Help with WinISD and XLS 12

[damo21]

I left out the lowpass filter on the graph of GD yes, I will try again and post a new one. I'm trying to make sense of the advice in Paul Spencer's group delay wiki page found here. If you have better advice on GD please update the wiki.

I wanted to explore audible GD because it seems to be the main difference between a vented sub and a closed sub... of course I would expect any vented sub has much larger GD than a closed one. But if it's possible to reduce GD in a vented sub, it would be better sounding bass for music reproduction correct?
The whole point is that you can get higher SPL for the same driver if you make a clever ported box, but at a cost of introducing unwanted delay.

So you are saying relative GD is what's important not the actual GD. How can I draw a graph of relative GD? Relative GD = GD / GD @ crossover freq ?

[damo21]

Originally posted by Rademakers
Over here (the Netherlands) most people seem to use, GD max= Frequency x Delay = <400, i.e. 10 ms at 40 Hz. The shape would matter here. Rather a smooth GD as one that is peak like.

The line as in the thumbnail above would suggest that the GD of about any 6th order bandpass would stay well below audibility while it's supposed not to be. Even more so, it would be rather difficult to create a sub with audible GD if this truly is the case.

Thanks for your comments Rademakers, here is a graph introducing your GDmax curve:

I would trust yours better because Paul's was just a rough guess. We are trying to create more accurate models so we can improve the simulations for everyone.

[rabbitz]

Originally posted by damo21
Vented 65L (tuned to 18.5Hz)
Port: 100mm diameter Pvc pipe x 1.0m long
Flares: 50mm each end

The vented box is too big and I say the Fb is too low. What will happen is the xmax will be used up quicker and you will lose efficiency.

With good flares, you don't have to go larger than 90mm ports.

[richie00boy]

It is promising to see GD being considered seriously in a design but focusing on this aspect alone is sure to be a bad idea. You need to consider a balance of parameters.

You will not have sufficient room gain to naturally EQ your design because it begins rolling off from such a high frequency. Room gain is a bit of a red herring IMO anyway. If you listen with a door open or you have a fairly open plan house it can negate a lot of the effect anyway.

Depending on the type of music you listen to and your tastes in sound you might actually want an anechoically flat design and take advantage of the extra bass from room gain. In my own case I like a bassy sound at times and I can set my sub to be anechoically flat to 20Hz. My room gain begins around 33Hz in my 17x12 ft lounge, so I can get a real impact when watching movies and when real deep bass comes in on music!

[damo21]

Interesting... So what are the tradeoffs for using EQ in a sealed sub? Surely something is lost by doing this.
Is there a simple kit I can buy from Jaycar or something for EQ? I am using a plate amp with L/R line inputs. I heard maybe a linkwitz transform circuit? How do you incorporate one of those into a plate amp?

[rabbitz]

I'm a vent head so won't comment on the sealed.

Jaycar do sell a nice sub processor kit KA1814 and works well in the right application such as in conjuction with a power amp, but will not help you here. The equaliser in it will not do the job and only the LT will help.

http://sound.westhost.com/project71.htm

Have a read of that and the links on the page as it's very informative... they go in at line level before the amp.

[damo21]

Ok, I read all that info and played with the Linkwitz-Tr spreadsheet calculator. It seems I'll need a maximum of 12dB of boost down the bottom end to roughly flatten the response curve down to 20Hz, in a critical q=0.5 box (35L).

When I tested this with WinISD it's telling me that the cone excursion will be up around 40mm with 200W passing through it! (surely will destroy the cone) Only 20Wrms gives the full recommended cone movement in the simulation down at 20Hz, but this only allows 96dB SPL at maximum output at 20Hz.

Does this mean I can't get a useful closed sub from this driver? Or is there something wrong with the sim?

[paulspencer]

Damo, please don't call it my curve regarding the GD!!! It is based on the study by Bauert and Laws, and a relationship was found with frequency and audible GD and it was related as cycles. It was not purely based on a linear relationship between frequency and GD, but rather there was a number of cycles which reduced as you went lower in frequency. John Murphy made a suggestion as to how you would extrapolate down further, and I merely calculated some values following his suggestion.

As for the other suggestion, I don't know where it came from.

I think you will find that 60L will be the minimum practical for XLS to get a vent that won't chuff. Also a 90mm vent may prove a challenge. Larger flares have more compression, so you should have the vent as large as you can manage. But it really comes down to balancing factors out. Try both 90mm and 100mm and keep in mind that a 1m vent won't easily fit inside a 60L box, and you will have a lot of bends. Upsize the box and the vent will then be shorter and you will have more room to fit it in as well. Lower tuning reduces vent velocity as well. Quite a lot of juggling is required.

I suggest designing with all the filters included. Once you start including filters, GD goes up a lot and you see a more complete picture of the performance. Try a sealed sub with a linkwitz transform and you will find if you match the response that the GD is about the same but that max SPL is 6db less. All you have gained is a smaller box!

It's easy to say you have a low GD vented sub until you add filters! As others have commented, don't compare while only considering one factor. There are a lot of opinions on GD, but there doesn't appear to be any credible well established research prove where it is actually audible.

What is certain is that high GD goes hand in hand with 20 Hz extension.

Don't count on room gain. If you live in a typical Australian house you may not have any at all. My room is ~4x5m and I have no room gain, just a room mode peak at 35 Hz. I understand a lot of US homes are more solid, have worse room modes and more room gain, hence many of their comments in this area won't relate as much to Aussie homes.

If anyone else has other recommendations on GD that is audible, please feel free to add it. But also include what it is based on.

[damo21]

For me, the size of the box is irrelevant, I just want a good sounding sub!

Thanks Paul, yes it appears that a closed box will end up 5-6dB quieter after the linkwitz transform for the same input power! And there is very little benefit in group delay after all the filters! I would prefer the extra 6dB at 20Hz, and I'd rather spend my efforts on building a nice box than fancy filters, (besides, my plate amp already has a rumble filter at 17Hz) so it looks like I'm building a vented one:

65L 100mm pipe x 900mm long. Massive flares > 60mm to reduce port noise.

[paulspencer]

Damo, assuming that you have the regular XLS (not the car version and not the XXLS), then 65L with 100mm vent with large flares and 900mm long looks like it will work well. I did a quick model and looked at response, excursion, GD and vent velocity. They all look ok.

It would be tuned at 19 Hz which with a 17 Hz rumble filter (3rd order?) ... will keep cone excursion no more below tuning than it is above.

GD is 50ms @ 20 Hz which I think is ok ... 35ms with the rumble filter off. Hopefully it is defeatable so that you can compare and see what you think. If it is a problem then for music you can turn it off.

Vent velocity peaks at 25 m/s which is ok with a 100mm vent if the flares are large enough.

A 2nd order filter set at 50 Hz gives you -3db points at 20 and 80 Hz anechoic, lower if you have the rumble filter off and more so with room gain potentially.