Wondering if a Classic 2.5 clone can be made with a series crossover on the woofers

[LOUT]

The cap-bypass on a woofer method is not supposed to work properly. There are threads here at PETT about just this idea in particular, and the greater minds than me said it does not work.

I'll probably try searching and digging around for more details then (fingers crossed there aren't tons of "series 2.5way" tagged threads to look through).
In the meantime do you happen to remember if the reason/s it didn't work were something that didn't show up in simulation or if the reason/s should become clear in a simulation after learning what to look for?


EDIT
Not sure if there were other reasons I'm missing, but a thread that showed up mentioned this:

Re: 2.5 way with series woofers?
It won't work with series drivers for one reason, you don't get the sensitivity increase with drivers in series.
2.5-way requires that the drivers in question be in parallel, so that you have the 6dB sensitivity increase at the lowest frequencies.
I'd go ahead and add a couple more woofers and do the MTMWW 2.5-way, as I did with the Schumakubin. You keep the impedance above 4 Ohm overall that way.

I see what they mean where the series version won't get the voltage sensitivity increase, but it can still get the full +6db from the surface-area and wattage-handling relative to the single woofer...just requiring an overall higher voltage all the way through (which is expected anyway since it needs to be an 8ohm build). It kind of feels like bulding down instead of up, but the end result between the pair of woofers relative to eachother seems to be about the same. Some woofers might make it difficult to stay at/above 6.6ohms through the mids after the lower woofer is shunted away, but the DC160-4 seems to work really well for this arrangement.

If you (or anyone else) finds any links or thinks of any other reasons why a series 2.5way shouldn't work, please speak up.
This looks totally legit to me, but I'm very new to all of this so I'm sure there's a lot I need to learn.

[Chris Roemer]

Hmmm ... not sure...

I KNOW Ben knows his stuff, but Lout's 5-element 2.5-way (bass section) SEEMs like it would work as hoped. I don't see anything weird in a sim.
If the parts were in my parts bin I'd certainly give it a go.

(interesting)

[Kootskid]

I like all the thoughts and the knowledge from the threads that have been investigated. Can a TMWW 3 way be built with the 2 woofers wired in series?

[LOUT]

Can a TMWW 3 way be built with the 2 woofers wired in series?

It can, but I'm pretty sure the DC160-4 can reach decently high into the mids and the DC28F can reach pretty low for a tweeter, so I'm not sure if adding a midwoofer between them would give much of a benefit.

That said, I have no experience with the DC160 aside from sims and manufacturer specs, and while the DC28F claims it can reach low and I don't personally notice it stressing out over it, that doesn't mean it's playing cleanly that low. I'm probably not nearly as sensitive to some kinds of distortion as many others....so I suspect I can happily head-bob away to lower XO points where folks with more experienced ears would likely cringe. I think I remember reading the old versions of the DC28 liked to be crossed way higher than the specs claim if it expected to stay smooth, but I'm not sure if that's the same with the current model.
AKA, I thiiiiink the 3way wouldn't offer much advantage over the 2.5way with these speakers, but if someone more experienced says otherwise I'd definitely follow their advice over my guesses.

[Geoff Millar]

It can, but I'm pretty sure the DC160-4 can reach decently high into the mids and the DC28F can reach pretty low for a tweeter, so I'm not sure if adding a midwoofer between them would give much of a benefit.

That said, I have no experience with the DC160 aside from sims and manufacturer specs, and while the DC28F claims it can reach low and I don't personally notice it stressing out over it, that doesn't mean it's playing cleanly that low. I'm probably not nearly as sensitive to some kinds of distortion as many others....so I suspect I can happily head-bob away to lower XO points where folks with more experienced ears would likely cringe. I think I remember reading the old versions of the DC28 liked to be crossed way higher than the specs claim if it expected to stay smooth, but I'm not sure if that's the same with the current model.
AKA, I thiiiiink the 3way wouldn't offer much advantage over the 2.5way with these speakers, but if someone more experienced says otherwise I'd definitely follow their advice over my guesses.

Just going by the PE provided frequency response (please see graph above) it doesn't look as if the 4 ohm version would be that great in the mid range, which to be honest isn't the DC160-8's strong point either. Great bass and it sounds smooth, but.

The DC160 works well in the Classix II and I really like how it reproduces less well recorded material. Paul put a lot of work into his crossover design!

Dennis Murphy of Philharmonia Audio released his 'Affordable Accuracy Monitor' based on the Dayton BR-1 kit which uses the -8 and the DC28F, I think his revision crossed the DC28 at about 1700Hz, it's detailed on his "MurphyBlaster" web site.

Curt's Tritrix crossed it a bit lower than that, I recall. I haven't noticed the DC28 sounding too stressed in that design, I find the DC130s bottom out before that point depending on the source material.

Geoff

[Wolf]

IIRC, it has something to do with the way the voltage works and how the current flows through the cap and adjacent woofer.

There is no gain in series woofers except gaining a more benign load. Sure, they handle more power total just as parallel does, but the doubling of SD is offset by the halving of current to make the gain null.

Wolf

[Chris Roemer]

Here's an XO suggestion (others can sim if they like), I think you'd find them "enjoyable". They'd use the same 1.1cf box as the 2.5 Classix, w/the DC28 tweeter. Fc near 2.4kHz. Around 85dB w/a slightly relaxed mid (like Paul's?). A few dB BSC. Using Lout's example for the 2".5".

3rd order HP: 4uF series cap, 0.30mH shunt coil (to gnd), 6uF series cap, and a 30n(ohm) resistor to grnd (across the tweeter). Skipping the shunt makes the DC28 a bit brighter. (Reverse polarity is "indicated" - but feel free to try it both ways (w/just the upper woofer?) to see which phase sounds best.)

LP: "notched" (tanked) 0.50mH series coil (notch is 4n resistor in series w/a very small 1.5uF cap, across that coil), then a 25uF shunt cap (to gnd), then a Zobel-like RC to gnd, using a 5n resistor and a 200uF(npe) cap, then both woofers in series w/an RC across the lower(in box) one using 2n+100uF(npe). The notch is for the (minor) breakup near 3k, not the (major) breakup an octave lower.

If you wanted to run a test (for topology feasibility) you could just put both woofers in a vented box (around 1cf), wired in series, w/a "Zobel-like thing" across the 2nd woofer (using a 2ohm series resistor and a 100uF series cap). In a sim, adding/removing that RC (to gnd - or across the bottom woofer) doesn't change the "summed" response, BUT - it does let the upper woofer play higher (to meet the tweet) while rolling the lower one off earlier (to reduce phasing issues w/the tweeter). Try it.

[Kootskid]

Chris,

Thank you for your time and effort with the x-over. I will try this. The only problem, is that it may not be right away. I have some other life irons in the fire and some audio projects slated for over the winter, so I probably won't be able to get to building until after Christmas. Since I am very new to this hobby, I appreciate all the help and knowledge that gets thrown around on this forum. "" I will try this"" First I need to draw the x-over and check back with the forum.

Myles

[LOUT]

Here's an XO suggestion (others can sim if they like), I think you'd find them "enjoyable".

Chris,
Thank you for your time and effort with the x-over. I will try this. The only problem, is that it may not be right away.

At least in simulation, that particular XO ends up at 4ohm instead of 8ohm. Kootskid, is your amp fine with a 4ohm load (you're just avoiding 2ohm), or would 8ohm be needed either for the current amp or an affordable AVR in the future?

[LOUT]

Chris Roemer, I tried to see how close to your XO simmed FR I could get while keeping impedance at/above 6.6ohms and would like your thoughts on this (particularly if there are any glaring flaws...like if the top woofer's response above 3K isn't cut down enough).

[Geoff Millar]

Here's an XO suggestion (others can sim if they like), I think you'd find them "enjoyable". They'd use the same 1.1cf box as the 2.5 Classix, w/the DC28 tweeter. Fc near 2.4kHz. Around 85dB w/a slightly relaxed mid (like Paul's?). A few dB BSC. Using Lout's example for the 2".5".

3rd order HP: 4uF series cap, 0.30mH shunt coil (to gnd), 6uF series cap, and a 30n(ohm) resistor to grnd (across the tweeter). Skipping the shunt makes the DC28 a bit brighter. (Reverse polarity is "indicated" - but feel free to try it both ways (w/just the upper woofer?) to see which phase sounds best.)

LP: "notched" (tanked) 0.50mH series coil (notch is 4n resistor in series w/a very small 1.5uF cap, across that coil), then a 25uF shunt cap (to gnd), then a Zobel-like RC to gnd, using a 5n resistor and a 200uF(npe) cap, then both woofers in series w/an RC across the lower(in box) one using 2n+100uF(npe). The notch is for the (minor) breakup near 3k, not the (major) breakup an octave lower.

If you wanted to run a test (for topology feasibility) you could just put both woofers in a vented box (around 1cf), wired in series, w/a "Zobel-like thing" across the 2nd woofer (using a 2ohm series resistor and a 100uF series cap). In a sim, adding/removing that RC (to gnd - or across the bottom woofer) doesn't change the "summed" response, BUT - it does let the upper woofer play higher (to meet the tweet) while rolling the lower one off earlier (to reduce phasing issues w/the tweeter). Try it.

I think I'm doing something wrong here with my Xsim attempt at translating this into a schematic: the impedance graph looks wrong, it should not be going as low as 4 ohms?

[LOUT]

It looks like the 25uF cap on your LP isn't connected to ground, but otherwise I was also seeing a ~4ohm load from that XO in VituixCAD, so I think that's by design. I'm guessing Chris was aiming for 4ohm because 2.5way builds are more commonly 4ohm (parallel 8/8ohm) and OP is likely aiming to avoid 2ohm moreso than actually needing 8ohm.
Though I do think (hope?) an 8ohm version might be possible without giving up too much for it...I think the sensitivity and general sound can be kept about the same, though I'd like to hear back from Chris for his thoughts on the ~8ohm XO to make sure I didn't do anything stupid with it.

[Geoff Millar]

Oops! Missed that mistake, thank you

As I know next to b- - - r all about crossovers and electronics (ordered a copy of Speaker Building 201 off the Internet but never received it!) I don't understand why a 8 ohm tweeter, with two 4 ohm woofers connected in series to make an 8 ohm load, make for a 4 ohm load when incorporated with the crossover....

Geoff

[Kootskid]

Hi Lout, Geoff, Chris,

I will have an amp that can handle 4 ohms. Also Lout you are correct about having between 4 ohms and 8 ohm system. 8 ohm would be great, but would be satisfied with any above 4 ohms. Following with interest even though i am just learning about x-overs. Lets hope Chris chimes in.

Myles

[Chris Roemer]

Lout, If you drop the "notch" cap from 1.5uF down to 0.47 (which pushes the 3k breakup down another -10 to -11dB) I think we've got a winner.
Nearly the same FR (even a bit flatter) with a Z-min of 7.4n, and THAT's on the tweeter section.
The need for that notch to be smaller comes from your use of a larger series coil (1.0mH vs my 0.50).

My vers. of Lout's LP topology (in MY sim) did indeed drop to 3.3n in the 300-400Hz range. (I didn't see that the OP requested/required an 8n load.)

[LOUT]

Lout, If you drop the "notch" cap from 1.5uF down to 0.47 (which pushes the 3k breakup down another -10 to -11dB) I think we've got a winner.
Nearly the same FR (even a bit flatter) with a Z-min of 7.4n, and THAT's on the tweeter section.
The need for that notch to be smaller comes from your use of a larger series coil (1.0mH vs my 0.50).

My vers. of Lout's LP topology (in MY sim) did indeed drop to 3.3n in the 300-400Hz range. (I didn't see that the OP requested/required an 8n load.)

Interesting. With the FRD and ZMA files I'm using (from PE, not measured) it looks like shrinking the LP notch cap from 1.5uF to 0.5uF slightly boosts the top woofer's response at 3000hz (though only by ~3db) while the decreases happen closer to 5000hz and above....Is this probably a quirk from slightly different FRD/ZMA files (yours might be newer or actual measured) or am I maybe misunderstanding something?
Not that it probably matters a large amount, since it sounds like the cap shrink is either significantly better (from what you're seeing) or mostly the same (from what I'm seeing in Vcad)....so I'm not arguing, just trying to understand why it looks different on my end.

If it helps anything, here's what I'm seeing:


OR




This is without any Zaxis/timing offset information added, nor any box impulse or diffraction/BSC information....just FRD and ZMA.


And, as Wolf warned earlier, there may well be something that goes wrong in a series 2.5way. The sim can't show everything, and there might be something sneaky that makes it sound wonky.
I'm hoping that it might instead just be a "telephone-game" effect of someone getting bad results a long time ago and the results/info getting a little scrambled along the way and time. I think the current flowing through the cap and adjacent woofer should act similarly to most series XO's...just with another woofer where the tweeter would normally be in a series 2way. I do remember hearing something about woofer sharing a box having different interplay when wired series VS parallel with something interacting physically with one of the woofers..so I'm guessing something electrically interacting might cause a similar effect in some situations. Though I can't remember what this effect was nor if it was even a vetted thing or just guitar player heresay.

[Kootskid]

Nice Work Lout, Chris, I have a few of questions:

1. For the tank circuit, can I assume the .68ohm & 16ohm resistors are in parallel and replace it with one resistor wired in series with the coupled 1mH inductor and 0.47uF cap. ?

2. What is the purpose/influence of the 75uF cap and is influence seen by both woofers or just the first one.?

3. On the schematic, D11 is the upper woofer and D12 the lower woofer. ?

3. Is there anymore items that need to be discussed before I give this a try. ?

Regards,

Myles