Hi
I have exhumed this unusual amplifier based on a patented idea: "constant power amplifier" by Nguyen Tan Tai:
Electronic power amplifier for delivering a constant power into a load impedance - United States Patent 4095189
Claims are no global feedback, load invariant output power, unexpected reproduction quality, and tube-like sound.
It is claimed that fifteen(!) 8Ω series-connected speakers do not alter the acoustic power.
I leave you to judge.....
Have fun!
I have exhumed this unusual amplifier based on a patented idea: "constant power amplifier" by Nguyen Tan Tai:
Electronic power amplifier for delivering a constant power into a load impedance - United States Patent 4095189
Claims are no global feedback, load invariant output power, unexpected reproduction quality, and tube-like sound.
It is claimed that fifteen(!) 8Ω series-connected speakers do not alter the acoustic power.
I leave you to judge.....
Have fun!
Attachments
great!
do you think it will work with modern output transistors without tweaking?
regards
Yeah, I have run a sim, but I haven't disclosed the results yet, as I wanted to hear unbiased opinions first
The THD figure shows that it definitely sounds tube-like, and even overdoes it by quite a bit
That's impossible.
EDIT: To be more precise: what's shown is a reasonable conventional poweramp.
Let's not forget that an amplifier can be seen as an audio modulated "valve" between the PSU and the speaker.
The amplifier will not create power which was not available at the PSU first.
Won't supply voltage higher than the rail(s) so when you rise load impedance you soon hit the rails, and can't supply current higher than what's available, so that quickly limits power into lower impedances, besides increasing losses in a horrible way.
That's why amplifiers have a rated "optimum" impedance; as an example 8 ohms, and can deliver reasonable power into 2X and 1/2X that (16 or 4 ohms) but to claim you can raise impedance 15X and keep power *constant* is ludicrous.
To be even more precise: that can be done, of course, with a multitapped output transformer .... but none such is to be seen at that schematic
Last edited:
You asked for it, here it is, but be warned: it is a real stinker!so bad?
I like this circuit
could you pickup asc file?
thank you in advance
First, maximum output before clipping (sort of), and THD
Second the clipping (sort of) behavior (yes, it happens as soon as ~14Vpp, and I have corrected some resistor values otherwise it is still worse)
Third the temperature stability (or instability rather) of the Iq and output offset.
I forgot to mention: the -3dB small signal frequency response is .... 14KHz! Really tube-like!
Attachments
Excellent work! This doesn't change it into a winner, but at least it has been rescued from abysmal depths.
Now, what about the load invariance claim? For me, it is simply a current output amplifier. I think the damping factor must be rather disastrous.
As long as tube amplifiers will require coupling capacitors and output transformers, this will go mostly unnoticed (except for the rare OTL designs)
Elvee, so let me know what is in your opinion the greatest advantage of this vetust amp?
BTW do you think that an amplifier must have an high dumping factor to sound good?
Francesco.
I am always on the look for original, clever, or unexpected schemes.
This one caught my eyes because of the "constant power" claim, and the bizarre attending circuitry (the one I removed in the improved version).
I said to myself: the claim is impossible to fulfill, at least with a linear circuit: a constant voltage output is linear, but the output power is inversely proportional to the load resistance, and a constant current is linear too, but has an output proportional to the resistance.
Constant power requires something in-between: a current proportional to the square root of the load. If it is done in real time (ie not with an AGC), it can only lead to a non-linear transfer function.
The "constant power engine" simply dumps the largest part of the drive current, removing in effect one layer of amplification.
That such a thing could be patented puzzled me: why go to such lengths, create a delusional mathematical demonstration for such a worthless scheme?
I sort of hoped there was more to it, but sadly, as is always the case, it is simply a big heap of bovine waste-product
Not necessarily, but a current output over-emphasizes resonances and is generally detrimental. But not always.
Some amplifiers (tube for instance) have a high impedance output, most decent solid state and tube amplifiers have a very low output impedance, and some amplifiers, tube or solid-state have a negative resistance created deliberately.
interesting vintage approach.
here the pdf-paper:
http://www.pat2pdf.org/patents/pat4095189.pdf
Do you know, in which commercial amplifier brand/model this topology is in use?
Maybe a small manufactur in France like YBA, Cairn-EZO or Lavardin
From the same Inventor there is this patent:
Patent EP0159257B1 - Electronic circuit for measuring very weak signals superposed on a high ... - Google Patents
here the pdf-paper:
http://www.pat2pdf.org/patents/pat4095189.pdf
Do you know, in which commercial amplifier brand/model this topology is in use?
Maybe a small manufactur in France like YBA, Cairn-EZO or Lavardin
From the same Inventor there is this patent:
Patent EP0159257B1 - Electronic circuit for measuring very weak signals superposed on a high ... - Google Patents
So I understand that you have posted the schematic not because you shared this idea but only for curiosity and pheraps to humiliate the author.
I'm sorry you feel that way.
I, however, I paused to read this topic just because I was attracted by the phrase "constant power amp" because of something like that I'm working on for a few years and this is a very serious thing for me. For a long time I thought I'd be alone to think so, then a few months ago a German guy sent me an article by Stanley White about his "POWRTRON AMPLIFIER" which shares a similar idea.
Since the relationship between the power from one side and the voltage, the current and the impedance of the other side are non-linear but quadratic relationship, then the big problem is to find a simple and fast circuit (no IC) which does not introduce distortions (no diodes) that implements such a function.
Now coming back to the idea of Tai Nguyen-Tan, I think it was good. Less good seems to me the real implementation of the same, because he dredge the bias current, making it variable, which for other needs, it should be quite stable. Another issue from the attached diagram is that the control loop goes too far away from the output, with associated problems to the inevitable delays in the process of adjustment.
My idea by such amplifier is developed in a different way.
Simple example i will post later.
Francesco.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.