Troels Gravesen high efficiency 3-way looks like a winner.

This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Troels Gravesen recently published a speaker design on his website that looks like it could be perfect for those with low-wattage tube and Class A transistor amplifiers. The speaker is a 3-way unit built with a ScanSpeak tweeter and Faital pro PA woofer and midrange. The crossover is a simple LR2 design at 500Hz and 2400Hz, and the cabinet is a bass reflex box of Baltic Birch which appears to be fairly easy to build.

Mr. Gravesen claims a sensitivity of 93dB/2.8V/1 meter. His measurements show a very flat frequency response which appears to be plus or minus 2dB from 100Hz to about 15KHz, and off axis it deviates from that only slightly at 30 degrees. Based on the raw driver specifications and the box size of the speaker, it appears that bass extension should be about F3=48Hz and F6=42Hz, enough bass for most varieties of music. Mr. Gravesen rates it as an 8 ohm speaker, and the impedance never goes below 5 ohms at any point, so it should be an easy load for most amplifiers.

I am no expert, but I believe this to be a remarkable design and I think it shows Mr. Gravesen's expertise clearly. If I were in the market for a DIY speaker this is the one I would build. I think it is perfect not only for the low-wattage folks but also for speaker building Newbies that want to get their feet wet. It can be found at
Last edited:
12" Faital 12PR320

...One designer's ported woofer -F3 ~ 50Hz in Troels Faital 3WC

...Is another designer's sealed midrange -F3 ~ 75Hz between a horn tweeter and a couple 24Hz sealed woofers

"Back in the day" Acoustic Research - AR - built a two-panel front baffle with a fixed baffle for the 12" woofer, plus a separate baffle board for the tweeter+midrange which could be rotated 90-degrees for cabinet side placement in large bookcases. Just add a baffle step boost ckt + switch.


Differences in baffle edge distances become less critical near walls, on desks, and in bookcases
Yes it looks very interesting.I too have reservations about using a typical dome tweeter with those PA type mids and woofers but it might be OK.From what I have heard bullet or slot type tweeters tend to integrate better because their inherent sensitivity is a closer match.
Does this design use BSC? The system sensitivity is very high, i.e. close to the sensitivity of the bass driver...

I've always criticised Troels designs for their apparent lack of attention to BSC. Either his quoted sensitivity figures are erroneous or he doesn't compensate correctly. His measurements aren't usually detailed enough to see exactly what's going on and his crossover values are hidden.

With some designs such as with a floor loaded bass driver or 2.5way implementation mean that BSC doesn't need to be traditionally applied, but Troels sensitivity figures always tend to imply that very little attention has been paid to it.

He also tends to crossover way too high to tweeters for optimal off axis performance.

This latest design of his has the woofer not that far off the floor and this is not one of the trendy narrow-baffle types either so perhaps there's nothing to worry about here.

He does usually show more information on measured cross-over details, perhaps he has limited the information to prevent easy-cloning by folk who don't want to be forced to buy audiophool caps and coils.
I must have measured and developed close to a 100 speakers using the Clio now. BSC just isnt such a big deal. I have used all sizes of baffles, and very small speakers require some BSC, but nothing close to what people seem to belive. Big baffles like the Faital-3WC I would guess needs very little, if any.

This construction is one that I would stay away from though. Troels has many excellent designs to choose from. F6 at 42Hz in a box I would call big already. No, make it bigger and use two woofers for F3 of 32Hz.
I've always criticised Troels designs for their apparent lack of attention to BSC.

If you look at his measurements, all bass units have a large dip at 500 hz which is not apparent on anyone else's measurements. This leads him to use a smaller than normal inductor allied with a notch filter to flatten the mid.

What would then give a 500 hz lift in most rooms plus his infamous left to right slope, to me, would make the speakers difficult to listen to.
I must have measured and developed close to a 100 speakers using the Clio now. BSC just isnt such a big deal.

I don't really see how a phenomenon caused by the law of physics isn't such a big deal. It's present in every stereotypical loudspeaker design in one way another and needs to be accounted for. Whether this means less or more of it needs compensating for depends on the design and intended application but to say it isn't such a big deal is very dismissive.

To me the final steps in balancing a loudspeakers tonal balance involve tuning the BSC and tweeter levels. If either are off by as little as 1dB it can make a design quite unpleasant to listen to. At least to my ears. Now I do have hyperacusis and that makes me more sensitive to speakers that are lacking. An overly forward sound, indicative of too little compensation, hurts to listen to.
In regard to Mr. Gravesen not listing the values of his crossover components, I read some time ago on his site that pirates were using his designs and passing them off as their own, thus his reluctance to "give away the store".

In regard to the slight downward slope of the frequency response (from bass to treble) of many of his speakers, I believe this tendency is supported by evidence that this is what most listeners prefer on axis (I don't remember the researcher or the methodology; I think it was Floyd Toole; someone please correct me if you know otherwise).

In regard to off axis performance, he reports that the Faital 6RS140 that he uses for the midrange has "excellent dispersion". The factory measurements show a very flat response on axis from 500Hz to 8KHz. The driver diameter suggests a rolloff that begins at about 2800. The ScanSpeak D2608/913000 tweeter has an Fs of 700Hz, and the response is flat down to 2000Hz. I don't know the specific measurement conditions under which he tests his speakers, but I see no reason to doubt the off-axis performance when crossing over LR2 at 500Hz and 2400Hz.

In regard to BSC and bass extension, he has designed speakers to specifically address BSC with wide baffles (Poor Man's Strad) and at least one with a downward sloping bass response meant to be placed against the wall (8008-Corner), so he does take BSC and boundary reinforcement into consideration in his designs. Though I have not analyzed all his speakers, it appears that most of them are bass reflex and transmission line designs, and I speculate that he may adjust the tuning of the cabinets to give a bit of a boost at the low end.

In the case of the Faital 3MC, the baffle width is 375mm, which suggests baffle step of -3dB at 300Hz. We don't know the DCR of the woofer crossover coil, but if we assume it to be .5ohm, a not unreasonable figure, it increases the .qts value of the Faital 12PR320 bass driver to give an approximately 1dB lift to the FR from about 150Hz to 60Hz in this cabinet. F3 becomes about 47Hz and F6 about 41Hz. This is well within the range of tone controls on preamps and receivers. The Faital woofer can handle the low EQ boost; the factory specifications suggest it can produce 108dB at 41Hz in a bass reflex cabinet at one meter. Considering that 41.2 hz is the traditionally lowest note on a double bass and a bass guitar, this should be plenty of bass response for most kinds of music with the exception of the lowest organ tones. This is a speaker for music, not movie sound effects. In addition to all this, BSC in practice may not be a concern for most people in most listening environments; I believe most listeners will place their speakers in proximity to at least one wall. For those with very large spaces, they could supplement the bass with subwoofers, or, of course, buy a different speaker. No one speaker can be all things to all people.

It is not my intention to be an apologist or champion for Troels Gravesen, but I believe in giving credit where credit is due, and I think this speaker is remarkable for a beautifully harmonized balance of performance, simplicity, and cost. I repeat, I think it would be perfect for those that need high efficiency and also for beginners that want an excellent, easy-to-construct speaker that does not cost an arm and a leg.
Last edited:
  • Like
Reactions: 2 users
Reads like an ad

I think that’s unfair. It was a good response to the preceding criticisms, providing more balance to the discussion.

I’m quite intrigued by this design and am interested to read more of everyone’s thoughts - not least because I learn from those with much more knowkedge and experience than I have.

Troels, in my view, has earned some respect. He has explored many different approaches so he has dealt with, first hand, many examples of the tradeoffs all speaker designs require. I try to learn a little of the why’s and wherefore’s from his articles.

5th element, you are also very experienced. I’ve picked up a lot from your writings.

I like mostly to design and build amplifiers, both tube and solid state. A good speaker for testing amplifiers I build needs to be that strangest of beasts - a chameleon suited to low power tubes and high power transistors, i.e. to be sensitive and have good power handling. Not being a large and expensive or difficult-to-build project makes it a lot more accessible. This latest from Troels may be a candidate.

What tradeoffs did he make that you would do differently? and where the time, effort and cost would be justified by the resulting improvement.
Last edited:
It came to me that I should post a disclaimer for those with an interest in this speaker: I have never built or heard one of Mr. Gravesen's designs, so I can not testify as to the sound quality of any of his speakers. Based upon what I understand about speaker design, this looks like a good one.

I will not post any more in this thread because I sense that it may devolve into a litany of nitpicking and personal comments. To those that are interested in it, I can only say that every speaker design is compromised in some way, and I think this one is a well-judged combination of those compromises. I think it will sound great for most people with most kinds of music. Again, I am not in the market for new speakers, but I will be interested to see if anyone builds it and posts a review.
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.