It`s something that bugs me for quiet a while now. I see a lot of amps with claims of 0.000x% THD and some even as high as 0.4%. But how much crossover distortion starts to be noticable ( and I don`t want answers from people claiming to have a golden ear 🙂 ).
The reason I ask is because I recently bought a Akai AM-M459, midi sized, nice looking little amplifier that has an STK4152II power stage. The service manual claims 0.1% THD but the ICs spec sheet claims 0.4% THD at it`s rated power. Also the amps rated power is 2x40W at 8 ohms with ±39.5V, but the ICs spec sheets claims 2x30W at 8 ohms with ±27.5V.
I was thinking of swaping out the 4152 with a 4151 (pin compatible but larger package) because it has a much lower THD of 0.02%, or even mod it with a discreet transistor amplifier.
Is it worth the time or will I notice no audible difference?
The reason I ask is because I recently bought a Akai AM-M459, midi sized, nice looking little amplifier that has an STK4152II power stage. The service manual claims 0.1% THD but the ICs spec sheet claims 0.4% THD at it`s rated power. Also the amps rated power is 2x40W at 8 ohms with ±39.5V, but the ICs spec sheets claims 2x30W at 8 ohms with ±27.5V.
I was thinking of swaping out the 4152 with a 4151 (pin compatible but larger package) because it has a much lower THD of 0.02%, or even mod it with a discreet transistor amplifier.
Is it worth the time or will I notice no audible difference?
IMO ---- you want less than 0.05% by a good margin. .02% THD is much better.
I would try going with the 4151. You will notice an improvement in most cases.
-RNM
I would try going with the 4151. You will notice an improvement in most cases.
-RNM
Thanks for the quick reply. What about moding it with a discreet transistor amp? Some say that STKs don`t sound very good.
Hi partyface86,
if you want to change the chip, better take a LM3668. It's not pin compatible but really audible better.
To build a discrete Amp that achieves the performance of the LM3886 is not easy.
Thanks for the advice, but I already have the STK4151. Moding it with the LM3886 for that I would need to make 2 PCBs from scratch for the ICs and my only option is perfboard as I don`t have the necessary things to make a decent PCB.
Also the discreet transistor amp is from an old Onkyo reciever and it only needs the power, signal and speakers to be connected to it, so moding it with this would be much simpler 🙂
Quite a few of us have built LM3886 amps dead bug style - parts soldered directly to the pins, no board. It is a very simple circuit. Just throwing in another option. 😉. It wants a decent heat sink so you stay well away from triggering the Spike protection circuits.
there are two types of distortion ..the bad one and the good one
the bad one is the one you describe above ....
The good distortion come from people that rely or believe simulators, implement circuits the wrong way , upgrade 741 op amps with OPA (version one million) with astronomical bandwidth without any change on compensation , forget terms like THD , pssr , cable rooting electrical and/or many other laws that apply in audio circuits ...
This typo of activity increases various types of distortion like the Op amp oscillation that often is described by the user/upgrader as : """ after the upgrade the high become unbelievably crisp """ and so on and on ....
But obviously you shouldn't bother about this type of distortion you can have as much as like since this one is the good type of distortion as said ..On the contrary you should focus if it is possible to obtain and original IC for replacement in your system .
With a tension for a bit of fun ...
Sakis
the bad one is the one you describe above ....
The good distortion come from people that rely or believe simulators, implement circuits the wrong way , upgrade 741 op amps with OPA (version one million) with astronomical bandwidth without any change on compensation , forget terms like THD , pssr , cable rooting electrical and/or many other laws that apply in audio circuits ...
This typo of activity increases various types of distortion like the Op amp oscillation that often is described by the user/upgrader as : """ after the upgrade the high become unbelievably crisp """ and so on and on ....
But obviously you shouldn't bother about this type of distortion you can have as much as like since this one is the good type of distortion as said ..On the contrary you should focus if it is possible to obtain and original IC for replacement in your system .
With a tension for a bit of fun ...
Sakis
I`ll consider that option too. Thanks 🙂
I still need more opinions on THD though. RNMarsh wrote that below 0.05% is always good.
I still need more opinions on THD though. RNMarsh wrote that below 0.05% is always good.
I`m very sorry... But I don`t quite understand what you are trying to say.
Well we dont all share the same sense of humor 😀
Distortion actually is very tricky consumer term At 99% of the cases the distortion refers to the main amplifier only ...Only very few manufacturers make that clear in their service manuals.
Obviously the total distortion of the total machine could be as ten time higher than the one described specifically for the amplifier unit /module
So in the remote case that you are able to locate original IC and replace it there is hardly going to be any noticeable change since the sound quality of such a machine is already compromised by the cost effective rules and consumer safety margins .
So in a way the all thing here for the specific application is either pointless or if possible doesn't worth the idea
Kind regards
Sakis
Now I understand 😀 I know that other components like opamps also ad up to the final THD, but I always tought they are negligible since even the cheapest opamps have very low distortion.
THD is probably not the best measure of amp performance. Look at some of the Pass DIY designs that have relatively high THD but still sound good. Amps with similar numbers sound different. My guess is what is more important is the relative levels of the harmonics. Low order even harmonics are more prevalent in musical instrument sounds and are less objectionable than odd harmonics.
Another factor is amp behavior at clipping. We clip our amps more often than most realize. There are a fair number of threads on this topic.
Another factor is amp behavior at clipping. We clip our amps more often than most realize. There are a fair number of threads on this topic.
Without wanting to spoil your musings about the audibility of 0.05% amplifier distortion, just one question:
Anyone having any idea how much your typical 2 way speaker (6-8" woofer, 1" dome) distorts at 90 dB SPL? An average one 5-10% (2nd and 3rd harmonic), and a "better" one (with improved linearity magnetic circuit, suspension etc.) still around 1 to 3%.
There are numerous readings and tests available on the subject, so go and see it with your own eyes, it is not my idea....
Anyone having any idea how much your typical 2 way speaker (6-8" woofer, 1" dome) distorts at 90 dB SPL? An average one 5-10% (2nd and 3rd harmonic), and a "better" one (with improved linearity magnetic circuit, suspension etc.) still around 1 to 3%.
There are numerous readings and tests available on the subject, so go and see it with your own eyes, it is not my idea....
IMO, distortion can't be summed up in one number. It depends on the spectral distribution of the harmonics. Lower order harmonics might sound "warmer" or "musical" (to use "audiophile" terms) while higher order harmonics sound harsh and discordant. So, a listener might find a higher amount of lower order harmonics to be more desirable or even pleasant in the signal.
The bigger bear in the works is intermodulation distortion, but that is really a non issue in most amplifiers in the last 40 some years. IC power amps never really had an issue with it. I think IM is more of an issue in transducers such as loud speakers and pickup devices.
Negative feedback lowers THD significantly but broadens the harmonic spectrum, so the THD should be made as low as possible. Personally, I think distortion should be no more than 0.1% across the power and audio frequency band (20-20kHz). Even a cheap LM1875 IC can beat that.
On data sheets, a higher distortion figure may be used to allow a higher wattage to be quoted. If provided, look at the distortion curves given on the graphs. These provide a much clearer picture of the distortion characteristics than a single number!
I've heard of blind tests conducted that showed listeners could not pick out the music reliably that had distortion on the order of several % added compared to a clean signal. I don't know the nature of the harmonics added or other details of the test, so take it for what it is worth.
The bigger bear in the works is intermodulation distortion, but that is really a non issue in most amplifiers in the last 40 some years. IC power amps never really had an issue with it. I think IM is more of an issue in transducers such as loud speakers and pickup devices.
Negative feedback lowers THD significantly but broadens the harmonic spectrum, so the THD should be made as low as possible. Personally, I think distortion should be no more than 0.1% across the power and audio frequency band (20-20kHz). Even a cheap LM1875 IC can beat that.
On data sheets, a higher distortion figure may be used to allow a higher wattage to be quoted. If provided, look at the distortion curves given on the graphs. These provide a much clearer picture of the distortion characteristics than a single number!
I've heard of blind tests conducted that showed listeners could not pick out the music reliably that had distortion on the order of several % added compared to a clean signal. I don't know the nature of the harmonics added or other details of the test, so take it for what it is worth.
Last edited:
1) I wouldnt lose dream or waste time with that .
Difference, if any, will be negligible.
2) as properly stated (by some who see the big picture) , power amp distortion is just a tiny bit of all the mistreatment the original audio signal suffers from original performance until finally reaching your ears.
That chain has at least 10 to 20 links ... and you are worrying about further polishing the one which arguably is the most transparent of them all.
As said above, .05% distortion pales when compared to 5/10% (or more) distortion in speakers and what's present on other elements.
Difference, if any, will be negligible.
2) as properly stated (by some who see the big picture) , power amp distortion is just a tiny bit of all the mistreatment the original audio signal suffers from original performance until finally reaching your ears.
That chain has at least 10 to 20 links ... and you are worrying about further polishing the one which arguably is the most transparent of them all.
As said above, .05% distortion pales when compared to 5/10% (or more) distortion in speakers and what's present on other elements.
I would not go too far overboard hyping the distortion of loudspeakers without some qualification about what you are talking about. In other words, while there is a grain (at least) of truth in the percent distortion numbers thrown about in this thread it is not accurate to say that these are an accurate depiction of the issue across the entire frequency spectrum that the loudspeaker produces.
I like to build loudspeakers, and I do so from a "measurements" perspective, meaning I measure performance of the drivers and system as a whole as I go. As an example of the distortion that I think you can expect from a good quality loudspeaker, I recently rejected using a certain midwoofer because below about 400Hz the distortion level exceeded 1 percent at about 85dB@1m SPL level. In my opinion, above 100-200Hz, this is a resonable cutoff between good and not-so-good driver performance. Distortion from a good tweeter can be much lower, in fact BELOW 0.1%, and that is just where the amplifier's distortion will be rising, so don't be so quick to say that the driver is much worse than the amp everywhere and all the time.
One place where driver distortion IS significant is in the bass range. Here the physics of sound reproduction by a direct-radiator requires that the cone moves, and significantly as the frequency decreases, if a reasonable SPL is to be achieved. Here is where the distortion climbs well above 1% and figures of circa 5% are not at all unrealistic. Even a high SPL capable, high performance subwoofer can easily have distortion exceeding 10% (!) below 30Hz if you are reproducing over 90dB@1m SPL. The motor and suspension of a driver are just not very linear as cone excursion increases, and this is a large source of the problem, although certainly not the only one. Furthermore, those cute 2-way systems with a 6" woofer trying to reproduce 40Hz are also likely pumping out a good amount of distortion. It's just what many have become accustomed to hearing, and so this largely goes unnoticed and unmentioned. One of the best "cures" for low frequency distortion is simply to use a larger diameter driver, thus requiring less cone excursion (all other things being equal), although a larger enclosure as well. Since small 2-way systems earn high WAF and are popular, this concept does not seem to be of much importance to most people.
I like to build loudspeakers, and I do so from a "measurements" perspective, meaning I measure performance of the drivers and system as a whole as I go. As an example of the distortion that I think you can expect from a good quality loudspeaker, I recently rejected using a certain midwoofer because below about 400Hz the distortion level exceeded 1 percent at about 85dB@1m SPL level. In my opinion, above 100-200Hz, this is a resonable cutoff between good and not-so-good driver performance. Distortion from a good tweeter can be much lower, in fact BELOW 0.1%, and that is just where the amplifier's distortion will be rising, so don't be so quick to say that the driver is much worse than the amp everywhere and all the time.
One place where driver distortion IS significant is in the bass range. Here the physics of sound reproduction by a direct-radiator requires that the cone moves, and significantly as the frequency decreases, if a reasonable SPL is to be achieved. Here is where the distortion climbs well above 1% and figures of circa 5% are not at all unrealistic. Even a high SPL capable, high performance subwoofer can easily have distortion exceeding 10% (!) below 30Hz if you are reproducing over 90dB@1m SPL. The motor and suspension of a driver are just not very linear as cone excursion increases, and this is a large source of the problem, although certainly not the only one. Furthermore, those cute 2-way systems with a 6" woofer trying to reproduce 40Hz are also likely pumping out a good amount of distortion. It's just what many have become accustomed to hearing, and so this largely goes unnoticed and unmentioned. One of the best "cures" for low frequency distortion is simply to use a larger diameter driver, thus requiring less cone excursion (all other things being equal), although a larger enclosure as well. Since small 2-way systems earn high WAF and are popular, this concept does not seem to be of much importance to most people.
Indeed, 2x 300W/8R 30cm 90db/W/m woofers per speaker at normal listening levels in an accordingly sized enclosure do pretty well. I don't go by WAF btw. She needs to have a decent S(peaker)AF instead 😉
Last edited:
Another meaningless thread. Search you already have a few on the subject.
BTW THD is more or less meaningless number.
BTW THD is more or less meaningless number.
- Status
- Not open for further replies.