[painkiller]

Interresting stuff. I think I have all needed parts at home. 225VA 2x15V trafo, 10mH inductors and of course resistors and caps. Got to try this.  aa

[Graham Maynard]

Hi Painkiller,

I wish you Luck, and do feel free to ask questions.

My own 2x 15" Beymas in parallel are presently being driven by the transformer based circuit which is now no more than a 500VA 2x40Vac transformer; the single 6.4mH choke - no resistor; 2x 470uF electrolytics in parallel with a one ohm resistor in series.

The Beymas are now rearward facing.  The upper one protruding through its baffle opening; the floor one surface mounted on the front of the baffle.  This is like having another crossover section, but without the phase delay in the lower mids, and without additional losses at LF.

Cheers ....... Graham.

[painkiller]

Holy Moses! 

I tried this circuit yesterday, as I said I was going to. I used the schematic in post #90. Values were 1 ohm parallell resistor, 1uF cap with 2 ohm series resistor, 10mH inductor (unknown rL) and 225VA 2x15Vac trafo. The drivers are Alpha 15A. I only had one trafo, so I could only make it for one speaker. Good for direct comparison though.

Anyway, the effect is remarkable. It's unreal how much more bass that one 15" makes. I can get away with 10-15dB less EQ now. It's deeper, it's louder and it still sounds awesome. I haven't really read half of this thread, and I have no idea of how this is supposed to work. But it works, and it's a true blessing for efficient OB bass.

So now I obviously have a lot of question, as I intend to implement this into my speakers. First of all, you use huge trafos. Will bigger be better in any regard? You also say you use no parallell resistor. Does that mean you connect the 0V tap directly to the (-) negative terminal? How to you calculate optimal component values? Will the same circuit be optimal for different drivers, or do the optimal values depend on T/S parameters? How about the polarity, does it change compared to a directly coupled driver?

Great circuit by the way! Really clever. Everyone should try it!

[painkiller]

Another question. This circuit opens up for using low Q bass drivers. How low can you go? Can it bring reasonable efficient bass out of a 0.3Qts driver like the Beyma, or is it better to aim for a moderate one around 0.5-6 like the Eminence Beta? My problem is matching the 99dB efficient midrange. The Alphas come close. Any hope of doing this with a lower Q driver?

[Graham Maynard]

Hi painkiller.

So good to hear of your success - and with higher Qes Alpha-15As too !

The first point I noted was your typed capacitor value. 
Here in Europe 1mF = 1,000uF.  1uF won't actually do anything in this circuit, but won't prevent it from working either.  If a typo then just correct it so that I can remove this observation.

Okay.  You are using the circuit in post#90.  This would not be my choice because it further increases the driver Qes, though in so doing the LF output also would be even further increased.

This version is my preferred circuit, though the post#90 entry might better lift single Alpha output to match a widerange driver's sensitivity.




The Alpha's higher Qes and 41Hz Fs give an SPL lift which covers about 35Hz to 100Hz but this could impart a LF monotony.
Try changing the value of that capacitor between 470uF and 1,470uF.  It can control energisation around driver Fs and thus modify the depth of driver generated tone (wanted/unwanted?).
 
You will very easily be able to try different value component combinations, and thus home-in on values which suit your own driver/baffle/room combination.  To do this I would recommend you run thick cables out to your listening position and connect different values whilst listening. (Very easy!)  The one ohm in series with the choke in my preferred circuit might not be necessary when an adequately rated transformer is being used.  I think it possible the the T-bass can overcome the Alpha's higher Qes induced bass monotony, but I do not have hands-on experience of these drivers so I cannot guaranteed such a claim.

Different drivers ?
I think you would need two low Qes 15"ers to equal one Alpha-15A for LF sensitivity, but they would then be able to go lower and louder.  However it would take MUCH MORE amplifier power not only to do so, maybe as much as 4 to 8 times, but just to acheive the same SPLs your Alpha already provides.

I regard the most important characteristics for OB bass as being (in order) displacement volume, high VAS, and low Fs.  A low Qes driver will have a better dynamic response and better LF definition, but will also be less sensitive unless you find one with a powerful magnet but NOT a large voice coil, as per the Madisound EM 1550, which I would prefer to the drivers available here in Europe.
A Beta-15A will not give much improvement over the Alpha-15A you already have, and yet will require different crossover arrangements due to greater output above 100Hz.

Yes your Alpha is now making better bass, but the improvement in the way it sounds is down to better phase coherence.  As soon as EQ is applied you modify waveform phase, and when the deep LF becomes phase (time) shifted with respect to upper bass and mids then the bass can still sound weak and muddy even when the cone is dancing.  A transformer-bass circuit retains better waveform phase integrity, hence the bass can sound louder and cleaner without the voice coil being driven from the magnet gap.

Large trafo ?  Yes; two reasons;-
1. to avoid core saturation at low AF frequencies - which could clip or damage a SS amplifier.  The trafo's ac voltage could probably best equal the amplifier's peak voltage out.
2. to keep the winding wires thick with low series loss resistance.

I would recommend 250 to 300 VA for a 15".  (yours is fine for the higher sensitivity Alpha), say 500VA for 2x 15", or 1kVA for a high X-max 21".
Just check what is available to you and choose the most cost effective.

Polarity ?
Simple + to +.  No phase changes to be considered.

0V tap directly to the (-) negative terminal ?
Not sure what you are asking here, so do feel free to e-mail me if you want to double check anything.  Removal of the one ohm resistor between the amplifier and a transformer designed for 50/60Hz line frequencies could cause amplifier problems in the post#90 circuit you are presently using.

Cheers ......... Graham.

[Vix]

Hi there!

Since this is the my first post here at AudioCircle, a few words: I've been reading this forum from time to time, but I was more active at diyaudio Pass labs and Fullrange forum.

Being infected with the OB bug, I see that there's no way back. I no longer feel the need for boxed speakers; however I am aware that by going OB route, I've opened up a Pandora box and now dealing with other set of problems. Despite initial disapointment, I am looking forward to improve my Open baffle speakers. I noticed that there are a lot of helpful diy-ers here. On the other hand, I hope that my experience will be useful to someone.

I started my OB's as a No-Box speaker from Visaton, with two differences: It was going to have active bi-amplification, and I bought Eminence Beta 15 instead of Visaton bass driver. Check here:
http://www.diyaudio.com/forums/showthread.php?s=&threadid=110452&highlight=

There were a lot of changes since then, mainly having to do with crossover points and slopes. Sometime later, I removed side winds, and, despite lighter bass, I liked the sound more, so I left it like that.
At the moment, Eminence Beta 15 is fed by Sony SS amp, active filtered at 80 Hz 12 db/oct low pass (Linkwitz-Riley). Visaton B200 is fed by Pass Labs Zen V9 amp, high-passed at 80 Hz, and with passive BSC consisting of 0.68 mH air-core inductor and around 7 ohm resistor.

While I am satisfied with the setup, I am looking for ways to improve it. There are two areas that I am looking to improve: 1. Low bass. I hope to achieve better performance using Graham's T circuit. 2. Better BSC or phase plug, or whatever that can make Visaton B200 sound smoother, but not duller.

Special thanks to Graham for going far into these problems and offering viable solutions!

Best Regards,

Vix

[Polarbear]

Nice work Graham

I would like to try the circuit on my Augies. Should I alter any values in the circuit for that purpose, or should I use it as it is?

Cheers

Bjorn

[Graham Maynard]

Hi Vix,

I guess it was the B200 which sent me on the path to passive boost at LF, it is a sensitive LS and simple passive circuitry cannot make for good sounding LF augmentation with readily available LF drivers.

As you say, there is no turning back from OB, and the problem becomes one of matching drivers, which can be hard before you buy/hear them.

This circuit is in series with the driver, so it is the driver Fs which limits improvement within our AF spectrum.  The Beta has 35Hz resonance and medium Q, so with this transformer circuit will sound very good and represent excellent value for money.

Suggestion.  Run the B200 directly from the Zen amplifier without any correction circuitry.
Run the Sony amp without any in-line crossover straight into the T-bass circuit, but put your now spare 0.68uH choke between the transformer output and the Beta.  This will allow the Beta to provide the BSC whilst the B200 has full reign, with increased sensitivity too.

The component values for use within the T-bass itself may be changed/adjusted according to personal choice and requirements, and it is worth noting that the best reproduction will not coincide with the flattest frequency/SPL response. 
To test the arrangement the choke value can be anything between say 4.7 and 10mH - whatever may be to hand;  then if you see potential you can check different values.  My 6.4mH is an unusual value, but little difference would be detectable between say 6 and 7mH.

This transformer arrangement can bring life back to OB LF in a way that other arrangements cannot, and when the B200 is run almost fullrange, save for say a simple series 47uF or 100uF capacitor, it is able to retain better wide band coherence.

A Zobel of 4.7 to 10 ohms in series with say 47uF may be connected across the LF driver in order to limit output at 'breakup' frequencies where both impedance and output can rise.


Hi Bjorn,

Try the circuit 'as is' first. 

Could be a shame if you buy in chokes and then prefer to use different values, but the choke controls the response slope more than the turnover frequency, with roll-off still being down to the driver plus its OB mounting, though being less steep with a low Qes driver.

Given that Darrell selected his Augie specifications for OB LF, they too are likely to sound very good when transformer driven.

Here I suggest you start without any conventional crossovers other than a 0.5 to 1.0mH choke in series with the transformer and the Augie, and 47 to 100uF in series with the mid/widerange. 
The reason for this being that the underlapping region where the LF driver becomes responsible for BSC, is much less phase/timing antagonistic with respect to the mid/widerange than when using normal crossover and BSC arrangements.

Cheers ....... Graham.

[Polarbear]

Thanks Graham

I run my Augies active, rigth of the amp, so no XO to allow for.
But since the Augie has a FS at 27Hz, I migth increase the inductanse in the choke some?

Cheers
bj

[Graham Maynard]

Hi Bjorn,

Strange as it may seem it is possible to have too great an OB LF weighting using this circuit. 
No harm in trying a higher inductance to see how it works; your ears will tell you when the balance is right. 
Higher inductance leads to a deeper cut around 100Hz with respect to a sharper Fs peak, the resultant characteristic will then not match baffle/driver roll-off.

My conclusion has become that a T-bass driven low Fs 15" OB driver with good X.max, front mounted on say an 18" square baffle with supporting top and sides having the same rear depth as the magnet (typ. 7"), will produce more coherent bass down to 30Hz than could a similarly sized aperiodic or tuned enclosure having the same external ~1.4 cu.ft volume.  (Note I am writing about quality listening here, and not just ultimate SPL.)

Cheers ......... Graham.

[Vix]

Hi Graham,

Thank you for the advice. I will reiterate here over some things that you suggested, to make sure that I understood you well:

1. Eminence Beta 15: I have to remove low pass active crossover. Then, after the Sony SS amp, insert a T-circuit first, then 0.68 mH inductor, then a zobel (4.7-10 ohm in series with 47 uF) across Beta's terminals. What I don't get here is that, by doing this, I will have just a first order low pass filter at almost 2 Khz. If you look at Beta's frequency response, you'll see that it has a peak around 2 khZ, and it's not pleasant at all. I have tried running Beta's with only a choke (I have tried 3.3 mH choke, I could also use this one instead of 0.68 mH) but I didn't like the sound because there was still a lot of midrange being heard. However, these were done without the T circuit, and with T circuit things may change in a way that I can't imagine right now.

2. Visaton B200. At the moment it is also actively high-passed at 80 Hz (2nd order LR) + passive BSC between a Zen amp and B200. So, you suggest that I remove the active crossover, and run Visaton full range, and remove the passive BSC circuitry. Or, do you suggest that I put a say 100 uF capacitor at the output of Zen amp, creating a first order high pass at around 300 Hz?

Anyway, I am willing to experiment. Thank you for the support.

Regards,

Vix

[Graham Maynard]

Hi Vix,

Your thinking is along the lines I suggested.

For the B200, the idea is to have it running as intended - widerange - with simple protection against LF either via a single capacitor or line level filter, so do feel free to try first with your already existing active arrangement. 
When the Beta is underlapping to cover baffle size loss, plus B200 drop in output below 1kHz the B200 should not need any passive compensation.  Series passive components can take some 'life' out of reproduction, and a directly connected tube amp is likely to be excellent here.

I do not have hands-on experience of the Beta-15A peak at 2kHz.  My view is that the transformer based circuit should be tried first, and if okay then deal with the Beta peak which arises due to its combinations of source/voice coil impedance and cone shape/baffle interaction.

The best form of electrical treatment for this peak would be a notching filter. 
Solution - if you use the 0.68mH choke between the transformer and the Beta, and a 10uF + 4.7R Zobel across the Beta, then simply parallel tune the 0.68uF choke by connecting another 10uF capacitor across its terminals.  The notch may then be deepened by increasing the value of the Zobel capacitor, up to 100uF without causing noticeable LF degradation, and this represents a very simple way of reducing upper mid cone breakup with a large augmentation driver.

Thank you for making me think about this - it could help with so many other drivers whilst having minimal effect upon low mid frequency reproduction.  I will even try this on my own shortly because it might further improve roll-off.

Thanks Vix.
This is one really positive outcome from the Forum based discussion here.  Something which might not have otherwise arisen because the Beyma 15s I am using here do not appear to exhibit this behaviour, so I had not thought about adding so simple an improvement.

The simplest form of mechanical treatment would be to reverse the connections and mount the Beta on the front of the baffle pointing rearwards into soft damping material.  This would not result in any further relative electrical phase change of waveforms going to the drivers.

The circuit needs to be tried first, and then fine tuned afterwards if deemed worthwhile.

Actually, I think the B200/Beta-15/T-bass combo has excellent potential, especially when driven by separate amplifiers.

Cheers ........ Graham.

[Vix]

Hi Graham,

Thank you for a reply. After reading some other threads, and especially the one about Alpha 15 and Visaton B200, I tried to just connect two drivers in parallel, directly, without any inductor, anything. Running the Beta 15 and B200 fullrange, the midbass and lower mids get some weight that is otherwise missing. However, two problems appear. 1. Midrange peak of the Beta 15 2. Low bass below about 100 Hz quickly "disappears" (both because of the baffle loss and falling sensitivity of Beta 15, + lowish Q compared to Alpha 15)

The first problem might be solved by a notch that you suggested, and 2. Adding a T circuit may bring the low bass back.
In the worst case, if none of these works, I will go back to my current setting 80 Hz 2nd order active both HP and LP.

I don't have time to try more things today, and from tomorrow I'll be on a vacation for around 10 days so most probably I won't be reading the forum. It will be a good time for thinking however, and a new idea may show up   

When I come back, hopefully with more fresh energy, I'll build a T circuit first, and inform you about the results.

Best Regards,

Vix

[Graham Maynard]

Enjoy the break Vix - will give you some thinking time.

My motto   "Never assume; just think slow and act fast"

Cheers ...... Graham.

[painkiller]

The first point I noted was your typed capacitor value. 
Here in Europe 1mF = 1,000uF.  1uF won't actually do anything in this circuit, but won't prevent it from working either.  If a typo then just correct it so that I can remove this observation.

Oh shit.  It wasn't a typo. I just read nF instead of uF.

This version is my preferred circuit, though the post#90 entry might better lift single Alpha output to match a widerange driver's sensitivity.

Thanks for the tip. I tried this one yesterday. 1 ohm resistor 660uF cap, 1 ohm resistor, 10mH inductor. This one gives much better tonality in the bass. Much tighter bass also. It's really impressive.

I'll try it on my other speaker also, and see how much bass I can get. What would happen if I parallel two Alphas with this circuit? Will there be problems with low impedances?

If I raise the capacitor value. What should happen in theory?

If I use a smaller inductor. What will the effect be?

Thanks for sharing this great circuit by the way!

[Graham Maynard]

Hi painkiller,

Looks like you are well on the way to never using boxes for bass ever again.

Two Alphas are going to start shaking your room furniture with some CD tracks.

When you parallel the drivers then take out the resistor in series with the T-bass choke.  No need to change the choke and capacitor values - these set the slope independently of driver impedance, though obviously you should fine tune the values afterwards to optimise your own setup.

Your 10mH value would be a little high for most lower Qes Pro drivers, however the higher Q Alpha will actually benefit from the slight extra cut around 65Hz which your present values are providing.  This might help correct the dynamic response, and provide the much 'tighter' response you already mention, though if you have two 10mH chokes then do try parallelling them for one channel when trying try two Alphas together, probably make for more but less deep bass.

Whether you can parallel up and run loudly is going to depend upon your amplifier.  Your existing transformer will handle the current, but few amplifiers are designed for low ohm loading and might be protected against outputting what this circuit will attempt to draw.  An amplifier capable of driving 2 ohm loudspeakers in an enclosure should cope with transformer driven 8 ohm drivers in parallel on an OB.

At gentle listening levels there should not even be any probs with a 4 ohm capable amplifier, but care must be taken regarding peak currents and heat dissipation if you start enjoying the bass.

Raising the capacitor value increases the cut in output where this circuit reduces unwanted output peaking due to driver Q and corner baffle placement, say around 65 to 150Hz.  If you increase the value too far you start reducing the bass boost frequency peak as well.

Increasing the inductor value lowers the frequency at which both boost and cut arise.  *However*, the driver is in series with the circuit, so the boost starts to fall again if the inductor value is increased beyond optimum, and the cut comes in too early at say 50Hz where bass output is still wanted.

I am very much hoping that you will be repeating those first two words in your last post - only this time when you are standing in front of two 15" drivers in parallel and you hear what they can do when driven by the T-bass circuit.

Folks everywhere are SPL simulating this and that crossovetr, baffle or enclosure, and aiming for deeply extended LF, but they will not be enjoying the tight, hard hitting, deep and colour free bass we here can with simple baffles and the T-bass circuit.

If a driver relies upon 'Q' or an enclosure to delay/change the phase of/re-use LF output, it might gain amplitude on continuous notes, but it still WON'T do drums or plucked bass right.  Sealed cabinets will, but then you end up with crossover phasing and timing difficulties as each cabinet has only a limited bandwidth at LF.

A couple of 15"ers on a minimum depth baffle and T-bass drive should cleanly get down to 25Hz at satifactory level.

Cheers ...... Graham.

PS.  My own experimental OB 'Spirit' project will be going into Mk2 with Di Apollito 10"-8"-10" above 2x 15", but even as it is, my wife is repeatedly asking "can I hear this" through it - to see what her CDs can sound like !  Same with my nephews and their mobile phones, they want to plug in and hear their music downloads through it.

[painkiller]

I see. My amp is "suitable for loads down to 3.5 ohms". And delivers "only" 90W (RMS continous sinewave) into 4 ohms. I've never been able to run them warm. 

Is series coupling an alternative for an easier load, or does that create unwanted interactions between drivers? I read something about small variations between drivers creating some sort of distortion. 

Could I parallell two different drivers, or is that just hopeless? The Eminence Beta is readily available here in Norway, and it's actually cheaper than the Alpha which is difficult to get hold of. The Betas have Qts=0.58, 98dB SPL, Fs=35Hz.


I'm definitely beginning to see the potential of high sensitivity OB bass without the need for lots of drivers. This is really exciting. I believe the T-bass circuit will enable one to successfully create a 100dB efficient passive OB speaker in a wife friendly package. 

I've ordered an extra 300VA 15V trafo, so in a few days I'll have a stereo T-bass.

[Graham Maynard]

Hi painkiller,

I suggest - try them in parallel, run at normal levels only and watch the amplifier for heating.
Transistors can run quite happliy beyond temperatures where you cannot hold your hand constantly on a heatsink, though this is where I would start being careful, and by then AF levels should be pretty loud.

It is the parallel driving of loudspeakers which controls the individually generated back-EMFs due to them being differently positioned on a baffle and with respect to the room loading and reflections.  Series connection allows the back-EMF from one driver to modify the input to another when the drivers are not identically loaded and positioned.

Each Alpha is going to be overdriven by about 40Wrms of LF when on a open baffle anyway, so it is just the amplifier's current driving capabilities which need to be watched.
Betas will probably need twice the power compared to Alphas, so running in parallel would lead to the Alphas over-excursing before the Betas are any way near producing useful output.

The Beta looks a good candidate, but you would definitely need to think about different amplifiers and transformers too;  The lower the Qes - the more power needed for LF output;  so its either parallelling up with the Alphas or a completely fresh start.
Betas/Deltas will work fine but as an 'upgrade' will not give much extra in the way of X.max, so the KappaLFs would then better for OB, with extra power requirements of course.

I'm not well up on the theoretical figures for how low 100dB or even (110dB) could be acheived, but dual 15s can easily be satisfactory for home use.

Cheers ...... Graham.

[painkiller]

Thanks for the reply. I guess I'll be trying to parallel an extra set of Alphas then. I'm quite happy with their sound, so I'm not looking for an upgrade, only matching SPL. I'm confident an extra pair of Alphas will do the trick.  aa

[painkiller]

With the T-bass circuit in stereo I get a bit better bass extension. With no EQ the sound is still very lean, but there's at least some deep bass. There's more mid bass than deep bass though. I still need around 8dB extra in the low end. Adding a couple more Alphas may introduce too much midbass and the need for additional filtering.