Severe distortion in home-built speaker

[murphyslabrat]

I have been working on this project for a few weeks now (as evidenced by my numerous posts), and I finally got the materials and everything all together, and built a speaker. It has an okay frequency response graph, quite biased towards the low-end, but I'll dig. The problem is that it has severe distortion across the board, to the point of certain tones being unrecognizable (2khz sounds closer to 3khz). I am going to describe the driver's construction below, and I would like to hear any suggestions on improvements or specific changes for my next prototype.

It is a regular canning jar lid (I think it's 2.5" diameter). I used molded a silicon pad into the diameter of the back and a 1/4" thick, to help reduce vibrational effects on the frame, and for sound-dampening. It is just jammed in the back, and can be removed. The diaphragm consists of two strips of packing tape (2" wide) crossed to completely cover the top (the end with the smaller opening). The coil is wound around the bottom end of a packing-tape tube (about 5/8" diameter, two and a half winds of tape). I cannot give a count on the number of turns in the coil. This tube has the top 1-1/2" cut into strips that are used to affix the coil to the diaphragm, leaving it dangling a little above and off center of the ring magnet (neodymium n-42, 1-1/4" OD x 3/4" ID x 1/4" thick). The coil is not perfectly round.

My first thought was make a spider by securing the tape to a 2.5" O-ring, and then secure the ring using silicone caulking (like I used for the backing).

So, any thoughts?

--P.S.--
It was unbelievable when I first heard it make noise. I haven't had that kind of exhilaration since my first "hello world" project in C.

 

[bobdole369]

Perhaps the coil is saturated (flux)? What is your core? Should be steel (or some other highly magnetic material) and the neo mag goes on the outside of that. That usually causes a phase shift or muddiness though?

 

[sjwaste]

Woah, can you post a picture of this? I'd love to see it.

 

[murphyslabrat]

actually, I don't have a "core". What is the purpose or function of a core, and what would the typical dimensions be relative to the magnet?

What would saturation of the coil do? As far as I understand it, if the coil is saturated, it will just have a higher resistance than is necessary, and be heavier than need be.

A little "muddiness" wouldn't hurt, as the frequency response between 300hz and 2Khz is like twice high as 7khz to 10khz, with a downward slope (more or less, it isn't smooth) in between. Though, the cool part is that the decline below 300hz is really gradual, and it is downright audible as low as 5hz. At this point it is has more distortion than GoreVision Glasses (WARNING: Attempted humor)

pic one: http://www.flickr.com/photos/21566228@N07/3448075642/
pic two: http://www.flickr.com/photos/21566228@N07/3448075538/

 

[bobsmith1492]

Hm, you need to work on your picture skills!

Measure the resistance across the coil; I would guess it is very low and your amplifier cannot drive it properly. You typically need >3ohms DC plus enough reactance to increase the impedance up to 4ohms.

What amp are you using? If it's a typical stereo, it might need more like 8 ohms (probably >6ohms DC plus reactance).

 

[herm0016]

i am going to guess that one problem is the diaphragm. it appears that you have built a flexible diaphragm, a good speaker needs a diaphragm that is very rigid with a very low resistance rubber or fabric surround that connects it to the frame of the speaker. the distortion is probably from the fact that every radius of the diaphragm will be moving with a different velocity and acceleration, giving you many different frequencies and phases.

 

[murphyslabrat]

It's plugged directly into my laptop's audio-out. I would estimate the coil as about 10' long, putting it pretty close to 15 ohms. Unfortunately, I don't have a multi-meter, so I cannot say for sure.

And, regarding the pictures, I thought they were pretty good, considering I'm using my laptop's web-cam.

--EDIT--
In response to Herm0016,
I understand. Packing tape seemed to me the most expedient diaphragm material. Would you suggest another material? Maybe signboard?

And, what would work well for the surround?

 

[herm0016]

anything rigid. I bet the lid from a can would work pretty well. then you could probably still use the packing tape for the surround. you want to keep the mass low but make it very rigid, you also want to make the surround very flexible. in the direction of the excursion of the speaker, while keeping the diaphragm centered in the frame. what size magnet wire are you using? I bet a 10 foot long piece is a lot lower resistance than you think. you really should pick up a cheap multimeter, you can usually find them for 5 to 10 dollars.

 

[murphyslabrat]

I'm using 37 AWG wire...and you're right, I was mistaken. 10' should come out to about 5 ohms resistance, as apposed to 15....meaning that, to get to 32 ohms, I would need more than 60' of wire.

For the diaphragm, what if I just put a ton of layers of packing tape together, would that be sufficient?

--Oh, and for the surround, you mentioned fabric. What if I cut up an old t-shirt for the surround. I could cut a ring out of the fabric.

 

[herm0016]

the t shirt would probably work great. more layers of packing tape will increase mass quickly, while adding little rigidity. a section of hard plastic, or metal in the tape layers may work fairly well.

 

[f95toli]

Originally posted by: murphyslabrat
It's plugged directly into my laptop's audio-out.

The audio-out from your laptop is probably a line-out output; it is not meant to drive a speaker directly (headphones are a bit different since they have much higher impedance than a speaker and do not require much power).
It could be that you are basically shorting the output which would explain some of the distortion, especially if the low impedance makes the short-circuit protection kick in (which is likely in this case).

 

[CycloWizard]

Originally posted by: f95toli
The audio-out from your laptop is probably a line-out output; it is not meant to drive a speaker directly (headphones are a bit different since they have much higher impedance than a speaker and do not require much power).
It could be that you are basically shorting the output which would explain some of the distortion, especially if the low impedance makes the short-circuit protection kick in (which is likely in this case).

This is a very good point. It's unlikely that the audio out has enough current to drive a coil.

 

[bobsmith1492]

At 5ohms, you could hook it to the output of a power amp made for 4-8ohm speakers. That would eliminate potential for distortion in the amp which, if it's a laptop output, makes me think that was the reason.

Is this by any chance the same laptop output that can't drive your headphones as loud as you wanted?

 

[murphyslabrat]

yeppers, the same one. The headphones in question were junk, and my solution was to just get a new pair (A friend's used AKG K27i's, $30 ^_^). However, regarding this project, I figured it would be pretty simple to just build another model. I could stretch wire across the room, which I know is 60 feet long, resulting in a length of wire with 30 ohm impedance. I have some homework to do tonight, but I could get to it either this evening or tomorrow morning.

--EDIT--
I got to it, and it still had distortion issues. I layered 8 strips of packing tape, and cut it into a circle, and placed that in the center. Unfortunately, I made the coil too narrow (I overcompensated for the larger coil diameter, and made the tube it wraps around too narrow), so it is a lot quieter than my previous build.