MiniDSP to fix time alignement in KHorns without biamping

[Bofkont]

It's been three years since I got a pair of 60th Anni KHorns in my basement. It has worked out well for me, thanks to info from this forum. Especially the FAQ Corner-Horn Imaging has been very helpfull (tnx CASK05!). Made me switch from a nice 300B SET amp to a FirstWatt F3 and tweak my room accoustics a bit further. Little things like putting the beast on spikes help, "sealing" the cabinet to the walls with isolation tubes helps etc. etc.

And now the miniDSP has entered. It is positioned between the input switch board and the Nelson Pass B1 preamp in such a way that it can be bypassed. So we can ABX. It is setup straight forward with crossover at 450Hz, 48 dB slopes where the High Pass signal is delayed 7.5 mS. Both High and Low pass signals are kinda shortcircuited and the sum signal is fed to the B1 preamp and stereo F3 power amp; so no bi-amping needed! It looses some 3dB but imaging has really improved. Female voices are overall nicer to listen to, instrument are behind the speakers, the stage is bigger and a piano is smaller so it can be pointed at. Somehow there is also more room information and the illusion of height is there.

Thought you guys might want to know or have a good laugh...... [H]

[Chris A]

Wow - that's interesting. If you get a chance, could you draw something like a circuit diagram? I'm not entirely sure of the configuration.

Good deal on the First Watt F3 and your room acoustics upgrades. They weren't subtle changes when I went with those two changes. I hope that they also have been similarly dramatic for you, too.

[Y]

Chris

[Marvel]

I think I see where this goes, and it's pretty clever. No need for multiple amps per speaker. But it sounds like you use the minidsp as a crossover and delay the signal, and then sum the signals back to a single output.

Bruce

[cfelliot]

It would seem that you'd be really screwing up the signal at the crosover point where both speakers are reproducing the signal. You'd have one signal delayed mixed with another undelayed????

[Chris A]

Well, that is the same issue that I saw, but I really don't know how the OP implemented it.

If the crossover is very, very steep, then the effects of "phase errors" (...actually bandpassed time delay errors of more than one wavelength summed together with the complementary bandpassed original signal...) would be minimized, but not exactly zero. It is clever if implemented successfully.

Chris

[Marvel]

I was thinking the same thing, that you could get some phase errors with a summed crossover signal feeding into a physical crossover. If that would work, though, I could mess with the signal coming into an audio application on my PC and do delays and such, feeding back out the soundcard.(Now my brain's trying to work, but nothin's happening...[*-)])

Bruce

[michael hurd]

You may want to check out the Rane white paper " Why not wye". The concept is ok, but there is an issue with summing both outputs. There is however a way to do it so you don't have the same issues.

[Chris A]

You may want to check out the Rane white paper "Why not wye".

http://www.rane.com/note109.html

[michael hurd]

[Y][Y]

You may want to check out the Rane white paper "Why not wye".

http://www.rane.com/note109.html

[Bofkont]

Thanks for giving my experiment some of your thoughts and time.

The reason summing up the signals does seem to work is probably because it is done after the Xover which is set at a very steep 48dB BW. So for the most you'r summing up a 0 V signal, which is inherently safe to do without resistors []. Just a bend cut iron nail in the original outputs. No need to say the nail was carefully selected for its audio quality and was bend at full moon, a blue moon actually.

My presumption was that around 450Hz you'd see some phase problems but that you suffer from that anyway given the difference in accoustical path between the bass bin and MF horn. Another thought was that keeping the original (split) X-over from the Anni KHorns in place would straighten some of the equalisation and phase problems that seems innate in the design. First I tried bi-amping with the Cayin 300B I had sitting around. Highly modded, really nice amp, but not as good as the F3 I'm afraid. So after that I decided to give it a try and shortcircuit the outputs. Still very happy with the simplicity of the solution.

BTW loosing 3 dB in the miniDSP appears to be a design choice. it has nothing do do with summing the (0V!) signals.

Best regards

Dirk

[Chris A]

My presumption was that around 450Hz you'd see some phase problems but that you suffer from that anyway given the difference in accoustical path between the bass bin and MF horn.

This is true. You have apparently traded for time alignment of the bass bin to the midrange for an apparently small price.

...but not as good as the F3 I'm afraid. So after that I decided to give it a try and shortcircuit the outputs. Still very happy with the simplicity of the solution.

Dirk,
I like your independent thinking, coupled with logical analysis and understanding of the physics to arrive at a simple solution. I also remember that Nelson Pass mentioned that he wasn't worried about the stability of the F3 under almost any load - input or output.

I'm also impressed that you also found the F3 amp to be really nice on the ears. I wonder why more high efficiency/low power folks aren't using these amps. There are also less expensive copies of the F3 amp available for those folks that don't like the price of the Pass-built units.

Chris

[mark1101]

Sounds like you could do something similar connecting the mini-DSP using a tape monitor loop.

[mark1101]

Can I ask how the final summing is done? Inside the mini-DSP, or with wires?

[Don Richard]

Interesting approach. You could use one of these and run the outputs of the mini DSP into 2 or 3 of the (stereo) mixer inputs which would take care of any level issues and allow analog knob adjustments of left and right channels simultaneously (knobs are quick and easy). Also serves as a preamp:

http://www.parts-express.com/pe/showdetl.cfm?partnumber=248-622

[Guest David H]

Interesting approach, although I am sure you would be better off by adding a ressonably priced amp for the bass section, then use the only a section of the passive crossover for hf.

Dave

[Bofkont]

Thanks for your interest. Here is a PDF with some pictures and explanatory notes. I hope this helps clarify.

Summary: the analogue input from relay board (LiteAudio, DIYClub) is fed to a toggle switch and than connected to either B1 preamp or miniDSP input. I use mostly Belden 8412 mic cable because it has been around as long as the Klipschorns. It is soldered directly to the little board. No plugs used. The outputs of the miniDSP are connected with 2 x 1” iron nails. Belden cable is again soldered directly to board outputs and connected to Channel 2 of the B1 preamp.

It is easy to switch between CH1 and 2 and in effect bypass the miniDSP. For experiments (bi-amping etc.) the outputs can be easily used.

Best regards, Dirk

miniDSP and B1 experiment.pdf

[Horatio]

OK, that's one for the books, for me, anyway.

I've been using a MiniDSP on my Khorns+EV HR6040's (w/Altec 288's), and a Decware SET to drive those 288's. Using the MiniDSP for over a year; added the SET last spring.

I have found that the LeCleach parameters worked spectacularly well for setting up the crossover. In this design, the low pass for the woofer can be set to something like 305 Hz while keeping the high pass for the tweeter (top end, however you call it) at 400 Hz (I think I have the LeCleach design set to 350Hz, which is why the lower end and upper end turnover points are as they are). The delay on the top is incremented a bit, and the slopes are butterworth, 18db/octave. Polarity on the top is inverted, also. Some have commented that they have tried this and that this doesn't sound any different than straight LR @400 Hz and 48db/Octave, but I beg to differ. The imaging simply pops into existence with the LeCleach parameters, at least for this combination of bass horn and top end. I just A/B'd this with the LR 48 db settings again yesterday, and it is undeniably favoring the LeCleach approach. Anyone who claims that horns can't image, or that Klipschorns can't image are in for a rude awakening with this system.

Your idea is clever! How did you arrive at 7.5 ms for the delay?

-M

[Arkytype]

Dirk,

Interesting project. I really like the concept of the MiniDSP! According to one of the MiniDSP forums, depending upon the plug-in loaded and filters enabled, the unit can have a latency of up to 2.5 msec.

I just happened to have finished the tedious task of measuring the acoustical path length thru one of my Klipschorns. The mean path length from the woofer side of the throat to the front edge (mouth) is.....wait for it.... 68 inches. That translates to a path delay of 5.014 msec. Add those two delays together and you get 7.5 msec. Of course, there will be some delay involved getting the woofer to moved after voltage is applied.

Anyone got a different path measurement for their Klipschorn?

Regards,

Lee

[Chris A]

Greg Oshiro experimentally determined 6.01 ms (bass bin to tweeter) and 4.4 ms (bass bin to midrange):

http://forums.klipsch.com/forums/p/156476/1651159.aspx

Look on the last page of the thread where he tried a different value for delay at my request, and found that his listed delays were correct.

I know you quoted bass bin delay, and I assume the you are referring to delay with respect to the front face of the bass bin, but for me, it's important to know the relative delay wrt the drivers, and to get this pretty close for the tweeter to midrange relative delay. I've found that 0.07 (one inch) ms error at the tweeter-midrange crossover point is very audible.

BTW: I picked up a Yamaha SP2060 today: once I find a fair 6-channel amp, I'll try out his tri-amp settings on my upstairs Khorns. I'm actually pretty excited to hear the differences.

Chris

[Greg Oshiro]

<snip>I picked up a Yamaha SP2060 today: once I find a fair 6-channel amp, I'll try out his tri-amp settings on my upstairs Khorns. I'm actually pretty excited to hear the differences.

Chris

Chris--

I hope those settings do something for you. Please keep in mind that those settings were based on measurements of one pair of Khorns in a specific environment. YMMV. If you want to go through the whole measurement process, I can attempt to walk you through it. I think REW can do all the measurements, but I've never used it.