Class D amplifiers

I had to buy a new hi fi amplifier last year since my old Sony amp died, it had a good innings though and lasted me 25 years. With an output of 45 watts per channel it was a big heavy beast like pretty much all old hi-fi amps.

For my new amp, I could not afford a top of the range model so settled for a tangent ampster BT. It was recommended by what Hi Fi and I had read some good reviews so I decided to buy it. It advertises 50 watts per channel, but you could fit four of them into the case of my old amp. I thought ok maybe they use a switching power supply, but what about the amp itself. Well, I did some research and as some of you have probably already guessed it was of class D design.

Now, I am very impressed with the sound of the amp, and it looks well made, but I have always thought of class D amplifiers being of inferior quality with their use reserved to things like PA systems where you need a lot of power, but sound quality is not too important. After some further research I discovered that with the exception of really hgh end equipment most audio amps are of class D design today.

Is it now possible to make a class D amp with a sound quality that is close to that of a class A or AB design, or are manufacturers sacrificing quality in order to make their amps smaller, lighter and more efficient?

I can't speak for your particular amp, of course, but it's certainly quite possible to build class-D amps these days with quality that matches or even surpasses class-AB or B designs. In my opinion, Hypex are among (possibly the?) leaders in these designs - the specs are quite astounding. If you visit their website, they have some interesting papers on their techniques.

I don't know how many hifi and consumer amps these days are class-D, but it certainly prevails at the extremes of power: Low-powered portable devices where battery life must be maximised, and high-powered PA applications (my end of things) where the power-limiting factor is your ability to remove heat from the amp, and also size & weight are critical.

It's impressive stuff...

I found no detailed spec's for the Tangent Ampster BT amplifier. The frequency response is 20-20kHz but without saying how much deviation, plus or minus 20dB or 2dB??
Power output 25W per channel into who-knows-what or 50W per channel into 4 ohms but with no mention about how much distortion.
Noise level? Unknown.
These missing spec's are used for cheap junk not for a hifi product.

audioguru said:

I found no detailed spec's for the Tangent Ampster BT amplifier. The frequency response is 20-20kHz but without saying how much deviation, plus or minus 20dB or 2dB??
Power output 25W per channel into who-knows-what or 50W per channel into 4 ohms but with no mention about how much distortion.
Noise level? Unknown.
These missing spec's are used for cheap junk not for a hifi product.

Click to expand...

I was disappointed about the lack of available specs, but I would not call it junk, I would love to be able to afford a £500 amp, but I cant

If you like the sounds of the amplifier then it is good enough.

Prototype said:

I was disappointed about the lack of available specs, but I would not call it junk, I would love to be able to afford a £500 amp, but I cant

Click to expand...

that's part of the problem with class D amps, most of them have no feedback loop, so there's no correction for distortion or management of the output impedance. there are some that do (such as Bang & Olufsen ICE amps and receivers that use B&O chipsets) but they are often more expensive than standard class AB amplifiers. without feedback, you will often see the amplifier specs not mention distortion, noise, or damping factor for class D amps. they will tell you how many watts (if it's Sony, they will use measurements of what AG calls "whatts" totally bogus numbers to get people to buy the thing). i was working at the service center for a large retailer, and a whole bunch of hype came out saying things like "'Big Iron' is going away" and "2010 is the year of the 'Ultralight'"... well, it didn't happen for the mid and high end stuff... all of the companies with good reputations for making good receivers stuck with class AB amps and heavy power transformers. Pioneer came out with a series of receivers using B&O ICE amps, but they kept the heavy power transformer. i got to make some measurements on those Pioneer receivers, and whatever method B&O used to produce feedback for their class D amps works, and works well. i got output impedance measurements only about 10-15% higher than that of a class AB amp in a comparable Pioneer receiver. that's better than i expected, but ICE amp chipsets are more expensive than the class AB amps they replace.

In hindsight I should have probably saved up for a class AB amp for a similar price and bought a decent DAC to go with it. I did spend what I thought was a decent amount of money on a DAC to use with my Sony amp before it died, but it was VERY noisy. I really don,t know if I would notice any difference with the speakers I have though, which are as old as my Sony, not to mention the acoustics of my little flat.

unclejed613 said:

most of them have no feedback loop

Click to expand...

I was going to question you "most" comment because I have never used one that was open loop.
I did find some that are open looped. If you can make 0.009% thd with no feed back then I guess there is no need. (head phone amp)
IT.com has a large selection of audio D amps.

Well I did some digging and found a German review for the tangent ampster BT. Well, the distortion curve is a little disappointing, but I listen to a lot of different types of music and I personally don't notice any colouration of the sound, but then my ears are 40 years old. Perhaps in a blind test alongside a more expensive amp I might be able to notice the difference, I don't know. However, listening to it, I would not have guessed the best it can do is around 0.1% THD, give or take.

Translation "About 50 watts of power creates the Ampster per channel, while the distortion is quite high, but not audible" According to google, that is

The distortion measurement is at only one mid frequency. Maybe it distorts like mad at low and/or high frequencies?
Does it roll off (attenuate) high frequencies to also attenuate high frequency distortion harmonics?

A man who is 40 years old has a significant mid and high frequency hearing loss so they turn up the treble (like my hearing aids do much more) to hear better.

ronsimpson said:

I was going to question you "most" comment because I have never used one that was open loop.

Click to expand...

Samsung, LG, Panasonic, and Sony "all in one" home theater systems use class D amps that either are completely open loop, or have very little feedback. class D doesn't have a lot of open loop gain, so there's not much of a gain margin when you do apply feedback.

Prototype said:

However, listening to it, I would not have guessed the best it can do is around 0.1% THD, give or take.

Click to expand...

0.1% is the point where most people begin to notice distortion (somewhere in the range of 0.1 to 0.5% depending on the person's hearing).

I was going to post the distortion of a TDA7492 class-D stereo amplifier since it is 10 times less that the one we are talking about. Its rated output power is about the same. It shows a graph of distortion that is minimum at 200Hz but is 20 times higher at 20Hz and 7kHz. BUT this distortion vs frequency is measured at only 1W! What happens at full power that they are hiding?

audioguru said:

I was going to post the distortion of a TDA7492 class-D stereo amplifier since it is 10 times less that the one we are talking about. Its rated output power is about the same. It shows a graph of distortion that is minimum at 200Hz but is 20 times higher at 20Hz and 7kHz. BUT this distortion vs frequency is measured at only 1W! What happens at full power that they are hiding?

Click to expand...

i found the data sheet for it.... have a look at the THD vs freq curves... it ain't pretty... on the first page the spec is 10% THD at rated output power... looking at the schematic on page 22, there is no feedback. the distortion mechanism in class D amps is in the linearity of the triangle wave generated for the comparator. the comparator and triangle source are the only linear devices in the whole amplifier (not including the output filters which are passive LC networks), everything else is a binary switch.

i think that chip is what is used in Samsung and LG home theater systems. one of the early Panasonic home theater actually used the amplifier's power supply (only the power supply for the output devices, the rails for previous stages remained constant) as the volume control. the output stage +/- Vcc voltages were variable from 0 to +/- 35V.

unclejed613 said:

i found the data sheet for it.... have a look at the THD vs freq curves... it ain't pretty... on the first page the spec is 10% THD at rated output power... looking at the schematic on page 22, there is no feedback. the distortion mechanism in class D amps is in the linearity of the triangle wave generated for the comparator. the comparator and triangle source are the only linear devices in the whole amplifier (not including the output filters which are passive LC networks), everything else is a binary switch.

i think that chip is what is used in Samsung and LG home theater systems. one of the early Panasonic home theater actually used the amplifier's power supply (only the power supply for the output devices, the rails for previous stages remained constant) as the volume control. the output stage +/- Vcc voltages were variable from 0 to +/- 35V.

Click to expand...

I looked, and yes it is a bit ugly. I remember reading a while ago that class D amps can go up to 10% THD at rated power, but I have never met anyone who drives their amps at rated power. I can't say whether what I read on that data sheet would have changed my mind about my choice of amp, but I don't have any regrets.

Nearly all audio amplifiers are rated at 10% clipping distortion to make the power number bigger. Other amplifiers are rated at "Peak or Maximum Power" which doubles the real power number.

Try this https://leachlegacy.ece.gatech.edu/lowtim/ as a nice amp. I built one in the 1980's with changes.

audioguru said:

Nearly all audio amplifiers are rated at 10% clipping distortion to make the power number bigger. Other amplifiers are rated at "Peak or Maximum Power" which doubles the real power number.

Click to expand...

Decent quality ones aren't, they give true RMS output powers at low distortion, and across the entire audio bandwidth.

Going back a VERY long time, I had a 'classic' Leak Stereo 70 amplifier, these were rated at 35W per channel RMS at under 0.1% distortion (as I recall), but were capable of well over 50W RMS per channel at higher distortion levels. I've seen many of them used as PA and Guitar amps, and used mine for PA many times.

That's how HiFi amps used to be rated, and it's still true today for decent quality ones.

Nigel Goodwin said:

Decent quality ones aren't, they give true RMS output powers at low distortion, and across the entire audio bandwidth.

Going back a VERY long time, I had a 'classic' Leak Stereo 70 amplifier, these were rated at 35W per channel RMS at under 0.1% distortion (as I recall), but were capable of well over 50W RMS per channel at higher distortion levels. I've seen many of them used as PA and Guitar amps, and used mine for PA many times.

That's how HiFi amps used to be rated, and it's still true today for decent quality ones.

Click to expand...

Yes that's something that really annoys me about the really cheap amps, they like to quote peak impulse power and rebrand it as something like peak music power. I had an old midi system that advertised 60W per channel, whats funny is the label on the back that stated the power consumption of 45W!!

That's marketing for ya

Yes, some National Semi (now owned by TI) amplifier ICs mention power only at all audio frequencies and at low distortion.
Some ST car amplifier ICs state power with a saturated squarewave input and output that is double the real sinewave output power.

audioguru said:

Some ST car amplifier ICs state power with a saturated squarewave input and output that is double the real sinewave output power.

Click to expand...

ST is the gold standard in Technical Marketing. Making bold claims that are so vague they cannot be disputed in court but, the claims seem substantial enough that equipmet manufacturers are willing to repeat in their own advertising. They have Learned well from PT Barnum and all the snake oils salesmen in between.