I agree it is bad, and seems like it's a result of keeping things cheap. I complain about it a lot on ASR in hopes the amp makers who sometimes join in the discussions start implementing designs without it.
Here you go AllenB, I was pointed to this thread when I was trying to diagnose the problem. There is some stuff in there that is beyond me, but one user states it's due to a lack of post filter feedback in the amps. I have no idea what that is but just some info to go off of.
https://www.audiosciencereview.com/...and-fr-into-complex-speaker-dummy-load.21682/
This is some data from another user on ASR who ran in to the same problem. I guess the blame really lies with the amps but these speakers are likely to be used with cheap class d amps so I wager a better solution is to create a xover load that doesn't present problems for the amps.
Also I'm sorry for derailing your speaker thread OP.
Here you go AllenB, I was pointed to this thread when I was trying to diagnose the problem. There is some stuff in there that is beyond me, but one user states it's due to a lack of post filter feedback in the amps. I have no idea what that is but just some info to go off of.
https://www.audiosciencereview.com/...and-fr-into-complex-speaker-dummy-load.21682/
This is some data from another user on ASR who ran in to the same problem. I guess the blame really lies with the amps but these speakers are likely to be used with cheap class d amps so I wager a better solution is to create a xover load that doesn't present problems for the amps.
Also I'm sorry for derailing your speaker thread OP.
Last edited:
Any amp that does this is not an amp I would bother with. Discussing it further also derails this speaker design thread.
It would be cool if another person built this XMachina designed speaker. It's a really unique piece of software (albeit some of the crossover results are a bit unusual compared to normative designs).
It would be cool if another person built this XMachina designed speaker. It's a really unique piece of software (albeit some of the crossover results are a bit unusual compared to normative designs).
If XMechanik wishes we can split this whole sub discussion to a separate thread since it has persisted.
I haven't read the whole thread but there doesn't seem to be a comparison between a resistive and reactive load. Either the issue is or isn't fixed by using a resistive load. Furthermore either the issue remains the same over different conditions or it varies.
In other words it looks like either you can fix this with an equaliser or you need another amp.
Impedance variations are smaller in case of Mechano22 than in case of C-Note, so the issue of load dependency should be less noticeable (but not non-existent). The situation can be improved by adding a 6.8R and 47uF compensation branch, see the picture. With this modification you can get 4.4ohm +-0.5ohm for f>150Hz, which should be ok for most class D amps.
Wow, how times have changed. 30 years ago, no one would ever consider doing this because it (1) lowers the impedance and (2) it wastes power. But now we have class D which handles low impedance well and can provide a huge amount of power. I am not saying this circuit is a bad idea, on the contrary it may be a great idea for class D. But certainly is a change from the days of MOSFET and tube amps.
edit : I may have spoken too soon, I think that Thiel did something like this to even out the impedance... so I should not have said "no one"
I have only used HF filter version with the 100 ohm resistor and it doesn't seem to cause any problems.
Even though the crossover is a bit complicated there was no problem to put it in the enclosure which is approximately same size as C-note. Some flexibility with mounting can be gained by splitting the LF and HF on separate boards.
The crossover is designed to deliver flat SPL and Power Response characteristics. But the front baffle used is different than in C-note. You likely won't get exactly the same characteristics if you use C-note front panel.
Last edited:
According to my bills, I payed c.a. 100USD for xover components for pair of speakers. But it was more than year ago and also prices in my country could have been different. Anyway, check if you're looking for right components. For example 3.7mH and 2mH coils are wound with a thin wire (intentionally high esr) and should not be very expensive.
You probably still have your C-note xover. Try some RC impedance correction (something like shown in #68) for better D-class cooperation and maybe you're done. C-note is not so equalized (as Mechano22) but it still sounds good according to many opinions.
I am constructing a two-way passive speaker, and I just paid $185 for crossover components for the pair. Some of the component values I needed were only available in premium name brand (i.e. Jantzen). If I could have found everything in Dayton-branded standard grade, I would have saved about $35, but some things were out of stock, and I did not want to wait.
A 3 way crossover can be be rather expensive, and spending $300 a pair for crossover components would not be unusual. In fact, I think that spending less than $200 per pair for a 3-way crossover would be unusual.
Active crossovers utilizing DSP are more expensive if you have to start from nothing and buy all the equipment and amps. But the cost differential between active and passive is getting smaller. If you already have some amps available, then it may make more sense to buy a miniDSP and go active. If I recall, a single MiniDSP 2x4HD will make a 3-way crossover for ONE speaker, so if I am right, you would need 2 miniDSP's. You could also go semi-active and use the MiniDSP for the lower crossover on both speakers, and then use a passive crossover for the upper crossover. It is the low frequency passive crossover that is most expensive... 8 mH low-DCR inductors and 180 uF caps are spendy.
There are software options for DSP crossovers, and if you always run a PC with your audio system anyway, this is an option. Then you don't need to buy a MiniDSP.
j.
A 3 way crossover can be be rather expensive, and spending $300 a pair for crossover components would not be unusual. In fact, I think that spending less than $200 per pair for a 3-way crossover would be unusual.
Active crossovers utilizing DSP are more expensive if you have to start from nothing and buy all the equipment and amps. But the cost differential between active and passive is getting smaller. If you already have some amps available, then it may make more sense to buy a miniDSP and go active. If I recall, a single MiniDSP 2x4HD will make a 3-way crossover for ONE speaker, so if I am right, you would need 2 miniDSP's. You could also go semi-active and use the MiniDSP for the lower crossover on both speakers, and then use a passive crossover for the upper crossover. It is the low frequency passive crossover that is most expensive... 8 mH low-DCR inductors and 180 uF caps are spendy.
There are software options for DSP crossovers, and if you always run a PC with your audio system anyway, this is an option. Then you don't need to buy a MiniDSP.
j.
I'd have to put the xover together again
I'll have to double check and see if I'm picking the right parts. It does look like I picked a few overkill inductors.
They do sound good, but having heard them with active filters similar to the tasks your speaker has on the woofer, the improvement is not small to my ears, they go from pretty good to great
Thanks for your great contribution to DIY speaker making and Xover design 
Freq response is King including off axis so right way to go for sure. But we should also remember that the system includes the room, so it is in room response and off axis resp which is King
Now following thread
Cool designs, but the rather complex Xover design will easily come to cost as much as the drivers ....
Don't seem to be possible to download the XMachina program at the moment
https://xmachina-ai.blogspot.com/p/current-version-info.html
Freq response is King including off axis so right way to go for sure. But we should also remember that the system includes the room, so it is in room response and off axis resp which is King
Now following thread
Cool designs, but the rather complex Xover design will easily come to cost as much as the drivers ....
Don't seem to be possible to download the XMachina program at the moment
https://xmachina-ai.blogspot.com/p/current-version-info.html
Thanks. Also from my experience (not very extensive tbh), when you equalize not only system spl but also the Power Response and all that directivity related characteristics, the sound gets on another level.Thanks for your great contribution to DIY speaker making and Xover design
Freq response is King including off axis so right way to go for sure. But we should also remember that the system includes the room, so it is in room response and off axis resp which is King
As for me, I do not make any hard assumptions that the xover components costs should not reach or exceed the price of drivers. If this makes the system reaching expected performance, then why not after all.Now following thread
Cool designs, but the rather complex Xover design will easily come to cost as much as the drivers ....
link
Hi Xmechanik, I'd like to built your Mechano23 as I own an SB13PFC25-04 (not rounded) and some components, would you help me if needed?
I also have a clean self-made box, HxWxD of 250x200x240mm for about 9L - I just have to design and add its front panel.
Thanks in advance and congrats for your nice work : )
Nicolas (France)
I also have a clean self-made box, HxWxD of 250x200x240mm for about 9L - I just have to design and add its front panel.
Thanks in advance and congrats for your nice work : )
Nicolas (France)
Last edited: