What if.... Let's try the current amplification

I was an happy camper with my horns and whistles.
Then I looked at my amplifier: one Quad 303, perfect for the job, a laid back sound that my 50 years old ears appreciate alot.
But I always felt like I need more "control" and "speed" in the lower end.
Money is an issue, amps are heavy money. I thought about some cheap pro amp, so I hooked up one Carver M1.5t I own from the beginning of this passion, to see if it was possible to drive a 105 dB speaker with one 350 W amp.
In my case, it is not.
The volume control knob is working too close to the beginning, where some non linearity is always possible.
Then, surfing around, I bounced on this Pass and Creazzo  papers about current amplification, and so I decided to try this possibility.
First thing to do, make the impedance response of my speakers flat.



I added to the crossover (red trace) an RCL serie in parallel to the midhigh input, as described in another post. This is already a good result if one plans to use this speaker with a tube amplifier (the blu curve). But it's not flat enough for a current drive, so I added an R in parallel to the woofer input (black curve). This is ok for a current drive, a bit too low for conventional amplification!






So what happens to the frequency response?
In red you see the system's frequency response when conventionally amplified with an equalized Quad 303.
In black you see the very same system, but with a Carver amplifier in current thanks to an added R to the output of the amp of 54 Ohms.So both conditions refer to the red impedance curve.

 1. Red - Quad303, Black - Carver M1.5t + R54Ohms

What surprises, even if expected, is to verify that the spl between 100 and 500 Hz of a 28W RMS amp like the Quad is the same of a 350W RMS when driven in current.

What happen if we add the RCL cell like in the blue trace of the impedance plot?
Here it is:

2. Red - Carver + 33R, Black - Carver + 33R + RCL

After adding the 5Ohms R in parallel to the input, like in the black trace of the impedance plot, the impedance went to a flat 3 Ohms +/- 1 Ohm, linear enough for the current amplification. This allowed to use a 33R instead of 54R, respecting the common sense of a 10 times greater R on the output of the amp respect to the impedance of the speaker.
The result is in the following RTA, where in red we have the conventionally driven Quad303, equalized through the AEQ function of my Behringer DEQ2496 (same as the red trace in graph nr.1), while in black you can see the Carver, similarly equalized and driven in current through the 33 Ohms resistor and the flattened impedance curve. Levels are arbitraries.

Red - Quad, Black - Carver

I don't want to describe my sensations, I just want to say that in this way the sound is much cleaner, with a prompt and steady bass, refined highs...Ok, just read the curves, they are saying the same! So, I don't really know how much this is an advantage of the current drive respect to the system's modification, but I could reasonably say that I feel free of some muddiness in the lower end, that I would like to associate to the back EMF issue (or its lack) but I am not even sure if this system (which is not, we must remember, a from-birth current drive system) is solving this matter properly... Anyway, when the ears tell you "now it's better than before", we can go to sleep thinking "mission accomplished", and that's what I'm going to do right now.
No, not really, there is one, big, issue in this work: the SPL is much lower than before, too low for a real slobber. I need more juice, because right now the volume knob of the preamp is going to almost the end, and the Carver's LEDs tell that 350W are thrown in the 105 dB speakers, and I need more: the level with the Quad was higher.
So, what to do? Difficult to come back, once tried. A 700 W amp instead of the Carver? It's only a 3 dB difference, should this be enough? Again I'm thinking to bi-amping...let's see what the future will bring.
Suggestions are welcome, as usual!

Aggiornamento

Some little frills, after all it's springtime.
I tried soft sanded glass covers, instead of plan plywood.
Nicer, both on the bottom (first pic) or on the top of the mid horn.
When on top, it allows to place the tweeter on the same line of the mid mouth. 
This affects a lot the phase emission and coherence between mid and tw, considering we have a passive filter not compensating for this. I have tried to listen carefully to the difference between the tweeter on axis from mouth or throat point of view, but with no final results.
Next step will be a 2 way amplification, one for the woofer with adequate delay (anticipation in this case) and one for the mid-high, with the tweeter again in its pristine position, like in the first picture.
Alcune piccole modifiche, sia dal punto di vista estetico che di fine tuning.
Cominciamo con il look:

Ho aggiunto un piano in vetro acidato, "per vedere di nascosto l'effetto che fa". (Grande Jannacci, R.I.P.).

Al solito per mantenere il trend cialtrone la foto è una pessima immagine presa col telefonino.

In un primo momento ho posizionato il vetro inferiormente, come piano d'appoggio per la tromba:

Poi ho provato invece a posizionare il vetro al di sopra, in modo da creare un piano sul quale fosse possibile avanzare il tweeter a filo con il medio, considerando le rispettive bocche:

A me piace di più così.
Certo si crea un problema nient'affatto secondario, ovvero, dato che il sistema è filtrato tutto passivamente, al ritardo di emissione del woofer sul medio-alto si aggiunge anche il ritardo del tw sul medio.
La soluzione che adotterò sarà una biamplificazione tra woofer e medioalto, con regolazione del ritardo e taglio passivo con arretramento del tw in modo che l'emissione alla gola del driver del tweeter e del cono del mid sia sullo stesso piano. Dovrò valutare se e quanto la superfice piana del vetro influisce sull'emissione.
Comunque, nell'attesa di verificare quanto questi sfasamenti di emissione possano nuocere all'ascolto (vi confesso che le prove fatte sulle due posizioni del tw rispetto al medio non ha dato esito) vi pubblico il grafico della risposta attuale con l'aggiunta dei sub woofer (curva verde).
Il basso è prominente, come piace a me, mentre la curva cala secondo Moeller, anche se dal grafico della risposta L+R in mono sembrerebbe molto di più, ma è la rilevazione che è fasulla in alto, basta vedere la risposta del singolo canale (in giallo nel primo grafico) rispetto alla somma dei due più sub (ciano e RTA)
Dovreste vedere le facce di coloro che vengono a sentire il risultato finale...
And here again the actual response of the system, with two subs in mono and equalisation from the DEQ2496 only above 125 Hz, as you can see from red and yellow curves on the first graph. This kind of measure is not trustable above a certain frequency, I would say around 5kHz, because it is made with the two speakers in mono feeded with the same signal. So, the cyan curve as well as the RTA ones are good only to "roughometrically" evaluate the integration between subs and system.

Rosso L no eq, Giallo L eq, Ciano L+R+sub

Rosso - L+R , Verde - L+R+sub