Modify amp from voltage to current feedback for low Qts drivers

Hi,

I have built TQWT enclosures for fostex fe 166en (low Qts) working with a sansui AU-217 amplifier. The problem is that this combination is overdamped, bass is low and over-controlled.
So I was thinking changing the voltage feedback loop of my amplifier to a current feedback loop by modifying the power output stage.
I plan to take feedback from resistor (about 8 ohm) placed in series with the loudspeaker and take feedback from there.
Has somebody already tested such a mod?

before :
[IMGDEAD]https://lscrrg.dm1.livefilestore.com/y1ppFtQfX4pAwQTyvqU56cw0CLwaTuDqDagjuUM20PTtpJK4SbqWiaopsikdicjbb-PDsmS0XonD3_WT0k_w9sT8TaZbzVl2z_c/AU_217_before.JPG?psid=1[/IMGDEAD]

after :
[IMGDEAD]https://lscrrg.dm1.livefilestore.com/y1p9ia3FHucZh-IFv0EdJFSEMr5LZ-31G_IjC7egLabszkkChv_AtMZBkF_Epee13xe81qLSgd0bpTsn4mtyNrK-5fx0-1q8KWx/AU217_after.JPG?psid=1[/IMGDEAD]
 
Last edited:
I have not tried to do so with this topology (ltp input) but with a much simpler small class A. It worked well for me. You will also have to make a provision for managing dc offset since the servo action of the ltp will be greatly reduced, such as making one of the ltp load resistors adjustable. This is true current feedback, not sure why feeding output voltage to a low impedance node is widely accepted as current feedback when its still just voltage feedback. Choose your resistor based on speaker impedance and desired gain. For instance, 8 ohm speaker with 1 ohm resistor t get gain of 8, .5 ohm for 16 etc. Using this type feedback makes the amp operate as a constant power output, increasing and decreasing voltage output relative to speaker impedance at a given frequency. I would also add a resistor from the output to your new feedback resistor so the amp doesnt go haywire without a load, say 100 ohms.
 
Last edited:
Hi Steve,

Many thanks for the link. I think I will try to play with the "Mixed Mode Feedback" to obtain a better loudspeaker/amp match.
Like this "By adjusting the impedance of an amplifier, the total Q (Qts) of a loudspeaker can also be altered, so driver behavior in a given sized box can be changed"
:)

Nicolas
 
I have not tried to do so with this topology (ltp input) but with a much simpler small class A. It worked well for me. You will also have to make a provision for managing dc offset since the servo action of the ltp will be greatly reduced, such as making one of the ltp load resistors adjustable. This is true current feedback, not sure why feeding output voltage to a low impedance node is widely accepted as current feedback when its still just voltage feedback. Choose your resistor based on speaker impedance and desired gain. For instance, 8 ohm speaker with 1 ohm resistor t get gain of 8, .5 ohm for 16 etc. Using this type feedback makes the amp operate as a constant power output, increasing and decreasing voltage output relative to speaker impedance at a given frequency. I would also add a resistor from the output to your new feedback resistor so the amp doesnt go haywire without a load, say 100 ohms.

The OP is using the original term for current feedback which is feeding back a signal depending on the output current. This has no relation to the node where you feed the feedback. This type of feedback tends to keep the output CURRENT constant, instead of the output voltage as with the usual voltage feedback (feeding back a signal that depends on the output voltage).

We discussed that in another thread with respect to the term 'CFA' in opamps where the feedback node is a low impedance and that is ALSO called current feedback. I think it is incorrect, because when you actually check the current into that low impedance feedback node you will find that it tends to be very, very small, a few 100nA max. But we are stuck with this situation that gives rise to confusion as in this case.

jan
 
Hi,

I have built TQWT enclosures for fostex fe 166en (low Qts) working with a sansui AU-217 amplifier. The problem is that this combination is overdamped, bass is low and over-controlled.
So I was thinking changing the voltage feedback loop of my amplifier to a current feedback loop by modifying the power output stage.
I plan to take feedback from resistor (about 8 ohm) placed in series with the loudspeaker and take feedback from there.
Has somebody already tested such a mod?

Before you start hacking up your amp, have you considered what such a modification will change about the frequency response of your system?

Essentially when you convert from "voltage drive" to "current drive" you are creating an driver-impedance-dependent gain. As you surely know, this will cause the low frequency response to rise, because the impedance peak at resonance will result in a larger voltage drop (around resonance) than above resonance, where impedance falls back to near Re. But the same effect will cause the high frequency response to rise as well, because of the voice coil inductance (rising impedance with frequency). Both bass and treble will be increased, or saying something similar, the midrange will be depressed. This will totally change the tonal character of your speaker! Is this also an effect you want???

Instead of making a modification to your amplifier that might make it unstable or otherwise do harm, why not just build (passive or active) a contour network that can be placed before the amplifier? This will give you more control on the boost - where, how much, etc.

-Charlie
 
There are some graphs posted by speaker dave on one of the threads, which showed that current drive lowers 3rd harmonic distortion due to flux modulation in drivers. Might be worth looking into.

OK... but:

Here's a quote by the same "speaker dave" in which he says his experiments show no difference in distortion with high or low driver impedance (voltage or current drive):
I've also mentioned on a thread or two that I've done distortion tests with variable output impedance amps and at low frequencies there was no difference in distortion level with high or low driving impedance, as long as the comparison was otherwise equalized to the same response curve.

Taken from this post:
http://www.diyaudio.com/forums/mult...river-qts-you-cant-tuna-fish.html#post2743861

Will the real speaker dave please stand up...

-Charlie
 
...and here's another post where speaker dave more emphatically makes his point that there is NO decrease in distortion:
http://www.diyaudio.com/forums/mult...ver-qts-you-cant-tuna-fish-3.html#post2745854
Sorry, but distortion at resonance is insensitive to output impedance, high, low, or negative. Of course negative output impedance will allow you to achieve very low apparent Q.

In every case, when discussing distortion we must compare outputs of the same level. That is where the confusion comes from. If you simply increase output impedance and allow the bass to peak up the distortion will rise as well, but only because bass output (excursion) is up.

In my testing I used a compressor loop to maintain a flat output level, regardless of amplifier input. Doing so showed that the distortion did not change with driving impedance.

Edit: but read on... the mystery deepens and speaker dave changes his tune!
 
Last edited:
OK, here's an update. I found the plots. Sure enough, as ra7 states, speaker dave's experiments showed that 3rd order distortion was reduced by about 10dB in the midrange (300-1k Hz). But no significant change in the bass region. Here's think to the measurement post:
http://www.diyaudio.com/forums/mult...er-qts-you-cant-tuna-fish-10.html#post2755619

In a follow up post in that thread, he states:
In the end, the midrange hysteresis effect is an instantaneaous nonlinearity of impedance (occurs every cycle of a sine wave) so, since F = BLi, constant i means lower distortion than the constant v of low source impedance.

That's very interesting stuff, actually. But I'm not becoming a convert to current drive just yet...

-Charlie
 
The OP original plan seems to be to pick an impedance at which the base will be optimized, not a full blown current drive - it may make the top end rise some, but that can be compensated.(and could even be an improvement if done right or with the right driver). I haven't read speaker dave's thread yet but it is on my list.
 
The OP original plan seems to be to pick an impedance at which the base will be optimized, not a full blown current drive - it may make the top end rise some, but that can be compensated.(and could even be an improvement if done right or with the right driver). I haven't read speaker dave's thread yet but it is on my list.

The "high impedance source" in speaker dave's post was 57 ohms. Not exactly current drive.
 
The "high impedance source" in speaker dave's post was 57 ohms. Not exactly current drive.

True, but the usual impedance for getting reasonable Q from a low q woofer is about 1 to 8 ohms. And for a normal speaker system, 57 ohms would be much closer to current drive then voltage drive.
Using a high wattage 1 to 8 ohm resistor in series to see the effects would be a good way to check the system out before modding your amp. It has been done in the past - I recall some bass reflex systems with improved bass performance with a small resistor in series, and the original Cizek speaker had a Q switch on the back which just inserted a resistor to bring up the system Q. (Result of suggestion from Dick Pierce when measuring the prototypes.) YMMV