Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Jeff Schoonover1]

Here are graphs of each side of the X-over separately.  All have 1/12 smoothing applied.

This is with the active crossover engaged.
https://photos.app.goo.gl/WRbCrIePue3m3qoX2

This is with the passive crossover.
https://photos.app.goo.gl/BobnJELYpVaffZX22

This is all four sweeps together. 
https://photos.app.goo.gl/GOy7BeBs2sZvoRi63

I'm working on full range samples, and how to export them as .zma files...

[Jeff Schoonover1]

Apparently, it's quite a chore to make impedance measurements and thus .zma files in REW.  If I gave you the model names of the drivers can it be done?
Supposedly the frequency measurements can be done as exports, though.
See if these work
https://drive.google.com/open?id=1J0Z8C9Xg8qXQsul62WjRpA4rNbj8YG6G

[Chris Grimshaw]

Apparently, it's quite a chore to make impedance measurements and thus .zma files in REW.  If I gave you the model names of the drivers can it be done?

The FRD files work fine.

For the ZMA, you really do need the measured impedance - a HF driver + horn combination will have a specific impedance curve that the datasheets won't tell you. Same for LF driver + cabinet.

Easy enough to make an impedance jig, though. I got a pair of cheap guitar cables, cut them in half, and soldered the correct connections with a 100ohm resistor in the middle. Chop one "head" off and put a SpeakOn on there, job done. 10 minute job to throw that together if, like me, you happen to have boxes full of misc. cables.


I'm not sure if you've posted it, but if you've got a crossover schematic labelled with all the component values you've used (preferably measured values), that'd be great - I can start by simulating what you've got and then suggest tweaks and show before/after curves.


Chris

[Jeff Schoonover1]

The FRD files work fine.

Glad to hear, that part is easy.

For the ZMA, you really do need the measured impedance - a HF driver + horn combination will have a specific impedance curve that the datasheets won't tell you.

This is because of how the driver couples with the horn?  The acoustic load changes the impedance? I know this is true with cones, I just didn't expect it with a CD.

Easy enough to make an impedance jig, though. I got a pair of cheap guitar cables, cut them in half, and soldered the correct connections with a 100ohm resistor in the middle.

* EDIT * Found a graphic, easy enough, I'll make one.

Chop one "head" off and put a SpeakOn on there, job done.

That might be a problem, see the drawing I'll post.  It needs at least three poles to make the right connections.

10 minute job to throw that together if, like me, you happen to have boxes full of misc. cables.

I have millions just waiting for something like this.
So then, do I make this measurement with/through the crossover? On each driver separately or together with X-over? I'm assuming this should be done with the drivers in the box?  Sorry for the hand-holding.  I've noticed that many folks who post instructions about topics like this make it unnecessarily complicated to understand.  I'll keep reading though, your words give me a better picture of what needs to be done, so maybe I can decipher their instructions better.

I'm not sure if you've posted it, but if you've got a crossover schematic labelled with all the component values you've used (preferably measured values), that'd be great - I can start by simulating what you've got and then suggest tweaks and show before/after curves.

Yes, and the values of the caps are very close to stated values. I measured all of them.  The inductors are exact to .1mH.

Switch:
https://photos.app.goo.gl/LqaIuZ0fJNVfAsaq2

Crossover:
https://photos.app.goo.gl/px4Gk9Q5RI9H5CMn1

Note - I omitted the L-Pad in the real thing.

[Chris Grimshaw]

Hey Jeff,

I'd like the impedance of the drivers in the boxes, please. No crossover stuff. Yes, the horn changes the impedance curve. Just like putting a cone driver on a horn, a HF unit will show impedance changes, too.

We all start somewhere. I'm happy to post up screenshots of XSim so you can follow along.

I know what you mean about having a million little projects. I've got some nice Fostex full-range units that need boxes, plus I'm doing some low-profile PA subs, and there's the pile of 2.5" full-range units that'll become a column speaker eventually, and the big pile of amplifiers that need repair work...
It never ends.


In the mean time, try this.

Here are some measurements which you can import into XSim. The HF unit will go down to 1kHz with a minimum 12dB/octave slope. See what you come up with. Assume the amplifier is capable of driving 4ohm just fine, but impedance drops below that should be avoided - in an emergency, you should be able to run two of these on a single 2ohm-capable amplifier.
The midbass driver is 6.5", so crossing over below 3kHz would be good - it starts beaming up there.

https://drive.google.com/open?id=12hWL4d06UqV-RSzMKDr9lpjK0GR-FDPo

My crossover for those looks like this, but try on your own first:


Though I went to 8ohm instead of 7 for the final build as space was tight and the 7ohm resistor was made up of a 4R7 and a 2R2. Didn't change the HF level much.

Once you've got to grips with XSim, it's very easy to make the relevant measurements with REW or similar and import them to design a crossover. When it comes to doing that, my recommendation would be to have some 0.5uF and 1uF caps on-hand, plus a couple of 0.1mH inductors, so you can increment values without having to order yet more components.

Chris

[David Morison]

Here are graphs of each side of the X-over separately.  All have 1/12 smoothing applied.


This is with the passive crossover.
https://photos.app.goo.gl/BobnJELYpVaffZX22

I know they're 30-odd dB down from nominal but it looks like there are at least 2 low freq peaks in the response of the HF driver visible at 30 & 60Hz - those should definitely not be there. Something might still be up with your wiring. Might be worth readjusting the vertical scale of the window to see more of what's going on down there?

Edit - Or: at least check the coherence to see if they might be from some kind of background noise.

[Jeff Schoonover1]

I'd like the impedance of the drivers in the boxes, please. No crossover stuff.

Took me all afternoon to figure out the routing and ASIO drivers with my digital mixer and a little more to make the jig, but I * think * I got the measurements you need.  Check these out.

https://drive.google.com/file/d/1QR9XHpxILlpV1I5DLdJGbVK0ucqLnrAw/view?usp=sharing

In the mean time, try this.

I can just use my own drivers now I suppose, right?  Still, I'm curious to see what you'll come up with now that we know the real values.

[Jeff Schoonover1]

I know they're 30-odd dB down from nominal but it looks like there are at least 2 low freq peaks in the response of the HF driver visible at 30 & 60Hz

I live in a small apartment, it's not exactly quiet.  Those could have been my cat walking by. 

[Chris Grimshaw]

Hi Jeff,

I downloaded the ZMAs, they work fine.

I had a look at your crossover schematic and found it difficult to follow without tracing it all out, so I ditched it and started from scratch.

Here's attempt 1:

4th order highpass, 2nd order lowpass. The 2ohm resistor on the HF side drops the HF level by a dB or so. I set off with a passive crossover calculator to get some rough values, and then deleted half of it and re-tuned the rest. Results weren't good to start with.
After a while, I got this.



I wasn't too happy with the overlap region. While the two bands do seem to sum together quite well, the overlap is coming up to around an octave, which will give some weird polars.

So, attempt two.
Third order highpass, second order lowpass. 3rd order is much easier to get right than 4th order, and still offers a useful amount of tweeter protection. A second order highpass should be avoided, IMO, as a 12dB/octave slope produces constant excursion as frequency decreases. Of course, when you add in the acoustic slope of the tweeter things will get better. Still, it's quite a lot of out-of-band energy.

I did try higher orders on the LF section, but invariably found that cutting out the extra components made life easier.



A couple of notes on this one:

R1 drops the Q of the crossover a little. Without it, there's a bit of a peak from the interaction between the inductor and capacitors. Limiting the low-frequency short-circuit that the inductor presents smooths this out quite well.
R2 is an optional 2ohm resistor that alters the HF level. I'd bypass it with a reasonably heavy-duty switch and call it a "bright" switch. The level difference is around 2dB across the kHz band. If you've got a fairly "dead" room, then shorting that resistor will bring the HF level up a bit, and conversely, if the room is quite "lively", then dropping the HF level a touch will be useful.
Both frequency response curves (with/without R2) would be perfectly acceptable IMO, and should be chosen according to the use case.

Both R1 and R2 see up to 50% of the power going into the HF driver, and should be sized accordingly.

Chris

[Jeff Schoonover1]

I downloaded the ZMAs, they work fine.

Excellent!  It was a good exercise learning how to make it happen.

I had a look at your crossover schematic and found it difficult to follow without tracing it all out, so I ditched it and started from scratch.

I can get you an actual schematic if you like.  That was a layout map factoring the physical sizes of the components and hookup leads to the switch and input.  The x-over is exact to the online calcs, they're both LR-4 with passband @ 1K3.

So, attempt two.
Third order highpass, second order lowpass. 3rd order is much easier to get right than 4th order, and still offers a useful amount of tweeter protection. A second order highpass should be avoided, IMO, as a 12dB/octave slope produces constant excursion as frequency decreases. Of course, when you add in the acoustic slope of the tweeter things will get better. Still, it's quite a lot of out-of-band energy.
I did try higher orders on the LF section, but invariably found that cutting out the extra components made life easier.
A couple of notes on this one:
R1 drops the Q of the crossover a little. Without it, there's a bit of a peak from the interaction between the inductor and capacitors. Limiting the low-frequency short-circuit that the inductor presents smooths this out quite well.

What is the value of R1?  I can't quite make it out.  Are there any phasing issues with using a 3rd order and a 2nd order?  Is that accounted for with X-Sim?
  See how on the first measurement pix I sent that BOTH active and passive x-overs result in that 7K peak?  Why would the active do that also, if it was an issue with the passive X-over?
See the nulls at about 250 and 500?  I was kinda guessing that was room response or the box, but it looks as though those happen on your impedance graphs too? Will that need to be eq'ed?

R2 is an optional 2ohm resistor that alters the HF level.

I think that since this is a bass rig, the tone controls on the bass or preamp will have vastly more influence on the signal.  Still, it's interesting to note.

Both R1 and R2 see up to 50% of the power going into the HF driver, and should be sized accordingly.

Is there a way to see a ratio or percentage of the power division?  I know the highs see a small fraction of the lows, but those resistors still may need to be quite substantial in power handling.

[ProSoundWeb Community]