Hey all,
I am designing my first "proper" 3 way where I don't let the mid just fall off naturally on the low end. I notice with this band pass filter that my impendence is sent to the moon. I am very new to band pass filters. Is there anything super obvious I am missing or is this all just a very poor design? I have thick skin so have at it. I am not super proud of this or anything and am just a poor hobbyist amateur trying to make some HIFI for my own enjoyment.
The drivers:
Planar tweeter: https://www.parts-express.com/HiVi-RT1.3WE-Isodynamic-Tweeter-297-421?quantity=1
Mid: https://www.madisoundspeakerstore.com/approx-4-midrange/sb-acoustics-sb12nrxf25-4-4-midrange-4-ohm/
Sub: https://www.parts-express.com/Peerless-830667-8-Paper-Cone-SLS-Subwoofer-264-1102?quantity=1
What I have done in the past is to run the mid and woofer in series to deal with impendence. I can't really do that with the band pass filter. Or can I.....? Please let me know. I am entirely open to suggestions.
I am designing my first "proper" 3 way where I don't let the mid just fall off naturally on the low end. I notice with this band pass filter that my impendence is sent to the moon. I am very new to band pass filters. Is there anything super obvious I am missing or is this all just a very poor design? I have thick skin so have at it. I am not super proud of this or anything and am just a poor hobbyist amateur trying to make some HIFI for my own enjoyment.
The drivers:
Planar tweeter: https://www.parts-express.com/HiVi-RT1.3WE-Isodynamic-Tweeter-297-421?quantity=1
Mid: https://www.madisoundspeakerstore.com/approx-4-midrange/sb-acoustics-sb12nrxf25-4-4-midrange-4-ohm/
Sub: https://www.parts-express.com/Peerless-830667-8-Paper-Cone-SLS-Subwoofer-264-1102?quantity=1
What I have done in the past is to run the mid and woofer in series to deal with impendence. I can't really do that with the band pass filter. Or can I.....? Please let me know. I am entirely open to suggestions.
Attachments
OK, opened and jiggled your file. You have an issue in your woofer impedance data (unless you found a 50 ohm woofer).
Why did I think this in the first place? Given that all 3 branches are in parallel, the woofer branch should have been pulling the impedance down in its range regardless of what the other branches were doing.
Why did I think this in the first place? Given that all 3 branches are in parallel, the woofer branch should have been pulling the impedance down in its range regardless of what the other branches were doing.
I fixed the impedence and now basically have to start over. Oh well. Here's where I am
Mtidge,
That looks a whole lot better than where I am. What's the trick for getting the band pass to spread out over a wider range of frequencies? I seem to always get this super peaky spike for mid instead of a nice spread like you have there.
Mtidge,
That looks a whole lot better than where I am. What's the trick for getting the band pass to spread out over a wider range of frequencies? I seem to always get this super peaky spike for mid instead of a nice spread like you have there.
Attachments
The old school says 8X to 10X between the low and high pass or the two filters interact making the spiky response. Don't trust what you see in my sim, it using traced data, you need to measure everything on your baffle. Have you fixed your woofer impedance data yet? Not sure about your design but I built a 3 way stand mount with the same mid plus a 8 inch woofer and a ribbon tweeter. I crossed at 450 and 4.5k before converting to active.
I am using traced data. I have always used traced data. I have only done one where the baffle was simulated. Reality showed that to be quite wrong.
I don't currently have an acoustic mic but will be buying one. I currently use spectroid on my phone. Its suprisingly accurate and great for testing systems on the fly. They really pack a lot into these phones. You need a very good phone obviously but it very much works. It can measure outside of human range accurately which I find astounding.
My last 3 way which was off of traced data measured dead flat from 30 hz to 15k hz. That's my best one yet but I just let the full range mid play out to the bottom. Something I won't get away with using this SB Acoustics mid.
I don't currently have an acoustic mic but will be buying one. I currently use spectroid on my phone. Its suprisingly accurate and great for testing systems on the fly. They really pack a lot into these phones. You need a very good phone obviously but it very much works. It can measure outside of human range accurately which I find astounding.
My last 3 way which was off of traced data measured dead flat from 30 hz to 15k hz. That's my best one yet but I just let the full range mid play out to the bottom. Something I won't get away with using this SB Acoustics mid.
Just a bit of an update on this. I am working on getting a DATS and I just order a calibrated mic so I can get proper measurements with this system.
I switched software and I have been just playing around with traced data just for practice. I think I have this a bit more figured out now.
Let me know if anyone sees any major flaw in my system. I have an ohm load of 4 at the lowest and I get to 13 ohm at the high end.
I am not sure if the impendence can be brought down any lower on the high end. If there is a way, please let me know. Thanks
I switched software and I have been just playing around with traced data just for practice. I think I have this a bit more figured out now.
Let me know if anyone sees any major flaw in my system. I have an ohm load of 4 at the lowest and I get to 13 ohm at the high end.
I am not sure if the impendence can be brought down any lower on the high end. If there is a way, please let me know. Thanks
Attachments
You can add a Zobel to flatten the rising impedance of a driver, it is just a series cap and resister parallel to the driver. Try one on the woofer and the mid, the rising impedance on the tweeter might not matter. It will make you filters work better. I don't know the drivers you are using but you might want to cross from the woofer to the mid higher especially if you are using a small sealed mid enclosure which will act like a high pass circuit.
I do plan on using a small sealed "cabinet" for my woofer. It will be just a small half sphere that I am going to 3d print out of PETG. I have found the PETG, which is slightly flexible, printed large with nearly no infill, works amazing for enclosure. Basically like a double wall half sphere with air trapped inside that acts as a dampener.
The woofer is good down to 100 but yeah, you are maybe right and I should cross is more at 200 or so
I tried the Zoebel and it didn't have any effect on the impendence, only the frequency response. I may be doing it entirely wrong. Please let me know.
The woofer is good down to 100 but yeah, you are maybe right and I should cross is more at 200 or so
I tried the Zoebel and it didn't have any effect on the impendence, only the frequency response. I may be doing it entirely wrong. Please let me know.
Attachments
You don't need special software to measure impedance. Many do it using the same software they take response measurements with.
Is that not what you are trying to do?
Any half decent solid state amplifier will handle impedance swings. Keep the minimum above say 3 ohms and your focus should be elsewhere in the design.
You're using traced data. Have you done the following?
1. Applied baffle diffraction / step response to manufacturer curves? This will also apply Y rolloff caused by beaming (so you can leave this as zero in VCad)
2. Set "use minimum phase" check box on all 3 drivers in the driver tab
3. Entered a suitable Z offset in the driver tab for midrange and woofer, assuming you are designing for the tweeter axis
Your overall system sensitivity should be targeting about a 83 - 84dB level, considering the woofer is 87dB above baffle step (allowing for 3-4 dB baffle step losses).
Once you do the above, be prepared to "start again" on the crossover. But it's worth it. Otherwise garbage in -> garbage out (and an expensive passive crossover parts order).
You're using traced data. Have you done the following?
1. Applied baffle diffraction / step response to manufacturer curves? This will also apply Y rolloff caused by beaming (so you can leave this as zero in VCad)
2. Set "use minimum phase" check box on all 3 drivers in the driver tab
3. Entered a suitable Z offset in the driver tab for midrange and woofer, assuming you are designing for the tweeter axis
Your overall system sensitivity should be targeting about a 83 - 84dB level, considering the woofer is 87dB above baffle step (allowing for 3-4 dB baffle step losses).
Once you do the above, be prepared to "start again" on the crossover. But it's worth it. Otherwise garbage in -> garbage out (and an expensive passive crossover parts order).
I have measurement equipment on the way. Right now I'm just playing with traced data so it gets me close
That being said, I have made a few systems so far that measure nearly flat with other people's equipment and those were made with traced data. I wouldn't discount it straight off. I don't know why everyone frowns upon it so harshly. Its certainly better than what most people do. Just buy some speakers and some off the shelf crossovers and hope it all works out.
I haven't figured out how to do Z offset yet. Will try to figure that out at some point
Will also click the Minimum phase tab and see what happens. What does that do? I've been using Xsim up until now
That being said, I have made a few systems so far that measure nearly flat with other people's equipment and those were made with traced data. I wouldn't discount it straight off. I don't know why everyone frowns upon it so harshly. Its certainly better than what most people do. Just buy some speakers and some off the shelf crossovers and hope it all works out.
I haven't figured out how to do Z offset yet. Will try to figure that out at some point
Will also click the Minimum phase tab and see what happens. What does that do? I've been using Xsim up until now
You're tending to the impedance beside one driver yet you're looking at the system impedance. You have to isolate the issue to see it properly. Disconnect all non-relevant parts before you design the impedance compensation for the woofer rise.
If you are able to loopback your computer output to your input, then yes, the recorded phase with an XLR mic (as long as you don't change the timing marker so that measured phase includes the flight time of the measurement) will be accurate and not require any minimum phase nor Z offset setting at all (i.e. leave minimum phase checkbox per driver UNCHECKED and do not enter any Z offset in either driver nor crossover tabs). If you are using REW for measurements, I'm unsure what settings you use or simply set the loopback option.