Hi Gary
Do an Impedance Sweep.
1) two woofers in parallel only
2) two woofers in parallel + Low Pass Filter
Mike
Do an Impedance Sweep.
1) two woofers in parallel only
2) two woofers in parallel + Low Pass Filter
Mike
What are you using for inductors? Many, many years ago, I wound inductors for a crossover, and had really nasty distortion because of the core material.
Was the tweeter crossover still wired up for this measurement? (ie did you just disconnect the tweeter)
If so, you might try measuring again with R6 or C7 removed to rule out problems due to inductive coupling feeding HF content from L8 into L2.
If this is the cause of the distortion, you should orient the inductors to minimize coupling as shown in Fig. 7 here:
Inductor Coil Crosstalk Basics | Audioholics
Plots above are only for left speaker, but the right is nearly identical. Seems unlikely both C4 caps could be bad.
And of course, I don't have a replacement on hand.🙁
There will be a fair amount of coupling with coils oriented as shown in the photo.
Closely spaced air-cores need to be oriented nearly exactly at 90 degrees to avoid coupling.
OK. I'll move them around and see what happens. Small one will be easy to rotate. Could put everything on a longer board to increase distance too.
Rotated L8 so it's 90 degrees vertical and horizontal relative to L2. That also moved it about 3 cm further away.
No change in plot. Still < -20 db THD.
Last edited:
Definitely change the orientation of the vertical coil so that it is at 90 deg to the other one! See troels tests on coils --> Placement of coils in crossover networks.
For the electros, I would consider replacing them with polyprop film caps (what did zaph originally use?) If zaph used electros you may also need to add some series resistance. perhaps .1/2 ohm to 1 ohm. 30uF axon's are $6.95 from parts express https://www.partsconnexion.com/capacitor_film_axon.html 🙂 that's probably close enough, or you could add a 1uF in parallel..
edit: just noticed the comment that the change in coil location made no difference 🙁
Tony.
For the electros, I would consider replacing them with polyprop film caps (what did zaph originally use?) If zaph used electros you may also need to add some series resistance. perhaps .1/2 ohm to 1 ohm. 30uF axon's are $6.95 from parts express https://www.partsconnexion.com/capacitor_film_axon.html 🙂 that's probably close enough, or you could add a 1uF in parallel..
edit: just noticed the comment that the change in coil location made no difference 🙁
Tony.
Note that the resonant frequency of the cap and the coil is 673Hz. Try putting a 1 or two ohm resistor in series with the electro and see what the effect is.
Tony.
Tony.
With something that's NOT a "bipolar electrolytic".
Short term fix - if C4 is an Electrolytic, try biasing it with a 9V battery (carefully!) and remeasure
See here, here
See also "Capacitor Sound" by Cyril Bateman, published in the Electronics World in July 2002 - January 2003, and the artices by Jung & Curl
Another thought. Can you do a distortion measurement at the speaker terminals. You will need a suitable attenuator so that you don't fry your sound card.
Also if you have another amp available try doing the acoustic measurements with another amp. It may be that your amp does not like this particular load.
Tony.
Also if you have another amp available try doing the acoustic measurements with another amp. It may be that your amp does not like this particular load.
Tony.
Wintermute ----
"For the electros, I would consider replacing them with polyprop film caps (what did zaph originally use?)"
John spec'd the same caps and coils I'm using. I ordered the parts from his list.
Zaph|Audio - Bargain Aluminum MTM
"If zaph used electros you may also need to add some series resistance. perhaps .1/2 ohm to 1 ohm."
I don't understand, theoretically, how either a cap or coil could introduce THD into an AC signal. How would increasing resistance reduce it?
"30uF axon's are $6.95 from parts express . . ."
Are electros more prone to distortion than polys? Why?
"edit: just noticed the comment that the change in coil location made no difference 🙁"
Nope. Not even 1 db.
"Note that the resonant frequency of the cap and the coil is 673Hz."
How did you determine that? What should it be?
"Try putting a 1 or two ohm resistor in series with the electro and see what the effect is."
I can try that.
"Another thought. Can you do a distortion measurement at the speaker terminals. You will need a suitable attenuator so that you don't fry your sound card. . . . Also if you have another amp available try doing the acoustic measurements with another amp. It may be that your amp does not like this particular load."
I don't have an attenuator, but I do have a couple other amps. Will try it.
Someone else suggested doing an impedance sweep. Perhaps the impedance is dropping very low in that 100-500Hz band and eliciting distortion from the amp. Will do that today.
Thanks for the suggestions!
"For the electros, I would consider replacing them with polyprop film caps (what did zaph originally use?)"
John spec'd the same caps and coils I'm using. I ordered the parts from his list.
Zaph|Audio - Bargain Aluminum MTM
"If zaph used electros you may also need to add some series resistance. perhaps .1/2 ohm to 1 ohm."
I don't understand, theoretically, how either a cap or coil could introduce THD into an AC signal. How would increasing resistance reduce it?
"30uF axon's are $6.95 from parts express . . ."
Are electros more prone to distortion than polys? Why?
"edit: just noticed the comment that the change in coil location made no difference 🙁"
Nope. Not even 1 db.
"Note that the resonant frequency of the cap and the coil is 673Hz."
How did you determine that? What should it be?
"Try putting a 1 or two ohm resistor in series with the electro and see what the effect is."
I can try that.
"Another thought. Can you do a distortion measurement at the speaker terminals. You will need a suitable attenuator so that you don't fry your sound card. . . . Also if you have another amp available try doing the acoustic measurements with another amp. It may be that your amp does not like this particular load."
I don't have an attenuator, but I do have a couple other amps. Will try it.
Someone else suggested doing an impedance sweep. Perhaps the impedance is dropping very low in that 100-500Hz band and eliciting distortion from the amp. Will do that today.
Thanks for the suggestions!
Last edited:
Early on I said something about dividing the enclosure to isolate the drivers. The shape of yours isn't conducive to that strategy tho would help regardless of what others claim.
I just noticed that all the prominent distortion components are odd (ie 3rd, 5th, 7th, 9th).
Perhaps this indicates a current limiter/protection of some type in the amplifier being activated, flat-topping the signal. Is there a lower volume setting at which the distortion suddenly jumps to a lower level?
Perhaps this indicates a current limiter/protection of some type in the amplifier being activated, flat-topping the signal. Is there a lower volume setting at which the distortion suddenly jumps to a lower level?
Yes. I remarked on that in the original post.
No (unless that threshhold level is very low). I tested at several different levels and mic gain settings. THD pattern is the same.
Also just tested using a different amp. Pattern is identical.
Hi Garry, The comment about film caps and electrolytic's is basically based on electro's being more prone to causing distortion, admittedly though the levels you are seeing, would indicate faulty caps (if indeed it is that caps that are causing it) rather than something inherent to electrolytics in general.
The main advantage of using film caps is that they tend not to degrade over time (which electrolytics do).
Electrolytic capacitors have a much higher ESR (effective series resistance) than film caps do. a 33uF electrolytic may have an ESR of around 1 ohm. A 33uF film cap may have an ESR of 0.01 ohms.
When you put a coil in series with a cap they can resonate (not desirable in a crossover as it can cause response peaking). Some resistance in series with the cap can damp this resonance (desirable). Electrolytics naturally have this resistance built in, but film caps do not (and therefore may need it added in the form of a series resistor).
It's probably unlikely it is a resonance issue as it seems to be pretty constant across the board below that freq rather than centered around it, but I thought I'd make the suggestion just in case.
Do do the impedance sweeps. They can show up problems. If you haven't seen it check out Claudio's page for an easy impedance jig. Cables I've started using REW (which you already are familiar with) for doing impedance sweeps. It works very nicely. I'd used speaker workshop previously. From my similar thread --> http://www.diyaudio.com/forums/mult...ource-distortion-peak-my-mtm.html#post3582441
Tony.
The main advantage of using film caps is that they tend not to degrade over time (which electrolytics do).
Electrolytic capacitors have a much higher ESR (effective series resistance) than film caps do. a 33uF electrolytic may have an ESR of around 1 ohm. A 33uF film cap may have an ESR of 0.01 ohms.
When you put a coil in series with a cap they can resonate (not desirable in a crossover as it can cause response peaking). Some resistance in series with the cap can damp this resonance (desirable). Electrolytics naturally have this resistance built in, but film caps do not (and therefore may need it added in the form of a series resistor).
It's probably unlikely it is a resonance issue as it seems to be pretty constant across the board below that freq rather than centered around it, but I thought I'd make the suggestion just in case.
Do do the impedance sweeps. They can show up problems. If you haven't seen it check out Claudio's page for an easy impedance jig. Cables I've started using REW (which you already are familiar with) for doing impedance sweeps. It works very nicely. I'd used speaker workshop previously. From my similar thread --> http://www.diyaudio.com/forums/mult...ource-distortion-peak-my-mtm.html#post3582441
Tony.
Last edited:
- Status
- Not open for further replies.