DIY home theater system

[Ethan]

Hi, this is my first post but I've been reading this forum for some time now. There's a mind-boggling amount of information here, I don't think I'll ever finish reading it all...

I had the insane idea of building a complete 5.1 set for my home theater. My goal is to get a better overall sound than my current commercial setup, and have fun in the process.


For what it's worth, my current setup is :

- Linn Ninka Front (closed box MTM 6.5" poly woofers, 3/4 soft-dome, 2250Hz xover)
- Linn Trikan Center (closed box horizontal MTM 5" poly woofers, 3/4 soft-dome, 2500Hz xover)
- Linn Kan Surrounds (MT ported 5" fiber woofer and 3/4 ceramic tweeter)
- Energy 18" paper cone closed box sub

My Tools :

- ETF 5
- Behringer 8000 mic with m-audio mobilePre on a laptop
- freeware/shareware I can find


My Skills :

- minor experience with DIY speakers. I read the loudspeaker cookbook about 10 years ago, made a ported sub, a closed box 2 ways, and a ported 2 ways. Didn't have any mesuring tools at the time.
- electronics skills and experience to design an elaborate active xover
- plenty of time to build and experiment, I give myself about 6 months to finish the project


Here's where I am right now in my criteria and decisions :

- Primary goal is to improve on Imaging, Details and Voice intelligibility compared to my current setup. I realised these qualities were lacking on my system when I went to the Montreal Home Entertaiment Show. The best speakers I heard at the show last year was a compact MTM design with what looked like a modified Mundorf AMT tweeter, and Skaaning midbass. A 5.1 kit of those was the cost of a luxury car. I hope to reach toward this sound quality with a DIY design.

- I want a Closed box design. I'm leaning toward closed box because, subjectively, I always prefered the bass of closed box. The gain of a ported design isn't necessary if I have a sub, and the sub integration should be better with a closed box. Dipole is beyond my skill level and would be difficult as a 5.1 system with only 32" height under the screen.

- Analog active crossovers. This is fun to build and play with, and also very inexpensive to modify and improve. I am more confident in my ability to design a line level xover than a passive one.

- Drivers must be available in Canada, because of the lower shipping cost and to avoid random problems with customs (had too many of those with UPS).

- maximum SPL : I usually watch movies at most around -10db of Dolby Digital level, usually lower. That is 95db peak at 12 ft, so about 107db at 1 meter.

- I want two subs to allow a vertical center channel, and to get more options for placing subs and adjust the room response.

- I want 5 identical speakers, so I can be assured of a perfect match.

- I have 32 inches of clearance under my projection screen for the center. Drivers for the center must be as close as possible to the screen, so I think an MT design is the best choice. An horizontal MTM would be better, but I understand it requires extreme mesures to acheive a good dispersion. A 2 way MT design with sub should be enough to reach 107db@1m.


Drivers :

Mundorf AMT are too expensive, so are the Skaaning drivers.

I first decided on the following drivers because of all the praises on the forums. There are plenty of DIY projects using those, so I would have a lot of references available, and I could be merely modifying an existing design :
tweeter : Millenium
midwoofer : W18E or W22
subs : Peerless 830845
system would be five identical MT on stands, with 2 subs.

However, being a little more daring, I would prefer to try the Eton ER4 with the W18E. It would be a lot more work on the crossover side and a more risky design for my level of experience, and the rear chamber will need a lot of experimentations, but I think I'm up to it.

Is it a bad idea to try pairing the ER4 with the W18E ?

I'm playing around with The Edge right now, is it reasonably accurate with rectangular tweeters ?

Anything in my criteria that do not make sense or possibly ill informed decisions ?


Thanks,

[Habs4life]

Hi Ethan , If you have not already seen it I would recommend reading Siegfreid Linkwitz's website Linkwitz Labs. Check out his Orion dipole design.It uses completely active crossovers and the superb SEAS and Peerless drivers you mentioned. (you can get them from Solen in Montreal). It is a pretty ambitious project but it sounds like thats what you are looking for.

[opt-e]

Ethan stated that he wants a closed box design, which rules out the Orion.

[Habs4life]

Yes Opt-e he did ,But if he reads Linkwitz,s strong arguments in favour of dipoles he may want to try that option. Anyway, just reading through that site is a great education .

[Ethan]

I did read Linkwitz website with great interest, but as I said specifically above : "dipole is beyond my skill level and would be difficult as a 5.1 system with only 32" height under the screen."

If someone manages to design a matching center to the Orions, under 32" height, then I found my project. Also, my center has to be at most 24" from the front wall, I think this is too close for a dipole.

[Habs4life]

Ethan, If you search his site he does have a smaller matching dipole centre channel that uses just the midwoofer and tweeter. 24" may be ok. Using Orions as rear speakers would certainly be a bit overkill. I know you said Dipoles were beyond your skills but the enclosures do not look that difficult to build ,the design has already worked out. The electronics on the other hand could be tricky depending on your experience. Anyway I just thought I would suggest it to you as an interesting option.

[Ethan]

thanks habs, I'll think about it

[Ethan]

Habs,
I reread the FAQ and according to Mr. Linkwitz, the speaker has to be at least 4 ft from the back wall, I beleive this is true for any dipole, so a dipole design is still a no-go for me :-( Another problem is the number of amp channels needed plus the cost of drivers is over my budget.

[Habs4life]

Yah Ethan its certainly not a simple project although I would like to build a pair someday.

[cjd]

I see two options for you: find a project using the drivers and techniques you want that's out there, OR change your requirements a little. If you can model and develop the active filters, you may be able to get target slopes for a passive design someone has done. Or, maybe you can find someone willing to work up an active topology for you based on an existing passive design. Or you could simply go passive.

The trouble is, unless you've got the equipment to measure accurately (looks like you do), it's *really* tough to pick your parts and find someone to whip up a design for you. Active may be even tougher, though it may be that fewer active XO designers speak up as much. I think we can help with cabinet layout, design suggestions, but I know I'm not up to trying to put together an active filter. Passive, sure. Could always try passive, figure out what the slopes are, and then target those slopes active. Which wouldn't require building the passive version, just designing it.

This doesn't mean you should give up. It just means that a project that fits the fairly narrow criteria you've set (some reasonable - like the limitations of the space and, for the most part, driver limitations to what's available to you - though it isn't always that much more to import from the US, sometimes it isn't worth it at all) may be tough to find.

The question I have a this point: are you able to actually design/build the active circuits to target specific transfer functions? Or are you limited to assembly, etc.?

Next step might be looking at box layout/design.

C

[Paul H]

Chris, you can build an active crossover - I did a 4 way last year, so I know you can do one.

Ethan, when you get started on this and get some measurements you may need some software simulations done by one of the gang here to help with the crossovers. That shouldn't be a problem. I'm in the midst of moving now, but will be settled in a month or two and will have a little time to help then.

One of the beauties of active crossovers is that you only need to get reasonably close with the first attempt -it costs almost nothing to adjust parts and tweak the design.

Paul

[Habs4life]

Yes if you meet certain conditions your chances of success using active crossovers is very good even without the aid of measurements.( although they would be a big plus). But getting good results building a passive xover without measurements and simulation software is impossible.1) the drivers should have relatively smooth response an octave or so beyond their intended xover point. 2)use relatively steep slopes such as 4th order L-R lowpass and hi pass networks. 3) add some delay to the tweeter either by using a slanted baffle or by adding an allpass network to the hi pass section. 4) have a way of adjusting the levels of each band to match their relative sensitivities. Following these rules and using Hi quality drivers and amplifiers the results can be quite rewarding.

[Dennis H]

Yes if you meet certain conditions your chances of success using active crossovers is very good even without the aid of measurements

Sorry to be a contrarian but..... No good speaker has ever been designed/built without the use of measurements. Ain't gonna happen. But the good news is LspCAD is equally happy working with active or passive designs. All it takes is good measurements of the drivers in the finished box and LspCAD can juggle opamps just as well as coils and caps.

[Habs4life]

Certainlly Dennis to completely optimize a design good measurments and simulation software are absolutely essential. But I still believe amatuers lacking these tools have a better chance of making a good sounding speaker using active rather than passive crossovers. Might I add that I am very new member to this forum (and the AVS forum) and have been impressed by how willing the experienced guy's with the test equipment and software are to help those that don,t to optimize their designs.

[Habs4life]

[QUOTE=Dennis H]Sorry to be a contrarian but..... No good speaker has ever been designed/built without the use of measurements. Ain't gonna happen. ]
It might be against the forum rules to mention the products name but a few years ago a multi thousand dollar speaker got a Class A rating in Stereophile and they published the measurements and boy were they terrible. The graph looked like a roller coaster and there was such a deep null around the mid tweeter xover that the two must have been out of phase. But some how it still got a Class A rating. Go figure.

[Dennis H]

But some how it still got a Class A rating. Go figure.

(Cynic mode on) Shrug. Stereophile's ratings have little to do with quality. More like who pays the most for advertising space and who bribes the reviewers with free products. (Cynic mode off)

[Habs4life]

(Cynic mode on) Shrug. Stereophile's ratings have little to do with quality. More like who pays the most for advertising space and who bribes the reviewers with free products. (Cynic mode off)

Yes I am inclined to agree with you on that one ! That would be the only explanation such a piss poor product got a high rating.By the way, you mentioned LspCAD is this what you use and is it reasonably priced. I am looking for some more up to date software at this might do it for me.Thanks Fred

[Ethan]

Paul,
Thanks, if I can get a little help from you guys with the xover that would be wonderful :-) Something I like about active filters is that everything is buffered, I can solve one problem at a time, it's one building block at a time.

Chris,
Well, my understanding of analog filters is limited to the usual building blocks (LP/HP, pass band, shelvings, all pass) and their relationship with FR and phase, so I think I'm able to make anything that fits my need after I mesured the output. However my understanding of xover theory is limited, I can look at the slopes and find something I can work with, but don't ask me WHY LR4 and LR8 filters are phase coherent and flat, I have no idea.

[Ethan]

Okay, so, this is my todo list for the next 6 months :

1. Choose the drivers. So far it's the W18E with ER4, but I'll give myself a few more weeks to look around and consider an existing design with passive crossover.

2. Think about a usable crossover, then go back to drivers choice if necessary : I have to be as far as possible from the resonance at 4.5k, and the ER4 starts distorting under about 1.5K, the few ER4 designs available are at 2k or above. I'll first try 1.8k with an LR4.

3. Simulate box response, ensure 107db reachable down to at least 80Hz

4. Design a baffle and simulate in Edge. Design a box.

5. Buy a pair drivers, test them nearfield, reconsider my options. I'll need to find a software that can do a sweeping THD plot, my current software (ETF) doesn't support that. I'll check what I can do with "R+D" as I have a free license along with ETF.

6. Make an MDF test baffle and mesure each drivers on the baffle. The response should match the sum of nearfield and baffle simulation (except the bass), if not, go back to step 4 and try to understand what happened, try a different baffle.

7. Mesure with different thickness of felt behind the ER4 to see what happens.

8. build a box.

9. Add the crossover : start with an LR4 at 1.8k, try something else later

10. Adjust gain between woofer and tweeter

11. Add Notch filter for the woofer resonance

12. Add All-Pass filters to align the tweeter around the xover frequency.

13. Add Shelving low pass for the baffle step

14. Mesure, then consider additional filters to flatten the FR, try other xover freq

15. Finish all 5 speakers :-)

[Paul H]

When you're getting into this let me know - I've got a 4th order 1800Hz low pass with a notch filter on my current W18's in a 4-way dipole.

Paul

[Ethan]

Okay, I finally have a nice box design that I really like. It's a 400mm x 250mm baffle with heavy bevels on the four corners (45mm), and a beveled top edge (25mm). Much inspired from the avalon and m8ta, but smaller, and less nightmarish to build 8O I didn't put the surfaces of the bevels in the simulation, I assumed they didn't count.

By having the speakers at a 400mm height, it will allow me to eventually place a 830845 under all three front speaker, and add an active crossover to make them 3-ways.

With passive xover for the mid/tweeter I woudn't need to buy another amp, so I decided to continue with the Millenium because this opens up the possibility of using an existing crossover design (passive) without too much modifications, I think, well I hope




edge-25mm



edge-ER4

[Paul H]

Which passive crossover design were you looking at using?

Paul

[Ethan]

I'm not sure yet it depends on my woofer choice. One thing's sure, I'm pretty much set with the box design and the tweeter choice. Now a 6.5" or 8" driver should be a little easier to pick : closed box of 12L and maybe up to 24L with lot's of depth, and I can crossover very low because of the millenium. So a good looking metal cones is prefered

I assume any xover done for a ported designs can be used closed box with a smaller volume if the woofer has a high enough Qts, and I keep a similar baffle width and same drivers.

I'm looking at the many options Jon did for the Modula MT right now... I never heard of CE crossovers before reading that thread. The Modula MT is already close to what I want, there's a xover version that should work with the millenium, I could maybe order the RS180 from the US. Should work well with my box design even with the tweeter centered, the trapezoidal shape is taking care of the upper band diffraction ripples pretty well. Maybe it could be adapted it for the W18. But again, I can get away with a lower xover with the millenium, so the Modula MT xover might be overkill for me (that is a seriously expensive xover)...

Another option I consider is using the M8a in my closed box, with plenty of depth I could get up to 24L. So maybe the M8ta crossover could be usable, but the main lobe would be a little downward even if I slant the baffle Also, the baffle might be too small and narrow to work withoud modifications. But I like the price of the M8a compare to the W18.

By using the W18 and Millenium, I could use the L18 and 27tdfc for the surround channels to save some money, should need to remove or shift the notch on the xover, not sure what else...

I also checked the "Seas Trym". It's a 2400Hz xover with a W22. Something's not right about that. It goes against every recommendations for the Excel woofers

Any suggestions ?

thanks !

[cjd]

I disagree with steps 9 on...

I would say, after building the box:

install drivers
measure in-box at one to two meters, each driver (or driver set) separately.
Spend some time modeling
Build
listen
measure
tweak

You get fewer components than your process (where you're doing crossover, notch, all-pass, baffle step all separately) and in the end cleaner. Plus, you can do a LOT of experimenting with different potential topologies.

Speaker Workshop is free (though a bit rough to learn) and does most of the measurements you want, as well. So, IMHO, no excuse.

[Paul H]

Chris,

I believe those steps were put together with active crossovers in mind, which would mean cheap and easy crossover modifications compared to passive.

Paul

[cjd]

d0h...

forgot about that part.

Not sure if that can be modeled up free, but your'e absolutely correct.

C

[Ethan]

Oh, didn't know about speaker workshop I downloaded it and played with it, with time I should be able to tweak a passive design with this.

I'm already rethinking my choice of drivers, maybe I should just pick something and get over it !

[cjd]

Well, lots of folks happy to gander at passive work, including me. Can just share the SW files as often as not. So, if you're stuck or want a second opinion or whatever else...

I think you're playing on the pricy end of the scale driver-wise, but not unreasonably so.

Your core driver arsenal is awfully close to what Brian chose for the 3-ways I worked up a crossover for. If I had bi-amp (and possibly active control, though bi-amp would do it well enough passively), I would have strongly considered RS265HF's in the same box ported to some low frequency (don't remember what it was is all).

Take your time, evaluate the different options. It's sometimes tough to be patient but in the end it's worth doing.

C

[Ethan]

Yeah that's what I was thinking, it's getting a little expensive considering I need 5 of them ! I'm not sure how much I gain with W18/T25CF002 compared to an L18/27tdfc. But again I want to be sure I'll get something better than my current speakers...

Thanks

[cjd]

Have you looked into what it ultimately costs to get the Dayton RS drivers to your door? (they'll ship USPS, you just have to ask)...

Peerless nomex (ori s it Exclusive) are also turning in some very good numbers I think (distortion-wise).

[Ethan]

yep thanks, rs180 is on my list and I added some more dayton, peerless, Hi-Vi and Scan-Speak. I decided to eliminate the L18 since the rs180 is less that half the price and is supposed to be overall just as good (according to Mark K's tests and conclusions at least).

Now if I set a maximum of 250$(canadian) per speaker excluding xover and wood :

Tweeter considered :
- 27tdfc 40$
- 27tbfc 45$
- xt25 57$
- rs28a ~60$ including shipping
- SS9500 146$
- SS9700 172$
- SS9800 187$
- T25-002 257$ hmmm. maybe a -7$ woofer with this one ?

Woofers considered :
- W18E 225$
- 830883 90$
- D6.8 77$
- M8a 72$
- M6a 55$
- RS225 ~55$ including shipping
- RS180 ~40$ including shipping

Honestly, unless the dayton woofers are much better, I want to use the M8a or M6a, just because they look great The M6a has too little excursion. So I'm back with trying a variant on the M8a-mkIV I guess.

Tentatively : With the M8a in a 18L closed box I get a (slightly) high Qtc of 0.8, the FR looks adequate to me for a xover to a sub. Is there something wrong with a high Qtc ? I noticed most great designs are 0.5. I have a weird feeling that there's a basic concept I don't get, but I can't put my finger on it

[Paul H]

A few thoughts ...

Of the tweeters you've listed, the bottom 4 are very expensive. I have SS9700's and really like them, but they do cost a lot and are into (I think) the area of rapdily diminishing returns.

I also have XT25's, which is a good tweeter (at least until I heard my SS9700's ) but doesn't hold up so well in distortion testing below 3000Hz. I may re-use mine in a 3-way using a higher crossover some day, but I don't think I'd pick these for a two way given the others on the market now.

9700's are only sold in matched pairs incidentally (per Solen few months ago).

You won't go wrong with the Scanspeaks you've listed, but you might get similar/close performance for less $$ with the RS28 or 27tdfc.

On the woofers, there's been lots of good publicity and testing and reviewing here on this forum for the dayton reference series, but I don't have any (yet).

I do have M8N's, and they do look great, and sound good too. I also have W18's - IMO they're more expensive because they're just better - incredible detail in the midrange.


Paul

[cjd]

Q above .707 is officially into muddy and loose. Or so the theory goes. My limited experience suggests it's probably not inaccurate.

The quality of your low frequency addition will dictate what direction to go. There are (passive) designs posted here for both M8a and RS225 (if you include the M8a, you've gotta include the RS225...) 2-ways. Also a whole pile with the RS180. If you need broader reach down low, the larger drivers will be a bonus. But if your subs can get up to 200Hz comfortably, a steeper (8th order C-E would be great) filter will match up to a smaller mid-woofer, which makes crossing to the tweeter MUCH less of a pain (1200Hz for M8a/225 vs. 1400-1600 quite comfortably with the RS180). W18 is indeed a superb driver, but I am not sure it's worth the cost difference over the RS180. If you're down to that last little bit of trying to squeeze the last drops out of your system, it's a different game.

Of the tweeters, I would stick with the Seas or the Dayton. The Vifa (as mentioned) does well but is really rough to get sounding decent below 3kHz.

This is, however, just my own thoughts. I really like driving the ultra-dollar cars, but I own a compact (granted, one of the best options for handling off the showroom floor anywhere, when it was available). The streets I drive on every day let me enjoy this car as much as I think I could enjoy that McLaren.

C

[TacoD]

I would suggest Scanspeak 9800, in my experience it sounds more airy than the RS28 and Seas budgetline. The Seas millenium is also very good, but I think the 9800 is a better match for the RS180. If you're interested, I've posted on this board a 2-way cauer filter which uses the RS180 & 9800.

[Ethan]

okay, these are the options I like so far:

the HARD way :

W18 + 27dtfc :
- 3 subs on active xover
- xover at 200Hz, while leaving the surrounds on the receiver xover.
- possibly best possible midrange
- more xover design to do, both active and passive
- more expensive, need to buy another power amp

RS180 + RS28a :
- as above, but less expensive and easier on the passive xover (many many designs already made)
- worse midrange
- not as cool looking


the EASY way :

M8a + RS28a or 27tdfc or SS9800 with dual subs at 80Hz (mono)
- box between 22L and 24L + stuffing (~26L) would be a 0.707 alignment, should have adequate transients considering the xover at 80.
- with SS9800 xover of the M8a-mkIV possibly usable almost as-is, or with minor tweaks
- sub xover from the preamp, and no additional power amps needed
- upgradable in the future to 3 subs with active xover

RS225 + RS28a or 27tdfc or SS9800
- using the RS28a, can use Mark K's xover
- similar to above but again not as cool looking, you know I think I don't like black cones, YMMV

[cjd]

The only additional thing I would mention: The difference between "better" and "worse" isn't large (W18 to RS180)

I would also cross the W18 higher, or the RS180 lower - at least, the RS180 has a little more down low while giving up a bit more on the top. It also requires a bit lower crossover to the tweeter IMHO.

For HT, there is some advantage to black...

C

[TacoD]

The grey cones also sound better ...

[Ethan]

On the SS9800 datasheet, it says the linear xmax is 0.1mm, is that a typo ?

Can I really use the 9800 with a 1200Hz xover or I'm I better off with the rs28a if I want a metal dome ?

Thanks,

[dlneubec]

Hi Ethan,

Since you seem to be headed toward somewhat of an active/pasive hybrid and did express interest in dipole, have you looked at the NaO series? Check out this link if you have not. John K. might be able to help alter his designs to fit your height challenges. Just some food for thought.

------------------------------------------------------------------
Edit by moderator to remove direct link to a commercial website

[cjd]

I hadn't noted dipole as being feasable, or I think the Arvo Part would already have come up.

C

[Ethan]

I decided to go with the 8" 2way M8a/ss9800. The box will be 250mm x 400mm x 400mm, with a Q of 0.7. The xover will be designed once I have the drivers mesured in a finished box, and can simulate in speaker workshop. I might simply go with the M8a-mkIV xover... but I want to try things, otherwise where's the fun

Now for the subs, I want them compact and also match the satellite design. I admit I have a tendency to do a little "form over function" but as long as it's not too detrimental to SQ.

- will initially crossover mono with the preamp, will upgrade later to a higher freq active xover.
- 3 subs
- box will be a 400mm cube to match the satellites design, and that also happens to be the remaining space under the center channel
- That means a maximum of 40L closed box, so alignment will need to be adjusted with a Linkwitz Transform.
- will have bevels identical to the satellites
- must reach -10db from the dolby digital calibrated level. Excluding LFE. 95db*5 = 108db in room for the three subs.
- the amp I have available for the subs can drive 120W into 8 or 4 all channels driven, and 240/4, 160/8 with 2 channels driven.

Drivers I considered:

- Peerless 830845
- Dayton TIT320C
- Dayton RS315HF
- HIVI M12

All factors considered, the rational decision would be to go with the RS315 as it seems to be the best bang (and quality) for the buck. However, for obvious reasons, I'm hoping the M12 will work :
- It matches the M8a visually
- It isn't black :lol:
- It's the least expensive
- Can be order from Solen
- Should give great results with a higher xover
- 7.2mm xmax * 515cm2 Sd * 3 subs, should be plenty for the SPL I need, considering boundary gain and room gain.
- All the other subs are a little overkill, and their excursion wouldn't be reached without buying another multi channel amp (considering a 40L enclosure).

Thank you all for the insights :T

[Paul H]

I agree with the points noted in favour of the M12, except the last two.

A driver with 515 cm2 SD and 7.2mm Xmax can (per Linkwitz's sound level spreadsheet) hit a maximum of 86dB at 20Hz in a box speaker. This doesn't include room gain, but is still pretty low. The M12's will be very strained to play low-frequency bass, especially for movies.

If I was putting together a speaker with M8's, SS9800's (good drivers incidentally) and M12's, I'd do a three-way (TMWW) with 2 12's on the bottom and use the speakers full-range.

Ultimately (later, next project ) I think you'd still want to add a subwoofer to this setup (ie high-excursion 15").

I'm not suggesting larger drivers/more bass just to play continuous "tear-the-walls-down" volume, but more to take the load off the other speakers so they can play cleaner and clearer.

Paul

[ThomasW]

I have the M8a-MKIII (M8a and Focal TC120) in my bedroom and they make a very good bedroom speaker. For HT or main listening room use, I suggest nothing smaller than a MTM.

[Ethan]

I'm not sure I'm calculating this right but 3x M12 is as much SPL as a pair of Linkwitz THOR. I realise you guys like it loud, but I don't think I mentionned that I'm in an appartment :rofl: Anyway, I might add a sub later for the reasons you mentionned, cleaner bass is always a plus. Or I could keep my 18" for that purpose. Many options to improve later in that area.

A TMWW with those is a great idea, it's still possible to do later, I could simply stack another M12 box, I would end up with a modular 1200mm tower. Or even use one with each speaker and have five identical 3 way. I have the amp channels available anyway.

So the possibilities for future additions are :
- two additional M12s for TMWW Left/Right
- 200Hz active xover
- Additional long excursion subs
- 7.1 ?
- sell everything to a friend, new project ?

okay... I'll start with project number one. 8)

[Ethan]

I'm considering using the new D3004/6600, the FR and dispersion look incredible. It might work well with a low xover, but I haven't found any independent test yet. Do you think it's worth the risk ?