And during this you revealed how person you are. Not funny.
Again: learn writing! At least 3 errors in 2 words is too much for me to correct them.
2 serious mistakes in 1 sentence (as usual). Losses doesn't mean saturation limit is close. You ignored both DC copper loss and eddy current loss. The other mistake is that you ignored that I tested it not at average level but at more than the rated current. With music none of my cores heats up more than 40 C above ambient, most of them do 2...10 C during real usage with max volume. And to check saturation limit I make other tests and calculations.
You are not a good listener for a forum. And no matter how you may surprised, this is a forum, not a concert hall. Here the only information is via sentences and attachments. You can't listen to sounds, you have to listen to sentences. But unfortunately you can hear only what can fit into your false generalisations based on your limited experience and almost nonexistent theoretical knowledge base.
Very reasonable... best in planet... I guess you tried all. You must be some smaller god.
Best core? You didn't even specify the material, only size! You can't believe material is not important! I know you think about the -2 material, but your imagination is so hardly limited that you don't think there are other options. Like molipermalloy, multi-gapped ferrite, etc. Have you ever tried to achieve (or to measure) Q factor over 200 at 200 kHz? Or optimizing for high power density instead of big size? Do you think big size is a goal? I guess yes. Your Lina amp is large. Not powerful, but large indeed, so the name is correct. The specification is less correct. 2000 W comes from it at stable power supply (big, expensive, not diy) with very high distortion and switching freq decreased to audible range. I never let a full-range amp to switch at audio freq. And I make nominal power available with about 1% THD at dropped power supply voltage. What you call 2000 W is 1200 W specified by me.
Experience? Do you mean should I learn how to copy and paste? You shared no experience. You mostly implied (instead of explicitely stating) your false generalisations in badly composed and spelled sentences. Please list the occasion where you explicitely stated an experience? As I can recall your last post contained your only real experience (video showing sparks during OC testing). What else?
Experience is not priceless, it is the most expensive. It costs both time and money and mental effort. And an experience is personal by nature, it cannot really be shared. I can't take account your experience, since you told almost nothing about it.
You ignored the experience of the builder of your D900 schematic. Also my simulations. Also my explanation why changing input polarity is beneficial. And many-many other things.
I don't prefer sharing experience, because it is personal and specific, and very hard to transfer and interprete correctly. Theory and logic in contrast are able to be shared, and to be checked also. And (correct) theory makes possible to arrange the very high amount of experience into a usable structure that makes possible to predict the result of a future experiment. This way of thinking let humankind come out of the caves and bulding skyscrapers.
Sorry Pafi, too much blah blah but I can answer shortly:
1. Micrometal T157 is very specific product
2. My amps are tested both by dummy load and on live concert. 2000W at 4 ohm.
3. You do not need to share anything. I already smell you have no working amplifier at all
You have a good talent in long writing and unfortunatelly I hate reading so much.
End!
This way it is perfectly usable, however I apply a precisely set limiter to these kind of amps to really avoid the chance of downmodulation.
I agree. Even lower switching freq is possible at good quality with different modulator (for example with uniform sampling and pre-distortion). The theoretical limit is 40 kHz.
Such product not exist. It is a shape (Toroidal) and size. Material is defined with a -number. If you don't care what number, then use T157-26 and see if you can cook on it!
At what distortion? You've never measured it.
Your nose must have been entirely damaged by your burnt amps. My ClassD amps are in use by many persons since 2002 and many person find me by their recommendation both for building and for development and repair (for example the very hardly serviceable velodyne) also.
Don't read and don't write if you can't, but don't try to mislead everybody!
Yesssssssss
But if we ask for the pcb design
You cant share.
Anybody who wants to convince there is one way, publish a complete set of files.
In any other case i can find thousands photos of amplifiers to post here
KW OR MW
I think this recommendation is for you.[emoji12]
You only bluff and bla-bla [emoji38] . You demand for distortion measurements , but yourself dont show them.
Maybe you are clever , but also a great boaster. [emoji57]
I remind you that we are on an international page where we speak different languages and English is not the same, in my case I use a translator I hope they understand me the same. In my opinion they should put their circuits to do it and that way we can hear how they sound and if anybody can contribute some changes of how to improve the circuit well come on, that is called collective intelligence [/ B], but If they want to do only their circuit and not contribute anything so they are here then. Just to presume they know something.
This is not true. I never demanded distortion measurements, but I provided result of D900 distortion sim.
Asking at what distortion did he state 2kW output power is not a demand for making distortion measurement. It is a neccessary circumstance to make the power value meaningful. Without stating the way of this specification was made it is simply meaningless. Sine power at 1 %, or PMPO, or basic calculation assuming ideal amplifier are very different things. And nobody asked me to do any measurement here, so I had no reason to publish one. This is very easy to understand, and if your goal was not to attack me you also realized this.
And your next personal insult will be processed by moderator. To make things worse you are perfectly offtopic.
Carefully designed self destructing methods
like the very rough DC offset setting with some positive feedback and lack of overvoltage protection. While with fewer parts a correct (negatively fed back) and much more precise offset nulling could have be made. Or a well designed input stage could do the job without any trimming/matching.Again you bluffing
That sparks come from dirty contact at high energy. I am an electrical engineer of biggest electrical company in the world. My job is designing substation up to 500kV. Stop telling nonsense!
You can't accuse me bluffing without any proof! If you do this then you are bluffing.
Dirty or not, where the energy comes from? If your protection was fast, then only normal operating energy from output inductor and capacitor could give energy to the spark. (You don't think dirt was explosive, do you?) The sparks from your amp are too high for this. Protection must be slower then should be. Considering your xLina schematic at a sudden direct short current starts to increase by 6 A/us. At about 50 A for the low side and 40 A for the high the BJT starts discharging the 1 uF cap, but since current is not infinite it needs time. To discharge 1 uF from 12 V to 4 V with 400 mA you need about 10 us (I considered relatively high ESR of electrolytic). During 10 us the current grows further by 60A, thus the current peak before interruption is more then 100 A. This overcurrent accumulates the high amount of energy in inductor needed to produce such strong sparks. Fortunately semiconductor designers at IR were real engineers and they made IRFP4242 capable to withstand such abuse with some luck, but why torture them?
A correct overcurrent protection is much faster.
I will pray for indonesian substations.
To kartino
Tell me why didnt you reply me when i asked you "amps ocp dont work" what to do?
And
Pls tell me why you remove me from that next generation class d amp facebook forum?
Just because i laugh on you
Ihave just asked you for pcb design only.....
Let it be. You are GREAT. But not if you cant share.................?
Tell me why didnt you reply me when i asked you "amps ocp dont work" what to do?
And
Pls tell me why you remove me from that next generation class d amp facebook forum?
Just because i laugh on you
Ihave just asked you for pcb design only.....
Let it be. You are GREAT. But not if you cant share.................?
Carefully designed self destructing methods
This is a stupid statement. If you repeat this kind statement I will bring this to moderator.
My schematic has been tested for live performance
DC offset range is 200mV max. Trimpot shall be trimmed at middle post
This is not 1 statement, but at least 7 statements. This is all what relevant to stupidity here.
Sorry for kindness!
Please do it! At least you will be moderated. Do you think analisation of a posted schematic is against the forum rules???A schematic is a drawing. How you can test a drawing "for live performance"? Even if you ment the circuit, what is the relevance of live performance to DC offset? What do you mean 200 mV max? After offset nulling? Or allowed for live performance? Or at middle position of pot? Or 0...100 % setting of pot? Is it possible for you to write meaningful specification? (And what about correct spelling? Do you intentionally make even more harder to understand you? Or don't you care your own thoughts enough to read them?)
UcD xLina V16 contains 10k pot with 2x33k series resistor run from +/-12V. Its internal resistance + series 4k7 result in 24k source impedance. No-load voltage is +/-0.8V during min to max setting. This 0.8V/24k=33uA is injected to inverting input and on feedback resistor converted to +/- 0.7 V output voltage. This is already too much, but it wouldn't be a real problem if it was stable. With precise setting 10 mV precision was achievable, but there are more effects that can change the set value significantly:
- zener voltage drift. Asymmetrical changing of zeners cause almost as much offset drift as the asymmetry. 10-20 mV happens easily.
- ageing of 33k. 1 % difference error makes about 50 mV offset drift.
- positive feedback from supply rails. I will not calculate this here, but it is not negligible.
- Temperature dependency of LTP.
And DC offset in the order of n*10 mV is dangerous. When? I already wrote it. When a low impedance is applied to output without input signal. DC output current transfers current from one rail to other. Typical 10 mV 50 mohm is already 200 mA output current, which means 100 mA out from one rail and 100 mA into the other. It may survive the increasing supply voltage, may not. And at much more DC offset something almost surely dies.
And the solution of this problem is terribly simple. After understanding the effect chain I described every person familiar with analog electronics must be able to design the correct solution, but I let everybody to have fun with thinking about it for some days.
Maybe once you will be able to understand a post. This was not that time. I did not write anything about the existence of the spark. I wrote about its strength.
This is not 1 statement, but at least 7 statements. This is all what relevant to stupidity here.
Sorry for kindness!
A schematic is a drawing. How you can test a drawing "for live performance"? Even if you ment the circuit, what is the relevance of live performance to DC offset? What do you mean 200 mV max? After offset nulling? Or allowed for live performance? Or at middle position of pot? Or 0...100 % setting of pot? Is it possible for you to write meaningful specification? (And what about correct spelling? Do you intentionally make even more harder to understand you? Or don't you care your own thoughts enough to read them?)
UcD xLina V16 contains 10k pot with 2x33k series resistor run from +/-12V. Its internal resistance + series 4k7 result in 24k source impedance. No-load voltage is +/-0.8V during min to max setting. This 0.8V/24k=33uA is injected to inverting input and on feedback resistor converted to +/- 0.7 V output voltage. This is already too much, but it wouldn't be a real problem if it was stable. With precise setting 10 mV precision was achievable, but there are more effects that can change the set value significantly:
- zener voltage drift. Asymmetrical changing of zeners cause almost as much offset drift as the asymmetry. 10-20 mV happens easily.
- ageing of 33k. 1 % difference error makes about 50 mV offset drift.
- positive feedback from supply rails. I will not calculate this here, but it is not negligible.
- Temperature dependency of LTP.
And DC offset in the order of n*10 mV is dangerous. When? I already wrote it. When a low impedance is applied to output without input signal. DC output current transfers current from one rail to other. Typical 10 mV 50 mohm is already 200 mA output current, which means 100 mA out from one rail and 100 mA into the other. It may survive the increasing supply voltage, may not. And at much more DC offset something almost surely dies.
And the solution of this problem is terribly simple. After understanding the effect chain I described every person familiar with analog electronics must be able to design the correct solution, but I let everybody to have fun with thinking about it for some days.
I dont need to read all your post.
An experienced guy only tells in brief. Like your amp I can smell only from your picture because I am really an amp maker.
That stupid comments related ti my amp is wrong as the amp has been built many units and tested on real live performance.
That DC offset setting is working fine. Only require pre heating. Runs the amp for several minutes with closed box and fans ON.