Hello,
I am building a coaxial speaker with the following drivers:
Eminence Beta 12CX
PRV Audio D290Py-S
I built large cabinets with the following specs:
Vb - 3.67
Fb - 37.8Hz
F3 - 43.6
Ports - 2
Port diameter - 4"
Port length - 11'
For the crossover I used the following thread on this site to get started: https://www.diyaudio.com/community/...igning-crossovers-without-measurement.189847/
I used a 1000Hz crossover point. Based on the calculations I ended up with the following crossover (at this point I hadn't reversed the tweeter polarity as shown):
At this point I built my cabs and crossover, not knowing how to use xsim yet. I also used a L-pad knob to bring the tweeter level down, as there is a 12db difference in sensitivity. In general I LOVE these speakers, the horn driver especially sounds incredible...highly recommend... but due to a room mode I started making speaker measurements using HouseCurve and noticed a big dip at the 2000Hz region, to which I reversed the tweeter's polarity and got much better results. However I now get a large db bump from the 500-2000Hz region. I turned the tweeter knob down all the way which still left a 500Hz spike so that got me thinking its not just the tweeter level. I eq'd it flat and it now sounds incredible...but I would really like to get the crossover dialed in so I started poking around with xsim and finally got the basics down. I created measurement files for both drivers and got the following freq response which seems to line up with the problem I am facing:
As you can see my levels start to peak before my crossover point....so I started playing with my crossover. I ended up zeroing in on lowering the parallel resistor on the tweeter to 1ohm and it seems to be the kind of response I am looking for. Is this going to work? The driver responses look way off but the series curve (blue line) looks linear and nice...am I doing this right?
Thank you for any feedback 🙂
I am building a coaxial speaker with the following drivers:
Eminence Beta 12CX
PRV Audio D290Py-S
I built large cabinets with the following specs:
Vb - 3.67
Fb - 37.8Hz
F3 - 43.6
Ports - 2
Port diameter - 4"
Port length - 11'
For the crossover I used the following thread on this site to get started: https://www.diyaudio.com/community/...igning-crossovers-without-measurement.189847/
I used a 1000Hz crossover point. Based on the calculations I ended up with the following crossover (at this point I hadn't reversed the tweeter polarity as shown):
At this point I built my cabs and crossover, not knowing how to use xsim yet. I also used a L-pad knob to bring the tweeter level down, as there is a 12db difference in sensitivity. In general I LOVE these speakers, the horn driver especially sounds incredible...highly recommend... but due to a room mode I started making speaker measurements using HouseCurve and noticed a big dip at the 2000Hz region, to which I reversed the tweeter's polarity and got much better results. However I now get a large db bump from the 500-2000Hz region. I turned the tweeter knob down all the way which still left a 500Hz spike so that got me thinking its not just the tweeter level. I eq'd it flat and it now sounds incredible...but I would really like to get the crossover dialed in so I started poking around with xsim and finally got the basics down. I created measurement files for both drivers and got the following freq response which seems to line up with the problem I am facing:
As you can see my levels start to peak before my crossover point....so I started playing with my crossover. I ended up zeroing in on lowering the parallel resistor on the tweeter to 1ohm and it seems to be the kind of response I am looking for. Is this going to work? The driver responses look way off but the series curve (blue line) looks linear and nice...am I doing this right?
Thank you for any feedback 🙂
Your tweeter is playing too low. Generally the L-pad goes between the crossover and driver. Try simulating with the 4 ohm between L2 and R2 and with an 8 ohm parallel resistor. Then see what the crossover frequency is. If your tweeter still doesn't roll off at a similar slope to the woofer you may need to reduce L and C values in the tweeter circuit. If the tweeter is still too loud, then you start messing around with the L-pad values.
With the series resistor between L2 and R2 you can put a small value capacitor (maybe 1.5-2uF) in parallel with it to bring up the treble a bit. This is a contour circuit often used with CD horns.
With the series resistor between L2 and R2 you can put a small value capacitor (maybe 1.5-2uF) in parallel with it to bring up the treble a bit. This is a contour circuit often used with CD horns.
That is the more common way. The series resistance has an effect on flattening the tweeter impedance although it's usually a smaller effect than the parallel resistor.
The reason it was used that way in the tutorial is that it was found to be a little more predictable when used for tuning by ear alone.
The reality is that either way can successfully be used, but for achieving the same response either way, measurement and/or simulation may have to be used.
Ok I will play around with this. Thank you
I just realized something. Does using an l pad knob with that tutorial make sense? Since the knob works for 8 ohm drivers, wouldnt flattening/lowering the impedance make the knob not function properly?
If you mean that there will be variations in the response when you change the level, yes there will. They may not be a problem but they may be and can be dealt with. This is the same for the variable and the two resistor version.
The variable L-pad is a good tool in this situation. Yes it keeps things at their theoretical 8 ohms, but so would a two resistor version normally do but it doesn't have to. Sometimes an L-pad is made to emphasise impedance compensation and sometimes it is made to change the overall impedance higher or lower for one of several possible reasons.
The variable L-pad is a good tool in this situation. Yes it keeps things at their theoretical 8 ohms, but so would a two resistor version normally do but it doesn't have to. Sometimes an L-pad is made to emphasise impedance compensation and sometimes it is made to change the overall impedance higher or lower for one of several possible reasons.
Does this simulation include baffle effect? Any idea about off-axis behaviour? Woofer's response looks too pretty to be real... and it's off-axis above 2kHz drops fast.
1,5kHx xo should be good, but compression driver might be too stressed with high spl - steeper acoustic slopes by shifting electric xo of woofer lower and tweeter much higher?
1,5kHx xo should be good, but compression driver might be too stressed with high spl - steeper acoustic slopes by shifting electric xo of woofer lower and tweeter much higher?
I guess not!
The off-axis behaviour cannot be estimated without measurement, as Eminence has only published the on-axis data.
According to the specification, the compression driver was measured in a small exponential horn (45x45°). This can probably only be transferred to the installation in the Eminence Coax to a limited extent.
The cabinet seems to have already been built, so measure it and then take a look at the crossover.
The off-axis behaviour cannot be estimated without measurement, as Eminence has only published the on-axis data.
According to the specification, the compression driver was measured in a small exponential horn (45x45°). This can probably only be transferred to the installation in the Eminence Coax to a limited extent.
The cabinet seems to have already been built, so measure it and then take a look at the crossover.
It looks like you changed the 1 ohm to 2.4 and the tweeter crossover is still too low. If these are not meeasurements of the horn on the driver in question then it is not worth even talking about though. The horn vs driver loading on the CD will change things.
Yes I tried moving the series resistor after the crossover and played with the L and C values but wasn't getting the results I was looking for. Thank you for the input, I am still very much a beginner. I guess what I am learning is that you cannot properly build a speaker using manufacturer driver specs and I need to get some measuring equipment.
I agree that this one didn't feel right and I understand the difficulty. There are a couple of things. It can be helpful to show phase for each driver. Your slopes should probably be a little more distinct, and the baffle step is estimated. (I didn't consider the crossover frequency)
Thanks so much...I really appreciate your time. This simulation software has really opened up a new level to my hobby...I still am going to get measurement equipment to really get precise.
So to put it simply...the wider the baffle, the louder the HF sounds become, so you have to compensate on crossover design.
Also, I've been soldering my crossovers, but I'm beginning to see it might be better to just twist them until I get it how I want it, then solder.
So to put it simply...the wider the baffle, the louder the HF sounds become, so you have to compensate on crossover design.
Also, I've been soldering my crossovers, but I'm beginning to see it might be better to just twist them until I get it how I want it, then solder.
Or use decent alligator clip test leads between components. It looks like a pile of spaghetti, but it works. And you can always add more leads to put extra components in. It speeds things up significantly.
I check test lead resistance before putting them in use, and typically go through and solder or resolder all the alligator clips. It can be surprisingly difficult to find good test leads these days. My favorites from the last bunch I tried were these:
Digikey 314-1133, Mueller BU-00286
Digikey 314-1131, Mueller BU-00285
I check test lead resistance before putting them in use, and typically go through and solder or resolder all the alligator clips. It can be surprisingly difficult to find good test leads these days. My favorites from the last bunch I tried were these:
Digikey 314-1133, Mueller BU-00286
Digikey 314-1131, Mueller BU-00285
Last edited:
Are those Mueller clips soldered? I've been looking for decent clipleads, too. Ones that are soldered (or that I can solder, as some are clamped in a way that soldering well can't be done). And that don't have such tight and slippery sleeves that I can't squeeze them open wide enough without them squish out of my fingers. I've bought some on the 'bay, at higher than typical cost, that claimed to be soldered. Only to find them only clamped - and badly at that. I have a bag of clipleads but only a few (that I marked) that are trustworthy. I'm open to recommendations
Actually it is very easy to predict directivity with almost any modern freeware xo simulators. Just tell baffle dimensions and driver width(s). Of course this is just a rough estimate, but you'll get an idea about it.
Page 22
https://kimmosaunisto.net/Software/VituixCAD/VituixCAD_help_11.pdf
The most simple is The Edge https://www.tolvan.com/index.php?page=/edge/edge.php
Just move the mic gradually and save each graph
Yes, the Mueller ones are soldered (as of 2020 anyway). A couple of them didn't look great, so I touched those up, but that was much faster than soldering ones that are just crimped on.
I hate the slippery booted ones also (Adafruit 1008). And ones where the boots slide down because of odd shapes (Cinch 46-400). Or ones where the clip springs are too strong (EZ Hook 284M-12-5). It's funny how many different ways they can make a bad set of leads.
Last edited:
Juhazi, one of the drivers in question is a compression driver firing through a horn composed of the 12" driver's cone. You can estimate the off axis of direct radiators with a pistonic simulator like the edge, but not horns.
For the OP, if you put a small capacitor (maybe 1.5uf) in parallel with the 7.5 ohm resistor in AllenB's crossover, you may be able to lift teh high frequencies somewhat. Simulate and see what happens.
For the OP, if you put a small capacitor (maybe 1.5uf) in parallel with the 7.5 ohm resistor in AllenB's crossover, you may be able to lift teh high frequencies somewhat. Simulate and see what happens.