first sub -- cheap style

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
I just finished building myself my first subwoofer. Cheap *** style. I got two 6 1/2 inch subwoofer drivers from my old busted computer speaker kit. I bought an 200 Watts MTX car amplifier on eBay for $35, and a handful of medical-quality power supplies (the things you find on ebay!), totaling 160 Watts for 40$. I found a 50 year-old pine wood box being discarded from a dead man's apartment. I brought it in, took it apart, and rebuilt it airtight.

Altogether, I have something that does 50 Hz @-3 dB and cost less than $100 to make. yay.

164716575_c70872d0d2.jpg


164716579_ef5349a871.jpg


The drivers were gutted out of two Altec Lansing ACS45.1 kits. I measured the parameters of the mystery drivers myself, and pluged the results in Winisd. I was working with a 58 litters box, so I decided on a vented design with two 20 cm long ports, 2 inch in diameter each. Winisd predicted the minus -3 dB at 35 Hz.

The response curve I am measuring does not match the prediction at all. I am attaching a graph with both the predicted and the measured curve. I photoshoped the measured response curve (in blue) on the attached winisd screenshot.

I'm not sure how to explain this difference. Are the drivers just bad? Did I make a mistake measuring the parameters? Maybe the Radio Shack SPL is just crappy. Room effects? Or maybe it is something else I'm not thinking of.

What do you think? What should I try first to improve the response curve?

Params follow:
Fs = 34.0
Qms = 4.711
Qes = 0.551
Qts = 0.493
Cms = 0.000888
Vas = 34.38
 

Attachments

  • winisd.png
    winisd.png
    7.4 KB · Views: 656
First off, nice resourcefulness. As to the jagged response we'll need the dimensions of the room to be sure but it looks like room modes. Try moving the speaker and turning it at an angle and measuring, or if you can go outside and measure.

Believe it or not in a small room with hard surfaces/walls it is not uncommon to get that response. In addition the low resolution of the Fr test is going to mask some of the true peaks and dips, try doing one with 1-3hz resolution.

Also I didnt see any bracing in those pictures, those "wide" flat baffles are going to radiate alot of distorted sound and at a surprisingly high spl. I would actually recommend you start there.

Of course the box is already too small for the woofers (notice the peak in the winisd curve) so be careful with the amount of bracing you use and also scrounge up some stuffing to increase the apparent volume to the woofers.
 
nunayafb said:
First off, nice resourcefulness.

thanks

Believe it or not in a small room with hard surfaces/walls it is not uncommon to get that response. In addition the low resolution of the Fr test is going to mask some of the true peaks and dips, try doing one with 1-3hz resolution.

ok, I'll try that tomorrow.

Also I didnt see any bracing in those pictures, those "wide" flat baffles are going to radiate alot of distorted sound and at a surprisingly high spl.

Right now there is one airtight middle panel that divide the box into two. I still considering whether I want to turn it into a proper brace. From what I understand, the response of two half-boxes is the same as one big braced box. Right?

Of course the box is already too small for the woofers (notice the peak in the winisd curve) so be careful with the amount of bracing you use and also scrounge up some stuffing to increase the apparent volume to the woofers.

I'll try that next and report.
 
Ok, here is a new graph, take outside this time.

purple is the my previous sub reading, inside the living room, 14' x 15' x 9' large
blue is the winisd prediction
green is today's sub reading, outside
red is like green, with the amp's low pass filter on.

Less jaggy this time. I think I am about to blame the drivers for the rest of the unevenest, and declare it good enough for now.

The peak around 75hz is still a bit odd. I'll try finding some bafle and playing with the tuning.
 
hmmm, looking better as to amplitude of the peaks . Just for the sake of completeness you might try putting some clamps all around the speakers panels as close to the middle as the clamps will fit use blocks of wood if necessary, this will constrain the panels from vibrating and reduce box resonance output. If the box resonance is a significant problem (I suspect it is) then with the clamps it should produce a much flatter and cleaner output. If this happens you will know exactly what is wrong and have the option of more bracing or leave it alone.
If there is a resonance problem one solution would be to pick up some liquid nails and some 1/4" MDF or particle board and glue it to the outside of the box(constrained layer damping) since you suface mounted the woofers it would be easy and look nice. this would then require carpet paint or veneer though, but it will damp out those resonances and stiffen the box up nicely.
Sorry to suggest buying more stuff.

I doubt it is the drivers, to confirm simply remove the baffle and retest fr.

Good luck, have fun
 
Alright, here is another round of tweaking.

Thanks a whole lot nunayafb. Your suggestions are useful. They tell me where to look and what to tweak, otherwise I would be flooded by the multitude of possibilities.

Here what I did:

- Added another pair of power supply, from the same ebay vendor, bringing the total available power up from 80W to 160W.
- I dropped two frabric window dressing I wasn't using anymore into the sub, as stuffing.
- Lengthen the ports from 6 inches to 10 inches
- Fixed a leak in the weather strip (The darker side panel of the sub is not glued in, but held there with strews. A weather strip makes it air tight. This way I can keep on tweaking)
- Moved the sub to the corner of the room.

Total expense: 20$ for the supplies and 2$ for the pipes. Still cheap.

Good news, I am now going down to 23hz. Bad news, there is a +13dB mode around 95Hz. It is also much louder, thanks to the supplies.


In this graph:
dotted light blue is the original
dotted green is the original, outside
purple is the tweaked sub, inside
blue is the same as purple, at higher definition (you can clearly measure the size of the room on this one)

This sounds pretty good. Further tweaking will have to share with the next speaker pair.

I am curious to know what you think is causing the mode.

cheers
 

Attachments

  • tweak2.png
    tweak2.png
    17.4 KB · Views: 307
a couple o' things..

if Fs is 34 hz, you're tuning to 32.5 Hz (calculated via Helmholtz frequency)...precarious. I wouldn't tune any vented box that close to the Fs. think about it.. Also 95Hz is very close to 3*Fs. I think the 95Hz bump is due to the tuning and the enclosure..

Do you have access to large pipe clamps or similar? Try clamping the front to the back and the sides to each other. Or try a ratcheting strap large enough to go around the enclosure. protect the corners from the strap.

95 Hz equates to a 12 ft wavelength..so I think it's not your room. Precisely what are the internal dimensions of the box? You may also try a sealed box if possible (even tape over the vent with a piece of wood and duct tape or similar). Measure again and see the results. Don't worry about the efficiency for now. Are the woofers in series or ll ?Try each thing seperately and note the results. It could be a cummalitive thing. Pretty sure internal height of you enclosure is perhaps in the 4 ft range (or a division or multiple of that). Make sure all adjustments for bass , are flat if your amp has an internal Xover.
 
whats wrong with tuning to FS? many HIGHLY regarded subwoofers are tuned to FS, alot of adire designs that are tried and true tune their drivers at or 1hz away from FS. my sub is tuned 2hz below FS. I dont get whats so bad, is it an impedance thing? i can undestand how maybe the drivers impedance spike, and the ports could maybe overlap and turn intao a massive spike.
 
why so bad?...

I don't have any electrical numbers for the altecs. When I model them in WinISD, I keep coming back to the enclosure being tuned to 29Hz. I've recalculated the Helmholtz frequency 3 or 4 times. The box is tuned to about 29Hz based on the info provided. If the impedence peaks, and the amp cannot drive it or control the drivers they will continue to resonate .

I keep looking at the 95 Hz peak and I keep coming back to either :

a) the box itseld is resonating (and what a very beautiful old looking box it is). so I'd try binding the box with duct tape or clamps (protect the box obviously from either method,as suggested by nunayafb). Feed the speakers a low magnitude signal at 95 Hz and listen (use a stethescope if need be. Increase the power slowly.. listen to your room .

b) speaker placement. If you have the speakers located on a even order division within the room, or worse yet, an even order subdivision of all three dimensions, you excite bass within the room at frequencies that correspond to sub divisions of that dimension, for example..14ft X15 ft is not a very "good" room to begin with as the 2 dimensions (as per nunayafb's earlier suggestions) are quite close. There are free room analysis spreadsheets and programs out there..and "room gain " can account for significant increases in output at lower frequencies.

pull the speaker back out into the room, and measure, (better yet put the sub at about 1/3 of each dimension. This should reduce your standing waves (If the amplitude of the peak got bigger as you got closer to the room boundaries, there's the culprit. )..or keep dropping the input level (or amplifier gain) until it measures where ya want it.

and list goes on. By the way, the RS SPL meter, can be modified inexpensively, to give better results. They are known to read incorrectly. Someplace on diyaudio (I think), mods are listed (or do a search on the net) for it


As per xstephanx's comments. It depends whether you are using low signal power or high power . TS modelling relie on low power signals, but the behaviour changes as more power is applied to the loudspeaker, by about 20%. Unless WinISD has taken this into account the results will be skewed, so yes, a lower tuned frequency "shifts" to a more correct frequency as power input goes up. No secret there.


Everyone need not get too excited about the numbers though. If gmarceau enjoys the sound, regardless of what the theory tells him (or any of us), if it sounds good , it is :)
 
Re: why so bad?...

Nanook said:
Everyone need not get too excited about the numbers though. If gmarceau enjoys the sound, regardless of what the theory tells him (or any of us), if it sounds good , it is :)

unless of course, gmarceau himself gets excited about the theory. Yes the sub sounds great, but that's no excuse to stop tweaking. :)

I found the sound meter mod, thanks, I will look into it.

Can you give me a reference where can read more about tuning around Fs? I have the Loudspeaker Design Cookbook. Is it talked about somewhere in there?
 
LSDC and tuning..

in the 5th Edition Dickenson makes comment on dual woofer formats in 2.140 (Vented box LF Systems) and refers you back to section 1.9 in closed box LF systems for further reading regarding F3, Vb, etc:

Fs remains same
Qts remains as though 1 driver
Vas will be twice that of a single driver (so the Vb needs to be twice as big as a single driver)
Impedence halves if in ll, double if in series
Sensitivity is -3dB if in series and +3dB if in ll
cone excursion will be half of what it would be if using 1 driver( a real benefit as drivers often are more linear in behaviour when not pushed to the limits

and remember if you mount the drivers as absolutely close together as possible without having any part of the loudspeakers (drivers) touching, then a +3dB gain can occur.

you still need to address the room and boxes though.

so if you put the drivers in series, you lose 3dB but can gain it back if you mount the drivers very close together. This is of course assuming the drivers share the same box with not divisions within it except bracing. I am assuming that the woofers ae 4 Ohm, but perhaps they are 8?

As far as the reference to box tuning and changes in impedence (and therefore frequency response), see section 2.115: Dynamic Changes in Frequency Response of Different Alignments in the LSDC. Near the end of that short section is a discussion about the nonlinearities of response due to increased poer levels.

I think this is quite a bit to digest.

gmarceau, I hope this helps. Welcome to the "art" side of loudspeaker design.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.