The 'No' is fitting to what we can expect from theory.
But isn't a simple 'No' a little bit compressed?
The only thing which becomes better with high frequency is that one can design the filter with a higher bandwidth.
In case of post filter feedback the loop gain can be adjusted for faster step response. An even stronger commercial motivation is that the output filter with higher bandwidth is smaller.
Everything else becomes worse with increasing frequency (at least in implementations with traditional analog modulators, where you do not loose information by lower switching frequencies).
But isn't a simple 'No' a little bit compressed?
The only thing which becomes better with high frequency is that one can design the filter with a higher bandwidth.
In case of post filter feedback the loop gain can be adjusted for faster step response. An even stronger commercial motivation is that the output filter with higher bandwidth is smaller.
Everything else becomes worse with increasing frequency (at least in implementations with traditional analog modulators, where you do not loose information by lower switching frequencies).
As the owner of a class D amp (the CLASSic-D) from silicon chip, the output is dead silent. Totally absolutely dead. Had to double check for power.
There IS a tiny bit of much higher frequency noise, way beyond the capabilities of human ears and tweeters.
I use a pioneer sc07 elite receiver which is as I think based on what is called class D ice power amplifier, why would a somehow respectable company like pioneer use a no "air" amplifier in its at the time one of their flagship receivers. As for sound, I believe "subjectively" as I can't compare to other receivers, that the sound of that receiver is very clean and feels "accurate".
At this same moment, I´m listening di.fm from internet, in a PC using some cheap speaker cabinet, and a self designed class d operating at 280KHz from the SMPS of the motherboard (Yellow and blue, +12 and -12V lines). Previously, the same setup except a STK series amplifier (Listening level about 1-2W no more for a job office).
The conclusion is that Class D is more clean and better stereo than class AB. Some of my co-workers said the same.
The conclusion is that Class D is more clean and better stereo than class AB. Some of my co-workers said the same.
The Devialet (highly praised in the press and in which I had great expectations) I heard at an audio fair last year, had that effect that Chocoholic described here in a post early this year: An overly “analytical“ top end and vieled in a frequency range just below that. A very strange (and annoying) effect I never heard before.
Well, if you mean that there was plenty of way overpriced and at the same time mediocre sounding stuff presented by BS talking clowns, you are certainly right.
On the other hand it's an opportunity to hear for yourself and there was some good stuff (Stax and Sennheiser headphones, refurbished Quad electrostats, Garrard turntables, etc.) too.
On the other hand it's an opportunity to hear for yourself and there was some good stuff (Stax and Sennheiser headphones, refurbished Quad electrostats, Garrard turntables, etc.) too.
I'm guessing that living in a time capsule where Quad, Stax and Garrard are revered, there is little interest in new stuff?
I have a Quad II tube outfit and Thorens TD124II that have been gathering dust on a shelf in my basement for 30 years and they were old then!
Would I be correct in thinking that it is a younger generation that finds a mistique in an idler driven record player?
I'm addicted to these little amps, they are so inexpensive that the comments of the bat eared experts are hardly necessary, I just gift myself and form my own opinions, it's fun.
One would assume with such a high sample rate high powered Class D amplifiers (Lab Gruppen, Powersoft, PKN) offer, Class D amplifiers would replicate high frequencies just as good as the other classes. I would imagine there is more to it than having a wide enough bandwidth needed in order to offer the high frequencies many are claiming Class D is lacking.
Harris and others have been manufacturing switching AM transmitters for many years now. In fact, tubes are non-existent in new MF AM broadcast transmitters. There are a few ways of implementing the PA stage. You can use a switched stairstep approach (basically a power DAC). You can do a symmetric PWM (extremely difficult to do in practice). Or, you can do class E, which is popular among radio amateurs (which I am) for homebrewing AM transmitters. This is the simplest way to build an AM transmitter: you have a MOSFET PA stage which runs at the transmit frequency (in the 160M or 80M bands, since getting higher freqs from regular switch MOSFETS is problematic due to their lead/bond inductances, although IRF DirectFETS can probably do 40 meters or even 20 meters just fine). Duty cycle is fixed at 50%. Modulation is performed by directly modulating the power supply with the audio. This can be done the old way, using a transformer, but is more commonly done as a PWM power supply where the vref is mixed with the audio to modulate the output voltage. The output from the PA stage is fed through a LPF to eliminate the harmonics from the PA square wave, and fed to the antenna. Despite the complexity, these are actually easier to build vs a linear SS amp for SSB or FM.
I did a switcher for a mobile tube amp. I ended up making it a voltage doubler output since I couldn't fit enough turns in the toroidal core I was using. Back then, HV SiC Schottkys didn't exist, so I had to use FREDs.
For home tube amps, I still prefer tube rectifiers with resonant choke input filters. If it doesn't have to move, I could care less how much it weighs, nor do I care about the extra watts burned
It's called phase noise. Low phase noise oscillators are easy enough to make, so it shouldn't really be a problem. In RF systems the PLLs used to synthesize higher frequencies are often the cause of excess phase noise, but class D audio amps never use PLLs, so this cause of phase noise is absent. I've built a few clas D amps, and they have all been silent. Well, audibly silent, anyway. RF emissions are a whole 'nother matter, in which pretty much every implementation of Class D amps in existence suck due to inadequate filtering.
I have several d-amps and they all sound very, very good to me.
Perhaps I am lucky to have been born without "golden ears".
Well, your speakers and listening environment also play a big part. I have heard the best high end switchers, and IMO linear tube and SS amps still sound much better.
I can attest to this. I have run 1MHz, and it is not a picnic.
Last edited:
I have been enjoying a $17 TPA3116 board for the last couple of weeks with a 12ft+ long speaker wire and I hear no noise at all connected to internet radio ,FM, CD, and MD.
I never, ever, listen to AM but connected a portable radio with AM and could detect no noise.
The TPA3116 is just the bare board placed on a plastic mat, I have added no ferrite beads or other suggested filters.
So, what is this noise you guys are discussing, is an omnipresence that can only be heard by true believers?
Actually, the reception on my old Sony 5000F FM tuner is much better when connected to the D-amp than when it was connected to a McIntosh SS amp. Go figure.
Now, I could not use FM on TA2020 D-amps, it was terrible.
I never, ever, listen to AM but connected a portable radio with AM and could detect no noise.
The TPA3116 is just the bare board placed on a plastic mat, I have added no ferrite beads or other suggested filters.
So, what is this noise you guys are discussing, is an omnipresence that can only be heard by true believers?
Actually, the reception on my old Sony 5000F FM tuner is much better when connected to the D-amp than when it was connected to a McIntosh SS amp. Go figure.
Now, I could not use FM on TA2020 D-amps, it was terrible.
...ugh, this thread is really messing with tons of different effects.
Must be in the nature of 'air' discussions.
TI has spend quite some money in R&D to get the disturbances and the EMI right. Their efforts were in their silicone design as well as in their PCB proposals. Radiosmuck's comment sounds like this is now showing fruits.
Regarding data sheet values we have to keep in mind that they are not always comparable, even if they look like and are intended to be comparable.
Must be in the nature of 'air' discussions.
TI has spend quite some money in R&D to get the disturbances and the EMI right. Their efforts were in their silicone design as well as in their PCB proposals. Radiosmuck's comment sounds like this is now showing fruits.
Regarding data sheet values we have to keep in mind that they are not always comparable, even if they look like and are intended to be comparable.
Guys, there may be different opinions about noise in Class D - good or worse than linear amp? BUT:
One thing should be made clear to everybody: "My amp is dead silent" and the like does not count, if the efficiency of the tweeter as well as the listening distance is not mentioned. At some point there WILL be noise. Through directivity and impedance transformation, good horns can produce up to 116dB /1W /1m
Air Array
Get your ear close to this horn and you would hear noise down to single picowatts.
One thing should be made clear to everybody: "My amp is dead silent" and the like does not count, if the efficiency of the tweeter as well as the listening distance is not mentioned. At some point there WILL be noise. Through directivity and impedance transformation, good horns can produce up to 116dB /1W /1m
Air Array
Get your ear close to this horn and you would hear noise down to single picowatts.
...old designs... hm, class D technology has its roots in the 70s, but at that time the switching devices were so poor that no good sounding full range solution was possible.
Since the late 90s classD is taking a restart and has evolved good enough that the differences are less obvious - often hard to notice at all.
Still classD is not a mature technology and will evolve further.
Very true, so we should stick to a simple bandwitdh limited measurement of the output, resulting in simple number of xx uVrms noise.
Or use a sound card and RMAA to get the spectral density.
Using our ears as a comparable measurement instrument would need more than just the efficiency of the speakers and listening distance: The sensitivity curve of the listeners ears.
Dead silent is simple for me (43 years) with my Dynaudios (87db/W).
For properly designed classD amp, I need to connect my headphones (beyerdynamic DT311 / stoneage) directly to the amp output and need a silent room in order to get an audible impression of the nature of the noise.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.