A place to discuss member's DIY audio projects & post pictures/schematics. NOTE: There is a limit of 2MB per attachment, and a maximum of 3 attachments per message. If you need to post more than 3 attachments, just add another message.
February 28th, 2024, 6:34 pm
The crossover slope is also highly dependent on the radiation patterns of the drivers. If they have the same patterns at least through the crossover points and hopefully through most of the range, then high slopes work best. Lobing is minimized with high order slopes. If they do not have the same radiation patterns through the crossover region, lower order slopes work better. With different radiation patterns, radical shifts in room contribution can occur with high slopes. The sloppy lower order slopes allow overlap through the crossover region, reducing the rapid changes in room contribution but resulting in more lobing. I noticed this when I used a Heil AMT1 with a 7" cone midrange. The different radiation patterns at the crossover points resulted in weird spatial effects with 48dB/octave slopes. I had to go to 12 dB/octave slopes to make them sound reasonable, but not optimum.
February 28th, 2024, 7:01 pm
Since the 1.7's planar drivers are arranged horizontally not vertically, the lobbing would be interesting to visualize.
I though about making a lazy susan style speaker rotator to do the testing, but not had time to build it and then mark the degrees on the base.
I have seen some commercial speaker rotators, but not inexpensive to purchase.
February 29th, 2024, 10:37 am
HAL wrote:Cogito,
I have experimented with different slopes for the DSP crossover. I do not use the original style filters that Magnepan developed since they are series mode filters. Triamping these with DSP is the way to go after the experiments with the MG10/QR's.
I like either 4th order or 8th order Linkwitz-Riley since they are 0 degrees at crossover. That is either an acoustical or electrical alignment depending on the driver. I typically use 8th order, as it allows the planers to have more power handling in band. I also high pass the planars and time align them to th subs. That solves the typical planar problem of sounding congested when trying to reproduce low frequency signals. Easy enough with the DSP.
Rich
Thanks. How would you describe audible difference between 2nd order and 4th order slopes?
February 29th, 2024, 10:52 am
Cogito wrote:HAL wrote:Cogito,
I have experimented with different slopes for the DSP crossover. I do not use the original style filters that Magnepan developed since they are series mode filters. Triamping these with DSP is the way to go after the experiments with the MG10/QR's.
I like either 4th order or 8th order Linkwitz-Riley since they are 0 degrees at crossover. That is either an acoustical or electrical alignment depending on the driver. I typically use 8th order, as it allows the planers to have more power handling in band. I also high pass the planars and time align them to th subs. That solves the typical planar problem of sounding congested when trying to reproduce low frequency signals. Easy enough with the DSP.
Rich
Thanks. How would you describe audible difference between 2nd order and 4th order slopes?
Using 2nd order L-R slopes usually requires one driver's polarity to be inverted due to the 180deg phase change at crossover. This has always sounded odd to me. Lots of possible explanations, but I avoid it. Using either electrical or acoustical 0 degree (360deg, 720deg, etc) phase at crossover has always sounded better to me. Just depends on the rolloff characteristics of the drivers as to which I choose.
Tom's explanation of beam patterns from speakers with different crossover slopes comes into play as well.
February 29th, 2024, 11:14 am
tomp wrote:The crossover slope is also highly dependent on the radiation patterns of the drivers. If they have the same patterns at least through the crossover points and hopefully through most of the range, then high slopes work best. Lobing is minimized with high order slopes. If they do not have the same radiation patterns through the crossover region, lower order slopes work better. With different radiation patterns, radical shifts in room contribution can occur with high slopes. The sloppy lower order slopes allow overlap through the crossover region, reducing the rapid changes in room contribution but resulting in more lobing. I noticed this when I used a Heil AMT1 with a 7" cone midrange. The different radiation patterns at the crossover points resulted in weird spatial effects with 48dB/octave slopes. I had to go to 12 dB/octave slopes to make them sound reasonable, but not optimum.
So, lower order slopes are good to addressing the mismatching of radiation patterns by overlapping the frequencies around crossover point.
February 29th, 2024, 12:45 pm
A good analogy of using lower slope crossovers to ameliorate radiation pattern problems is using a pain killer when you have a broken leg. The leg is still broken.
February 29th, 2024, 2:03 pm
tomp wrote:A good analogy of using lower slope crossovers to ameliorate radiation pattern problems is using a pain killer when you have a broken leg. The leg is still broken.
March 2nd, 2024, 5:49 pm
First 6 pairs of locking bananas from P-E arrived yesterday to make the 14awg cables from Belden 1309A cable to the 5-way binding posts for drivers to the amps.
The 3 pairs of 5-way binders on the connector plate are close, so locking bananas plugs will work well.
- Attachments
-
![LockingBananas.jpg (35.78 KiB) Viewed 6495 times LockingBananas.jpg](./download/file.php?id=2582&t=1)
March 6th, 2024, 6:07 pm
Have the first 1.7 rewired for testing with the locking banana connectors.
- Attachments
-
![1_7LockingBananaCables.jpg (473.97 KiB) Viewed 6408 times 1_7LockingBananaCables.jpg](./download/file.php?id=2585&t=1)
-
![1_7LockingBananaTest.jpg (265.85 KiB) Viewed 6408 times 1_7LockingBananaTest.jpg](./download/file.php?id=2584&t=1)
March 21st, 2024, 8:30 am
The Magnepan 1.7 with the dspNexus 2x8 DSP crossover is now running. This is the one speaker with all three planars working for measurements.
From the passive crossover design and first measurements, the DSP crossover is working. Now to make individual driver measurements to fine tune the crossover. The DSP crossover is nothing like the original passive crossover in design.
The three Orchard Audio Starkrimson Monoblocks with the 300W GaN SMPS are triamping the planars with the 2x12 in servo sub on the left. The amps were in my room when the ground water problem happened. I bought the new SMPS as the originals were water logged and most of them died.
The super tweeter driver is padded with a series resistor since its resistance is 1.4 Ohms. This makes it an easier load to drive for any amp.
- Attachments
-
![Magnepan1_7NewDSPXO.jpg (384.7 KiB) Viewed 6104 times Magnepan1_7NewDSPXO.jpg](./download/file.php?id=2586&t=1)
Powered by phpBB © phpBB Group.
phpBB Mobile / SEO by Artodia.