Budget active 3 way crossover - buy, eval, mod

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
The solder-averse budget method...

If you can bring yourself to consider a PC based DSP, these could be done as cheaply as a PC that will run DSP software (such as J River Media Center); you can get PCs well < $200 that have HDMI out (such as Pipo X9), feed HDMI to a cheap 5.1 or better HTR and you have, effectively, a multi-channel DSP to use as you like.
 
Nothing against DSP approaches, its often the only way to do certain things. I have used my PC with APO equalizer's convolution engine and Rephase to generate filters for both the room and the speaker. I also have another thread trying to do a similar thing on an RPI which is where I would prefer it. After I get basic EQ and a convolution engine working on the RPI, I plan on trying a s/w crossover as well. Maybe both will be used, a hardwired filter is safety from rogue s/w.
 
Last edited:
Distortion

The following are measured with REW and a PC soundcard (24b). The graphs are scaled so that the max line in = 0db. So I adjusted the output such that the max signal into the soundcard is -3db to avoid clipping. It also means that the soundcard output can't be maxed because of the gain at the active XO input.

In all cases the TH is -80db which I'm good with. Its hard to determine if its the XO or the PC sound card that's the limiting factor here but I suspect the PC. I also need to clean up some switching noise from the SMPS. Then I might try a different setup that would allow me to figure out the actual noise floor on the XO and if I can push the TH a little lower.
 

Attachments

  • LowBand Distortion.jpg
    LowBand Distortion.jpg
    100.2 KB · Views: 513
  • MidBand Distortion.jpg
    MidBand Distortion.jpg
    102.8 KB · Views: 512
  • HighBand Distortion.jpg
    HighBand Distortion.jpg
    97.9 KB · Views: 500
Last edited:
Noise

I don't have equipment or a method of proper noise measurement other than time waveform and an oscilloscope. There is some SMPS HF noise present, and the XO board lacks HF decoupling. This test was with the inputs shorted to ground and a 10K load on the outputs.

The on board regulators attenuate some of the lower freq SMPS variation. Pic#1 show the ac component of the +15vdc rail with the mid outputs (M) as a wide view. Pic#2 shows the zoom in of the ac components of the +15VDC and the low (L) outputs. Pic#3 is a zoom in of the ac components of the +15VDC and the high (H) outputs. Pic#2 and #3 show the onboard regulator helps for some variations, but the HF goes through everything. A power filter will be added after I test a few other related pieces of equipment.
 

Attachments

  • 15VDC and mid out wide view.jpg
    15VDC and mid out wide view.jpg
    104.8 KB · Views: 472
  • 15VDC and LF out.jpg
    15VDC and LF out.jpg
    87.5 KB · Views: 468
  • 15VDC and HF out.jpg
    15VDC and HF out.jpg
    83.6 KB · Views: 80
packaging

Finally got around to putting it in a box. The fan will eventually something like a LM317 based speed control to make it quieter. There is enough room left for additional boards. The current FR is shown with the mod'd sub bandpass.

I'm getting closer to deciding on the final 3way XO points that I want. Each driver would handle 2.5 octaves and behave well in the XO overlaps.

sub : 20-100Hz
woofer : 100Hz-600Hz
mid : 600-3.6Khz
tweeter : 3.6Khz-20Khz
 

Attachments

  • DSCN5240 reduced.jpg
    DSCN5240 reduced.jpg
    587.4 KB · Views: 174
  • DSCN5234 reduced.jpg
    DSCN5234 reduced.jpg
    560 KB · Views: 201
  • DSCN5233 reduced.jpg
    DSCN5233 reduced.jpg
    640.4 KB · Views: 205
  • Sub bandpass mod + stock 3way.jpg
    Sub bandpass mod + stock 3way.jpg
    95.9 KB · Views: 187
The woofer will have a sealed chamber, so it will have a -12db/octave acoustic rolloff at 100Hz. Still, the slopes are different (sub -24db, woofer -12db) so a small bump could occur. The sub's volume (gain) and XO point are adjustable for this reason.
 
How do you tailor the individual pass band's output level to adjust for different driver sensitivity? The crossover board has no provision for output volume.

I am building a crossover trying to use this board.
Preamplifier Crossover 2-Way Electronic Frequency Divider board Free adjustable 831230323184 | eBay
But I cannot get the seller to provide a schematic or to send instruction on how to set the jumpers and trimmers after I got all the other parts and a case sized for this board. I may need to look at this.
Linkwitz-Riley 2-way electronic crossover at 2.2KHz stereo (changeable by user ) | eBay
 
That's correct, this XO is fixed gain, but there are a few other places you could adjust the channel gains. It depends on your system and what's more convenient for you. My master volume control is in the DAC, and I'm using Sure class D amps with "fixed" (preset) gain that is set via dip switch (MF, HF) or a pot (LF).

1) select the amp fixed gain (some Sure class D amps have [25.6, 31.6, 35.1, 37.6 dB] settings
2) use amp variable gain (some Sure amps have a volume control/attenuate board) adjust for the diffs and leave it
3) use amp gain offsets, often in HT amps if you are repurposing one, it allows a delta adjustment per channel
4) put an L-pad on the tweeter
5) EQ the channel gain differences at the signal source (using EQ-APO, or FIR filters), you may end up needing this due to unruly drivers, user FR curve preferences, enclosure or room behaviours.

Sure Electronics AA-AB33184 4x100W TDA7498 Class-D Amplifier Board
AA-AA11117 Digital Volume Control Kit
 
I have 3 old fashion linear power amplifiers (6 channels). They have no input volume control. The 2-way board I purchased does have LP, HP and master volume control. I will wait a few more days to see if the seller will provide more information.

Another option is this fully assembled crossover box. The seller promises customized crossover frequency.
2 Ways 24dB/Oct Linkwitz-Riley Active Crossover Electrical Frequency Divider | eBay
 
Last edited:
Usually, boards like the one you ordered have master volume and tone controls. There is not enough info to tell what the extra controls are actually doing.

It's very easy to add a simple 2 resistor signal attenuator (fixed gain) to the HF output of your XO before the amps. Or you could add a pot between the XO and amp input if you wanted it variable attenuation.
 
Reprogrammed XO points

I finally got a chance to reprogram this to want I need. Only the resistor values were changed, .... lots of resistors. Luckily you can buy small 1% kits with a limited [1K-47K] value ranges to suite this.

List is per channel, so double this:
[R30,R34,R50,R51,R52,R53]=4K7
[R31,R32]=9K1
[R33,R35,R36,R37,R40,R42]=7K5
[R41,R43]=15K

New XO points [725Hz, 4550Hz] will allow domes midranges and small tweeters.
 

Attachments

  • Reprogram 3way XO [725Hz, 4550Hz].jpg
    Reprogram 3way XO [725Hz, 4550Hz].jpg
    109.5 KB · Views: 397
Linkwitz-Riley 4th order 2-way

After assuring myself it wouldn't blow up my sound card. I ran a quick test through REW to plot the FR. No surprise it crosses at 300Hz and 3Khz. It also appears (no surprise) to be an LR4, pass flat with -24db/octave attn.
I just downloaded the REW. Can you point me to a section/page of the REW user's manual on how to create the line level FR plot you show.

I just finish building a 2-way crossover using the PCB from a seller in China. It sounds fine first connected. It is time for some measurement.
divider-crossover-linkwitz-riley-filter-2200hz-24dboct-class-a.jpg IMG_4753.jpg IMG_4754.jpg
 
Nice. What is the XO point set to?

First a wording of warning. The XO has gain of 4x (12db) if it's the same base design as mine (looks like it) and its output can go to 7V rms so you need to make very sure the signal level going into the XO is set to a low level. There is gain control on the soundcard output (via PC) as well as in REW (signal level). Start with both on low (-20db) setting and gradually increase to avoid clipping. REW allows you to check input levels. Assuming you are connecting your soundcard's Line Out or headphone (green) to the XO inputs, and the soundcard's Line In (blue) to the XO outputs.

The curves I show are actually 3 measurements. I sweep 20Hz-20Khz connecting the soundcard Line In Left to the XO output I want to test. REW requires both the XO's output (connected to line in left) and its own test signal Line out right (connected to line in right) to make the measurement. Set REW to output the test signal on both Line Out L&R, reference Line In right, measure Line In left.

First configure your settings via "preferences" section and then make the measurements via "making measurement" section of the help file (manual).
 
The filter is text book Linkwitz-Riley 4th order set at 100 Hz for subwoofer. The seller did the crossover frequency mod for me and used the correct R and C values that Linkwitz's calculation show. For my board, I have R=11.3K and C=100nF, with corresponding 2R and 2C. Component quality and construction look pretty good for me.

I drive the LR4 with a Harmon Kardon AP2500 pre-out.
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.