Bagby Mandolin DIY Speaker Review

[tktran303]

Hello mac,

good to see you here. You introduced me to Hypex over a decade ago.

still a fan of open baffle speakers I see.

I still have dipoles in a second house, but moved back to monopoles because they do more things right more of the time.

 

[tktran303]

here’s a speaker that gives a very good spinorama

https://www.audiosciencereview.com/...ds/genelec-8030c-studio-monitor-review.14795/

At 96dB the distortion is dirtier than the Jeff Bagby Mandolin 2. But interestingly not labelled as “unhappy” I wonder what that means

 

[richard12511]

I think there’s a pretty easy (and quick/dirty) test to get a real feel for that.

Just set your system volume for whatever you generally like to listen at.

then unplug all but one speaker.

then go sit at 1m to that speaker, then play this sweep

https://www.audiocheck.net/testtones_sinesweep20-20k.php

Using an SPL meter (or equivalent eg. iPhone with MobileTools or favourite SPL app- these will get you in the ballpark +/-1dB)

wind up (or down) your main volume control until you read 96dB-

You may be surprised that 96dB @1m from a single speaker is quite loud- near the limits of what you can tolerate.

A sine sweep at 96dB is way louder than 96dB music, though, which was kinda my question. I have no doubt I could barely tolerate a sine sweep at 96dB, but 96dB with music is easily tolerable, as the upper treble is 30-40dB down. Sine sweep plays all frequencies at equal volume. That's why I asked about how bass frequencies specifically(where speakers seem to distort the most).

96dB at 80Hz may not be that loud at all, and easily could be the average level I listen to. In that case, the 96dB sweep is very relevant for me(and others who listen mid to high 80s). I don't know, though.

I think I need to try what @ROOSKIE suggested with a spectrum analyzer to see what my average levels are at ~80dB.

*Edit: Grabbed a phone app to track(not sure how accurate). Used my usual -12 on the AVR. With an average level of 86.1dB(c), the average @ ~80Hz was 94.6dB(c). Usual volume for me is around -12, but sometimes I push it to -5 or so.

 

[gopher]

Welcome to ASR!

While clearly some nice woodworking, any further discussion of these clearly belongs on a new thread. Feel free to start one and I (and anyone else who is willing), will help evaluate.

Thanks @Rick Sykora - Started new thread here.

 

[Xyrium]

Not horrible overall. I don't feel as if the low sensitivity is a problem, though it is a hair below average probably. Just 3 more dB would have been very nice. I also like the shelf above 5kHz, I tend to like such a rolloff.

Easy fix to chamfer those edges though, whether on a table saw, or router, even after the box is glued.

So, Rick, I see you got the S2000, and a handful of other near DIY designs, any consideration for the Carrera? I've pondered it a gazillion times, but even finishing up a flat pack cab would be annoying for me right now.
https://meniscusaudio.com/product/carrera-speaker-kit/

 

[Everett T]

Not currently, but if I did it would be LXmini.


When drivers are selected for low distortion in their operating ranges and sufficient power is available, it's easy to listen at loud levels.

My BE tweeters, Eton midranges, and Satori woofers are very low distortion and it's still very loud to me. I just was pointing out that most, in average listening rooms, wouldn't be listening near that loud.

 

[Rick Sykora]

Not horrible overall. I don't feel as if the low sensitivity is a problem, though it is a hair below average probably. Just 3 more dB would have been very nice. I also like the shelf above 5kHz, I tend to like such a rolloff.

Easy fix to chamfer those edges though, whether on a table saw, or router, even after the box is glued.

So, Rick, I see you got the S2000, and a handful of other near DIY designs, any consideration for the Carrera? I've pondered it a gazillion times, but even finishing up a flat pack cab would be annoying for me right now.
https://meniscusaudio.com/product/carrera-speaker-kit/

Maybe if membership funds, but have done a bunch of small 2-ways at this point, so am looking for something different if it comes out of my pocket. For the price, I need something with better bass extension than little woofers produce.

 

[ctrl]

At 96dB the distortion is dirtier than the Jeff Bagby Mandolin 2. But interestingly not labelled as “unhappy” I wonder what that means

This probably means that Amir is human, has good and bad days and has subjective views and perhaps a little reservations about DIY speaker

Your example is very well chosen and shows impressively why objective data is so important and why subjective comments on objective data can be very manipulative. But this has been discussed many times, is Amir's decision to make and should be discussed further here.

The Genelec woofer is clearly worse than the SBAcoustics woofer of the Mandolin in terms of harmonic distortion.
This becomes especially clear when one looks at the frequency range 50-150Hz (which still plays an important role for all music genres), where the Genelec woofer shows significantly increased third (blue curve) and fifth order (green curve) harmonic distortion.
Especially the strongly increased fifth-order distortions are definitely hardly masked anymore (similar applies to the third order distortions) and should be audible.

Mandolin

Genelec

 

[Zvu]

but lots of research into THD has shown that it alone doesn’t correlate with sound quality-

how does it sound to the ear? @Zvu

For years i've wanted to make something of a volkslautsprecher. I don't care how much hours i must invest in measuring and crossover development because i believe in free stuff. My goal was to look at measurements and parameters to find the best midwoofer that will not cost over 50€. Naturally SB17NRXC fit perfectly. I don't like the sound of plastic cones so MFC was out of the question and NAC was more expensive. Dayton was at the time available only in US. While now it is available in Europe, prices are something like >40% higher (which i can explain only by greed of local sellers since shipping from Taiwan shouldn't be that more expensive for Europe, or at all).

Anyhow, i wanted to do a passive ported two way with waveguide loaded tweeter, linear measurements, low distortion, without port issues for about 200€ in material. Copyleft of course. I've tried to make them work for a long time. Two versions of loudspeaker cabinets were tested - circular and slot port.

Fact is that when i compared it with two other loudspeakers (both of which i've designed) it fell short in midrange clarity quite a lot. It shocked me because it measured good and build quality of those woofers is superb. We are not sensitive about low frequency distortion which doesn't mean that we don't hear it but rather that we aren't too bothered by it. I wasn't too bothered by it also, but midrange clarity was very bad compared to say my two way with Seas L15.

It's Serbian Easter today so im visiting my family but as soon as i get back home i'll post some pics.

I am not saying that Mandolin is a bad loudspeaker when compared to what is available at this moment on the market. It has a great value. Depending on type of music it can sound superb. But feed it some piano or music with lots of male voices and flaws begin to show when A/B with other loudspeaker.

 

[Ilkless]

Considering this was an older mainstay of (RIP) Jeff's extensive body of work, no excuse needs to be made for it. This is a very good design. For a flat baffle speaker in this form factor, that DI is really good, barely any of that ugly peak-dip you see in less competently-designed flat baffle speakers (which is imho more impressive than just looking at the spin). Nonetheless, tuning the on-axis so that LW is as smooth and similar to the ER and PR as it is also demonstrates Jeff's keen ear for voicing.

 

[Rick Sykora]

For years i've wanted to make something of a volkslautsprecher. I don't care how much hours i must invest in measuring and crossover development because i believe in free stuff. My goal was to look at measurements and parameters to find the best midwoofer that will not cost over 50€. Naturally SB17NRXC fit perfectly. I don't like the sound of plastic cones so MFC was out of the question and NAC was more expensive. Dayton was at the time available only in US. While now it is available in Europe, prices are something like >40% higher (which i can explain only by greed of local sellers since shipping from Taiwan shouldn't be that more expensive for Europe, or at all).

Anyhow, i wanted to do a passive ported two way with waveguide loaded tweeter, linear measurements, low distortion, without port issues for about 200€ in material. Copyleft of course. I've tried to make them work for a long time. Two versions of loudspeaker cabinets were tested - circular and slot port.

Fact is that when i compared it with two other loudspeakers (both of which i've designed) it fell short in midrange clarity quite a lot. It shocked me because it measured good and build quality of those woofers is superb. We are not sensitive about low frequency distortion which doesn't mean that we don't hear it but rather that we aren't too bothered by it. I wasn't too bothered by it also, but midrange clarity was very bad compared to say my two way with Seas L15.

It's Serbian Easter today so im visiting my family but as soon as i get back home i'll post some pics.

I am not saying that Mandolin is a bad loudspeaker when compared to what is available at this moment on the market. It has a great value. Depending on type of music it can sound superb. But feed it some piano or music with lots of male voices and flaws begin to show when A/B with other loudspeaker.

Please enjoy your Easter and family, but when you are able, just want to know if your experience applies to the newer NRX2C as well the NRXC? Earlier you mentioned an edge resonance that appears to be the dip in Amir’s nearfield, but it also appears SB targeted that aspect for improvement in the newer woofer. Thanks!

 

[Zvu]

Thanks.

It was older SB17NRXC35-8. I didn't get the chance to use the newer version of it.

Maybe version 2 of the midwoofer is fixed in that regard.

 

[richard12511]

Not currently, but if I did it would be LXmini.


When drivers are selected for low distortion in their operating ranges and sufficient power is available, it's easy to listen at loud levels.

I see you live in the same city as @amirm. Would you ever be willing to drive over to his place to have the LX521 measured on the NFS? It would be a huge service for those of us(like me) who plan to build this speaker someday. It's what I would consider an end game speaker, so it would be wonderful to see exhaustive measurements.

 

[devopsprodude]

Preference Rating
SCORE: 5.4
SCORE w/ sub: 7.4

Sensitivity: 82.9dB (300Hz-3kHz ; spec: 85dB)
Frequency response: +/-2.5dB 80Hz-20kHz

​

​

Wow, that's one of the better testing speakers to date in terms of preference score.

 

[Francis Vaughan]

SL designed all of his speakers with priorities surrounding room interactions. If you read his writings you'll understand that his objective was to make the room part of the music experience instead of fighting to eliminate it as most audiophiles do.

One might argue it was even more scientific than even that. His designs are clearly important and it isn’t as if our favourite luminaries were not to unaware of his work.
Floyd Toole has this to say about SL on his passing.

Siegfried Linkwitz was a rare combination of music lover who could hear analytically, and an engineer who could design, measure and interpret the performance of loudspeakers.
He contributed much more than his well known crossover topology.
He created novel loudspeaker designs and candidly reported on their performance in objective and subjective terms. He respected the scientific method. His insights and observations will be missed. RIP

IMHO there is a huge benefit in measurement of any and all of his speakers. But I agree. The standard ASR quick review procedure is inappropriate. CEA-2034 cannot be used to generate in-room response, as it uses a weighted average of the squares of the power in each direction. By definition it does not take phase information into account, and will cheerfully calculate a massive peak where no such peak exists. Nor does it take path lengths into account, and thus basically cannot successfully predict a dipole in-room. This has been a problem previously, so recognising the limitations early is important. The impact of the crossover's correction of di-pole cancellation will similarly end up incorrectly in the predicted response.

We don’t really want a review. We want a set of careful measurements. The importance of the LX521 is such that it deserves more than being lumped in with all the me-too $300 bargian two ways. The spinorama will be of extraordinary interest, as it will provide a basis for understanding the speaker and advance the state of the art. If we could convince Klippel to deliver phase information in the spinorama there is much more that could be learned. Understanding from a quantitative point of view how the SL speakers work in a room would again advance the state of the art.

It would be reasonable to reach out to Floyd as well. I can’t believe he would not be interested in participating in the investigations in some manner.
If transport is a problem I bet more than a few ASR members would be prepared to chip in. I’ll be happy to.

 

[Everett T]

One might argue it was even more scientific than even that. His designs are clearly important and it isn’t as if our favourite luminaries were not to unaware of his work.
Floyd Toole has this to say about SL on his passing.


IMHO there is a huge benefit in measurement of any and all of his speakers. But I agree. The standard ASR quick review procedure is inappropriate. CEA-2034 cannot be used to generate in-room response, as it uses a weighted average of the squares of the power in each direction. By definition it does not take phase information into account, and will cheerfully calculate a massive peak where no such peak exists. Nor does it take path lengths into account, and thus basically cannot successfully predict a dipole in-room. This has been a problem previously, so recognising the limitations early is important. The impact of the crossover's correction of di-pole cancellation will similarly end up incorrectly in the predicted response.

We don’t really want a review. We want a set of careful measurements. The importance of the LX521 is such that it deserves more than being lumped in with all the me-too $300 bargian two ways. The spinorama will be of extraordinary interest, as it will provide a basis for understanding the speaker and advance the state of the art. If we could convince Klippel to deliver phase information in the spinorama there is much more that could be learned. Understanding from a quantitative point of view how the SL speakers work in a room would again advance the state of the art.

It would be reasonable to reach out to Floyd as well. I can’t believe he would not be interested in participating in the investigations in some manner.
If transport is a problem I bet more than a few ASR members would be prepared to chip in. I’ll be happy to.

It will be a difficult speaker to measure with the Klippel, if I understand the methodology. I've heard two of the Orion versions and thought they were both fabulous and the mini, which was very interesting.

 

[amirm]

If we could convince Klippel to deliver phase information in the spinorama there is much more that could be learned.

All computations in Klippel NFS are with complex numbers so phase is there. Indeed I can plot contours of phase in the visualization. I also have a license for the expensive option to export data with phase. Problem is what to do with a massive dump of numbers. Spinorama does not specify anything with respect phase so no way of "cooking" the data in a presentable way with phase.

 

[Francis Vaughan]

I also have a license for the expensive option to export data with phase. Problem is what to do with a massive dump of numbers.

Oooh. That would amazing. Perfect challenge for the ASR masses. There are enough tools, open source and not that could be brought to bear. It won’t be easy, but I do think there is valuable science to be done here.
What sort of size of data is involved?

 

[sutantoroy]

Im not sure if driver location is meant to built like the one reviewed, what i see at the meniscus picture it has a non center tweeter location.

 

[Rick Sykora]

As promised, checked up with Meniscus on the design discrepancies for this speaker and here is what I found:

1. The tweeter position was changed by Jeff and so the drawing is correct and the review speaker is as he intended in that regard. This also means I do not plan to change the speaker I built.
2. The box tuning difference is going to require some additional research and should have a response tomorrow.

In any case, it is a solid performer for the price and, as it is DIY, the port length can be adjusted to taste. Once Amir returns it, will try the longer port and do a nearfield measurement comparison.