[Graham Maynard]

(Moved from my B200 mod thread.)

This 'T'-bass is a 'clip-in' supplementary circuit to go between a SS amplifier and the series choke or crossover circuit feeding a low frequency driver (not between a choke/crossover and the driver) to improve sound reproduction from a baffle or a 'U'-frame which is already being used to augment the low frequency output from a wide-range driver.  It may also be used (carefully) with full aperiodic and properly damped line/reflex LF designs, and is called "'T'-bass" because it is the the Transformer which provides for increases in both low bass output and reproduction quality below a driver/system's resonant frequency. 

I expect it to particularly improve LF reproduction when driving the likes of an Alpha 15 on a plain baffle with another FR driver, though it does need a 4 ohm (SS) amplifier to drive an 8 ohm LS, and, whilst it reconstitutes some LF otherwise lost due to phase change and roll-off at frequencies below LF driver resonance, it cannot help with regard to driver excursion limitations.
See;-

http://server6.theimagehosting.com/image.php?img=T-bass.PNG

                     
At low bass frequencies the inductor has low impedance - this makes the transformer step-up amplifier output (15-25Hz).
At high bass frequencies the capacitor behaves like a short circuit - this makes the transformer pass amplifier output without boost (>65Hz).
Capacitor and inductor values can be chosen and adjusted to match a specific driver's resonance characteristics, though this is not highly critical.
The resistor value may be adjusted to set the level of low frequency boost;  from zero to generate maximum sub-resonant drive (punch), to 2.2 ohm for subtle (non-tiring) improvement.

To appreciate the difference this circuit makes it is simply a matter of swapping over the LF loudspeaker circuit feed from the amplifier output to the transformer output. The boost peaks <25Hz, so a large transformer core is wise though not absolutely necessary, and yet it need not have a hi-fi specification either because it is shunted above 100Hz.  Of course there will be wire and core losses which introduce a slight overall loss, but even this can be useful as many bass drivers are too 'efficient' above their resonance anyway.  Transformer voltage windings, core size and VA rating should be generous when compared to those used with PSUs, though any specially constructed component then needs be no more than an easy to make bifilar wound choke with the windings connected in series.

When powered by an non-EQ'ed amplifier the reproduction is as if full use of rear of driver cone output has been made, only without the massive cabinet volumes normally required for porting/venting or inducing a long horn/transmission line delay.  Thus the very low frequency output from a baffle can be improved without there being additional rear of driver pressure reactions impacting upon forwards reproduction, and the slight loss in output >80Hz can advantageously remove some of the upper-bass peak which is beyond passive damping control and can so often contribute towards tiring reproduction unless designed out.  I have tried this circuit driving 8" Peerless bass drivers in small U-frames and on baffles, and as a result I will never go back to direct wiring again.  The sound is as if a second driver has been added to come in below resonance only, without there actually being another - unless higher low frequency SPLs are wanted.

SS amplifier damping at resonance normally controls and limits a bass driver's response, whereas this little circuit not only increases lower bass output, but appears to correct reproduction phase change due to driver characteristics around the mechanically induced resonance.  Thus the 'T'-bass circuit really does improve reproduction.  We know that the output terminal of a SS NFB amplifier can better control and damp a driver plus any air-spring energy storage induced reactivities, but when this circuit is added with suitably matched 'L' and 'C' values and there are *no* air-springs modifying transduction, then baffle reproduction characteristics become more like those enjoyed with current driven (non-NFB tubes) as opposed to voltage drive, and a single SS amplifier per channel might then become more of a realistic option.

Of course many listeners prefer to use tube amplifiers from the outset, which means I must honestly state up-front, that this circuit will not work as well when driven by a non feedback tube amplifier - unless that amplifier is fitted with large tubes, has sufficient power rating and a 4 ohm driving capability.

Anyone adventurous enough to give this a try ? 

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

[mcgsxr]

I will be the first to admit that I find this challenging to follow, as my science is likely not up to the task.

Can I clarify one part - this is to be used with a full range input, not the input following an Xover?

I use an active Xover (to send 150Hz and up to my SE EL84 amp, and 150Hz down to my sub amp), so I would not look to implement this?

Sorry for the noob question, I am just trying to contextualize.

Thanks for sharing!

[Graham Maynard]

Hi mgcsxr,

You would tag this circuit onto the output of your sub amp, and then your LF driver on to the circuit output.

If one amp was being used the circuit would go between the amplifier output and the LF driver choke/crossover components only, whilst the FR circuit would still go straight to the amplifier.

The secondary for my test transformer was 2 amp rated.

Cheers ....... Graham.

[Graham Maynard]

Other questions asked of me;-

Are folks free to use this ?  Anyone:  Including commercial.

Have I actually tried it (by someone who thought maybe I had derived this via simulation) ? 
This design was purely empirical and not written down until finished.
 
If you could see all of the components etc. scattered on my lounge floor at the moment - you would realise this is not a software design.
(Just need to make sure I get them all tidied before wifey gets back from her short break in warmer climes.)

I have however simulated the circuit since making it, and this shows the sort of amplitude boost which can be achieved by re-using rear output via a long (12ft) line, only without all of the other parasitic peaks, and without the audible path delay between the first quarter cycle and the boost being perceived !!!

If I were to bring this a product to one of the loudspeaker meets the first to hear would want to buy !

Yesterday I tried it with an already equalised B200 running solo, and it so lifted the low bass that it became a delight to listen to.  Of course the B200 could not then be driven as loudly, but it was so much nicer at lower levels the wish did not arise !

No one got junk box parts lying about to give this a try ?

Cheers ........ Graham.

[Graham Maynard]

Another aspect I should like to raise here relates to all of the SPL/frequency plots we are presented with relating to bass reproduction.

Steady sine plots arise *after* all of the initial dynamic driver and 'air-spring' energy storage mechanisms have been energised, self stabilised and re-combined.

FACT:- 
A driver on a baffle generates a much more substantial low-frequency leading edge wavefront (the peak amplitude at 90 degrees of the cycle) compared to when the same driver is enclosed within a box, due to the work the back of a cone is obliged to do when compressing air within a cabinet during a first half cycle.

The boxed driver then goes on to output with greater amplitude during the 180 to 360 of a first cycle than it did during 0 to 180 degrees as the initially stored energy (air-spring) then re-inforces on-going drive, especially at and around resonance.

Thus large drum reproduction from an initial hit via a boxed, resonant, vented, cardioid (single drive), TL arrangement etc., becomes modified as front/stored rear/external energies become summed in music time, with reproduction often sounding invertedly thumpy due to the second half cycle having larger amplitude than the first (!!!), whereas often the leading low frequency half cycle should have the greater amplitude.

Thus a driver on an open baffle is more likely to reproduce low frequencies more accurately and thus less tiringly, even if it does not pressurise a room to the same degree as a monopole, and it has a slightly inferior LF amplitude response.

High power, heavy coned LF drivers can behave similarly on a baffle due to the electrical energy being consumed/stored more through generating/controlling mechanical motion than in linearly moving air.

Thus frequency response and SPL plots only tell us what the measured responses are like after all mechanical and waveform effects have stabilised and re-combined, not how a drum waveform will be reproduced, nor what all of the associated amplitude modulated components will sound like as they too recombine differently in music time.
_______________________________________ _____________________________

This 'T'-bass circuit cannot compensate for high Q driver resonances, but when used with a plain baffled driver the first half cycle is boosted from outset due to the inductor being like a short-circuit at 20Hz, and thus an open baffled driver can sound more natural as the boost compensates for compliance induced roll-off.

It is also possible to fit a damper resistor of up to 2.2 ohms in series with the capacitor, but I felt that the LF advantage of the circuit, especially when driving a higher Q LF driver with optimally matched 'C' and 'L' values, would be equally useful without it.


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

[-Richard-]

Hi Graham ~

It is important to remember that there are cancellation issues with OB's... the back-wave comes around the baffle and cancels some of the front wave information... that should be "factored" into our all-over thinking.

Also... I have found that too much tampering with additional add-on-circuits to "tame" driver behavior actually detracts from the lovely 3-dimensional, surreal quality that is there with the use of very simple circuits. Circuit add-ons kill and flatten the sense that what we are listening to sounds like "real."

"Perfection" of this or that holds no interest for me... human and real does.

Warmest Regards ~ Richard

[Graham Maynard]

Hi Richard,

Thanks for your post.

All I can say is that if anyone would care to try this, they might enjoy their music even more than they do presently, for this circuit releases that last bit of low bass below driver resonance lost due to spider and surround stiffness unavoidably damping movement and modifying the LF phase response.

Some say you cannot hear phase change at so low a frequency;  well maybe those people have not experienced the fundamental appreciation that develops when the loss of coherence is demonstrated by it being overcome.

This circuit is not about perfection - loudspeakers don't do perfection at all well - it is about enjoying the music and revealing the ambience captured within recordings - by overcoming that 'round the baffle' cancellation you mention.  In other words it flattens and extends the LF response of a bass driver beyond its normal cut-off, with a very worthwhile improvement in reproduction.

The first baffle listener to try this circuit will be back here remarking at the positive improvement in LF reproduction.  I've tried many things with LS for over 40yrs, and wonder why such a simple mod as this has not come up before.


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

[Badwater]

Graham, please pardon my questions,  I am a complete noob on this.....

Can you suggest a transformer?  What does o/c mean? If you  use a ST70 would you drive the speakers using the 4 ohm speaker taps?

Thanks for your time.

Bill

[Graham Maynard]

Hi Bill,

Questions welcomed.

I have many transformers here.  I tried this with a 2x 12V @50VA rating and this is good enough up to 20W listening.  These being inexpensive general purpose low voltage PSU transformers.

ST70 will be okay on the 4 ohm tap, and although it will not have the same maximum driving power as a SS at the low frequencies, it does have decent LF response.

o/c means open-circuit.  In other words this was a mains (line) transformer with the primary windings left open circuit.

Do you have any transformers about the place which might be pressed into service, either twin winding PSU types, or say a spare amplifier output transformer with 4 and 16 ohm taps, or LS line or matching transformers ?  The primaries are not used.

Cheers ........ Graham.

[-Richard-]

Hi Graham ~

Thanks for the further clarification. Much appreciated. Please keep us informed as you discover more "solutions" to any OB problems.

Warmest Regards ~ Richard

[Badwater]

Hi Graham,

Thanks for the clarifications. The info helps a lot.  All I need to do now is find some transformers and have a go at it.   i am very intrigued by your suggestions and wish to give it a try when I update my EVSP12B OBs.

Thanks,

Bill

[D OB G]

Hello Graham,

Looking at your T-bass circuit, if I am reading it correctly, at low freqs, where rL=0.5R, is the amp seeing the 0.5R, plus the 1R, plus the resistance of the transformer between the 0V and 12V taps (might be, say, 0.5R), in parallel with the speaker (plus the series resistance of the 12V to 24V tap)?
i.e. is the amp seeing something less than 2 ohms, even with a nominally 8 ohm speaker, very typically 6 ohms, ending up with, maybe, 1.5 ohms.

With trepidation,

David

[Graham Maynard]

Hi David,

Yes you would be correct at say 20Hz if the inductor and transformer were perfect and a series resistor was not being used.

Those who like low bass or want HT like sound and have a 2 ohm capable amplifier might well enjoy a further increased LF response with very low resistance transformer and choke components fitted and without the series resistor, however for normal audio a 4 ohm amplifier is fine. 

Remember that the transformer winding resistances are in series with both input, output and the common (input-output) capacitor/inductor node, and that this is not a pure ratio-resistance situation.

So, with most bass being in the 30 to 100 Hz region where bass drivers already have a much higher impedance, and the circuit already introduces a slight loss where natural driver efficiency (resonant region) already introduces a slightly un-natural sounding reperoduction modification, a 4 ohm output really is adequate for real world drive.

Thanks for bringing this up. 
I had already thought about it, but not mentioned.

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

[Graham Maynard]

Nobody tried this yet ?    It is so simple !

Writing an explanation here in Mark's thread today-
http://www.audiocircle.com/index.php?topic=53649.0

made me realise what it is about this circuit which makes dipole bass sound better, and especially so when lower 'Q' loudspeaker drivers not specifically designed for dipole use are used on an OB.

Originally I established by ear that this circuit made a worthwhile improvement to dipole listening.  Then before I started this thread I simulated circuit operation with an equivalent loudspeaker circuit and saw that the waveforms would indeed be modified in exactly the same way as I had heard in real life. 

Thus I saw that a competent 4 ohm amplifier would indeed be adequate - the main requirement being that the transformer needs to have a 2 or 3 times higher than the 50/60Hz VA rating when called upon to drive low audio frequencies.

Now I have looked at the impedance of the transformer output which drives the LF loudspeaker, and as I noted when empirically changing the capacitor, choke and resistor values, the driving impedance also becomes modified.

This circuit not only boosts the very low AF frequencies but raises series impedance at the capacitor/choke crossover frequency, this providing a slight peak in the drive impedance which is similar in impedance to that of a triode amp without NFB, though only at the low frequencies bounded by the value of capacitor and choke.

Thus the 'T'-bass circuit not only provides boost, it increases the source impedance seen by the driver around resonance where the output voltage from the circuit is actually slightly reduced, which allows a low 'Q' driver to transduce more naturally and more efficiently without enforced damping, with the degree of boost and the damping being adjustable by changing series resistances.

The quality of reproduction is much improved compared to what can be achieved using active or passive crossovers or any amount of amplification or an increased cone area, this because the impedance change coincides with driver resonance where otherwise reproduction would sharply roll into a final very sharp cut-off with un-natural sounding phase change.

I would be sad if no-one else tried this, but alas I can do no more.

Cheers ....... Graham.

[D OB G]

Hi Graham,

I did indeed try the T-bass circuit, and it seems to work as you have said.
I'm using DEQX (not to be confused with the Behringer products) to tri-amplify.
By introducing the sub 20 Hz hump, it is much easier for the convolution filter to be derived, and for room equalization to be applied.
I played around with different C and L values, but I also tried a "T" resistor in place of the transformer, on the basis that I couldn't see how a sencondary winding with no magnetic coupling to the primary (please correct me if I am wrong) wasn't acting as a resistor divider (or is that a voltage divider?).
Anyway, with your circuit as described, I did indeed have the amplifier see 1.5 ohms (high driver impedance at low frequncies is irrelavant), so I wasn't prepared to run that configuaration for very long.
With the resistor divider in place, and 4.7R in place of your 1R, a sub 20 Hz bump is also produced, but at a safe amplifier load.
I'm probably missing something with your use of the transformer, but at least I've tried it!!

David

[Graham Maynard]

Hi David,

Good to hear of a test result.  Thank you for trying.

You appear to say there is an improvement, but you still have concerns about the amplifier loading.

(I too have tried EQ but it does not produce the same 'liquid' sound due to the way a directly connected amplifer imposes LF phase change upon a driver as it peaks the damping currents - especially after EQ.  I suppose this is swings and roundabouts, and of course an EQ stage can do more at other frequencies, but this simple circuit is intended to work without any need of having those additional EQ and amplifier chassis!)

First off, the transformer is the two windings ( 0 to 12, and 12 to 24 ) connected in series ( or 0 to 24 with 12 centre tap).  The other (third) winding should be left untouched.  Placing anything across the other winding will merely load the amplifier more.

I am curious as to how you measured the AC impedance as 1.5 ohm, or was this the plain DC resistance through the transformer winding and choke + R in series without music playing ? 

1.5 ohm of DC resistance is already in series with the amplifier load due to the choke and series resistance alone, so the AC (music) load simply *cannot* be lower than this, even at the single frequency where the capacitor and choke tune (circa 40Hz to counter driver resonance) and even with the most perfect of components and LS.  Real world transformer and loudspeaker circuit losses are then added into the equation, these ensuring that the dynamic impedances driven by the amplifier are always higher !

The DC measured resistance of the transformer, choke and R cannot load the amplifier as long as there is sufficient inductance in the transformer and choke, and this is why we can connect a transformer across the output terminals of an audio amplifier even though one might have a DC measured resistance of only 0.1 ohm.

Also given that music duty is not the same as constant VA sine loading, and 4 ohm loudspeakers already present dynamic dips of 2 ohm (and lower) to a 4 ohm amplifier,  it might be worth trying a little more.  Most 4 ohm audio amplifiers are constructed with dynamic music induced impedance dips in mind, and they already sail right through these without anyone knowing, though they are not intended to run constantly (100% sine duty) into such a low impedance.

Actually if an amplifier current clips on impedance dips then I would not consider it to be worthy of use in an AF system anyway.

(My own 4 ohm amplifier will even drive music waveforms into less than 1 ohm loads without it or me noticing, though it would get hot or blow fuses if I attempted to drive steady sine through the same resistance at constant maximum amplitude.)

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

[D OB G]

Hi Graham,

Last things first.

The amp is a Naim NAP250 (trannie), which is rated to run at 4ohms, and in the past I have run three nominally 8ohm speakers in parallel off it (2 ohms at 1kHz).

Yes, I measured the DC resistance through the winding (0-12V) plus choke plus R in series.

I'm not sure how this is in series with the amplifier load (the speaker?).

I would have thought it is in parallel.

If the DC resistance is 1.5 ohms, what would you expect the AC reactance to be in this instance, when the inductor is essentially all pass at the very low freqs we're talking about?

The reason I stopped the experiment was that after a while, the woofer started "clicking".  No visible cone motion.  The amp was not hot at all.  I presumed that the amp may have been becoming unstable?

When I substituted with the resistor divider the clicking stopped.

Its hard to say what the audible improvement was, since I doubt that the material I was using had anything below 20Hz.

What I did get though was a measured hump, through the measurement equipment provided as part of DEQX.

It seems that my problem is a fundamental lack of understanding of how transformers work!

David

[Graham Maynard]

Hi David,

The AC load as seen by the amplifier is the transformed loudspeaker load in parallel with the capacitor, all this in series with the choke plus resistor.

Your Naim should be able to drive a 2 ohm resistive load at full power but this does not guarantee will drive reactive loads !  It is an 8 ohm design, and clicking suggests current protection sensing operating at some part of a voltage cycle. 
(My amp does not limit trip at any angle and carries on regardless.)

Inductors have a specific impedance at any frequency.  If the winding also has resistance then electrically this is in series with that AC impedance.  This is why we need to feed woofers via thick wired inductors - to minimise resistance/power losses.
Choke winding resistance is in series with the magnetic field, not parallel to it, and thus any ratio transformed loudspeaker load is also in series first with the individual transformer windings, and then in this in series with the choke plus R resistances.  So the alternating load as seen by the amplifier cannot be less than the DC resistance, with a lowest impedance at the series tuning AC impedance between the capacitor and choke.

These capacitor and choke values need to be adjusted to match a driver, and an additional 1 to 2.2 ohm resistor can be connected in series with the capacitor can reduce current phase angle variation without decreasing the LF boost, though my amp did not need this - as mentioned early on.

If you were already applying an EQed signal then the circuit will not be operating as intended, because the most significant improvement does not arise at 20Hz but about driver resonance circa 40Hz.  So I wonder if you were EQing before driving the circuit, in which case the series capacitor and choke would be storing much more energy than intended, and a resistor in series with the capacitor would become essential, maybe even up to 4.7 ohm in value.

The T-bass circuit is intended to make non-EQed systems behave similarly to those which are EQed though without needing the extra EQ chassis and separate amplifier.

Cheers ........ Graham.

[D OB G]

Thanks Graham,

Yes...Naim and reactive loads.

They will drive capacitive loads OK (one of the few amps that Quad recommends for ESL57s - which I used to have!)

But there is no inductor in the output section, and the amp relies on the inductance of the speaker wire for stability- therefore Naim's recommendation to use their own, known, wire.

However, in this scenario, I suppose we are looking at inductive loads? (plus I use Naim speaker wire).
So is there a reactive load problem? (and therefore, why the circuit protection activation, which has never occurred before?)

The driver resonance is 23 Hz, and I was determining the convolved filter before applying the T-bass, and then EQing afterwards.

My particular interest in the circuit is that the EQ works by applying sum and difference signals between room resonant nodes, i.e. room equalization, as well as any convolved "linear" EQ i.e. 6 dB per octave compensation, and allowance for low Q (0.23).  This means that if I can establish a room node below 23 Hz, I have a lower reference for the accuracy of the EQ.
(This is all mitigated of course by JohnK's demonstration that dipole bass can't be sustained below a room's lowest fundamental, which in my case seems to be somewhere in the 20s).

As you said in your first post, there may be a need for me to breadboard the loungeroom floor to establish the optimum values.

David

[Graham Maynard]

LOL

The  ?joys?  of audio !

You could try 1 to 4.7 ohms in series with the capacitor.

And possibly 22 ohms across the amplifier output terminals in case the current is becoming too 'unreal'.

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

PS.  Seems like you have the kit for this -
http://www.rememberthatnight.com/
http://www.davidgilmour.com/dvd.htm