Applied impedance going pretty low. Time to learn something. Teach me.

maynardg

I'm using a Hiquphon OW1, which has an Re of 5.8 and a minimum impedance of 6.8. However, the impedance sweep drops well below 3 ohms (2.3 ohms) for the full 3-way speaker at about 4K, which can be rectified by increasing the resister in the tweeter XO, though I do not need further padding.

The XO is third order electrical with a 2,2 ohm resister first, followed by a 12uf cap, then a shunt of .15mfd, then another series cap of 15uf. Close to a LR4 target at 2400.

What are the possible causes?

jr@mac

What's the full schema?

DrewsBrews

The smaller the shunt coil the more impedance drops compared to the driver impedance above the xo point. Probably why I don't really use ones that small unless it is a high xo point.

I don't specifically know why, but I suspect it has to do with the coil inductance getting close to the driver's inductance?

maynardg

@jr@mac said:
What's the full schema?

Woofer - 2 X Usher 8955A - 7mh with 100uf shunt

Mid - Seas U18 - 2.7 ohm - high pass is 164uf with 4.5mh shunt - low pass is 1.2mh with 18uf+1ohm shunt.

Tweeter - OW1 - 2.2 ohm - 12uf - .15mh shunt - 18uf

dynamo

I’m far from great at crossovers but even before I read Drewsbrews reply that inductor struck me as small for the size of the caps. In my experience though you are on track with the resistor out front, at least in my experience it’s easier to get a friendly impedance with it there.

ScottS

One conductive path goes from Amp + through the 2.2 ohm resistor, then the 12 uF cap, then the 0.15 mH inductor to ground. That conductive path is essentially a notch filter with a center frequency right around 4 kHz.

DrewsBrews

That would explain why the issue doesn't change much if the 2nd cap is there or not. Cool.

jhollander

To bring up the resistance you would need to bump up the shunt inductor then adjust the cap to get the knee right. Try 2.2 ohms, 9.1 uf, .2 mh and 18 uf.

maynardg

Thanks fellas. I need the .2mH. Or, I'll keep digging around here. Things tend to turn up. Or, since I'm impatient as heck, I'll unwind a .39 I do have. Hate to pay $10 shipping for a $5 part.

I'm not an engineer, really NOT. I'm a psychologist, far from a discipline that would be helpful in understanding.

Can anyone offer an intuitive description of how this works as it does? Scott's comment gets part way there. How does that effect impedance at this point even though the response curve changes very little?

Also, I understand that low impedance dips demand more of the amplifier. That would manifest how? Distortion? If the amp is not approaching clipping, does it lose some control of the drivers earlier?

Anyway, thanks again for getting me on track for a practical modification.

Wolf

The resonant circuit created by the 2 components will result in a low impedance.

DrewsBrews

I think it is a case of constructive interference vs destructive interference depending on the tune of the circuit. One way the impedance will rise (high-pass/low-pass), the other way the impedance will fall. A "Series LCR" notch is in parallel with the driver and will shunt the frequency to ground that it is tuned to within the Q. In your case the values begin to line up with a series LCR tune.

I don't know much of the minutia regarding amplifiers.

a4eaudio

Nobody has mentioned this, so maybe I'm way off. But what about playing with resistor placement.... Maybe the resistor in series after the components or a resistor in parallel as part of an L-pad? I know a resistor in series after the shunt coil can help with minor phase adjustments, does this help with raising the impedance at all?

ScottS

In this case, moving the resistor after the components would be a bigger problem, because then the connection between amplifier plus and ground would drop to almost zero ohms at the resonant frequency.

What you really want is a smaller cap to replace the 12 uF and a larger inductor (don't start unwinding your 0.39 yet) to get a similar cutoff frequency. The problem is that you will wind up with a sharper rolloff which you will then have to mitigate. The mitigation could be a resistor in series with the inductor in the shunt leg, which will solve your impedance problem. Then you'll likely need to tune or remove the other cap.

jhollander

The resistor, cap, inductor, and cap create a filter transfer function that can be compared to another transfer function with different component values. You can get the same values I came up with using any tweeter impedance file in XSim.

maynardg

Thanks again everyone for getting me on a path to a practical solution. I still use Winpcd and the model and measurements track really well. I ended up with 8.2uf, a 0.22mH shunt, and a 22uf. This nicely maintained the frequency response and brought the impedance up to a safer level. I was allowing myself to get stuck working with the 0.15 coil, because I had it. Scott, I unwound that 0.39. :-)

ScottS

I'm glad you found a solution. I just didn't want you to unwind it without trying some other values for that inductor. I speak from experience...

PWRRYD

@maynardg said:
I was allowing myself to get stuck working with the 0.15 coil, because I had it.

I've found myself there before too

jr@mac

I've been there before - get set on using a certain value component and pull my hair out trying to force an outcome.

Now I just stock a couple of pretty much every value available.