"ok, I got it, too late: a capacitor is after the transformer, is it?"
exactly, that way you do not draw too much current from first amp, it only drived midrange and up
no se low power tube amp is capable of driving woofers properly, so it sounds great, yet on its output is normal full range output, just like line level signal, before the cap, goes to ss amp and drives the woofer
believe me, or do not believe me, try it! sound heavenly!
exactly, that way you do not draw too much current from first amp, it only drived midrange and up
no se low power tube amp is capable of driving woofers properly, so it sounds great, yet on its output is normal full range output, just like line level signal, before the cap, goes to ss amp and drives the woofer
believe me, or do not believe me, try it! sound heavenly!
jnb said:
I think the hump in the baflfe response is lower in my article because I used the floor to "extend" the dimensions of the baffle. If you look at my Figures 6 and 11 you will see how I used this hump to interact with the electrical crossover to optimize the acoustic crossover. It is both significant, intentional, and an integral part of the OB design process. If you want to design OB speaker systems you need to move past the EDGE program, it is a great piece of software but it only tells a part of the story.
I'm surprised that you have been able to get away with such a simple crossover.
I see that you've brought together the various responses to produce a flat sum. None of the component responses was too extreme, and this seems good. But what if they had been?
What I'm trying to ask is, would it be acceptable to use, say, a 6dB baffle peak to compensate for something else? I'm wondering whether resonance, or off axis response or something else will become a problem.
I see that you've brought together the various responses to produce a flat sum. None of the component responses was too extreme, and this seems good. But what if they had been?
What I'm trying to ask is, would it be acceptable to use, say, a 6dB baffle peak to compensate for something else? I'm wondering whether resonance, or off axis response or something else will become a problem.
jnb said:
Why are you surprised?
What exactly do you mean by an extreme response? What would cause an extreme response at this low a frequency in decent quality drivers? My experience has been that full range drivers can produce nasty looking measured SPL responses above ~5 kHz. But even if an eye popping peak or dip is present, the sound of the driver playing music may still be excellent.
My only concern with the smaller Fostex drivers is about the high frequency response, I am not at all concerned with the baffle or crossover design. To be honest, I am not even too concerned about the Fostex FE-103E since I am basically an optimist and have had good luck with Fostex and the manufacturer's data sheets.
A good OB design, and any good speaker enclosure design, will use all of the properties of the driver, baffle, and crossover response to yield an optimized result. What is the "something else" you are concerned with and why do you think "something else" needs to be addressed? I don't understand.
I just got through designing a 3-pole bridged-T plus two extra poles to EQ and cross a dipole woofer. What I probably meant to say is: I'm impressed with how easy you make it look.MJK said:
That in the Fostex region, the floor bounce and baffle contributions are largely within +/-2dB.
I'll try to put it a different way.
If I were to do an alternative design, I might try to offset a falling mid-drivers low end with the open baffle peak, which I'd want to exaggerate, or to fix a dip in the midrange with some other baffle peak.
Does putting the driver in the centre of a baffle to enhance the OB peak, create a time-smearing resonance? and will putting the driver close to the edge of the baffle give off axis response problems?
Since I have no idea what a "3-pole bridged-T plus two extra poles to EQ" is, I guess my idea must look pretty simplistic. My two way is at the upper limit of my capabilities. I like simple and minimal, although my minimal is probably much more than the full range driver purist would be able to tolerate.
Yes, but remember the driver is only coming into the picture at a relatively high frequency for floor bounce, The fundamental floor bounce cancellation for this geometry would be much lower in frequency. Couple that with the rising attenuation provided by the carpet on the floor and a little directional behavior of the driver and floor bounce is reduced to a manageable level. My guess is that the ripple shown in the plot is a little overstated.
An alternate design is definitely possible, but how much better could you make the frequency response over what is shown. Granted at different listening distances and off axis it might change some but for the standard 1 m on axis position that people use to assess speaker designs it seems to work very well.
The driver is offset only a little bit, I found it improved the response. It is not offset dramatically so I guess I am not too worried about "a time-smearing resonance". Considering how the SPL response in the room will change from the nice flat response calculated, I tend to not worry too much about every minor nuance in the calculated response. How it sounds is the final judgment.
Martin,
I cant find the layout for the FE166/Alpha15 you produced awhile back, but this baffle seems to be very close to the dimensions I remember. And I believe the Alpha needed about 2-3db boost for that one.
So can the baffle be used for both drivers or was the other design somewhat different?
Also, Im wondering if a passive line level xo might be a good way to go. The cost is far less than dealing with the inductors, although a second order PLLXO seems more difficult to design. Might offer a less expensive way of biamping and being able to use a valve, T-amp etc on the Fostex and a bigger amp on the bottom. I realize that this wasnt your design goal, but it might be a nice deviation to allow some flexibility.
Have you played with any PLLXOs and if so, what are your thoughs on them vs active XOs.
amt
I cant find the layout for the FE166/Alpha15 you produced awhile back, but this baffle seems to be very close to the dimensions I remember. And I believe the Alpha needed about 2-3db boost for that one.
So can the baffle be used for both drivers or was the other design somewhat different?
Also, Im wondering if a passive line level xo might be a good way to go. The cost is far less than dealing with the inductors, although a second order PLLXO seems more difficult to design. Might offer a less expensive way of biamping and being able to use a valve, T-amp etc on the Fostex and a bigger amp on the bottom. I realize that this wasnt your design goal, but it might be a nice deviation to allow some flexibility.
Have you played with any PLLXOs and if so, what are your thoughs on them vs active XOs.
amt
amt said:
Yes, the Alpha needed 2 or 3 dB of boost. The baffle was a little bit larger if I remember correctly.
I doubt it, that was a different design and it was bi-amped. We are talking apples and oranges comparing that older design with the design being discussed in this thread.
I have not worked on PLLXO at all. I have the capability to tri-amp and would go down that path if I decide to do another active design. The goal of this design was low cost and passive. I think a new thread would be warranted if you want to continue with a bi-amp OB design discussion.
One important aspect of Martins paper is the integration of baffle response and crossover design. That allows for the response rise of the baffle with out resorting to a notch filter, and lowers values / cost of the components.
Bi-amping with LLPXO is an option, but why complicate the signal path if it's not required?
Geoff.
Bi-amping with LLPXO is an option, but why complicate the signal path if it's not required?
Geoff.
Re: Re: Re: FX120
Yes. Measured efficiency across some 20 pairs indicate a typical pair is between 89 & 91 dB (the same or a tiny bit more than a typical FE103. FR should be broadly similar to older 103s. I suspect the latest 103 may have some small differences -- it is the 1st major change in the cone/basket design in some 40 years.
I expect that developmental details would not differ much from those required between Martin's theoretical 103e and a real one, just in (perhaps) different directions.
dave
tubesguy said:
Yes. Measured efficiency across some 20 pairs indicate a typical pair is between 89 & 91 dB (the same or a tiny bit more than a typical FE103. FR should be broadly similar to older 103s. I suspect the latest 103 may have some small differences -- it is the 1st major change in the cone/basket design in some 40 years.
I expect that developmental details would not differ much from those required between Martin's theoretical 103e and a real one, just in (perhaps) different directions.
dave
Yes, the complexity is simply moved but also the cost of experimentation is quite a bit lower. I have a dozen amps and lots of small caps and resistors. The cost of the specific inductors and big caps is about $140 so I had merely been looking for a less costly route to try the basic design.
But as Martin has stated, this was designed as a simple, passive system.
amt
But as Martin has stated, this was designed as a simple, passive system.
amt
I took my lead from LinkwitzLabs and it's more complicated than I normally like to have.
From my mentor and friend (since passed away) " the hardest thing you can do is to make something simple".
At the lazy B ranch (Boeing) we once designed a wheelbarrow. It could self level and even distribute loading, in the end it really didnt perform better than something you buy at Wally World, besides who wants a 30K$ wheelbarrow.
I'm surprised that you have been able to get away with such a simple crossover
^ see above.
ron
From my mentor and friend (since passed away) " the hardest thing you can do is to make something simple".
At the lazy B ranch (Boeing) we once designed a wheelbarrow. It could self level and even distribute loading, in the end it really didnt perform better than something you buy at Wally World, besides who wants a 30K$ wheelbarrow.
I'm surprised that you have been able to get away with such a simple crossover
^ see above.
ron
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.