The Darbari: new speaker project

Driver mounting screws

Got some really excellent driver mounting screws from TheAudioCrafts.

I prefer wood screws in place of T-nuts and bolts. T-nuts and bolts are great for situations where you expect to remove and replace drivers many times. But I feel that they are not too suitable for long-term mounting because they are, in essence, bolts being used in a vibration-prone setting, and bolts used in such settings always need lock nuts, or else they will come loose. Basic mechanical engg. And you can't fit lock nuts on a T-nut+bolt setup unless you can reach the back of the front baffle after the drivers are mounted. Well, this is not possible in any of the speakers I have built.

Therefore, I prefer wood screws. And these ones I got are all wood screws.

I got copper finished 4mm dia screws (really luvvly) for the tweeter and midbass, and 5mm ones for the woofer. These screws have heads designed for screwdrivers with hex tips. This means that the screwdriver head will go into a hole at the head of the screw. This means that the screwdriver will not slip out accidentally when tightening, which means it removes one of my biggest causes of heart attack when fixing drivers. (Imagine puncturing a hole in the cone of a ten-thousand-rupee driver. :mad: )

I wish these AudioCraft chaps had also kept a stock of 2-inch-long and 3-inch-long screws. Would be useful for large and heavy woofers. I would have felt better if I had got them even for my Darbari's 10-inch woofers, though I guess a one-inch screw is within the borderline of acceptable length.

If all goes well, I intend to use these copper-finished screws for all my speaker projects henceforth. They look super-lovely. Thanks to a FM for pointing out this vendor to me -- my dealing with them was super-smooth, and they have interesting items.
 
Impedance of Dayton RS270

I measured the impedance of one of the RS270 10" woofer drivers which I will be using. Measured thrice keeping the driver on the floor, facing upwards, and then measured thrice more holding the driver "vertically", facing in the horizontal direction. Got slightly different curves in the region below the Fs.

And there are two clear, consistent kinks in the impedance curve, at about 100Hz and 1.2KHz. Are these problems?

I have yet to measure the second RS270.
 

Attachments

  • Dayton-RS270-free-air-horizontal.jpg
    Dayton-RS270-free-air-horizontal.jpg
    14.8 KB · Views: 200
  • Dayton-RS270-free-air-vertical.jpg
    Dayton-RS270-free-air-vertical.jpg
    16 KB · Views: 200
Hi tcpip,
Very nice project; actually, projects! I've gone through your site with your photography and cooking interests. Very nice of you to share your speaker build details over there. Your site is a MUST read for a DIY-er, especially for the trouble shooting sessions; I believe that your perseverance and the passion has taken you places. Thanks!
BK
P.S: By the way, I tried to post my comments over there itself, but somehow, I'm being getting rejected. Hence, this message here! :)
 
Yes indeed, I can vouch for his "foodie" side. A recent trip to Delhi where we both partook of chicken "Kobiraju" Cutlet had him going into its history in some depth. You could say he knows his subject well!
 
By the way, I tried to post my comments over there itself, but somehow, I'm being getting rejected. Hence, this message here! :)
Thanks for the nice comments. :) You are right about the problem in posting comments on my website -- it's apparently some bug in Drupal, which I am too lazy to patch. Am very happy to interact with you all on hifivision, therefore nothing's lost. :)
 
Yes indeed, I can vouch for his "foodie" side. A recent trip to Delhi where we both partook of chicken "Kobiraju" Cutlet had him going into its history in some depth. You could say he knows his subject well!
Aaah... will I ever be allowed to live down the "kobiraji cutlet" discussion? :D
And it is "kobiraji", not "kobiraju". :indifferent14:

But GeorgeDa --- it was a glorious evening, wasn't it? Being hosted by India's foremost speaker gooroo on a chilly November evening in Delhi under the stars has its own charms. :)
 
Re: Impedance of Dayton RS270

... And there are two clear, consistent kinks in the impedance curve, at about 100Hz and 1.2KHz. Are these problems?

The notch around 1.2k may be the cone breakup. Could the one at 100Hz caused by an unbroken spider? :confused:
 
Re: Impedance of Dayton RS270

And there are two clear, consistent kinks in the impedance curve, at about 100Hz and 1.2KHz. Are these problems?
.


Looks like some kind of cone/surround resonance kicking in around the kinks. These kind of usually mitigate when driver is mounted firmly. You have to check on that.
 
Re: Impedance of Dayton RS270

The notch around 1.2k may be the cone breakup. Could the one at 100Hz caused by an unbroken spider? :confused:
Yes, the 1.2K kink seems "normal" -- I too felt we can ignore it. I expect to cut off the driver sharply (4th order or more) at a max of 300Hz anyway, so the cone breakup will (I hope) be inaudible.

The 100Hz is interesting. I will take measurements of the other driver and see if it is found there too.
 
Re: Impedance of Dayton RS270

Looks like some kind of cone/surround resonance kicking in around the kinks. These kind of usually mitigate when driver is mounted firmly. You have to check on that.
Cool. Will see what the in-box impedance graph is like -- will post here when the drivers are mounted (a month away, I think).
 
Re: MiniDSP next steps

re: your point about volume control; the minidsp has a digital volume control. If you are anal about "losing bits" then you'd not use the volume control and use something downstream. I use the MSB 8 channel control (which is what SL also uses)- but I use it more from a gain structure mitigation perspective.
Hi Kapvin,
I have been wondering about the "loss of bits" myself, but I just posted a question on MiniDSP's own forum and got some answers, and it appears that the DSP module feeds the DAC with 24-bit data.

If this is the case, there are plenty of "spare bits", and hence even a volume control done in the digital domain would have 16+ bits going to the DAC at low volumes. Therefore, there should be no audible drop in sound quality with a digital volume control. This "loss of bits" becomes an issue if one is starting with, say, 16 bits of data, where a 60dB cut in volume would leave you with quantisation noise in the output.
 
Re: MiniDSP next steps

Hi Kapvin,
I have been wondering about the "loss of bits" myself, but I just posted a question on MiniDSP's own forum and got some answers, and it appears that the DSP module feeds the DAC with 24-bit data.

If this is the case, there are plenty of "spare bits", and hence even a volume control done in the digital domain would have 16+ bits going to the DAC at low volumes. Therefore, there should be no audible drop in sound quality with a digital volume control. This "loss of bits" becomes an issue if one is starting with, say, 16 bits of data, where a 60dB cut in volume would leave you with quantisation noise in the output.

With the noise floor of the 4x10, you should have about 18-20 bits of resolution which is much more than enough. The problem with Loss of bits come when there is a mismatch between the output level of the crossover and the input level of the amp.

Ideally if you could ensure that 0dbfs digital = DAC output voltage = input sensitivity of the power amps you are okay. But if there is a mismatch say 10 db and then when you listen at a volume of say -30 db then that's 7 bits you have lost there. Some people may perceive a difference. Get your gain structure sorted and you should not feel a difference... ( with my DCX the issue was exaggerated. So maybe I am being extra cautious)
 
Re: MiniDSP next steps

Ideally if you could ensure that 0dbfs digital = DAC output voltage = input sensitivity of the power amps you are okay. But if there is a mismatch say 10 db and then when you listen at a volume of say -30 db then that's 7 bits you have lost there. Some people may perceive a difference. Get your gain structure sorted and you should not feel a difference... ( with my DCX the issue was exaggerated. So maybe I am being extra cautious)
Help me out a bit here, please. Am trying to see if my understanding is correct.

Typically, we generate analog output from our CD player, and that's a 16-bit DAC process. After that, we do whatever we need in a typical analog preamp, and feed it to the power amp and speakers. Here, the volume control is after the DAC therefore there's no loss-of-bits issue.

If I am implementing a volume control before the DAC process, I get loss of bits, because I usually listen at a volume level far lower than 0dbFS. However, if the stage before the DAC generates a 24-bit digital output, then even after cutting volume by 8 bits, I am still left with 16 bits. And 8 bits of attenuation is about the max that most volume controls need to do anyway, therefore it's unlikely that I'll be left with less than about 16 bits.

In that case, is there any danger at any point? And does the noise floor of the 4x10HD matter? Noise comes only after the DAC, and we're discussing cutting bits before the DAC step, right? (I am assuming that all equipment we play with will have at least 15-16 bits of SNR, to be able to handle CD-quality signals.)

Are you worrying that if the gain of my power amps is very high, I may have to cut more than the "normal" 8 bits in the attenuation stage? If that's your worry, then I can see why you're worried. I am, too. :cool:
 
Last edited:
It is an inspiration to see from close quarters how your design philosophy has evolved over the past decade in your typical bohemian way.

You do it like the best Litti Mangsho and Kobiraji ;)
Sir, your encouragement is like food to a hungry Bengali stomach (the kind of stomach which never seems to get enough). But I am now desperately trying to finish the project, and my carpenter has run away, and it's become a classic case of "the last 5% taking 50% of the time." :sad:

Once I get any half-way decent carpenter to come and finish the remaining bits, all of you are invited for music and kosha mangsho at my place. Even single malt thrown in, if any is left when you come. :D
 
Haanji, perhaps with a palang tor paan thrown in after the Kobiraji? I'm quite looking forward to my trip to Jamshedpur/Calcutta later this year to meet up friends, visit Tolly and RCGC and sample some specialities!
 
Re: MiniDSP next steps

Are you worrying that if the gain of my power amps is very high, I may have to cut more than the "normal" 8 bits in the attenuation stage? If that's your worry, then I can see why you're worried. I am, too. :cool:

In a nutshell, yes. should probably be less of an issue than we make it out to be, but hey, we live in an era (and forum) of cable risers and bespoke power cords that make all difference to the sound. and if those things are important, then this loSS of bits should be sacrilegious.

:rolleyes:
 
What is the logic behind using a driver upto the region which shows cone breakup peaking and then applying notch filter as band-aid to correct it? :sad:
I don't think I understood your comment. Do you have any other suggestion about how such drivers may best be used?
 
Follow HiFiMART on Instagram for offers, deals and FREE giveaways!
Back
Top