Rogers LS3/5a 1100065A crossover PCB minor modification

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Ex-Moderator
Joined 2003
Years ago, I modified my Rogers LS3/5a and thought it made them sound clearer. Note that this modification does not change the frequency response, but today I measured the crossover inductors, modelled the circuit in LTspice, then compared before and after responses to the drivers, with the result shown in the graph.

As can be seen, without the modification, low frequency attenuation to the T27 tweeter is limited to 50dB.

The modification moves the negative wire going to the B110 from its existing point on the PCB to the incoming negative. That's it. It makes a difference because a narrow PCB track having 17 milliohm resistance is common to both drivers and the B110's current develops a low frequency voltage which is distorted by the T27.
 

Attachments

  • Rogers LS35a modification.gif
    Rogers LS35a modification.gif
    15.3 KB · Views: 466
Many years ago, I modified my Rogers LS3/5A crossover too. In so doing, I discovered that one of it's key principles was the attenuation of a huge peak around 1kHz in the woofer, and the much of the crossover board implements a notch filter that is very sensitive to small changes in LCR values. That bugged me a lot, I'd always though that drivers should have basically flat response in their passband. I ultimately switched to use Revel M20 speakers, which I consider about as good as a small box speaker can be, based around LR24 or similar crossover with aluminum woofer and tweeter in very heavy ported box. Revel quit making that incredibly superior and underpriced speaker (IMO) a decade ago and most have forgotten about it, but the 3/5A (which sounds very flawed to me) is still a hot item.
 
I did duplicate that finding in Visaton Boxsim. By connecting a common high resistance track ( I used 1 ohm to test it...) between the -ve of the tweeter and the -ve of the tweeter coil carrying the bass return signal. It ruined the low frequency isolation of the tweeter filter! The last image.

Common mode crosstalk and inductor crosstalk is always tricky. Hence the 90 degree alignment of coils by good practice. :confused:

Earthing unbalanced speakers and amplifiers is actually extremely problematic. Even if you star earth the crossover within the box, you still have the impedance of the return lead of the speaker cable to consider. But, yes, you have star-earthed and it does something.

One way round that is to use a series crossover. Here is an extremely elegant Bud Fried circuit which has no real earth point, the Fried Model H. It uses the KEF B200, B110A and T27. Bit shallow on rolloff, so you best choose well-behaved drivers with inherent rolloff, but you can use higher order series filters too. If you work out a target response for a parallel circuit, it is usually possible to duplicate that with a series network. The only real difference is completely different impedance.

Does it sound better? IDK! :D
 

Attachments

  • Crossover Bud Fried Style.PNG
    Crossover Bud Fried Style.PNG
    7.8 KB · Views: 468
  • Bud Fried Freq Response and impedance.PNG
    Bud Fried Freq Response and impedance.PNG
    23.6 KB · Views: 327
  • Bud Fried Phase.PNG
    Bud Fried Phase.PNG
    28.5 KB · Views: 362
  • BBC LS3-5A crossover.PNG
    BBC LS3-5A crossover.PNG
    35.9 KB · Views: 423
  • BAD EARTHING.PNG
    BAD EARTHING.PNG
    9.6 KB · Views: 483
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.