How low is very, very, very low? 0.1R?, 0.05R?
One thing to keep in mind is that a conventional emitter follower (with parallel devices to provide the necessary load current) will perform really BAD with such a low load impedance - the gm doubling distortion will be horrendous.
Do you care about the distortion? If so, and if you only wan't 10W or so, class A is a easy solution to the gm doubling issue into such a low load impedance.
You will also need good control of the DC offset at the output to prevent large bias currents flowing through your transducer. A DC servo would be a good idea.
I can not thank you enough for your interest in this project
I'm a member of AES (Purdue Chapter) and have a relationship with some of the faculty at Herrick laboratories. They've expressed interest in this project, however time in the laboratory is a scarce resource. We'll see what I can manage. I'd like to do testing in the Anechoic chamber as well as access their Laser Vibrometer (I was told it cost over $500,000
) for displacement measurements.Are there any other tests that members would like to see conducted with regards to this project? I have a RAAL 140-15D available for testing. I can not make any promises, but I can always ask.
Nor I, as that would be entirely missing the point. If the reason for doing this is the fun and challenge of design, go nuts. If it's predominantly to avoid 'transformer distortion' it might be worth confirming it's a significant factor.
wow ! they never raked you over the coals , even Tinitus got in ...
Shoot for a 150-200 @`1 ohm , this way i can get some info without having to duck ...
A high power amplifier may not necessarily be such a bad idea. Although levels will probably exceed that of average listening with full power, such an amplifier would afford the opportunity for a significant fatigue analysis, of which I have seen very little research conducted.
Last edited:
Hi thadman,
I need to ask for more input. What upper limits have you set on things like dissipation at idle and distortion? Is this for home use or testing?
Here is what I have so far. 0.055% THD at 10kHz with 10W out into 0.25 ohm load. Maximum output of about 50W. Dissipation at idle is 75W.
I need to ask for more input. What upper limits have you set on things like dissipation at idle and distortion? Is this for home use or testing?
Here is what I have so far. 0.055% THD at 10kHz with 10W out into 0.25 ohm load. Maximum output of about 50W. Dissipation at idle is 75W.
With regards to dissipation, as long as sufficient heat transfer solutions are available for the application, there is no limit. I am not concerned with the efficiency of the amplifier. Providing an unadulterated waveform to the transducer is paramount.
It may get a small amount of exposure in a home environment, however its significant role will be served in the laboratory for transducer analysis. As such, distortion should be minimized.
If you read thisI'm interested in building a DIY ribbon loudspeaker in my free time over Christmas break and would like to drive it directly. Following the resolution of some meetings with faculty in the engineering department, I might be able to gain access to the Universities Laser Vibrometer and Anechoic chamber for measurements (and comparison with the RAAL 140-15D)
I do not wish to use a transformer, as I believe it may contribute excessive distortion. As a result, I will need a special amplifier to drive the ribbon. The impedance of the ribbon will be almost purely resistive, but very, very, very low. Most of my knowledge in the field of loudspeakers is related to acoustics and the physics of transducers...I haven't yet been able to extend it to amplifier design.
My primary goal for the project is reproducing a Dirac Pulse/Square Wave, so impulse response will be weighted with the highest significance.
Are there any designs available? Can anyone recommend any texts?
How difficult would it be to build a direct drive 10 watt amplifier with negligible distortion (linear and non-linear distortion <20khz)?
Thanks,
Thadman
The revolutionary Ultrasound amplifiers
and this
Current driving
and this approach isn't a fake, then I would try the follow project from masterdeveloper Nelson Pass:
The Complementary Zen
http://www.passdiy.com/pdf/zen-v5-hires.pdf
Please note, that you need 1 pair of MOSFETs for 8 ohms, 2 pairs for 4 Ohm and 20 pairs for 0.4 ohms and 30-60A idle current by approx +/-8V, that means 500-1000 Watt loss power. Therefore I would diy my own heatsink like this proposal:
http://sound.westhost.com/articles/diy-heatsink.htm
Do you want to build a two way full range ribbon loudspeaker like Apogee
http://news.cnet.com/i/bto/20070816/apogee_540x405.jpg
or do you prefer only a ribbon tweeter like ESS AMT
http://hifi4me.de/img/-ess amt monitor-0_1166313911.jpg ?
In the last case the needed amplifier would be more smaller
Last edited:
How do you measure such low impedance anyway ?
Ahh, maybe measure through a bigger series resistor ?
And measure at least 10 times to have a guess on the right one
I suggested to build a bigger ribbon/planar
It could manage down to 300hz, and work fine in 2way with woofer
1-2ohm would make sense
A small ribbon could have impedance close to zero
You may have a very nice tweeter
But you will still miss the most important, midrange
But sure, I would be interested in this amp too, to drive 1ohm safely
But with classA theres a disadvantage of low power, or bias would be exstreme, I suppose
Ahh, maybe measure through a bigger series resistor ?
And measure at least 10 times to have a guess on the right one
I suggested to build a bigger ribbon/planar
It could manage down to 300hz, and work fine in 2way with woofer
1-2ohm would make sense
A small ribbon could have impedance close to zero
You may have a very nice tweeter
But you will still miss the most important, midrange
But sure, I would be interested in this amp too, to drive 1ohm safely
But with classA theres a disadvantage of low power, or bias would be exstreme, I suppose
Last edited:
With that as a general guideline, here is what I have so far. At 10W output into a 0.25 ohm load and dissipation at idle of 160W. This is the output impedance and THD at 20KHz, 10KHz and 1KHz. Would this meet your low distortion criteria?
output_impedance_at_V(6) 7.337045e-008
v6#Input_impedance 4.722100e+004
Transfer_function 9.953144e-001
Fourier analysis for v(6):
No. Harmonics: 10, THD: 0.000674031 %, Gridsize: 200, Interpolation Degree: 1
Harmonic Frequency Magnitude Phase Norm. Mag Norm. Phase
-------- --------- --------- ----- --------- -----------
0 0 0.000994329 0 0 0
1 20000 1.57901 -1.175 1 0
2 40000 8.20142e-006 -158.27 5.19403e-006 -157.09
3 60000 5.23556e-006 94.6582 3.31573e-006 95.8332
4 80000 1.78915e-006 -76.338 1.13308e-006 -75.163
5 100000 2.00522e-006 -75.951 1.26992e-006 -74.776
6 120000 1.63168e-006 -78.777 1.03336e-006 -77.602
7 140000 1.75428e-006 -101.41 1.111e-006 -100.24
8 160000 1.62417e-006 -68.897 1.0286e-006 -67.722
9 180000 1.73192e-006 -95.629 1.09684e-006 -94.454
Fourier analysis for v(6):
No. Harmonics: 10, THD: 0.000326382 %, Gridsize: 200, Interpolation Degree: 1
Harmonic Frequency Magnitude Phase Norm. Mag Norm. Phase
-------- --------- --------- ----- --------- -----------
0 0 0.000993506 0 0 0
1 10000 1.57913 -0.58735 1 0
2 20000 3.83563e-006 -161.82 2.42895e-006 -161.24
3 30000 2.62513e-006 92.4543 1.66239e-006 93.0416
4 40000 7.98668e-007 -79.339 5.05764e-007 -78.752
5 50000 9.63861e-007 -62.039 6.10374e-007 -61.452
6 60000 8.57829e-007 -63.521 5.43229e-007 -62.934
7 70000 1.05071e-006 -53.682 6.6537e-007 -53.095
8 80000 8.62622e-007 -73.442 5.46264e-007 -72.854
9 90000 8.99739e-007 -55.917 5.69768e-007 -55.329
Fourier analysis for v(6):
No. Harmonics: 10, THD: 4.8926e-005 %, Gridsize: 200, Interpolation Degree: 1
Harmonic Frequency Magnitude Phase Norm. Mag Norm. Phase
-------- --------- --------- ----- --------- -----------
0 0 0.000942919 0 0 0
1 1000 1.57912 -0.056677 1 0
2 2000 2.75538e-007 163.073 1.74488e-007 163.13
3 3000 4.42357e-007 80.1838 2.80128e-007 80.2404
4 4000 1.57562e-007 32.8675 9.97779e-008 32.9242
5 5000 1.65976e-007 22.8461 1.05107e-007 22.9028
6 6000 1.90366e-007 24.5163 1.20551e-007 24.573
7 7000 2.55109e-007 16.2734 1.61551e-007 16.3301
8 8000 2.65485e-007 15.8933 1.68122e-007 15.95
9 9000 3.18022e-007 18.8298 2.01391e-007 18.8865
Those distortion figures look fantastic! We'll see what the actual impedance ends up being. I'll have to run some numbers on the foil I'm able to source. However, I expect the resistance to be slightly higher than the RAAL 140-15D (.1R), as the ribbon element may end up being narrower. There appears to be some interest in testing a linesource alignment (same width, 90" length) as well.
I read an AES article recently with regards to the radiation impedance of pistons and arrays of pistons. For flat pistons of uniform velocity, a diameter of wavelength/6 provided the highest radiation impedance. A separation of 5/4 wavelength provided the highest radiation impedance in an array configuration. Of course, this is quite a dissimilar comparison to a ribbon transducer...however I believe some parallels may exist between the piston diameter and the width of the element.
Last edited:
The proposed "current driven" amplifier approach looks interesting. What are members thoughts on the aforementioned approach? What disadvantages do current driven amplifiers possess? The benefits appear quite favorable (lack of power compression, cancellation of inverse current).
Maybe, or just what it would need fore the one you suggested
Is it low voltage, 12V, 15V, or
And 160watt at idle seems a lot, which I suppose would take higher wattage trafos
You have specced it to do 10watt into 1/4ohm
Would be nice to know specs into other loads to get a picture of what it is about
Number of output devices etc
Until thadman find his subject cleared its his thread, but maybe other peole would find such amp useful fore slightly more human loads
Is it low voltage, 12V, 15V, or
And 160watt at idle seems a lot, which I suppose would take higher wattage trafos
You have specced it to do 10watt into 1/4ohm
Would be nice to know specs into other loads to get a picture of what it is about
Number of output devices etc
Until thadman find his subject cleared its his thread, but maybe other peole would find such amp useful fore slightly more human loads
Last edited: