Eva, you seem to communicate only with certain members in this thread trying to make others who communicate directly with you look like fools. Maybe your level of knowledge, which seems to rise above the rest, prevents you from doing so, but that's fine with me. I will get my answers elsewhere from people who are able and willing to communicate in a normal way, such as some here in this thread have allready done.
you no complete wrong glene but more nearing if playing that way. Your efficiency improve A is now H you seeing, you modulate too and less smart way you still need more supply and heatsink at power or limiting.. maybe you make up for with smps or oh.. ok. Class D do having crossover distortion of kinds to self, one of kinds not zero crossing but HF spray event. Silly thinking feedback and filter will attenuate this to non matter some class D run hot air. About distortion capability of D amp, hello from Jupiter giant gas ball of high density, your amp playing my field. You know you discuted party thinking all amp same sounding if below 3% THD and kilo power clipping at 200hz anyway. I agree you saying A+D silly thing for other academic dead weight lift. Pollyannas this form filled thinking best of both planet, realisty find it worst of them. One not crutch for other, D potential standing for self fine even if most not good that bad design like A+D, but no laugh me with efficient A OK. That all gas till ing this day. Now this can go back to saying how it sound for all importance, Pollyannas want cracker.
courage said:
What you should figure out from what I write is that there is absolutely no point in making such an hybrid amplifier. You will find little sources because the people that attempted this were motivated by:
a) They didn't know exactly what they were doing.
b) They were just trying to take advantage of the meaning of "class A" in high end audio marketing terms.
The classic advantage of class A over class B or AB is a continuous open-loop transfer characteristic with no jumps, dips or slope changes near zero current crossing. Class D already provides this without adding anything.
The reduced THD of an hybrid approach is achieved with just class D alone and increased negative feedback. The efficiency lost is never recovered. The increase in complexity is never paid for.
Eva said:
LOL! :bs:
You really need to brush up on your linear amp theory. ~0.001% THD at 1kHz was done in a basic 50W "blameless" with 30dB loop gain at 20kHz (giving ~55dB at 1kHz) and ancient (now no longer made, dating back to the 70's) 2MHz power transistors in the output.
With two pole compensation the loop gain is easilly boosted significantly and the THD-1kHz drops well into the noise floor.
As for 1980's MOSFETs, Bob Cordells EC MOSFET amp did 0.0006% THD at 20kHz (immeasurable, <0.0001% at 1kHz) in 1984. Look up the paper on his website!
Modern RET's allow power output stages with -3db open loop bandwidth >4MHz (VMOSFETS even higher) and closed loop bandwidth in excess of 1MHz in basic "voltage feedback" designs.
Linear power output stages are also easilly wrapped up in local current feedback stages, which allow unity loop gain frequencies approaching or exceeding the -3dB bandwidth of the power output stage itself, reducing the distortion by an order of magnitude or more BEFORE being wrapped up in the global nfb loop of a conventional voltage-feedback amplifier block.
And there are still further tricks (two pole comp, transitional miller comp, NDFL) to boost the loop gain even further. Read the last half dozen pages of the Negative Feedback thread on the SS forum and look up the PGP amp (<0.0001% THD at any frequency and power below rated maximum), for just one example.
Eva said:
However, there are new MOSFET families soon to be released with lower gate charge, lower capacitances and lower Qrr body diodes for given Vds and Rds-on requirements. These will allow to use higher switching frequencies and more negative feedback (like 500Khz-1Mhz with rails well over +/-45V) without efficiency loss, and do the same: "Blameless" class D
The point that I was trying to make is that in class D you have a much more linear open-loop characteristic to start with than in class B or AB because the parameter that drives the output is time rather than voltage or current (which are subject to all kind of transistor non-linearities).
Riight.
This person who started this thread proposed a hybrid class A/D idea for improving the linearity of a Class D amplifier.
Doing so offers no practical advantage in terms of heat dissipation and zero performance advantage over that achievable with more advanced linear class B design.
At least you seem to agree that such a contraption is pointless.
Also, please let me know when you've got your class D MOSFETs switching fast enough to return a THD-20kHz of less than 0.0001% at rated power.
Eva said:
My attention was drawn to the information in this website:
http://www.cruz-digital.com.
It seems they carried on where Audio Physic left with their no longer available Strada amplifier. I'm not sure wether the Cruz amps are a combination of A + D topologies, but the amps seem to be quite expensive and rarely known. My knowledge to judge wether it made sense to follow the approach they have taken is limited, so I cannot comment on that. What I however conclude is that their approach makes for a pretty complex design.
Plainly stated; I am looking for a power amp with the Class A sound, little dissipation/distortion, a wide bandwidth (open loop if possible) and enough power to drive my Infinity Renaissance 90 speakers which occasionally dive to 2 Ohm. Most class D amps I've seen specify their power ratings at 8 or 4 Ohm. I also notice that if load halves, the power rating isn't really doubled. Now these are of course just figures, but I think important to know what happens at loads of 2 Ohms.
Thanks for your input
All that stuff with THD in the 0.001% range used +/-35V supply rails or lower. There were no high power amplifiers (like +/-120V) with so low THD. Nowadays you can do 50W class D with uniform THD well below 0.01% (there are no strong non-linearities to correct, it's mostly noise what increases THD, not harmonics). See ICEpower 250A or UcD180 THD plots.
Cordell's error correction are just another 40dB of negative feedback on top of the main loop to patch the absolute lack of linearity of MOSFET class AB. I love figure 9 on page 10 of his paper, it tells with an image what I'm trying to explain with words. At least, bipolars produce a flatter plot. MOSFET are just switches without those 40dB of correction.
Cordell's error correction are just another 40dB of negative feedback on top of the main loop to patch the absolute lack of linearity of MOSFET class AB. I love figure 9 on page 10 of his paper, it tells with an image what I'm trying to explain with words. At least, bipolars produce a flatter plot. MOSFET are just switches without those 40dB of correction.
Eva said:
An extra set of prehistoric output devices and a rail boost would give you 100W with the same ~ or < 0.001% THD. Make it a bridged amp and you've already got a 80V peak amp that would blitz any comparably (power) rated class D.
I just looked up the UCD180 datasheet. Nice, simple compact design. But it is a 180W amp and is already well over 0.02% at 50W (1kHz / 4ohms). At 100W it is approaching 0.2%, and that is well out of the noise. No strong non-linearities?
But at 15W it's 0.005% flat to 20Khz. And this is achieved with minimal negative feedback. This is the linear region I was talking about. It ends at 20W or so for that amplifier. Since it's a current dependent penomena, double these power levels for 8 ohm (so you get very close to 50W "clean" watts, and some headroom for parties ).
Check this one too : http://www.icepower.bang-olufsen.com/files/solutions/icepower250adata.pdf
But it does not have frequency independent THD.
).Check this one too : http://www.icepower.bang-olufsen.com/files/solutions/icepower250adata.pdf
But it does not have frequency independent THD.
Eva is absolutely correct, the output devices now will not switch as
efficiently at more then 100 volts. The body diodes are slower and the
gate charge/ capacitance become tougher to drive quickly.
There is also the fact that the resolution of the amplifier decreases
as the rail voltage increases for a given switching frequency.
ie 350khz at 75W or 200W of output.
btw I am using +- 43 volts on the rails of my amplifer
this is good for 225 W rms into 4 ohms
efficiently at more then 100 volts. The body diodes are slower and the
gate charge/ capacitance become tougher to drive quickly.
There is also the fact that the resolution of the amplifier decreases
as the rail voltage increases for a given switching frequency.
ie 350khz at 75W or 200W of output.
btw I am using +- 43 volts on the rails of my amplifer
this is good for 225 W rms into 4 ohms
G.Kleinschmidt said:
Because it works with almost no heatsink and no bulky banks of output devices and uses half the power because very little is wasted... It can be made small and light weight and has half the power supply requirements...
One of my prototypes was doing 100-120V bass on 2 ohms for almost 5 hours last saturday night. Heatsink never got more than 22ºC over ambient. 50W THD is .00x (below my measurement capabilities). You could call it 50W hi-fi with a *lot* of headroom
it's not really much bigger than 50W class AB.
Glen is true d amp can be as very best, immeasurable, and already this day. To saying could be better than amp of same caliber of other class, more efficient yes, otherwise equivalent, still with all having own problem set per class, unless selling me something then all gravy.
Still saying you not so wrong, so if true other can no be 100% anything. You least keeping more point and you not even d guy. Problemm you seeing in discusting is opponent here for sale. Almost every post in almost every thread as record with worn line my new prototype my new prototype my new prototype, no can see but can buy, my new prototype call me. Someone saying new prototype? Who having new prototype? Thought heard something for new prototype somewhere.. no can remembering where.
Still saying you not so wrong, so if true other can no be 100% anything. You least keeping more point and you not even d guy. Problemm you seeing in discusting is opponent here for sale. Almost every post in almost every thread as record with worn line my new prototype my new prototype my new prototype, no can see but can buy, my new prototype call me. Someone saying new prototype? Who having new prototype? Thought heard something for new prototype somewhere.. no can remembering where.
G.Kleinschmidt said:
Who was arguing about efficiency? I was under the impression that class D was setting new, unmatched standards in linearity
It's not setting new standards, it's coming closer and closer to the existing ones
probably to the extent of replacing linear amplifiers because the smaller and smaller linearity improvement is no longer worth the expense.
http://ap.com/products/peripherals/apx581
http://ap.com/downloads/file/148
I am not sure that the way used nowadays to measure THD in class-D amplifiers is fair.
http://ap.com/downloads/file/148
I am not sure that the way used nowadays to measure THD in class-D amplifiers is fair.
Glenn,
I've done some extensive reading and also consulted a friend of mine who is a succesfull high-end audio designer on the issues regarding class D and class A amplifiers (not necessarily a combination of these two!).
I have decided to go for class A and Andy C has kindly sent me the AES papers you mentioned earlier.
Once more thanks for your valuable input.
I've done some extensive reading and also consulted a friend of mine who is a succesfull high-end audio designer on the issues regarding class D and class A amplifiers (not necessarily a combination of these two!).
I have decided to go for class A and Andy C has kindly sent me the AES papers you mentioned earlier.
Once more thanks for your valuable input.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.