[FallenAngel]

Hey guys,

I just built a tiny Modified Linkwitz Crossfeed.

It works great as a crossfeed, I hear the channel blending when the switch is ON, don't hear it when it's OFF, still not sure if I can tell the difference between High and Low settings, but that's fine.

On to my little problem though, I wired it internally with SPC and have a built-in output plug, also SPC. When I tried it with my Pimeta, I noticed ALL THE BASS IS GONE! Aaah! That's a very scary thought considering I was using bassy AD843's opamps and DT770 80Ohm! Without the crossfeed the combination is very nice, great bass when I'm playing racing games, can really hear the roar of the engines and it sounds great with hiphop / dance music. I put on some 2Pac and I can barely hear any bass at all, which totally threw me off.

The full config is this:

Audigy2 ZS (ASIO) -> Magami StarQuad (copper) Mini-Mini -> Crossfeed (SPC) -> Pimeta with Cardas internal wiring.


Do you guys have any idea why the bass was suddenly sucked out with this crossfeed (I'm thinking it's because of mixing SPC wiring/built-in interconnect with a copper interconnects)? Is it normal to notice less bass when using a crossfeed?

 

[FallenAngel]

Well, I built this tiny little thing on a 10x6 pin breadboard with the resistor values above.
I decided to go with one 0.1uF cap for C2.

It sounds kinda funky, you can certainly tell that the channels are mixed and the whole 3D effect , and it's much quieter. I can't tell the differnce between High Crossfeed (2.2K) and Low Crossfeed (1.5K) at all.

The scary part is though, I think my Pimeta sounds quite a bit more detailed and clearer without the Crossfeed all together. Maybe this is because of my implementation of this crossfeed, maybe it's that I don't like the resulting sound effect. In either, I'm not likely to be listening to this thing.

 

[GregVDS]

The mica caps are not recommended, I think, try to find orange drops Vishay-Sprague 200V 0.1µF and 0.022µF.

Do you use the Crossfeed between your source and the amp, or between the amp and your headphones? the first solution is the best one!

I cannot hear a very obvious difference in the highs between with and without position. In the other way, I would preferably use it all the time, as the effect is perfect for me, and let me listen to music hours long without being sick.

Try maybe to fiddle around the values, I cannot tell you the precise effects of yours, I used the precise ones ordering the parts. There is an xls file letting you play around with the values and look at curves to evaluate the effect on delay and levels of the mixing. Search around in the library of Headwize. Let me know if you don't find it.

All the best,

GregVDS

 

[FallenAngel]

From what I heard, Mica isn't bad. I'll consider getting the board and making one using exact specs, but that woun't be until I finish all the other projects first.

I'm using the crossfeed between the amp and source.

Not sure exactly what the crossfeed does except for mixing the channels, so maybe this one will be just another little box on my desk at work.

 

[mb3k]

Maybe you can try changing, I belive its now R2 to 910R instead of 1kR, straight from Chu Moy's update (link here)

 

[FallenAngel]

Well, after reading this quote from cmoy "In these types of circuits, it is amazing how much difference a hundred ohms this way or that way makes!", I'm strongly much reconsidering making a crossfeed on a professional board by Tangent with the new values. Maybe I'll build another amp along with it as to not order just small stuff and wait for delivery.

Thanks

 

[GregVDS]

Quote:

Originally Posted by mb3k Maybe you can try changing, I belive its now R2 to 910R instead of 1kR, straight from Chu Moy's update (link here)

I think it's not R2 but R3 that is 9.09K (or 910R depending on the high Z or low Z you do). R2 is the resistor linking the output to ground, this one is 1k (or 100R). R3 is the resistor in serie with the 0.02µF (or 0.2µF), its value being 9.1k (or 910r). This Chu Moy update dates from 2001, and is called the Modified Linkwitz Crossfeed filter, and it's this version explained thoroughly by Tangentsoft on his website. R3 in serie with C1 defines the amount of highs going through the cap, hence going from 10K (1K) to 9.1k (910R) will give more highs.

at least, that's what I understand.

FallenAngel, not only the crossfeed mixes a small amount of each channel in the other one (through the 3.3K resistors), but this part has a dephasing frequency dependent added to it. This part is fairly high in volume on the lows, and less and less going up in the frequency, hence simulating the unidirectionality of the lows, and the masking of the head of the sound coming to opposite ear from a virtual speaker (ideally placed on the isocele triangle base = height). Now the delay added is very small in the lows, and goes up to build to something around 300µsec, the delay the opposite ear hear the signal compared with the direct ear. With some training, your brain begin to decipher this thing, and rebuild the position of the audio source somewhere in front of you, rendering a stereo scene more natural, more like the one you can experience with a hi-fi system, not a headphone.

Another psychoacoustic thing is that your brain, when receiving a sound from one ear, "hopes" to hear it from the other one, in order to process the difference in time and spectra, so as to situate the source in 3D (where is the frightening lurking monster under the bed, or behind the curtains?). So after some times recieving totally different signals from both ears, your brain begin to hurt, and you have to take off the phones, hence the crossfeed!

Hope this helps a little bit.

ps, there is also an alternate value for the 1.5K, being 6k, if the other (2k) is always connected, so to change the settings, you just have to connect the 6k in parallel ( (2*6/2+6) = 1.5k). this is perfect to have a silent change, without cutting off momentarily the signal.

GregVDS

 

[mb3k]

Quote:

Originally Posted by GregVDS I think it's not R2 but R3 that is 9.09K (or 910R depending on the high Z or low Z you do). R2 is the resistor linking the output to ground, this one is 1k (or 100R). R3 is the resistor in serie with the 0.02µF (or 0.2µF), its value being 9.1k (or 910r). This Chu Moy update dates from 2001, and is called the Modified Linkwitz Crossfeed filter, and it's this version explained thoroughly by Tangentsoft on his website. R3 in serie with C1 defines the amount of highs going through the cap, hence going from 10K (1K) to 9.1k (910R) will give more highs. at least, that's what I understand.

Aha! You're correct, I just briefly skimmed through threads on Headwize not thinking throughly.
Modified Linkwitz Crossfeed = The link in my post above (low z version)

 

[spendorspain]

Hi

I’ve used the same Modified Linkwitz crossfeed circuit, built using the PCB Tangent sells (with the rotary switch option between high and low crossfeed settings). However, it dulled too much the sound and in some recordings the soundstage was pushed back too much. I thought the crossfeed level was excessive for classical music recordings, where there is already some real ambience captured by the microphones (unless it is a artificial spot-miked recording like some old “4-phase stereo” DECCA or some DG). So I decided to experiment and change only the main-channel resistor R1 to 1kOhms, so there is more direct sound and less mixed from the other channel: now the crossfeed is more subtle, but more to my taste, and the treble level is more natural. Moreover, to avoid to spoil the PCB with this experimentation, I opted for hardwiring the caps and resistors I had at hand, soldering them point to point (it is easy following the schematic). This combination of hardwiring, no PCB, smaller R1 and no switches in signal path, made the crossfeed circuit a lot better for me… However, it’s ugly to see, I know (see pictures).

BTW, I’m using now an AD744 (output bypassed) based headphone amp and I feel the crossfeed is less necessary with it than with the previous LM6171 amp, so currently I don’t use it

Regards
Jose

 

[FallenAngel]

New question, built a crossfeed to exact specs for resistors and Panasonic ECQP caps and SPC internal wiring, but the bass is totally gone because of it. Please see first post for full details.

Thanks!