Print

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

[Jeff Schoonover1]

Here is the design I have in the box right now.

https://photos.app.goo.gl/Qt5koyKLkFLWvkfQ2

[Chris Grimshaw]

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

Hi Jeff,

I'm not sure this is in any particular order, but...

R1 is 1.8ohm.

XSim accounts for driver & crossover phase. A textbook crossover only works properly when both drivers have a flat frequency response and impedance curve around the crossover frequency, preferably for half an octave each way, minimum. When we're dealing with real drivers with impedance wiggles and a non-linear frequency response, a little more creativity is required.

I didn't realise this was for a bass rig. In that case, yes, the tone controls will have much more effect. I'd skip the resistor leading into the HF section.

If you go under "Add Graph", and then "More", you can get a graph of power dissipation by component, and then choose whichever component you like. Double-click on the amplifier to ramp up the voltage. Each resistor dissipates roughly half of what the compression driver would be dissipating, so I'd recommend something reasonably heavy-duty. Maybe aluminium clad and bolted to the SpeakOn plate - handy little heatsink there. That rather depends on how much power you're planning on putting into this cab - if it's just bass guitar, you might well get away with a fairly low-power resistor as bass guitar rarely has much content above a couple of kHz compared to lower down. If it might see use as a full-range cabinet for vocals or monitoring, then I'd make sure the resistor will take at least as much power as the compression driver.

The impedance wiggles associated with the 500Hz peak/dip says to me you've got a cabinet resonance. Maybe a standing wave. It could also be midrange leaking out of the bass port.

The narrow 7kHz peak doesn't show up on the FRD files, so it looked to me like it was a measurement artifact that's since been rectified. Any ideas?

Chris

[Jeff Schoonover1]

XSim accounts for driver & crossover phase. A textbook crossover only works properly when both drivers have a flat frequency response and impedance curve around the crossover frequency, preferably for half an octave each way, minimum. When we're dealing with real drivers with impedance wiggles and a non-linear frequency response, a little more creativity is required.

So it account for phase issues as well.  I understand phase pretty well, and how it relates to frequency curves.  I get that they work hand in hand.  I'm wondering if an improved crossover will actually sound better.  I know the freq. curve is only half of the story, but it seems to me in looking at these graphs that the improvements with optimal over what I have installed won't be immediately audible - at least not unless you could A/B two identical boxes side by side.  Am I wrong?

I didn't realise this was for a bass rig. In that case, yes, the tone controls will have much more effect. I'd skip the resistor leading into the HF section.

Agreed.  There is very little above 5K I'm interested in any way. A little EQ will work if I ever used it for something other than bass.

If you go under "Add Graph", and then "More", you can get a graph of power dissipation by component, and then choose whichever component you like. Double-click on the amplifier to ramp up the voltage. Each resistor dissipates roughly half of what the compression driver would be dissipating, so I'd recommend something reasonably heavy-duty. Maybe aluminium clad and bolted to the SpeakOn plate - handy little heatsink there. That rather depends on how much power you're planning on putting into this cab - if it's just bass guitar, you might well get away with a fairly low-power resistor as bass guitar rarely has much content above a couple of kHz compared to lower down. If it might see use as a full-range cabinet for vocals or monitoring, then I'd make sure the resistor will take at least as much power as the compression driver.

Cool, I'll play around with that.  Nice to know it has that ability.

The impedance wiggles associated with the 500Hz peak/dip says to me you've got a cabinet resonance. Maybe a standing wave. It could also be midrange leaking out of the bass port.

500-ish is very likely with a box this small.  That wouldn't show up on WinISD either...  It could even be bouncing back through the cones.  I took all the stuffing out before I did this new X-Over.  I had put it in there by ear, maybe I ought to measure with it.

The narrow 7kHz peak doesn't show up on the FRD files, so it looked to me like it was a measurement artifact that's since been rectified. Any ideas?

Yep, it may have been a resonance off the floor tile or something.  I will move the box somewhere else, maybe put it on the carpet or something further away or closer to the mic and measure.

[Chris Grimshaw]

So it account for phase issues as well.  I understand phase pretty well, and how it relates to frequency curves.  I get that they work hand in hand.  I'm wondering if an improved crossover will actually sound better.  I know the freq. curve is only half of the story, but it seems to me in looking at these graphs that the improvements with optimal over what I have installed won't be immediately audible - at least not unless you could A/B two identical boxes side by side.  Am I wrong?

Agreed.  There is very little above 5K I'm interested in any way. A little EQ will work if I ever used it for something other than bass.

Cool, I'll play around with that.  Nice to know it has that ability.

500-ish is very likely with a box this small.  That wouldn't show up on WinISD either...  It could even be bouncing back through the cones.  I took all the stuffing out before I did this new X-Over.  I had put it in there by ear, maybe I ought to measure with it.Yep, it may have been a resonance off the floor tile or something.  I will move the box somewhere else, maybe put it on the carpet or something further away or closer to the mic and measure.

Hi Jeff,

I recently had a similar problem - choppy midrange from a ported box. You can see some graphs here: http://www.diyaudio.com/forums/pa-systems/271132-chris661s-pa-system-thread-9.html#post5355295

In the end, it was quite a lot of wool underlay to get the results I was after, but the cabinets sound exceptionally clean now. I think there was around 30mm of material directly between the port and the driver to get the "final" result.


With regards to the audibility of changes to the crossover, I suspect the difference will probably be audible with music playback. With bass guitar, maybe, maybe not. Bass/guitar cabinets are usually chosen for their "sound", rather than having a flat frequency response, so if you're happy with how the cabinet sounds as-is, I'd be inclined to keep it. The second-order lowpass takes out an inductor (with its associated DCR) from the woofer side, so you'll get a smidge more power delivery, but that's about it.

Interested to see if the new measurements work out better.

Chris

[Jeff Schoonover1]

I recently had a similar problem - choppy midrange from a ported box. You can see some graphs here: http://www.diyaudio.com/forums/pa-systems/271132-chris661s-pa-system-thread-9.html#post5355295
In the end, it was quite a lot of wool underlay to get the results I was after, but the cabinets sound exceptionally clean now. I think there was around 30mm of material directly between the port and the driver to get the "final" result.

Hi Chris,
I actually read that very thread about 6 months ago when I was trying to understand why it's no longer valid to stuff LF/sub boxes. I ended up with a 2x18 box with 18-sound drivers which is very nearly flat outdoors to 30 Hz with no stuffing.  I think you're exactly right about issues caused by unwanted sound reflecting back through the ports and cones.  I had stuffed this box by ear before. Subjectively it sounded better.  Time to play around with stuffing and measuring. 

With regards to the audibility of changes to the crossover, I suspect the difference will probably be audible with music playback.

Now I wish I had two of these boxes.  I suspect that while there are small differences visible on the graphs, it'd be pretty tough to hear.  I think going after those big scoops in the midrange will be more audible.  Since I can, do you think I ought to phase align between the high and the mid when using the active x-over in my amp? each driver is on its own channel, meaning I could add delay to either.

Bass/guitar cabinets are usually chosen for their "sound", rather than having a flat frequency response

My rig has always been about combatting that nasty distorted fart-bass made by store-bought rigs once they rise above 85dB outside the store. I want the reproduction to be as accurate as possible, like an ideal PA.  I'll make the 'sounds' with my instrument, preamp, and compressor.

Interested to see if the new measurements work out better.

  Me too.  I'll try to do some more today.

[David Morison]

Hope you don't mind a few thoughts from the peanut gallery...

Regarding the spike at 7kHz, I notice that it is much less pronounced in the FRD file than in the REW screengrab posted earlier. Was the FRD file extracted from that exact measurement, or did you re-measure the cab to generate that perhaps? If the latter, did you notice if the 7kHz spike was still there 2nd time round?
If you measured the cab in a different location, it is perhaps possible that some at least of the spike was some kind of reflection. Likewise, if any of the midrange troughs were reflection/room based, then simply moving the cab could fix them (or indeed maybe just move them to different frequencies!).

I had a play with XSim too, and have come up with a slightly different take.



Definitely not saying it's better than Chris's, but perhaps it'll show some of what is possible (and the tradeoffs involved too).
Working on the assumption that the 7kHz spike is real and needs a bit of help, I've tried to flatten that range a bit more.
This does make the circuit more complicated, and results in a greater impedance swing, so it would be marginally harder to drive if your amp didn't like inductive loads. (Modern solid state amps should be fine, but if you got a retro valve amp at some point, it might become an issue.) It also has slightly higher insertion loss than Chris's.

I've done a lineup of the original circuit, Chris's and mine on the same graph so you can hopefully see the differences a bit easier too.



Cheers,
David.

[Jeff Schoonover1]

Hope you don't mind a few thoughts from the peanut gallery...

Not at all, that's why I posted...

Regarding the spike at 7kHz, I notice that it is much less pronounced in the FRD file than in the REW screengrab posted earlier. Was the FRD file extracted from that exact measurement, or did you re-measure the cab to generate that perhaps? If the latter, did you notice if the 7kHz spike was still there 2nd time round?

I did not notice it.  Yes, it's a different measurement.  Excellent - chalk that up to room anomaly.

If you measured the cab in a different location, it is perhaps possible that some at least of the spike was some kind of reflection. Likewise, if any of the midrange troughs were reflection/room based, then simply moving the cab could fix them (or indeed maybe just move them to different frequencies!).

I am going to measure in a different location in a few minutes.  Also, I'll play with stuffing a blanket in there and see what happens, I will post results.

I had a play with XSim too, and have come up with a slightly different take.
Definitely not saying it's better than Chris's, but perhaps it'll show some of what is possible (and the tradeoffs involved too).
Working on the assumption that the 7kHz spike is real and needs a bit of help, I've tried to flatten that range a bit more.
This does make the circuit more complicated, and results in a greater impedance swing, so it would be marginally harder to drive if your amp didn't like inductive loads. (Modern solid state amps should be fine, but if you got a retro valve amp at some point, it might become an issue.) It also has slightly higher insertion loss than Chris's.

I have tons of solid state power available, so that isn't an issue.  I'm still wondering how much the real audible difference would be.  I know that when I flip from this passive design to the active crossover and bi-amp, there is a difference - it's a bit clearer and defined, but the difference is small and only noticeable with music, and then only if you flip back and forth.  If we're designing for theory and optimal, sure it's good to improve, but for practical use?  I may be to my goal already.
Which brings me to a second realization.  In the first measurement I posted, that 7K spike was there with the active X-Over too, meaning it's a really good bet it's a room anomaly, and not my design.

[Art Welter]

If we're designing for theory and optimal, sure it's good to improve, but for practical use?  I may be to my goal already.
Which brings me to a second realization.  In the first measurement I posted, that 7K spike was there with the active X-Over too, meaning it's a really good bet it's a room anomaly, and not my design.

Jeff,

Your passive crossover is actually reducing the midrange from 2kHz to 5kHz by about 5 dB in both your initial woofer/HF test and the later tests of the HF only. That difference is not a "room anomaly".

The low end of the passive crossover is "safe", which could not be determined by the initial woofer/HF test. The passive crossover's midrange attenuation could be considered a "loudness contour" feature.

"Good enough", EQ to taste..;^).

[Jeff Schoonover1]

So here it is WITH a blanket stuffed in the box:

https://photos.app.goo.gl/W4he4YkTJRiIkvWu2

Here it is without:

https://photos.app.goo.gl/ae6nv2z5FRGNDyJC3

These two are with EQ from about 2.5 K and up.  Without that EQ though, unfortunately, there is still the "nasty 7K" even though I moved the speaker a few feet and facing a different direction.  The good news is that if I get some actual stuffing (instead of my favorite blankie) and play around with it, the lower half can be improved significantly.

[Jeff Schoonover1]

Your passive crossover is actually reducing the midrange from 2kHz to 5kHz by about 5 dB in both your initial woofer/HF test and the later tests of the HF only. That difference is not a "room anomaly".
The low end of the passive crossover is "safe", which could not be determined by the initial woofer/HF test. The passive crossover's midrange attenuation could be considered a "loudness contour" feature.

This is just now, the attenuation between active and passive is about 5 dB across the spectrum.
https://photos.app.goo.gl/jXrQAcXpCRvnNleK2

"Good enough", EQ to taste..;^).

LOL, Okay, point taken

[ProSoundWeb Community]