Talk:Class-T amplifier

advert tag
An advert tag was placed on this article in late may 2007.

The amps in this class are commercially significant, somewhat different than other PWM amplifier designss, and are quite differetn from digital amplifiers. The distinction is sufficintly important to merit notice in WP. Though I would suggest not more notice than this small article whih links to more general articles related to the design approach.

If phrasing is not improved by WP editors in the near future, I suggest that the advert tag be removed as the commercial ad content of this article is not excessive as it currently stands. ww 03:34, 26 May 2007 (UTC)

Merge Proposal
The merge proposal is a good one, however, there is an issue with the Tripath entry redirecting here to Class T amplifier. Instead of merging with electronic amplifier, I propose renaming this article to Tripath Technology and placing this content as a section of an article about the company. Then with appropriate linkage, methinks this would be a great solution.
 * There's a difficulty with this approach. Tripath declared bankruptcy and its intellectual property assets (ie, the amp designs, etc) were acquired by Cirrus. So an article about Tripath will be an obituary, rather than dealing with an amplifier design approach. so I oppose this suggestion as overtaken by events. ww 00:52, 2 August 2007 (UTC)


 * Tripath now redirects to Cirrus Logic then. :) — Omegatron (talk) 03:07, 5 June 2008 (UTC)
 * Tripath should not redirect to Cirrus since they have only the ownership of the technology. As long as i am aware of, they do not produce or sell these specific products, only similar. Tripath had its own story which is well worth being separately stated.

Requested move

 * The following discussion is an archived discussion of the proposal. Please do not modify it. Subsequent comments should be made in a new section on the talk page. No further edits should be made to this section. 

No consensus to move. Vegaswikian (talk) 05:36, 3 November 2010 (UTC)

Class T amplifier → Tripath Technology — 71.167.62.91 (talk) 19:24, 26 October 2010 (UTC)
 * can you please provide an explanation.--76.66.180.54 (talk) 01:58, 27 October 2010 (UTC)


 * Oppose this article is not about the company. 76.66.203.138 (talk) 05:08, 27 October 2010 (UTC)
 * Speedy close no rationale given. 76.66.203.138 (talk) 05:08, 27 October 2010 (UTC)
 * The above discussion is preserved as an archive of the proposal. Please do not modify it. Subsequent comments should be made in a new section on this talk page. No further edits should be made to this section.

Requested move 18 February 2017

 * The following is a closed discussion of a requested move. Please do not modify it. Subsequent comments should be made in a new section on the talk page. Editors desiring to contest the closing decision should consider a move review. No further edits should be made to this section. 

The result of the move request was: not moved. Primefac (talk) 13:35, 13 March 2017 (UTC)

Class-T amplifier → Class T – The article is about a trademarked product. There is no such thing as a "Class-T amplifier", Class T is simply a trademark for this company's design of a class D amplifier. This is clear and there is no dispute. Because of this, we should be referring to "Class T" as a trademark, and class D as the design, both in the article and in the title. Also, I believe it should be "was a trademark" as I don't see a live trademark, and the parent company seems out of business. Best I can tell, the dash wasn't used either. Second choice is move to company name. Current name is misleading. 71.76.228.224 (talk) 02:58, 17 February 2017 (UTC) Relisted. Jenks24 (talk) 06:54, 27 February 2017 (UTC)


 * And not that it should matter, but I am not the IP above that made the previous request, not even the same area. 71.76.228.224 (talk) 19:44, 18 February 2017 (UTC)


 * Apologies about the template removal. You inserted the malformed tag (as you could've seen when you put in the template) which caused this to appear under "Malformed requests" on the RM page. I've gone ahead and fixed the template for you. :) -- QEDK ( 愛 ) 21:05, 18 February 2017 (UTC)
 * No problem, mistakes happen, I assumed it was an innocent misunderstanding. I had subst: the command but it borked for some reason.  I stumbled across the article, and I wonder if it should be a redirect, but even if we are generous, I just want it to have a title that is consistent with WP:MOS and the facts. 71.76.228.224 (talk) 21:32, 18 February 2017 (UTC)


 * Noting that the target location is currently a dab page. Jenks24 (talk) 06:54, 27 February 2017 (UTC)
 * Oppose Class-T amplifier meets commonname, is clearer and the more widely used. It's also consistent with our articles on Class-D amplifier etc. If there was a use of the same Class T approach for a "Class-T oscillator" or a "Class-T turboencabulator", then there might be more reason for it. I can't begin to understand the OP's claim, "There is no such thing as a Class-T amplifier". Andy Dingley (talk) 09:47, 27 February 2017 (UTC)

User:Andy Dingley, if you would have read, you would have understood. There isn't a class of amplifier called "T". There is a Class A, B, A/B, C, D, E, F, E/F. Class T is a trademark only, it isn't a class of amp. The trademark doesn't use the word "amplifier". Common usage doesn't use the name amplifier as part of the name, only as part of the description. The trademark isn't current, but the registry is still easy enough to find. If you are strictly following WP:COMMONNAME, (which I was) then shortening the name is absolutely called for. 71.76.228.224 (talk) 12:00, 8 March 2017 (UTC)
 * Sorry that I was slow to respond, working two days user:Jenks24, but I feel it was already covered in the nom, and I have covered it here again, so please reconsider the close or consider leaving open. You can't compare Class A to Class T because one is a real class, the other is a trade mark.  It is confusing to do so, demonstrated by the exchanges above.  There is NO such accepted class as T.  It is a trademark for a modified class D.  As such, the article can only possibly be about a trademarked product, and should be titled as such.  This isn't hard to verify. 71.76.228.224 (talk) 12:09, 8 March 2017 (UTC)
 * I've reopened the discussion and will leave it for another admin to close. I still think this is a clear no consensus case, but we'll see if anyone else feels differently. Jenks24 (talk) 15:13, 8 March 2017 (UTC)


 * A Class-T amplifier is an amplifier. It is of a distinct class, supported by industry-wide sourcing. It's not one of the original A-E classes (but then neither was AB!), but it is one today. Nor is it a trademark: it was, but after the disappearance of Tripath it's now generic. It's also distinct as a subset of Class-D designs, with specific and well-documented features.
 * Mostly though, why would you seek to rename "Class-T amplifier" (if there's no such thing as "Class-T") into "Class-T" alone? It's still just as much an amplifier! If you wanted a rename to "Tripath amplifier design" that would at least make some degree of logical sense. But COMMONNAME still seems to override all other arguments here. Andy Dingley (talk) 15:42, 8 March 2017 (UTC)
 * Because Class T is the common name, although I'm willing to bend on the title here. It was a unique implementation of a class D amp that was just different enough to pass muster at the USPTO office.  The manufacturers even said it was a modified class D amp.  We can't seem to get passed that. Class A/B doesn't compare here as it is combining two accepted standards using a phase inverter, T is using class D with patented add-ons, Class T is a trademark for a neat design, but it is not a unique, accept amp class It is unquestionably class D.  The manufacturers patented it as an extension to class D and never claimed it wasn't class D. I promise Andy, if you look it up, you will see. If you really think "Tripath amplifier design" is better, I would accept that as a 2nd choice. Again, the goal is clarity for the reader (who we are accidentally giving bad info to). 71.76.228.224 (talk) 00:43, 9 March 2017 (UTC)
 * So what are you going to rename Class AB to? Because your only reason here seems to be that it's not one of the original canon classes A..E, but a derivative. Yet both of these derivatives are clearly distinct and notable. Andy Dingley (talk) 10:16, 9 March 2017 (UTC)
 * Class A/B is an accepted class within a class. One of most popular classes, in fact, as most higher end tube guitar and audio amps use it.  T isn't a class at all, it is not that distinct, certainly not to have it's own class, and it doesn't.  It is a trademarked name and a few widgets on top of a class D amp.  Had it been a class, it would have followed the current nomenclature in naming and been given the next letter, or an extension to the D name.  It wasn't.  It is an improved class D amp with a fancy name.  BTW, you know that patents mean very little as 50% of patents challenged in court are thrown out.  That aside, if we treat it as a distinct class of amp, we are misleading the reader.  In the end, that trumps everything.  The title and prose should reflect the reality that this is modified class D amp with a trademark and patent to prevent anyone else from using it.  71.76.228.224 (talk) 12:25, 9 March 2017 (UTC)


 * Oppose. The IP appears to not understand descriptive name titling.  On Wikipedia, there is a big difference between "Class-T amplifier" and "Class-T Amplifier".  The current title is fine.  Even if there was a problem, it would have to be Class-T (amplifier), because a Class-T is never introduced as such unless already in the context of amplifiers.  --SmokeyJoe (talk) 04:39, 13 March 2017 (UTC)


 * The above discussion is preserved as an archive of a requested move. Please do not modify it. Subsequent comments should be made in a new section on this talk page or in a move review. No further edits should be made to this section.


 * I understand it quite well. We don't have articles on "Ford cars".  If the article is on the trademark, it shouldn't have the word amplifier in the title.  If it is on the technology, it should be renamed or simply merged.  As it stands, it is simply inconsistent with WP:MOS, but what do I know, I'm just a bloody IP. 71.76.228.224 (talk) 00:19, 15 March 2017 (UTC)
 * You've accused me of not having read the article, and now you're accusing me of opposing you simply because you're an IP (rather than having to repeatedly present a reasonable case against renaming). Either raise it at ANI, or cut it out. Andy Dingley (talk) 00:59, 15 March 2017 (UTC)
 * The comment wasn't directed at you specifically, Andy, and surely you know the place to raise a question wouldn't be ANI, it would be a content dispute issue. 71.76.228.224 (talk) 22:38, 15 March 2017 (UTC)

technology
Can this article explain the class T technology, separate from any implementation or company? That is, enough to make the need for a separate article from Class D? Otherwise, if Class D covers all switch-mode amplifier technologies, then it might as well redirect there. Note that Classes A, AB, and C are defined pretty much independent of the implementation technology, only on what part of a sinusoid the output devices are conducting. It does occur to me that it should be possible to make a switch mode amplifier other than with one or the other size always conducting. (Specifically, with zero output not generating a full power 50% square wave into a low-pass filter.)  Gah4 (talk) 19:36, 13 May 2019 (UTC)
 * Before Class-T there were only a handful of Class-D amplifiers. They were expensive, obscure niche products for people who wanted a hi-fi that was "different". There were a few as industrial amplifiers (I used them myself, driving shaker tables) but these were frequency limited to barely the audio range and only a narrow band within this,
 * Then Class-T happened. Suddenly they're everywhere. Either as Class-T, as Tripath's own chipsets, or badged as Class-D to distance themselves, but still relying on the distinctive features of Class-T over Class-D without this. Of course Class-T is still Class-D, it's a subset not anything distinct, but Class-D without the Class-T features has still gone nowhere.
 * The specific features are in the patent, but their implications weren't clear at the time. Using such a high sampling frequency gives high quality, but it also makes the PSU cheaper (and removes the need for the expensive and heavy transformer of other designs). The high frequency gave exceptional power rail noise rejection, to the extent that switch-mode power supplies could be used, regulation barely needed and reservoir capacitors shrunk to a fraction. Andy Dingley (talk) 22:27, 13 May 2019 (UTC)


 * That sounds about right. In 1979, I knew about Class D amplifiers in EE labs, maybe for student projects. As well as I remember, the output filters were a big problem. And yes, I would have been the person who bought something that was different, just for that reason.  But instead, I bought a Kenwood KA9100 that I still have, but don't use because it has too much hum.  If I get around to it, I will find out which capacitors need replacing. It seems that I did a search on Amazon for them, and now it keeps sending me suggestions to buy one. There are many small amplifiers with reasonable amounts of power for the size.  It always seemed to me that they needed to reach economy of scale to get cheap enough, and get cheap enough to reach economy of scale.  It is now 40 years later, and they are out there, but still not mass market items.  The class D article doesn't mention much about building filters, but does about switch timing. As for power supplies, as far as I know it is much the same for class AB. With appropriate feedback, you can make a class AB amplifier pretty much independent of supply voltage, until you clip.  Also, I believe it is possible to make one more efficient by adjusting the supply voltage depending on output level. You could build a SMPS that follows not so much more than the needed output voltage, into a AB output stage, with appropriate feedback. (I might remember someone doing something like that.)  But still, what is it that class T amplifiers have special? Gah4 (talk) 03:50, 14 May 2019 (UTC)
 * What Class-T amps have is two-fold. Firstly they have the more sophisticated algorithms (read the patent) because they're built at a time once it's practical to put a lot of digital signal processing onto the chip, thus permitting this. That would have been unworkable 20 years earlier (when Class-D esoterica appeared) and still unaffordable 10 years earlier, once it was possible.
 * Secondly, they run at an inordinately high switching frequency, once again because the cheap silicon now permits this. That has several advantages - the output filters are tiny: little ferrite cores, whereas the first Class-Ds had things the size of the PSU transformer. Secondly, this changes the effect of Q. Audio frequency is now a tiny, tiny band compared to the sampling frequency, so the filters have a flat response right across it  (these filters don't even notice a band so narrow). The Class-Ds I used around 1990 were powerful and efficient, but they had a very narrow bandwidth, because of limitation in the output filter design and in how close the sampling frquency was to the output frequency (in octaves).  You just can't make a passive filter work with such a square pass band.
 * Also the sampling frequency makes them immune to PSU noise because they effectively track that variation as a slow change and compensate for it.  Unlike your Kenwood, where it's reliant on passive filtering to remove that variation from the supply rails first.   You can "make an AB independent of supply voltage" but that's not the same thing: if that "variation" in the supply rails is in the audio band, you'll hear it.  With the Class T, you won't. Andy Dingley (talk) 10:46, 14 May 2019 (UTC)


 * OK, but higher switch frequency is mostly due to changes in transistor technology, and not so fundamental. Also, as far as I know, with appropriate negative feedback, you can mostly isolate AB amplifier output from power supply variation. The complication is that we are especially sensitive to some variations in sounds, so it isn't easy to do. Even more, it might affect the specifications in inaudible ways, but buyers might still not buy them, though if you build one it is still class AB. That is, there are many patentable improvements that will still result in being class AB. Some manufacturer could decide to market one as class Y or class Z, though. Also, I suspect one could use similar digital technology to track power supply variations in an AB amplifier. How easy it would be, I don't know. Gah4 (talk) 20:59, 7 July 2019 (UTC)


 * higher switch frequency is mostly due to changes in transistor technology, No, it isn't.  At least, not for the time period since Class-D was practically available.  VMOS FETs were available in that time and would easily run this fast, at high power and good efficiency. Andy Dingley (talk) 21:47, 7 July 2019 (UTC)


 * I am not trying to claim that there isn't enough technology there, but only that the extra technology doesn't get a new class. One can build AB amplifiers with fancy power supplies, and probably DSP control, too. Maybe digitize the exact curve of the output transistors, and with complicated DSP techniques feed in the exact base current needed, given the instantaneous power supply voltage. But it will still be class AB. (Maybe I should quickly patent that!) Any amplifier where the output device (transistor or vacuum tube) conducts through the whole cycle is class A. That is independent of how complicated the rest of the circuit is, or how much digital technology went into it. I haven't followed them so exactly, but I do know that there have been many tricks used in AB amplifiers over the years to improve efficiency and/or power output, I believe including switching power supplies that adapt to the input signal. Still AB, though. Gah4 (talk) 14:55, 15 July 2019 (UTC)

On the subject of amplifier classes and whether or not there is value in calling a Tripath implementation "Class-T".

Amplifier classes have historically been assigned based on the mode of operation of the output stage (Class-A linear, Class-B push-pull, Class-C resonant, Class-D switching, etc). The front-end architecture is not (to my knowledge) considered when assigning a class, nor is the operation of the power supply. A class "G" or "H" amplifier is then simply a class A/B design (generally) with variable rails. So does the Tripath architecture call for a new class? I don't think so. I don't think anything would be gained by doing so, nor is anything really lost by not doing so.

Class designations may have been informative and useful seventy years ago (when the most recent class was defined), but the designation of an amplifier as a certain "class" today tells you almost nothing about a modern amplifier in total. An LM1876 is very simple to implement compared to a discrete 30W design from the 1970's, but it comes with significantly improved noise, distortion, thermals and protection, and all at less than a quarter of the price. To describe them both as "A/B amplifiers" would be absolutely correct, but not entirely informative. The same is true with Tripath amplifiers and traditional Class-D designs - yes, they both have switching output stages with filters, but that's about as far as it goes. My wife's 250 HP car and the Ford Model-T both use inline 4-cylinder engines, as well.

Most of the things that Tripath brought to switching amps - input-modulated variable switching frequencies, offset switching frequencies, adaptive idle setpoints, adaptive/predictive clipping, simple gain adjustment, etc - were all made possible by the highly integrated front-end processor. So, disregarding the front-end, you could say if it's operating in switched mode then it's a Class-D amp and I would not argue the point, even though I worked at Tripath for most of its (admittedly brief) history. Internally and to our engineering peers we referred to out products as switching amps. To most of our customers and the press we referred to the architecture as Class-T, mostly in order to distinguish ourselves from the not-so-pretty history of Class-D audio designs. We felt, I think justifiably, that our product's performance so significantly improved upon those designs that a new designation was worth considering. — Preceding unsigned comment added by 2600:1700:65A0:A050:BC07:ADA1:216C:99B3 (talk) 07:08, 25 December 2020 (UTC)