Fig1 - 3 Way Electronic Crossover Speaker System ( Front view)
Fig2 - 3 Way Electronic Crossover Speaker System (Rear view)
Fig3 - 3 Way Electronic Crossover Speaker System Author Muhammad Hasib bin Md. Nor June 2002 Any Comment And Question Email to : hasib_tangkak@yahoo.com
|
|
