Announcement

Collapse
No announcement yet.

Help with a variation on Pete S. 3Way Peerless Design

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • shudson105
    replied
    I did make quite a few changes to the crossover and I'll have to post it later...not perfect yet though. Some frequencies are perfectly solid in the center when doing dual mono pink noise test tracks but some others are defuse. You can also tell this when listening to voices as the entire voice isn't perfectly centered. Mostly though, these are the best speakers I have made or own! Soundstage width can be wall to wall when recording allows for it. Image depth is pretty good but I suspect it'll get even better with better amps - running the Akitika GT101 Chip amp now. Bass is solid down to 40Hz or so, still benefits from a subwoofer. The impact from the three 6.5 drivers though is very nice, it greatly adds to the reproduction of music. I find myself unable to listen to bookshelf speakers anymore lol.

    Leave a comment:


  • shudson105
    replied

    Leave a comment:


  • shudson105
    replied

    Leave a comment:


  • shudson105
    replied
    fractured my right wrist today, gonna be out of commission from this for a week or so until the swelling goes down and i can finger peck on the keyboard. thanks for the help thus far RPB!

    Leave a comment:


  • rpb
    replied
    The frd that I'm using doesn't have a tail added, or minimum phase , etc. My mid / woofer summation will model a little different than yours. When I build a 3-way, there's a lot of trial and error to get the x-over done. I usually end up with too many parts in the x-over to be practical. Also... the sum will be different at different heights. Odds are, we are worrying a little too much about it at this stage. I have lots of parts on hand. I just try things until I get the response I want.

    Leave a comment:


  • rpb
    replied
    Got to sleep now. I'll look at your new sim tomorrow.

    Leave a comment:


  • rpb
    replied
    Can you post the tweeter frd, and zma files. I'll see if I can shape the response.

    I only mentioned the off axis dip in case you were not aware of it. The tweeter can help fill it in, but if crossed too high, it won't help a lot. I've done 2-ways crossed above 4k with the 830874, and I liked it. It did sound a bit different than when crossed lower.

    Leave a comment:


  • rpb
    replied
    Man, I am so sorry. I didn't mean for you to try again. That was not my intended message. I was attempting to post the graphs, and had a mess. So, I "Tried again". and got everything posted, and then deleted the screwed up post. I didn't want to edit it, because half the time I edit, I loose the whole post. I didn't realize that it sounded like a slap in the face until just now. I'm really sorry. You have to type a few words to post. that "try again" was just to satisfy that requirement.

    I'm enjoying the challenge of this design, and I hope to be helpful. You obviously are very good at this as well.

    Leave a comment:


  • shudson105
    replied
    Also, without distortion measurements of the tweeter, who knows what problems exist lower in freq...

    Leave a comment:


  • shudson105
    replied
    What am I not accomplishing with my previous sims? With your help, I got my newest sim to reflect much steeper roll offs, all of my drivers are in phase now, with strong nulls when reversed (suggesting optimal phase at the crossover point). Is this a significantly better design now with twice the parts count, much less overlap and higher (midrange and treble) impedance vs some of the early models with similar overall frequency response, but much overlap, higher efficiency and flatter (but lower) impedance? What made the earlier models less desirable?

    Regarding your statement that the 830874 rolls off steeply off-axis, wouldn't more overlap help with that? What does more overlap from identical drivers even do? Does that smear the imaging or would it add too much energy on or off axis that isn't seen in these models?

    The ESS AMT5 tweeter is (I think) a hard to work with speaker. Its natural response on an infinite baffle is slowly rising from 1.5KHz to about 4, then very rapidly rises nearly 10db at 5KHz and remains flat to 10KHz before falling off again. This is the main reason I am doing the mid/treble crossover at 4-5KHz. I tried making a notch filter to shelve down the 5-10Khz range but I don't know what values make a notch filter work; trying random values on a parallel circuit did not accomplish anything productive.

    Thank you for your help RPB, this has been an enjoyable exercise so far. If you would like my frd/zma files, both with and without the box model I would be happy to provide them. What do you think of this latest sim?
    Attached Files

    Leave a comment:


  • rpb
    replied
    I've never tried a filter like the one using the 700uf cap. I've no idea if this is a good idea, or a bad one. The resistor might need to be a high powered one, and it might get real hot. Just don't assume it's OK. Maybe test the woofer section while keeping an eye on an external x-over.

    Leave a comment:


  • rpb
    replied
    The mid has a peak that is not shown in the Peerless literature.

    I think the speaker can be made close to these sims, but I suspect that it will take some trial and error. I hope you have extra parts, on hand.

    Leave a comment:


  • rpb
    replied
    Try again. Click image for larger version

Name:	874 3 way two.jpg
Views:	45
Size:	438.2 KB
ID:	1408316Click image for larger version

Name:	874 3 way mid peak two.jpg
Views:	45
Size:	408.6 KB
ID:	1408317Click image for larger version

Name:	874 3 way mid filter two.jpg
Views:	47
Size:	485.8 KB
ID:	1408318Click image for larger version

Name:	874 3 way mid peak two.jpg
Views:	47
Size:	408.6 KB
ID:	1408319

    Leave a comment:


  • shudson105
    replied
    So here is a revised crossover starting with your values RPB and adjusting. It did come out quite flat! I was surprised at some of the high values. I generally prefer a downwards trend in my responses, a lot like how Sonus Faber voices their speakers...darkly beautiful....so my final design may reflect this, assuming these new values are in the right direction.
    Attached Files

    Leave a comment:


  • rpb
    replied
    I usually round off part values, but I got in a hurry on posting this.

    Also, the tweeter in my sim is different than yours, and I'm using my measurements for the 830874 in a modest size stand mounted design. 9 x 17 baffle.

    The 830874 drops off axis at 3k pretty sharply. I'd cross at 2.5k if the tweeter can handle it.
    Last edited by rpb; 03-26-2019, 11:44 PM.

    Leave a comment:

Working...
X