Go Back   Home > Forums > >
Home Forums Rules Articles diyAudio Store Blogs Gallery Wiki Register Donations FAQ Calendar Search Today's Posts Mark Forums Read

Class D Switching Power Amplifiers and Power D/A conversion

iraud200 short story
iraud200 short story
Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 7th November 2017, 03:04 AM   #71
mgshightech is offline mgshightech
diyAudio Member
 
Join Date: Nov 2012
In case there's confusion ... this data point will help clear things up ... the heat sink on the 2092 on my 1 kw board with irfb4227 has roughly 15 to 20 times more surface area than the chip, and it's hot enough to be slightly uncomfortable to keep your finger on it ... so that's around 50C. So, around 20-25C in temperature rise ... multiply that by even 10, and you get too hot for a chip to survive. Multiply it by 15 and you get ... definitely a toasted chip. I wouldn't try this board without the heat sink.
  Reply With Quote
Old 7th November 2017, 09:16 AM   #72
rafsky is offline rafsky  Italy
diyAudio Member
 
Join Date: Oct 2007
Send a message via MSN to rafsky
Quote:
Originally Posted by mgshightech View Post
In case there's confusion ... this data point will help clear things up ... the heat sink on the 2092 on my 1 kw board with irfb4227 has roughly 15 to 20 times more surface area than the chip, and it's hot enough to be slightly uncomfortable to keep your finger on it ... so that's around 50C. So, around 20-25C in temperature rise ... multiply that by even 10, and you get too hot for a chip to survive. Multiply it by 15 and you get ... definitely a toasted chip. I wouldn't try this board without the heat sink.
With all due respect, electronics is part of the whole of science, it is not based on randomness, empirism, or even worse than the probability that something works or not.
The irs2092 has a pilot stage that defines miserable is to give it a compliment, it was designed and built to drive lightweight mosfets Qg <60nC, all if you want an amplifier that works according to class D rents with a frequency of switching is acceptable, if we are satisfied with a malfunctioning craze with high distortion rates, huge heatsinks on components that should not be dissipated and other devils, though you are free to pinch Chinese stuff that originally does not work and modify it to make it malfunctioning, a single amplifier that works for a year does not make statistics, following that philosophy many of those modules are going to fail, so if you do not want to run the risk of losing your $ 20, change the mosfet with something more consistent with the purpose, do not challenge the fate.
  Reply With Quote
Old 9th November 2017, 03:20 PM   #73
MerlijnD is offline MerlijnD  Netherlands
diyAudio Member
 
Join Date: May 2011
Quote:
Originally Posted by mgshightech View Post
In case there's confusion ... this data point will help clear things up ... the heat sink on the 2092 on my 1 kw board with irfb4227 has roughly 15 to 20 times more surface area than the chip, and it's hot enough to be slightly uncomfortable to keep your finger on it ... so that's around 50C. So, around 20-25C in temperature rise ... multiply that by even 10, and you get too hot for a chip to survive. Multiply it by 15 and you get ... definitely a toasted chip. I wouldn't try this board without the heat sink.
Hey, ive imported a version of the Iraud200 directly from china for 10$.
No idea which version this is, but it looks alot more **** than yours.
Can you tell me anything about it?
Click the image to open in full size.
Click the image to open in full size.
Click the image to open in full size.
  Reply With Quote
Old 9th November 2017, 11:32 PM   #74
mgshightech is offline mgshightech
diyAudio Member
 
Join Date: Nov 2012
Default science?

"With all due respect, electronics is part of the whole of science, it is not based on randomness, empirism, or even worse than the probability that something works or not.
The irs2092 has a pilot stage that defines miserable is to give it a compliment, it was designed and built to drive lightweight mosfets Qg <60nC, all if you want an amplifier that works according to class D rents with a frequency of switching is acceptable, if we are satisfied with a malfunctioning craze with high distortion rates, huge heatsinks on components that should not be dissipated and other devils, though you are free to pinch Chinese stuff that originally does not work and modify it to make it malfunctioning, a single amplifier that works for a year does not make statistics, following that philosophy many of those modules are going to fail, so if you do not want to run the risk of losing your $ 20, change the mosfet with something more consistent with the purpose, do not challenge the fate."

It appears that someone misunderstands "science".

There is a stunningly long list of serendipitous discoveries that changed the world. Here is a small sample. Science and serendipity: famous accidental discoveries | New Humanist

The compton effect is an extremely noteworthy one not on the list. Compton scattering - Wikipedia

I'm actually not an EE, but an information theorist or systems theorist. My background is roughly six university degrees. Basically .... nobody can beat me regarding paradigms on how to process information. And you have provided almost no information excepting engineers intentions. Science is in fact highly empirical. Interesting you would bring up the 60nc issue though, as the 4227 is rated at 70nc typical. OMG 17% extra load for 2000% more surface area to heat sink away.

reminds me of the dam someone built high up in a canyon .. made of huge boulders of solid granite, cemented together. Well, it was too old to have been engineered by a certified engineer, so they had to blow it up ... so they took explosives up to blow up the dam, and the explosives couldn't blow it up. So they had to get a gigantic load of explosives to blow it up. After the second try, they did get it to come down .. guess it never occurred to anyone that big granite rocks are far stronger than the concrete they use today. Compression and Tension Strength of some common Materials What really happened is that every engineer worth his salt would have known that the dam was over-spec. They just didn't want to put their signature on it for various political reasons.

That heat sink has 20x more surface area than the chip. You think that chip is going to go bad for driving <30% extra load... you're up in the night. You don't need any statistics to know that the potential output of pretty much any semiconductor will go up by more than 30% with a (relatively) big heat sink on it.

While we're talking heat sinks though.
The heat sinks from a single old ant miner similar to my current setup could easily handle six of these boards. What I have put together is overkill. The iraud200 probably needs a little more heat sink to do 2 ohms than what it's born with, but don't think it requires a whole lot.

There's no reason to expect high distortion rates. And from listening, I can't detect any (other than the hum caused by the goofy 200k feedback resistor) . Reality is what it is ... not what people intend. That's why we have "discoveries".

You do have one great point to your credit though.

This board is surely not the optimal solution. It's a fixed board that didn't work. If there were a comparable option out there that didn't need any work done on it, that would be preferable. The L30D, which is comparable to amp9, and does what you are suggesting with gate drivers has .. unfortunately .. a bad reputation. Sure electronics doesn't offer a comparable board for 2 ohms. Their 2-2.5 kw board won't pass any more current than the iraud200 because it uses the same output inductor. Unfortunately, it's 3 to 6 times more money.

And BTW .. people tend to overdrive these inductors. They'll put them on boards that can pass as much current as the inductors DC rating rather than rf rating. Then, what you discover after doing your homework is that the rf rating is really just limited by heat build up. .. so heat sinking fixes that.

If you want to drive a lot of power into 2 ohms .. this is the only good DIY solution out there that exists.. that I know of. Unless you want to design and build your own board. I'll be keeping my eyes open for other solutions.

Also BTW ... that's actually 2 boards that have functioned flawlessly for a year. I'm thinking of doing 4 more, but having a hard time scratching the time together for that, which is why I need to cut out now ... tata

:-)

Last edited by mgshightech; 9th November 2017 at 11:39 PM.
  Reply With Quote
Old 10th November 2017, 12:36 AM   #75
mgshightech is offline mgshightech
diyAudio Member
 
Join Date: Nov 2012
Default That board?

That is not the iraud200, but the iraud200-s

and you didn't reveal any info. about what transistors it is sporting. However, it appears that the output inductor is the 2 ohm version. If your transistors are 4227, that would confirm.

Your board is missing the overtemp protection, and it looks like the power supply transistor was moved onto the board, so it's not on the heat sink any more. You have a slightly different circuit. I don't know exactly how it will behave. It's possible that some chinaman read my docs and built a new, corrected circuit ... I don't know. But if its driving 4227, then the 2092 should have a heat sink on it, and turning that thing on makes me nervous because it has not overtemp protection. At least the best you can do is make sure the output transistors are greased, and get a heat sink on the 2092.

I don't see any output parasite dampener, so it will be needing that, .. and I can't tell if the feedback delay is there or not. But the easy way to find out is to go back to their claims document. If they are claiming 50 khz frequency response, then they took it out. You don't want them to take it out, so you would have to identify the feedback loop and mod it as this thread shows at the beginning. If it's only claiming 20 khz response, then they left it in, and you're good.
  Reply With Quote
Old 10th November 2017, 12:47 AM   #76
MerlijnD is offline MerlijnD  Netherlands
diyAudio Member
 
Join Date: May 2011
Quote:
Originally Posted by mgshightech View Post
That is not the iraud200, but the iraud200-s

and you didn't reveal any info. about what transistors it is sporting. However, it appears that the output inductor is the 2 ohm version. If your transistors are 4227, that would confirm.

Your board is missing the overtemp protection, and it looks like the power supply transistor was moved onto the board, so it's not on the heat sink any more. You have a slightly different circuit. I don't know exactly how it will behave. It's possible that some chinaman read my docs and built a new, corrected circuit ... I don't know. But if its driving 4227, then the 2092 should have a heat sink on it, and turning that thing on makes me nervous because it has not overtemp protection. At least the best you can do is make sure the output transistors are greased, and get a heat sink on the 2092.

I don't see any output parasite dampener, so it will be needing that, .. and I can't tell if the feedback delay is there or not. But the easy way to find out is to go back to their claims document. If they are claiming 50 khz frequency response, then they took it out. You don't want them to take it out, so you would have to identify the feedback loop and mod it as this thread shows at the beginning. If it's only claiming 20 khz response, then they left it in, and you're good.
Thanks for your reply, it contains the IRFB4620, even though i bought the 4227 version so I got scammed a bit. And yes it also says 50khz frequency response. Im going to rebuilt a subwoofer amplifier thats completely broken and I want to put this one in instead. So will it really matter if I fix the feedback loop for the 20khz?
  Reply With Quote
Old 10th November 2017, 09:27 AM   #77
rafsky is offline rafsky  Italy
diyAudio Member
 
Join Date: Oct 2007
Send a message via MSN to rafsky
The way you're talking in italy is called "supercazzola" that translated assumes the following meaning: talking about everything but actually speaking of nothing.

"I'm not really an EE, but an information theorist or systems theorist. My background is roughly six university degrees. Basically .... nobody can beat me about paradigms on how to process information. engineers intentions. "

If you are not an engineer, but a theorist of the information (that is to say, I do not know what it's like to say), taking for granted that your path of study keeps you away from the world of engineering, I have the information given to you but you also have to knowing how to read in the datasheet to find the right information, science is absolutely based on the evidence, for empirical, of course I would refer more to a concept of randomness, if we empirically want to assume the concept of perfectly repeatable proofs, then we talk about scientific method (Galileo Galilei taught us something) and I do not see what the example of the canyon, granite and explosive is about, what we are talking about is class D, nothing else.

I have given a parameter for the mosfets used in coupling to the irs2092, Qg (gate charge) PREFERABLY <60nC, here we are not talking about percentages, how much is the data compared to what our reference model is, in this case irfb4227, however, I invite you to open the component datasheet, go to the fourth page and look at the Typical Gate Charge vs.Gate-to-Source Voltage (n.10), the power supply voltage of the pilot is for 99% of the time 12 volts DC, at that gate voltage 4227 exhibits a gate charge close to 85 nC, not 70 nC, going to bounce off the percentage calculation (of no scientific value) you did, you need to know how to read the datasheet and not just stop at the first page of the presentation.
Assuming that my 60 nC is a parameter to NOT exceed, as in the highway (at least in italy) I can not exceed 130 km / h, in engineering there are no aesthetic parameters and practical parameters, science should be understood as a whole laws that allow us or to understand the world and its limits.

"That heat sink has 20x more surface area than the chip. You think that chip is going to go bad for driving <30% extra load ... you're up in the night. the potential output of pretty much any semiconductor will go up by more than 30% with a (relatively) big heat sink on it. "

So according to this theory if I put a sufficiently large dissipator I can dissipate 10KW with a pair of TO220, the thermal inertia parameter might be useful in the calculation and it is pretty much the limit that the CPUs have come to, which is why protections have been inserted (thermal throttling), if the heat exchange surface is very small and has a high thermal inertia (and this is the case of 2092 with smd plastic package) as big as the heatsink will fail to dispose of the temperature it accumulates in junctions.

Now let's not forget that the main parameter to evaluate the chip heating is beyond the Qg also the switching frequency, i do not know your amplifier's frequency but without any doubt if your configuration try to set it to 350Khz will be destined failing with a percentage of 100%
Otherwise, lowering it to 150-200 kHz, there are good chances to resist, let's not forget that the quality of an amplifier is subjective and what is acceptable for me may not be for you and viceversa, everyone has their own qualitative standard that if not supported by instrumental scientific evidence (hence only the ear) have no value. For me a class D with high dead time and low switching frequency sounds bad, instead for you it's great, i don't know.

"The L30D, which is comparable to the amp9, is not the optimal solution, but it is not a good solution. , and does what you are suggesting with gate drivers has .. unfortunately .. a bad reputation .. Sure electronics does not offer a comparable board for 2 ohms. Their 2-2.5 kw board will not pass any more current than iraud200 because it uses the same output inductor. Unfortunately, it's 3 to 6 times more money. "

This card is definitely not a solution at all, since it does not work, and not everyone has the ability to tweak it to make it work at least.
I have never talked about the L30D, because I have never tried it, but it seems like another Chinese solution, but between the two might be the simplest to me, maybe they will use the wrong components but the basic idea is the correct one to use a pilot buffer, they will definitely have used the 2n5401 and 2n5551 to drive the gate and it can only work badly, talk about cost, but sorry for me is stupid speech, if I have to spend my money to buy something I expect this thing to work, not to come to my house wrong.
The Sure modules work, they are perfect copies of iraudamps, so they cost 3-6 times as much but I would say that the difference is between a bad thing and a working one, so your speech does not stand, the evidence of what I say is that the Chinese themselves have changed their mosfets into new versions using the most consonants irfb4620
  Reply With Quote

Reply


iraud200 short storyHide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump

Similar Threads
Thread Thread Starter Forum Replies Last Post
Iraud200 mgshightech Class D 57 2nd January 2016 11:51 PM
Blown power tube - a short story 50AE Tubes / Valves 4 25th December 2011 07:00 PM
Does anyone know the name of this story? johngalt47 The Lounge 2 7th September 2010 11:29 PM
My story Dj BASS AMP Introductions 14 16th March 2003 10:43 AM


New To Site? Need Help?

All times are GMT. The time now is 02:03 PM.


Search Engine Optimisation provided by DragonByte SEO (Pro) - vBulletin Mods & Addons Copyright © 2018 DragonByte Technologies Ltd.
Resources saved on this page: MySQL 15.00%
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2018 DragonByte Technologies Ltd.
Copyright ©1999-2018 diyAudio
Wiki