Best JBL driver for a HiFi 80 Hz horn

[more10]

My next project is to build a 80-400 Hz front loaded horn for HiFi use. After a lot of reading I have realized that the JBL transducers I have (e145, 2226h, 2205a) are not optimal for this. There are better drivers. I have been thinking 15", but maybe 12" is a good option?

2227h: EBP 182, BL 23, RE 4.7,
e140: EBP 168, BL 24, RE 5.5
e145: EBP 135, BL 16, RE 5.7

The 2227 should be very good, but is hard to find. Can a 2226 be reconed to a 2227?

Any suggestions are welcome.

There are of course other brands, at speakerplans they talk about EV 15", but after reading other threads on this forum, I rather stick with JBL :-). Precision Devices makes a few horn drivers too.

[1audiohack]

The 2227 is all it's own, it's cone assembly will fit nothing else. Have you run the numbers on the 2220? Your application is right in line with those as well.

[more10]

2220h: EBP 178, BL 22, RE 5,7. Seems to be a very good candidate. And it cannot be built from any other basket. Thanks 1audiohack!

2202h (12"): EBP 278! BL 22.5, RE 5.5. What will be the tradeoff going from 15" to 12"?

[yggdrasil]

I have done some horn modeling using JBL drivers, and the best I have found is the 12" 2202H. You could also look at 2206H and 2020H.

If my memory serves me right, the 15" 2225H models ok.

IMO a 12", in a home environment, on a horn, has the possibility of being a very good combination.

[Hoerninger]

My next project is to build a 80-400 Hz front loaded horn for HiFi use. ... maybe 12" is a good option?It is not that easy to develop a horn on your own. I would highly recommend a simulation with HORNRESPONSE by David McBean (look this forum) for getting an understanding of the interaction of parameters.
Do not care too much for rule of thumbs :no:. You will need the Thiele-Small parameters, not only a small subset. Maybe the simulation will not give you the exact real horn performance but close to it (I have done it :thmbsup: ).
For getting 400 Hz you should care for a light cone. So a 12 incher is a good idea. For home environment it will be loud enough, Isuppose. The simulation will show you.
____________
Peter

[more10]

Thanks for your comments Hoerninger and yggdrasil.

I am looking for 2202H or 2220H. I will do simulations for these using Hornresp quite soon.

I know that it is not an easy task to design and build a horn, it is quite fun though. And I got plenty of time.

[yggdrasil]

I tried a quick simulation with the 2202A parameters, which I had already registered in Hornresp.

The parameters are quite close to the 2202H to give you a head start. Here's an 80Hz horn with few/no size constraints, which you might want to try to play a little with.

[more10]

Thanks for parameters Johnny. I have made a simulation in Hornresp using 2202H. The simulation confirms that 2202H will work in an upper bass horn (80-400 Hz). Around 130 dB.

The horn is a bit long though, the default horn generated by Hornresp is 150 cm. It is too long for a straight horn. I need to figure out how to shorten it.

[yggdrasil]

Shortening the horn to ~60cm can give you a simulated response down to ~130Hz, and usable down to 150-200Hz. (I have drawings for one in the workshop... They have already been there for a couple of months, so who knows when anything will happen.)

When it comes to front loaded horns there are no substitute for physical dimensions, so if you want that combination, then the dimensions are more or less given. If a few cm's help you can compensate by increasing mouth area and/or rear chamber.

Another alternative is to bend/fold the horn if that makes it easier to fit.

If your target are only about 2 octaves (80-400Hz) you can try to look at a tapped horn variation. This can also be modelled in Hornresp. The horn won't be shorter, but it will have smaller mouth and it will have to be folded.

[more10]

The horn is a bit long though, the default horn generated by Hornresp is 150 cm. It is too long for a straight horn. I need to figure out how to shorten it.

Playing around with Hornresp I have learned that the bigger the driver, the shorter the horn! A 15" driver will be 25 cm shorter. Placing it on the floor to a wall will further shorten the horn by 35 cm, 90 cm in total. A 15" horn it will be!

[more10]

Shortening the horn to ~60cm can give you a simulated response down to ~130Hz, and usable down to 150-200Hz. (I have drawings for one in the workshop... They have already been there for a couple of months, so who knows when anything will happen.)

Thanks for he offer, but I really want the horn to reach 80 Hz since this will be the bottom of my sides. I will use a tapped horn below.



Another alternative is to bend/fold the horn if that makes it easier to fit.


If I bend it I will loose phase information. My current system is phase aligned down to 300 hz, I can pinpoint female singers, but not male. I would like to try to include this horn in the phase aligned part of my system.



If your target are only about 2 octaves (80-400Hz) you can try to look at a tapped horn variation. This can also be modelled in Hornresp. The horn won't be shorter, but it will have smaller mouth and it will have to be folded.

I will have timing problems with a tapped horn as well. Actually I have a tapped horn using a 2226 already, but it needs some improvements :-)

[more10]

Now I have simulated e145-8, 2220H, 2205H and 2227H in Hornresp. They all go up to around 300 Hz, 2227H just a little bit further, but in a 15 cm longer horn. e145 have a bit wobblier frequency curve. Horn curve is basically the same for e145 and 2220H, so these can go into the same box (except for rear chamber). Compression for 2220H and e145 is about 1:3 and for 2227 1:5. 2205A need a compression of 1:4, and doesn't go as far as the other ones.

First choice will be 2220H. Since I already have a pair of e145 I will use these until I can bet a pair of 2220H.

[Hoerninger]

With the appropriate rear chamber you can flatten the frequency reponse at the lower end.
A wobblier frequency response at the upper end indicates a horn which istoo short.
The compression ratio is to your choice. But have had best results with 1:3 most of times.
Do these 15 cm really do have an impact on the upper limit? (A screen shot could be helpful :) ).
____________
Peter

[more10]

Do these 15 cm really do have an impact on the upper limit? (A screen shot could be helpful :) ).


The reason the 2207 needs a longer horn is because of its compression ratio which is higher. I have not tried much to modify the design genereated by Hornresp. The 2227H goes higher because of driver parameters, not horn length.

I will add some pictures tonight, I didn't realize I could upload images, this feature is quite rare in forums.

[more10]

4737447375

[more10]

4737647377

[more10]

4737847379

[yggdrasil]

A few thoughts:
- 5:1 compression (2227) ratio can be too much (giving much stress to the cone).
- Vrc and Lrc is very small (might actually be hard to build, and higher Vrc will give better/more LF).
- Have a look at the group delay too (there's normally lots of group delay close to fc)

[Hoerninger]

Another thought:
When the horn is free standing (in other words: not within a large baffle)
shouldn't be taken a room angle of 4*PI ?
____________
Peter

[more10]

- 5:1 compression (2227) ratio can be too much (giving much stress to the cone).




The cone used in the 2227H is reinforced with fiberglass for added strength.


Most probably the reinforced cone is intended for higher compression.

The 2227H is a bit too extreme for my taste :-).

Yesterday I bought two 2202H. I am going to use these for my project. Maybe I can get another 2, using two drivers in one horn will shorten it by 35 cm.



When the horn is free standing (in other words: not within a large baffle)
shouldn't be taken a room angle of 4*PI ?


These simulations are for a horn placed on the floor, half space (2pi). When the horns are up in the air 4pi is used.