The Darbari: new speaker project

tcpip

Well-Known Member
Joined
Mar 1, 2010
Messages
542
Points
93
Location
Bombay
Am starting a new speaker project. After building a few pairs for friends, I am finally aiming to build these to keep for myself. I am calling this the Darbari.

Initial thoughts are here.

Have done a draft of the enclosure design. Each channel will have two boxes.

Comments and questions most welcome. I have ordered the crossover; will start work on enclosures any day now.

Wish me luck. :sad:
 
All the very best for your Build, will keep watching the build progress

i got to know that a three way speaker is ideal to distribute the work perfectly, and what we get to hear is a complete range ( Loss-less ), which intern eliminate the need to add a SUB woofer etc to get that deepest bass

Appreciate your knowledge on speakers, crossovers etc

All the best

Tanoj
 
All the best tcpip,

Some noob questions :)
~If one can find 'good' full ranger with 'sufficient' top end wouldn't two way with single crossover do ? Like full range + (sub)woofer. Not separate sat-sub, but one enclosure.
~ Is preamp analogue or is it with built in DAC. Instead of feeding low level signal through long cables from source, wouldn't amplified signals be better if sent from a active integrated amp ?
~ Enclosure design looks superb. Would the top two way enclosure benefit if we tapper it a little from bottom to top and tweeter little recessed to avoid defraction ?
~ How are the top (two way) and bottom enclosures joined ?
Regards
 
Am starting a new speaker project. After building a few pairs for friends, I am finally aiming to build these to keep for myself. I am calling this the Darbari.

Initial thoughts are here.

Have done a draft of the enclosure design. Each channel will have two boxes.

Comments and questions most welcome. I have ordered the crossover; will start work on enclosures any day now.

Wish me luck. :sad:

Congrats Tarun

Nice project, tell us more about crossovers, i assume you are going for actives with this project.
 
Last edited:
~If one can find 'good' full ranger with 'sufficient' top end wouldn't two way with single crossover do ? Like full range + (sub)woofer. Not separate sat-sub, but one enclosure.
It depends on whether you consider a crossover the bigger evil or the imperfections of a single full-range driver.

A single full-range driver has very large challenges to tackle, purely from the laws of physics if not from engineering aspects. For instance, if it is a mid-sized driver, it will beam too much at high frequencies (i.e. poor off-axis dispersion). If it is a small driver, it will beam less at high freq but will need to have huge Xmax to reproduce low frequencies adequately. And so on.

So, in theory, your assertion is right. If you can find a "good" full-ranger, then a two-way, or even a single-driver, will work well. In practice, I have yet to see a full-range driver I will be happy with. (I've lived with perhaps one of the best full-range drivers ever made, the Jordan JX92S. And I've often said that I'd like to use it only as a midrange in a 3-way.)

~ Is preamp analogue or is it with built in DAC. Instead of feeding low level signal through long cables from source, wouldn't amplified signals be better if sent from a active integrated amp ?

In my case, there will be no separate preamp. There will be a single digital xo box which will have a volume control at its output. And your question about long signal cables versus long speaker cables is independent of what I'm doing -- it's a generic question. The conventional answer to your question would be "no". Longer speaker cables are supposed to be a bigger problem than long signal cables. In general, anything carrying high currents is harder to handle well than something carrying maybe 5mA at 100mV.

~ Enclosure design looks superb. Would the top two way enclosure benefit if we tapper it a little from bottom to top and tweeter little recessed to avoid defraction ?

No, such taper would not give me any advantage. Rounded or bevelled corners and edges will help. I intend to do them anyway, as shown in the diagram -- I am intending to apply bevelling in many ways on the side edges of the front baffle.

~ How are the top (two way) and bottom enclosures joined ?

I don't intend to join them. I intend to place the top unit on top of the bottom unit and align their front baffles. This will make the whole thing easier to manage at home. I intend to put some damping material between the two boxes to prevent low freq vibrations from passing upwards to the top unit.
 
Last edited:
Nice project, tell us more about crossovers, i assume you are going for actives with this project.
Yes, I intend to use a digital active xo, the MiniDSP 4x10. :)

I don't know what the crossovers will be like, but I am expecting LR4 at 100Hz for the low-pass on the woofer and the high-pass of the mid. The crossover between mid and tweeter will probably be a Cauer-Elliptic at 1500-1600Hz, followed by a hand-tuned notch filter on the mid to aggressively cut out the cone-breakup peak of that metal cone.

In addition, there will be LT on both the mid and the woofer, to get their low-end right. The enclosures will be sealed, therefore the LT will be needed to tune the low end. I'm thinking of keeping a Q of 0.6 or even less for the woofer.

And finally, I will use up 2-3 channels of parametric eq in the crossover to get my far-field low-frequency response (150Hz and below) reasonably flattish.

That's the plan till now. Let's see how things shape up. :D
 
Last edited:
All the very best for your Build, will keep watching the build progress
Thanks for the kind words. :)
i got to know that a three way speaker is ideal to distribute the work perfectly, and what we get to hear is a complete range ( Loss-less ), which intern eliminate the need to add a SUB woofer etc to get that deepest bass
I too used to feel this way, but seeing some of the better MTM two-way designs, I am not sure there's anything lacking in the bass. You certainly don't need any subwoofers with a good 2-way MTM, even with the Asawari Mark II. :)
 
I've used the drivers you noted and they are really good.

this is a design philosphy that I also use - Metal drivers, sharp slopes, active filters. be prepared for a lot of tweaking though

RS150 is a good choice if you want to go down to 150hz, but even with a cauer-elliptic - 3khz might be a bit high for the -70db goal that you have for the first break up. crossover any lower and you are then compromising your design goals of a crossover free midrange :)

you might want to evaluate the peerless hds 6" nomex which is on sale at a parts express for less than 50$. it is hard to argue with the midrange quality of that driver and it can do 3khz with without the need for sharp slopes.

I use the Rs225 in my mains and I can tell you that it's a good choice. but it is a 8" not 10" (RS270). either way, if you are crossing over at 150hz, then you could safely use a proper sub; if you have the amplification; the RS210 or RS265HO are good bets for a small sealed box with lots of extension.

lastly, would love to hear you the minidsp 4x10 sounds like. I have the 2x4 (two of them) in my car for a 4way active setup.
 
Yes, I intend to use a digital active xo, the MiniDSP 4x10. :)

I don't know what the crossovers will be like, but I am expecting LR4 at 100Hz for the low-pass on the woofer and the high-pass of the mid. The crossover between mid and tweeter will probably be a Cauer-Elliptic at 1500-1600Hz, followed by a hand-tuned notch filter on the mid to aggressively cut out the cone-breakup peak of that metal cone.

In addition, there will be LT on both the mid and the woofer, to get their low-end right. The enclosures will be sealed, therefore the LT will be needed to tune the low end. I'm thinking of keeping a Q of 0.6 or even less for the woofer.

And finally, I will use up 2-3 channels of parametric eq in the crossover to get my far-field low-frequency response (150Hz and below) reasonably flattish.

That's the plan till now. Let's see how things shape up. :D

Sorry.. when i was replying, I missed this post.

LR4 @100hz? you might find a lot of plusses of RS woofer at the bottum are not being utilised. RS woofers are great fro low distortion below 1khz. Also, let me see if I can pull up the data, but afaik, some RS subs have lower distortion below 100hz than the RS woofers, at everything above nominal volume. (broader BL curves)

RS150 at 1.5khz? break-ups will not be a problem. but does that not compromise your design philosophy? also you might want to consider LR4 instead of LR8 if you are putting a passive notch .. might help your power response.

Will follow this thread with interest.
 
Last edited:
I've used the drivers you noted and they are really good.

this is a design philosphy that I also use - Metal drivers, sharp slopes, active filters. be prepared for a lot of tweaking though
Good to know.

Tweaking --- yes, I guess I will be prepared for that. My crossover is a tweaker's dream -- it'll be all digital. :D

RS150 is a good choice if you want to go down to 150hz, but even with a cauer-elliptic - 3khz might be a bit high for the -70db goal that you have for the first break up. crossover any lower and you are then compromising your design goals of a crossover free midrange :)
Frankly, I am not sure that one needs to treat the midrange extending up to 3K for human voice reproduction. I am comfortable putting a crossover at 1.5 to 2K, and I am not sure this will cause a problem. I'm a bit flexible when it comes to philosophical issues. Jokes apart, I feel that getting the fundamental frequencies of the human voice (150Hz and above) right are probably more important than getting the extreme high overtones of the female voice (2.5K and above) into the same driver.

you might want to evaluate the peerless hds 6" nomex which is on sale at a parts express for less than 50$. it is hard to argue with the midrange quality of that driver and it can do 3khz with without the need for sharp slopes.
Unfortunately, for this project, the drivers have been bought and I do not intend to change the choice. I will pay my dues and suffer the consequences, but I want to build this Darbari with this set of drivers and then listen to the results. I am 100% sure I will do a Darbari Mark II with the Jordan 4" as a midrange and a good tweeter (maybe a ribbon, maybe a high-end Scanspeak) as a tweeter. But till then, I will keep to the current set for the Mark I.

I use the Rs225 in my mains and I can tell you that it's a good choice. but it is a 8" not 10" (RS270). either way, if you are crossing over at 150hz, then you could safely use a proper sub; if you have the amplification; the RS210 or RS265HO are good bets for a small sealed box with lots of extension.
I guess a sub would have done, but I am inherently a bit suspicious about these sub-oriented drivers. The manufacturers seem to make these drivers for huge Xmax and SPL, at the cost of accuracy. That's why I deliberately chose a lighter, more general-purpose driver for the woofer. My apprehensions may be misplaced of course. In fact, Parts Express sells two flavours of 12" Dayton RS subwoofer drivers, I believe: the "HF" flavour for high fidelity, and the "HO" flavour for high output. Maybe I could have just taken the HF 12" drivers, but they were much more expensive than the 10" drivers I've chosen, so I let them go. :)

lastly, would love to hear you the minidsp 4x10 sounds like. I have the 2x4 (two of them) in my car for a 4way active setup.
Yes, me too. Let's see how they turn out.
 
LR4 @100hz? you might find a lot of plusses of RS woofer at the bottum are not being utilised. RS woofers are great fro low distortion below 1khz. Also, let me see if I can pull up the data, but afaik, some RS subs have lower distortion below 100hz than the RS woofers, at everything above nominal volume. (broader BL curves)
Well, I am using RS drivers (two of them) for the entire range from 1.5K to the bottom octave. Am a bit confused about what you are referring to here.

RS150 at 1.5khz? break-ups will not be a problem. but does that not compromise your design philosophy? also you might want to consider LR4 instead of LR8 if you are putting a passive notch .. might help your power response.
I am in fact inclined to try a Cauer-elliptic instead of a plain LR4 or LR8. Depending on how the measured SPL/impedance curves appear after I put the CE filter, I will see whether a notch is needed. And if it is, I will almost certainly put an active notch at line level, not a passive one. Unless you can educate me about some benefit of a passive notch.

Will follow this thread with interest.
Yes, it should be an interesting ride. :D

In my heart of hearts, I am nervous, but I am betting on the ability to measure and the flexibility of the digital crossover. Making changes will be easy, therefore I am hoping that I can learn and correct mistakes fast enough to arrive at something good in spite of my inexperience. :)
 
Well, I am using RS drivers (two of them) for the entire range from 1.5K to the bottom octave. Am a bit confused about what you are referring to here.


I am in fact inclined to try a Cauer-elliptic instead of a plain LR4 or LR8. Depending on how the measured SPL/impedance curves appear after I put the CE filter, I will see whether a notch is needed. And if it is, I will almost certainly put an active notch at line level, not a passive one. Unless you can educate me about some benefit of a passive notch.


Yes, it should be an interesting ride. :D

In my heart of hearts, I am nervous, but I am betting on the ability to measure and the flexibility of the digital crossover. Making changes will be easy, therefore I am hoping that I can learn and correct mistakes fast enough to arrive at something good in spite of my inexperience. :)

Dear Tcpip,

I was referring to the RS225. like I said, I use it too, in an active 3 way driven by a DCX2496 crossover. (Not as versatile as the minidsp, but i had purchased it 5 years ago, when minidsp wasn't on the scene and it has been lying in the box ever since).

I found the lower midrange quality of the RS225, rather undistinguishable versus an RS150. (i have got and used the 150, the 180 and the 225), at least upto about 800hz. but I guess if the drivers are done; then they are done :). RS225 +RS150 is a combination thats been used a lot in passive towers on htguide.

if you are purchasing the 4x10; two tips.

1. do get the umik, if you don't have a mic already. it's worth the money.
2. if you are importing a. please ask them not send by HK post (their default choice, which goes to speedpost in India and all the consequent nightmares). also consider asking them if they can give it in kit form; which may allow you to classify it for a lower duty structure because the 4x10, is basically the 2x8 + box +psu + additional I/O.

best wishes!

Kapvin


ps - impedance curves won't change with an active filter. i do not know how to model a CE in Minidsp, i guess it would be done with biquads, do share once you do it.

pss - i guess you would have checked out charlie laub's active crossover designer. it is tailor made for minidsp and gives the output in terms of biquad coefficients which you can feed into minidsp directly.
 
I was referring to the RS225. like I said, I use it too, in an active 3 way driven by a DCX2496 crossover. (Not as versatile as the minidsp, but i had purchased it 5 years ago, when minidsp wasn't on the scene and it has been lying in the box ever since).
Yes, I too had conceived of the Darbari when I did not know about digital active xo, and I was all prepared to do my modelling in SoundEasy and then build an active analog xo the classical way. :)

I found the lower midrange quality of the RS225, rather undistinguishable versus an RS150. (i have got and used the 150, the 180 and the 225), at least upto about 800hz. but I guess if the drivers are done; then they are done :). RS225 +RS150 is a combination thats been used a lot in passive towers on htguide.
Understood. I wanted to use the 6" so that I could cross over a bit higher, and get a tweeter to come into the picture relatively easily. And I wanted to keep the really high-excursion bottom two octaves out of the scope of work of the 6", so that the 6" could do the detailed lower mids free of any Xmax concerns.

1. do get the umik, if you don't have a mic already. it's worth the money.
I have a Panasonic capsule and I use it with Speaker Workshop for my SPL/phase measurements. I've been using this combo for a few years now. I hope that this will be enough?

2. if you are importing a. please ask them not send by HK post (their default choice, which goes to speedpost in India and all the consequent nightmares). also consider asking them if they can give it in kit form; which may allow you to classify it for a lower duty structure because the 4x10, is basically the 2x8 + box +psu + additional I/O.
I am asking them to ship to my brother in Singapore, and I'll get it from there to my place in personal baggage when I visit Singapore next. (My colleagues or I need to make occasional trips there on business.)

ps - impedance curves won't change with an active filter. i do not know how to model a CE in Minidsp, i guess it would be done with biquads, do share once you do it.
Impedance curves won't change with an active filter, true. But impedance curves will help me get the precise F3 and Q of the speaker driver in the box, if I do T/S parameter measurements using Speaker Workshop. These values will be needed to let me plug them into a formula to design the LT.

And you are right, both CE and LT are done using biquads. And the spreadsheet you mention is the one to use.

pss - i guess you would have checked out charlie laub's active crossover designer. it is tailor made for minidsp and gives the output in terms of biquad coefficients which you can feed into minidsp directly.
Yes, what I want to do can't be done using the GUI filter design options of MiniDSP. Those are too primitive. For instance, they don't allow me to string together two 2nd order filters one after the other, e.g. an LR2 at 2.5KHz followed by another LR2 at 3.1KHz. I have to choose an LR4 at one frequency. These are too primitive for real-world crossover designs once one starts with SPL and phase measurement. You need to have the ability to string together arbitrary building blocks, so that you can shape the curve precisely.

In fact, if biquad support and the spreadsheet were not available in MiniDSP, I would have been forced to go the analog active xo route, doing PCB design and soldering 2% tolerance capacitors and all that. :(
 
Enclosure building has started. Some pictures of the woofer enclosure bracing are available here. Work progressing slower than expected due to some mechanical problems with the router. I may have to buy a new one. :(
 
woofer-box-brace-top.jpg
woofer-box-brace-side-2.jpg
woofer-box-brace-pure-side.jpg
 
Enclosure building has started. Some pictures of the woofer enclosure bracing are available here. Work progressing slower than expected due to some mechanical problems with the router. I may have to buy a new one. :(


Very nice. Looks like it is really overbuilt; which can only be a very good thing.

a quick somewhat OT question - what measurement window to you use for your quasi anechoic measurements?

thanks
 
Very nice. Looks like it is really overbuilt; which can only be a very good thing.
Not overbuilt at all. If the enclosure sounds like solid wood, not hollow box, on your knuckle test, then you know it is built well. With all my bracing, I am still not 100% certain I will reach there. :sad:

a quick somewhat OT question - what measurement window to you use for your quasi anechoic measurements?
In my small living room, I get just about 5ms window.

Some more pictures and notes of construction posted here now.
 
The amplifier side of things

The amp side of the story is inching along.

I have finally decided to build the six channels of amplification using the LME49811-based compact modules from Panson Audio of Hong Kong. They're quite affordable. They're also quite low power (perhaps 50W into 8 Ohms), and have just one pair of OPS devices per channel. But they should be low distortion.

Got my first two samples of speaker protection circuit, straight from China (jims_audio on eBay). Each protection circuit works with two channels. So I will need three of these for six channels. Each module uses a dual-pole Omron relay. Hope they'll behave well.
 
Last edited:
Back
Top