New shelf box with AL170 and G25FFL (_WG)

[jama]

fine tuning (optional)

[nlcarbon]

OMG 23 componentes for a 2-way crossover...

Here my suggestion(-9 components)

[Willy_65]

OMG 23 componentes for a 2-way crossover...

No, but 19 components (without LS) and that for excellent results!
- frequency curve without large waves or peaks
- frequency response in the band 10dB = 32- and at 6dB 38-30 kHz
- perfect takeover at the crossover point
- very good frequency response directions to top - right and left
- ideal flat impedance line of 5 ohm for tube amplifier

Cost for 2 crossovers (including the PCBs) = 125 €
For a set of boxes that cost € 470, including crossovers but without housing, the quality must prevail. That applies to every part!

Otherwise, you might also equip a Porsche with a VW Beetle engine

[Willy_65]

The circuit board is designed, I am still waiting until after New Year to place the order.
If there are any comments in the meantime, please let me know in time.

[nlcarbon]

I added a option with less components in #32

Here a info how to place the coils:
http://www.troelsgravesen.dk/coils.htm

[Willy_65]

I added a option with less components in #32

Here a info how to place the coils:
http://www.troelsgravesen.dk/coils.htm

I have looked at your proposal, but then we are back to the point that we already reached a few days ago.
The impedance curve is fine for solid-state amplifiers, but less good for tube amplifiers: too low for 8 ohms and too high for 4 ohms. Moreover, the resonance at 60 Hz is not eliminated and tube amplifiers do not like that.

That is why I opt for Uwe's proposal, which means that the impedance curve is straight 5 ohms.
All this without compromising on the "flat" frequency characteristic and perfect phase transfer, although with a different concept and addition of some parts.

The link that you sent for placing coils on a PCB is interesting, but it teaches me that the induction deviation of 2 coils connected in series, gives a reducing of 3% when placed next to each other.
Good to know, but when it is difficult to do otherwise, it is also not so bad.

[Willy_65]

Via this link https://www.dropbox.com/sh/nox8xvdiz...pCIbl1yma?dl=0
you come in my dropbox and you can download files such as an excel containing all the necessary parts for building boxes.
I always order my Visaton parts via Reichelt and if not available there, via Visaton.
A hyperlink is provided to the Reichelt or Visaton website for each part. Each part is provided with the daily price on 2/03/2019.
With regard to the speakers, only those that were part of the design process are included.
If you want to edit the excel yourself to add, for example, speakers, you must download the excel.
Note: the last 2 columns contain formulas, so it is best not to delete those cells.
If you have inserted lines you will have to copy the formulas from, for example, the higher-lying cells with formulas.
At the bottom is the total amount of your order and the total value for the boxes. It is possible that you still have parts that you do not have to order, which explains why you have to enter 2x the number for a part and can get 2 different totals at the bottom.

[Fraggle]

Hello Willy,
I noticed an error in the BPJ file posted in #27 by UweG. Your modified version probably has the error too. The impedance data for both drivers are incomplete. Therefore, the simulation result is not correct.

You can restore the correct impedance data by reloading both drivers from their files. Then you will notice the response is not optimal. So you will probably want to revise the crossover. I hope you have not ordered all the parts yet.

I am attaching two suggestions. Version FR2 takes some extra parts, but version FR4 does not. The results are similar. I am not sure which one is better. The directivity index is slightly smoother for version FR4, while version FR2 maintains a better frequency response in the 10 degrees up and down directions.

[Willy_65]

Hello Fraggle,
Thank you for thoroughly reviewing the design.
No PCBs or orther parts have been ordered yet, so no problem adjusting anything. The next few days I have to little time to delve deeper into it, but I will certainly work it out.
The directions response is the most important in terms of top - right and left. To the bottom is in the current setup (shelf) less important.

@UweG:
Any idea why the impedance data from the speakers suddenly stops at 799Hz in your adjustment?

[jama]

Hello Willy,
I noticed an error in the BPJ file posted in #27 by UweG. Your modified version probably has the error too. The impedance data for both drivers are incomplete. Therefore, the simulation result is not correct.

effectively there’s a mistake seeing the f-gang
I recharged AL130 and G25 Ffl-WG
I will choose the version FR2

[UweG]

Hello Willy,
sorry for the mistake. It happened because I used a not published temporary version of Boxsim that had a bug.

To be honest, I do not like this additional capacitor in series with the tweeter, as it shifts the phase in the wrong direction und makes the energy response worse.

Attached please find my suggestion to repair the bug. I liftes up the directivity a bit and tried to smoothen the enery reponse.

[Willy_65]

Hello Uwe,

I agree that the additional capacitor in series with the tweeter, shifts the phase in the wrong direction and makes the energy response worse.

If you still want to optimize it, I'll send you the latest version of "SHELFI" with driver data updated - in attachment.

Thank you all!

[UweG]

Sorry, I did not see your post when I edited my last ine. This one has reloaded drvers.

[Willy_65]

Sorry, I did not see your post when I edited my last ine. This one has reloaded drvers.

Thank you Uwe and Fraggle for discovering the error and for the adjustments while maintaining perfect phases, good power response and good frequency response directions for the shelf speaker.

Uwe, I only realized this evening that you had already placed your adjustment in your previous post, too busy with a big paint job here at home.
One of the following days I will be able to free up some time to make a new PCB layout, let it be a few parts more, the final quality prevails!

And I love Boxsim 2.0, I have made almost 150 simulations (for testing / learning) since the launch. Only working with 2 instances of Boxsim at the same time is not reliable when you start optimizing one, you should try it.

[UweG]

Only working with 2 instances of Boxsim at the same time is not reliable when you start optimizing one, you should try it.

I'll definitely do that with the newst version 2.1, unfortunately available only in german language.