Peerless HDS 830883/SEAS 27TBFCG (aka: I hope Zaph does not kill me for this!)

[TheD]

Hello,
This is my first post (so be kind).

I have been looking to build Zaph's SEAS "all metal" system for some time now, I was all set to go till the SEAS importer (I am from Aus BTW) sent me an email saying that they do not have stock of the L18 and they are not going to get anymore in

So I am thinking about somewhat coming up with my own speaker, I have set my mind on using the Peerless HDS 830883 mid/woofer (because Zaph gives it a good rap) and I have been looking around for a tweeter to use but I have come back to the 27TBFCG, it also helps that Zaph tells you what needs to be done to it via the crossover (which makes it a bit like the "all metal").

I am looking at using a notch on the woofer of about -4db from 1500hz to 6000hz to get it level with the rest of the woofers output (before x-over) and a crossover of about 2000hz LR4, would this be ok?

This is my first speaker (that I am building) and I hope I have got the jist of what to do.


BTW: I can not get the SEAS reed cone in Aus so the SR71 is not an option and importing would cost a hell of a lot.

[tktran]

I understand your sentiment with cost of importing drivers, particularly since the AUS dollar has lost about 1/3 of it's value against the Greenback in just 3 months.

But rather than start a design from scratch and completely waste your money I think you build a design with locally available drivers, or just hang onto your money...

[Jonasz]

A completely different design but this may be an option using standard Seas drivers. To me a good designed 3-way gives a more effortless sound than a 2-way. At least an option for you to consider.

http://www.troelsgravesen.dk/3WClassic.htm

[Idefixes]

Contents of post removed by moderator. It's a good idea to read the forum rules before posting....

[TheD]

A completely different design but this may be an option using standard Seas drivers. To me a good designed 3-way gives a more effortless sound than a 2-way. At least an option for you to consider.

http://www.troelsgravesen.dk/3WClassic.htm

They are too costly and too big.

[tktran]

Hi. It seems to me that you've selected drivers based on John's tests, costs and local availability, and deciding to implement your own crossover?
It also sounds this is your first go with xo design or have had only a little experience with this.
Everyone's recommended a well documented and critiqued design but cannot take into account your own constraints.

Your Xo idea is a bad one. Please read the FAQ at both John Krutke and Troels Gravesen's websites.

[HareBrained]

Don't get discouraged about developing your own xover. It's not impossible but there is a learning curve. There are several free Excel-based tools that can help you design but there is some knowledge that needs to be gained to understand the nuances. Search this forum and the PE forum for PCD (Passive Crossover Designer), Response Modeler (RM) and Unibox. Most of the posts containing references to these tool are questions about their use (including a bunch from me). You'll also want to request the FRD & ZMA files (you'll find out why) or search for the SPL Trace tools. This is going to mean a little reading and then a lot of investigation. It's been 3 months since I started my designing and I still haven't built a speaker. (I'm getting close though.)

On the other hand, there are many find designs out there that could probably meet your needs. Click thru on the signatures on the forums and see what people are building. Be honest about what you're constraints and requirements are. Most designs that might set you back $500 as a DIY would be $2000 in a store. Personally, $500 DIY speakers would make me very happy as my listening environment isn't "clean" enough to hear the improvements of more expensive designs. After all, there is a point at which you spending a lot more money for very small finite improvements in quality.

[TheD]

Don't get discouraged about developing your own xover. It's not impossible but there is a learning curve. There are several free Excel-based tools that can help you design but there is some knowledge that needs to be gained to understand the nuances. Search this forum and the PE forum for PCD (Passive Crossover Designer), Response Modeler (RM) and Unibox. Most of the posts containing references to these tool are questions about their use (including a bunch from me). You'll also want to request the FRD & ZMA files (you'll find out why) or search for the SPL Trace tools. This is going to mean a little reading and then a lot of investigation. It's been 3 months since I started my designing and I still haven't built a speaker. (I'm getting close though.)

On the other hand, there are many find designs out there that could probably meet your needs. Click thru on the signatures on the forums and see what people are building. Be honest about what you're constraints and requirements are. Most designs that might set you back $500 as a DIY would be $2000 in a store. Personally, $500 DIY speakers would make me very happy as my listening environment isn't "clean" enough to hear the improvements of more expensive designs. After all, there is a point at which you spending a lot more money for very small finite improvements in quality.

At last..... someone who is not trying to shill other designs to me!

I have been using WinISD and http://www.diyaudioandvideo.com/Calculator/
But I will look into them.

Everyone's recommended a well documented and critiqued design but cannot take into account your own constraints.

Your Xo idea is a bad one. Please read the FAQ at both John Krutke and Troels Gravesen's websites.

I want to build my own!

You do know that I am not just dropping the woofers into the same sized box with the same crossover?????

I was just asking if the 1500hz to 6000hz notch was overkill.

[ThomasW]

I was just asking if the 1500hz to 6000hz notch was overkill.

Under normal circumstances one wouldn't consider a 2 octave wide filter to be a 'notch' filter

[TheD]

Under normal circumstances one wouldn't consider a 2 octave wide filter to be a 'notch' filter

I did think it might be going too far, the main peak would be something like -30 (odd) db down at it anyway
:lol:

[HareBrained]

I was just asking if the 1500hz to 6000hz notch was overkill.

Hmmm. If the LP is 4LR @ 2kHz, I don't see that you'll need a notch. It should be squashed by the filter. You may need to adjust the first coil to tilt the FR down but the 4LR should take care of it. It would be overkill.

On a related topic, I was worried about the hi-freq content of a small metal dome woofer, and tried to add a notch but it didn't work because the LP filter changed the shape and position. That's the cool thing about the modeling tools (PCD), they show you what's happening, interactively.

[Dave Bullet]

I second HareBrained suggestions.

Follow the instructions at Roman Bednarek's site: www.rjbaudio.com

The above explain how you can use measurements taken by someone else to then simulate your own crossover using PCD or whatever crossover design tool you like.

Another option is to find another design with the 830883 that has a crossover slope and target similar to your intended target. Likewise with the 27TBFC/G, then merge the two circuits together (assuming parallel crossover topology).

PS: The 27tbfc/g can handle a crossover point lower. My modeling (assuming your typical 5mm xmax midwoofer) shows 1700Hz is fine with the 27tbfc/g. Zaph went with 1450Hz in the BAMTM.

PPS: Zaph works quickly... he must be ok with this, otherwise you'd be dead even before posting this :B

Dave.

[TheD]

Thanks for the ideas.

[TheD]

Hmm...
I have been trying to use both PCD and speaker workshop to model a crossover but I have run in to a few problems.

With PCD I got it to work but when I tried a LR2 crossover the summed output was wrong (somehow it made all the dips and peaks of the tweeter twice as big)

And with Speaker Workshop I got it to model Zaph's "All metal" 27TBFCG and crossover and it was mostly fine but for some reason it had a lot of dips all the way to zero output but they are only 1-2 hz across. But now I have tried it again and it does not seem to be working anymore.

[Dave Bullet]

Can you post some screenshots of the problematic amplitude responses?

SpeakerWorkshop is buggy.. as long as you stay within the process of setting things up, it won't crash.

Can you post the sequence of steps you went through to generate the FRD /IMP files. Better still - can you put the .swd file up somewhere and I can take a look.

Dave.

[TheD]

I did not save it...... but its not just Speaker Workshop thats not working as it should, now PCD will not load ZMA and FRD files!

[Dave Bullet]

Reminds me of my favourite PC haiku:

"Chaos reigns within,
Resent, repent and reboot,
Order shall return"

[Dave Bullet]

How about uploading the files you are trying to work with? I can then import them in SW and have a look.

[TheD]

I think I might just give up.... and try and find the money to build the only other Zaph speaker that I can get the parts for, the ZRT.

I just wish I could have done it

It is just too hard to make anything when the programs just do not work.

[alias2]

link removed by moderator

Has useful information on the 830833 and is not as expensive as the ZRT.

There is enough information on that site to reverse engineer a different tweeter.

[Idefixes]

link removed by moderator

Has useful information on the 830833 and is not as expensive as the ZRT.

There is enough information on that site to reverse engineer a different tweeter.

Yes, I propose the design above. TheD why don't go on jay's design with HDS1" It will certanly cheaper than the ZRT???

Marc

[---k---]

Please refer to the forum rules with regards to Jay Wo's designs.

[Dave Bullet]

Those rules don't make sense. Probably need to be reviewed.

[alias2]

Please refer to the forum rules with regards to Jay Wo's designs.

what is the point of "banning" useful information ? which you can take or leave ...

[ThomasW]

Those rules don't make sense. Probably need to be reviewed.

They make sense if one knows the history of Mr Kim's very dysfunctional behavior when he was a member of this forum

what is the point of "banning" useful information ? which you can take or leave ...

sreten,

See my response to Dave. In addition, it's not your place to debate the rules. If you don't like them feel free to return to DIYaudio and post under your usual username

[JonW]

I think I might just give up.... and try and find the money to build the only other Zaph speaker that I can get the parts for, the ZRT.

I just wish I could have done it

It is just too hard to make anything when the programs just do not work.

TheD,

I just finished designing my first speaker (using the ZRT drivers, actually) and I used the LspCAD software. I didn't feel like fighting with all the other software. It was totally worth the US$200. It allows you to take measurements in your cabinets (with a calibrated microphone) and model the crossover using the measurements. I would imagine you can also import zma/frd files as well but I'm not sure. Very glad I bought that.

[ttan98]

TheD

If you know someone in the States to buy on yr behalf and then send to you via International Post this way you will save money on quality parts. Also you have more choices on high quality parts.

I use this way quite successfully.

cheers, happy building and listening.

[---k---]

They make sense if one knows the history of Mr Kim's very dysfunctional behavior when he was a member of this forum

This 2000+ poster agrees.

Now back to regular programming...

[TheD]

TheD

If you know someone in the States to buy on yr behalf and then send to you via International Post this way you will save money on quality parts. Also you have more choices on high quality parts.

I use this way quite successfully.

cheers, happy building and listening.

I would not save money at all!
The AUD is at 65 USD cents and I would also have to pay shipping (which is about $200 USD) and (if we are talking about the ZRT) I would also have to pay customs tax and fees of 15% of the price of the goods and shipping!

TheD why don't go on jay's design with HDS1" It will certanly cheaper than the ZRT???

Marc

The HDS tweeter costs about 3x (in Aus) as much as the 27TBFCG or the 27TDFC yet does not perform any better.

[Dave Bullet]

I think I might just give up.... and try and find the money to build the only other Zaph speaker that I can get the parts for, the ZRT.

I just wish I could have done it

It is just too hard to make anything when the programs just do not work.

Dont' be discouraged. If this is your first pair you are building - even building an existing design, you will still learn a lot. Plus you have the knowledge you will get something you will very likely enjoy and will be money well spent.

If the bug gets you, then you can on-sell those speakers and invest the money in a measuring setup and take things from there.

Your biggest hurdle is your willingness to not give up and persevere. Simulation will work for you - but you must give it time and practice. Simulating existing published designs so your curves match the designers is the best way to prove you are on the right path (where you cannot source the parts to build, measure and confirm).

I've been following this hobby for about 3 years, and I've found there is more I don't know now than when I started! Have I designed and built speakers I have enjoyed? - yes, but it will cost you to setup even the most basic of setups to be able to measure your speakers, then designing a passive crossover is another set of skills.

My view with shipping - is you get more bang for buck buying expensive drivers. I mean, a cheap driver will cost the same to ship as an expensive one, so you may as well justify the cost of shipping by buying more expensive drivers. In this case - it is worth building another design to maximise your enjoyment until you are skilled enough to give it a go.

Even then you'll have a reference system, and some nice parts to redesign in the future should you want. I'm sure the scanspeak drivers in the ZRT will still be highly regarded in a few years.

Cheers,
David.

[tktran]

Hi,

The low exchange AU/US rate and high costs of international shipping is difficult to swallow. Understandably you choose not to take this path.
However, this is here to stay, and unlikely to go away soon.

You also do not want to pay for the 810921, which you claim is 3 times the costs of of the 27TDFC in Australia. I'm not sure who your retailer is, but IIRC the price difference is closer to 2.5, which is not unlike the situation in the US, where the 810921 costs 2 times that of the 27TDFC.

This may seem like a lot, but the absolute difference is probably about $AU200. This is probably the absolute price difference for importing parts for the ZR71 kit. Having built multiple speakers, in the grand scheme of things (sourcing parts, cutting, building, finishing, building crossovers, installing drivers), $200 is a small expense...

You can design your own first crossover, but of course that requires some knowledge of electrical circuit theory, components and ability to make use of the available tools. If you are having problems with PCD perhaps you could give a more detailed description, or screenshots of your problems. This would make it easier for others to give feedback.

I think you are shutting out all the options, and good ideas that others have suggested. If you really want to build a DIY speaker, it's good to keep an open mind.

regards,
Thanh.

[TheD]

But the fact remains that the HDS tweeter does not perform much better than the TBFCG and costs about 3x as much and yes it is a lot more money if I am only going to have a tweeter that works about the same.

[alias2]

Your original idea to use the Peerless HDS 830883 mid/woofer and Seas 27TBFCG c/o
around 2Khz is perfectly do-able as long as you stay away from crossover calculators.
You can find a Peerless HDS 830883 / Seas 27TDFC crossover at www.diyaudio.com .

Search for 830883 tweeter, the two tweeters are very near interchangeable.

[tktran]

Well then buy the TBFCG tweeter then.

Round and round the merry go round we go...

[Dave Bullet]

TheD,

My good fellow, follow alias2 advice. You'll have a crossover design that will work with the drivers, so if you decide you want to really re-design a crossover for those drivers, and it sounds bad, you'll have a good crossover design to fall back on.

Experienced designers have stated the HDS requires a little more crossover work to shape the top octave response (if that is significant to you) to bring out its best. This is something not for a newbie, will require more (albeit lower cost) crossover parts, and the ability to measure to ensure you hit the shaping filter you need spot on (given effects the baffle may have).

Therefore the 27TBFC/G is an excellent performer, that is cheaper and easier to implement to maximise its return. I'm not saying it is better than the HDS.

Dave.

[TheD]

Hmm, I think know what I am going to do.

I will use the peerless woofer and one of the SEAS tweeters but instead of making a passive crossover I will get a Behringer DCX2496 and use it, that way I can try and try some more till I get the speakers how I want (without building and rebuilding crossovers all the time).

[alias2]

Having perused the DCX2496 user manual it does not seem to be a short cut.
It does not seem flexible enough to do the sort of things that you can simulate.
(you should be able to find all the simulation data you need quite easily)

http://www.rjbaudio.com/Audiofiles/FRDtools.html

the DCX2496 can be used if you know exactly what you are doing but like
crossover calculators it can give the wrong impression you are doing it right.

If you get to settings you like you will still have to go back to the FRD tools
to work out your real components, you can plug DCX2496 settings into one
of the crossover calculators as long as you fully zobel the drivers, but this
is not the most efficient way of building crossovers, see FRD tools.

[TheD]

How is it not?
Otherwise I would have to make at least a few crossovers which take a while to build (the DCX does not take long at all) and source the parts and try and design the crossover in unusable (because it is buggy to the point of crashing) software.

[alias2]

By using the information from other designs that have used those tools you
can get very near the sort of thing you want, and possibly bypass some of
the tools your having problems with.

If you want to do the whole thing by ear and the DCX fair enough, but
it does not help you with the fine tuning of passive components in a c/o.
You are stuck with its standard functions, Bessel, L/R, Butterworth etc

L/R for example should be an acoustic target, not an electrical function.

[TheD]

It is not going to just be "by ear" I also have a mic that I can use, it is not the best mic but it works (it is an audyssey/onkyo setup mic).

I know that a crossover slope does not have to be electrical.

What functions can it not do?
It can do x-overs of 1st 2nd 3rd 4th and 8th order, you can use the EQ to make strange ones, you can use the EQ to do what a notch filter does ect.

[JonW]

For whatever it’s (probably not) worth…

Regarding tweeter choice, it sounds like TBFCG might be your best bet if you want to go cheap. If you’re willing to spend some more and get a better tweeter, it sounds as though you should skip the HDS and maybe consider the 6600. I know it’s excellent and it seems to be priced comparably to the 6600 for you.

Regarding crossovers…

I got a DCX2496 and fooled around with it a little. It works and you get a speaker that plays. But without measurements and all that, you’re just shooting in the dark. It didn’t sound all too great. I then took measurements and designed a passive crossover. For my first speaker design I felt like I learned more about it all by going the passive route. Matching driver levels, baffle step, changing impedance, phase (a biggie for me), zobles, notch filters, etc. You can take all that passive info and then put it back into the DCX2496, I think.

On a similar note… modding an existing crossover or using SPLtrace, etc. with manufacturer’s specs, etc. will also get you a working speaker. Kind of like the DCX2496 tuned by ear. They will both work. But not nearly as well as if you do things the more rigorous way, with measurements in your cabinets and such.

So I’d recommend thinking about what your goals really are here. If it’s just to get a functional speaker and learn a little, you can take one of the routes described above. If it’s to get an excellent sounding speaker with minimal work, build a preexisting design. But if it’s to learn everything you can about speaker design, have a creative, new design, and also get an excellent speaker out of this all, you’ll have to go the slower, more rigorous route with true measurements and such.

Just saw you have a mic. Good. Make sure you get it calibrated, otherwise it could have a significant (up to 10 db) rise or dip at higher frequencies.

[minkuni]

LspCAD has a filter-box that replicates the abilities of the DCX2496, so if you have good measurements to work with, you can find the required adjustments on the DCX through simulation, rather than having to resort to trial and error. The final fine tuning will of course be reliant on some trial and error :W

Download the demo of LspCAD (here) and give it a try.

[alias2]

measurements are good, as long as they are accurate .....
simulation is good as a starting point .....
simulation as an end point requires good information .....
good measurements is the best way and the most rigourous .....

Regarding putting info into the DCX : Take Zaphs L18 example above :
What is the 2nd order electrical filter on the tweeter ? not DCX settable.
(You need to be able to set the Q of a second order high pass filter)
How do you model the baffle step inductor ? it is not 1st order low pass,
it might be if you zobelled the driver but the real driver is not zobelled.
The notch filter interacts with the above function, What Q is this notch ?

What software is it that is unstable ?

[Dennis H]

The DCX can do any Q crossover filter by adjusting the shoulder shape with a parametric band filter. It also has shelving filters for BSC, etc. As far as transfer functions, it can do anything a passive filter can do and much more. There are other 'problems' with the DCX but that isn't one of them.

That said, if you're going to use LspCAD or SoundEasy to model the DCX, you can use them to skip the DCX step and emulate a passive crossover with the sound card before you buy any XO components. But the first step is to be able to run audio software on the computer before you can do any of that. Sounds like a computer (or user) problem if everything that works for other people crashes.

[alias2]

a)The DCX can do any Q crossover filter by adjusting the shoulder shape
with a parametric band filter. b) It also has shelving filters for BSC, etc.

Having perused the DCX2496 user manual it does not seem to be a short cut.

I agree with a) but it is getting very complicated, poles, zeros and all that stuff.
For b) I cannot see that described in the user manual (in section 4.3.3 ?).

My point was there is not an easy way of going from DCX settings to a real c/o.