Yet another version of IR audio amp ??

Status
Not open for further replies.
Jump to Latest
#81
The isolation for the transformer is quite simple....Just use a CSA or UL shrink tubing rated for 1500V over a 1000V rated litz wire, that was enought to pass the hipot test of 1500V. And best, temp rise of the toroid did not reach the max temp of 125 celcius for shrink. Look at ''sumi pac'' for real shrink tubing. For all one who ask for the core number, just to know that you can use nearly any core with 3f3 material, just to maybe change the primary to reach the point of minimal aidle current in the switching. The way we work here is that is faster to try than to simulate, so we take a core, only bobin the primary and find the number of turn who give the minimal idle current in the switch. At this point, because it was an unregulated switching power supply, it was easy to determine the ration for the secondary winding. If you whant to regulate it, just give 10% more than the desired output voltage with regulation and everything will be OK.

Just to let you know that I hate simulation software, they always give wrong result and take more time to debug a bad from the base design than directly experiment with real world parts...

Anyways, every one have is best way to do!

Bye

Fredos

www.d-amp.com

PS if anyone go to Montreal, go to the Parking night club, you will hear 48KW of pure sound...I have install 6X 8000HVI power amp for the main room yesterday, they have some Crown before and I can told you that Crown sound like s**t....No EQ have been done and now the sound was amazingly better (and louder too...)
 
#82
Hi Fredos,

Do you never simulate?? What simulators have you tried.... dont' say EWB.

You have to realize there's a number of people like myself who enjoy the challenge this art provides, and don't have the resources to do with it what we'd like.

Parts can cost alot of money, many can be scavenged and the odd free samples are saviors. Making PCB's is not at all cheap, I haven't done one yet, and to this day have been unable to procure an oscilloscope 🙁

So for me, spice is all I've got. It does have certain advantages.

Regards,
Chris

PS: I'd like to know what "whole new way of doing class d" might be, even if it proved to not work. Drop me an email if you like. It is good to hear of creative thinking.
 
#85
It is possible.
 

Attachments

  • 3.jpg
    3.jpg
    21.4 KB · Views: 977
#90
Dear Oldmaelstrom,

what type of protection - current limiter are you using?
Are you using the original PSU, or another one?
What are the optimal values of R26-C23 in your prototype?
Are the diodes after U12 working?

Thanks for the answer, best regards,
 
#92
Hello Fredos,

I am intrerested abaout the final power stage but in an total N-fet
The mosfet witch I hawe is irfp260n; What kind of changes i need to make in your schematics to achive this? (replace p-fet with n-fet)
Any schematics with n-fet power stage is welcomed.
The pwm I intend to use is a copy of JBL car amplifier.
 
#94
fredos ?

What do you think does it would be possible to simply inject some clocking signal into modulation stage, as I drawn in sch, and perhaps I do not need at all to remove trm. pot. of 5k, insted of this I just use a higher clock than is self oscillation freq. to be able to sync. those two.
So what is your oppinion ?
BTW. how is progressing your new class AD design, any pics etc. . . .

Best regards
Zeljko Kaiser
 

Attachments

#95
The best way in this schematics to add reference clock is to remove positive input from ground and feed reference triangle wave to this point!

I have already start a news tread about my final project, look at the news DLS3000 tread! Product is finish and now in production.

I'm working too on a news design of phase modulated 2Kw heat sink less IGBT power amp...Look at this!

Fredos

www.d-amp.com
 
#96
ok, but then I do not need startup circuitry cos. forced oscillation will at start skip first positive cycle, and as soon as negative cycle occurs bootstrap cap will be charged and system will continue to work normally, but don't I need anymore c33,34 and 5k trim. ?

btw. is it best to use quartz oscill. and then transforming square into triangle wave ?

thank's for fast response, have a nice day!
 
#97
Hello everybody again !

I was just wondering is it possible to occur a bad feedback response if I use +-25V insted of +-50V which was originally suggested by IR ? I asking this cos. this my version is still in testing phase and I've seccessfully added crystal controlled clock to lock freq. to 447.443kHz
(3.579545 MHz / 2^3)
 

Attachments

  • dsc03652 (1000 x 750).jpg
    dsc03652 (1000 x 750).jpg
    92.9 KB · Views: 707
Status
Not open for further replies.
Top Bottom