Go Back   Home > Forums > Amplifiers > Class D

Class D Switching Power Amplifiers and Power D/A conversion

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 28th September 2013, 04:53 PM   #3711
diyAudio Member
 
andrewlebon's Avatar
 
Join Date: Feb 2008
Default cn sound d class

Sameerx1 completing second pcb
Attached Images
File Type: jpg 1376537_230414377115010_357625876_n.jpg (87.5 KB, 1181 views)
File Type: jpg image-003F_51ECE751.jpg (161.3 KB, 1163 views)
__________________
[image][/image]"It's Better to try and fail than to never know and always wonder"
  Reply With Quote
Old 30th September 2013, 06:58 AM   #3712
diyAudio Member
 
lorylaci's Avatar
 
Join Date: Aug 2008
Location: Dabas, Hungary
Send a message via MSN to lorylaci
Quote:
Originally Posted by mkuczma View Post
Hi lorylaci, I want to build iraudamp7 with my own pcb and I have a question:
You show your PCB with big ground plane on a double sided board but in IR note AN-1138 is wrote: The irs2092 does not accept single ground plane. And there is a picture show a 3 plane: analog ground, reference ground and power ground. What is your opinion about it?
Check this:
http://www.hypex.nl/docs/appnotes/EMI_appnote.pdf
Bruno reccomends one big ground plane. IR reccomends usually three ground plane. Why?

If all connections would have 0 impedance, then it would does not matter star groudn / ground plane / etc. However connections have impedance, sometimes PCB track inductacne matters, sometimes the capacitance, and sometimes resitance. Ground plane acts as a low impedance connection and also as a shield (and it also couples capacitively or inductively to tracks!).
One big ground plane on both sides usually gives the lowest possible connection impedances for the most important reference potencial. This is why I reccomend one big ground plane.
I also soldered elco's on both sides, to ensure short between top and bottom gorund plane everywhere.

IR reccomends three separate gorund planes for the following reasons:
- switching stage can inject high noise to the tracks around the FETs, and this noise can leak to input stage. Also IR deisgns are all half-bridge single ended design, so input ground must be separated. Separation makes most noise currents to bypass this region. Mine is a full-bridge fully balanced design, and input part is on the other side of the board.
- Due to the bootstrapped high side FET driving, this part is floating, and high side gate signal and its supply is referenced to this. This is why IR reccomend this refrence to has a small plane. This gives some noise immunity, however: this plane is floating. The bigger the floating part is, the higher lever it radiates! Also dV/dt of this must be controlled (under load conditions it can be hard, hard switching!), since high dV/dt gives a high frequency component. If the radiating area is comparable to the wavelength, then it can be an efficient radiating element.
Using one big and good gorund plane and good FET driving (turn-off transistor is right next to FET) decreases high side FET driving noise sesnitivtiy, and also shield the floating part.

I hope my opinion was throughout.
__________________
Olyan nem volt még, hogy ne lett volna valahogy
  Reply With Quote
Old 30th September 2013, 07:02 AM   #3713
diyAudio Member
 
lorylaci's Avatar
 
Join Date: Aug 2008
Location: Dabas, Hungary
Send a message via MSN to lorylaci
Quote:
Originally Posted by Cresnet View Post
my amplifier shows 192KHz at output, sine wave

even when I feed it with sine wave 1KHZ, still I get 192KHz at speaker output

any advice?
Haven't you heard of CARRIER FREQUENCY or MODULATION?!
First learn to use measurement instruments properly!
Then start with a begginer's project fitting your electrical knowledge! Without proper knowledge switching stuff is a p**n in the a**.
__________________
Olyan nem volt még, hogy ne lett volna valahogy
  Reply With Quote
Old 30th September 2013, 07:04 AM   #3714
diyAudio Member
 
lorylaci's Avatar
 
Join Date: Aug 2008
Location: Dabas, Hungary
Send a message via MSN to lorylaci
Quote:
Originally Posted by andrewlebon View Post
Sameerx1 completing second pcb
Is it intended to be a Class-D amp?
- If does, than it must be really bad. Typical, still thinking about Class-B layout. Forget it, make up your mind. Heatsinking shoudl not be priority. A good Classs-D amp, (even high power ones) should not need heatsinks that big.
- If does not, then why it is here?
__________________
Olyan nem volt még, hogy ne lett volna valahogy
  Reply With Quote
Old 1st October 2013, 05:15 AM   #3715
redmx is offline redmx  United States
diyAudio Member
 
Join Date: Jul 2011
Default Class d is easy

. lorylace seems that you do not like any amplifier or all are bad though your amp is really horrible not discouraged not so difficult to make a class d test with combining TL081 + 4049 + fan7392 is wonderful and can handle a pair of MOSFET of 100NC by side without any problems at 450 khz heard very good greetings to all who are not pessimistic and see not all bad I think your amplifier (Andrew lebon) looks very well and is a lot of power for good job Greetings to all
  Reply With Quote
Old 1st October 2013, 06:57 AM   #3716
Cresnet is offline Cresnet  Germany
diyAudio Member
 
Join Date: Nov 2012
Quote:
Originally Posted by lorylaci View Post
Haven't you heard of CARRIER FREQUENCY or MODULATION?!
First learn to use measurement instruments properly!
Then start with a begginer's project fitting your electrical knowledge! Without proper knowledge switching stuff is a p**n in the a**.
Problem was in the filter values, Also I know how to use instruments very well. and very expensive ones.

Beginners project may be good for you at this stage, I am sorry.

And yes without proper knowledge in switching power electronics, its pain.

Thanks for your help.
  Reply With Quote
Old 1st October 2013, 04:07 PM   #3717
diyAudio Member
 
Workhorse's Avatar
 
Join Date: Dec 2004
Quote:
Originally Posted by andrewlebon View Post
stereo d class amp kit
Andy that class-D amp is nothing but a joke
  Reply With Quote
Old 1st October 2013, 04:32 PM   #3718
diyAudio Member
 
lorylaci's Avatar
 
Join Date: Aug 2008
Location: Dabas, Hungary
Send a message via MSN to lorylaci
Quote:
Originally Posted by redmx View Post
. lorylace seems that you do not like any amplifier or all are bad though your amp is really horrible not discouraged not so difficult to make a class d test with combining TL081 + 4049 + fan7392 is wonderful and can handle a pair of MOSFET of 100NC by side without any problems at 450 khz heard very good greetings to all who are not pessimistic and see not all bad I think your amplifier (Andrew lebon) looks very well and is a lot of power for good job Greetings to all
Post some measurements, scope, efficiency data... Unless its hard to believe that that amp is good. Just few reasons:
- TO247 FETs whiel can dissipate a lot (however a good class D amp has good effciency, so it does not have to) have higher parasitic inductances. This coupled with Coss and other generates spikes, repsonsible for ringing etc...
- I don't see any snubbers on FETs (maybe there are SMD ones on bottom, if so then its really good, but I doubt it has them), this means that it either has lots of ringing/spikes etc high EMI, or switching is controlled, in this way SLOW.
- Slow switching either increases Integral(U*I) losses, or either comes with high effective dead-time (which gives higher THD and noise)
- Supply is bypassed very far from FETs (or there are SMD bypass capacitors on bottom, same applies as for SMD snubbers) this means that to FETs leg (and internal) parasitic inductance add up with all PCB track inductance. (it must also be kept in mind, that the length for the path of ripple current of the bypass capacitors should be minimised)
- toroids while have relative low leakage field, they have some, since leakage field minimalization needs totally symmetrical winding (ideal world assumption). If mounted this way, they couple to the ground plane (if there is ground plane, if there is no ground plane then its much worse), which acts as a looesly coupled short, decreasing SRF of the inductor (decreasing filter efficiency at high frequencies, so another EMI problem)
- mechanical relay: a relay for speaker potection is a really false hope for powers over 100W. Speaker can be inductive (or any kind of reactive) and a mechanical relay has a hard job disconnecting inductive loads under DC (which appears under proetection condition) due to arc formation. If speaker disconnection needed MOSFET relay must be used, or only speaker protection is needed then crowbar protection

Its MHO but much better amp can be made than (even without much better components). Just some ideas to see that I not olny tell you what not to do, but what to do:
- Forget easy heatsinking, a good class D amp only needs a heatsink 1/10th of this. Use L-profiles if needed. Put the FETs together (less path and also more freedom enables you to eliminate track LOOPs)
- Use TO220 devices, less paraitic inductance. Bypass supply as close to possible. RC snubber as close to possible. Gate driving buffer should be put as close as possible. Usually fast urn off and not so fast turn on is needed. Use 1 PNP transistor per FET pust just next to source and gate) to buffer gate turnoff. Gate turn on curren is smaller, so its path can be longer.
For gate buffer make a list of all transistors you can get. Search the datasheet. Search for current gain at high currents (PEAK is only enough, average current is slow). Don't be afraid of SMD, an MMBT2907A (SMD version of 2N2907A) can boost peak gate turn off current >3A (rated at 0,6A average).
- Draw the current path under all periods of switching cycle (att ny load condition, be aware of the 2 crossover points), see the loops, minimize then, don't put any sensitve part inside it.

If that was not enough I can contione. But if you just ignore my advices (just read application nots, white papers, see hypex.nl, International Rectifier - The Power Management Leader etc) and keep designing that way you will never improve your design.
__________________
Olyan nem volt még, hogy ne lett volna valahogy
  Reply With Quote
Old 2nd October 2013, 05:13 AM   #3719
diyAudio Member
 
Workhorse's Avatar
 
Join Date: Dec 2004
Quote:
Originally Posted by lorylaci View Post
Post some measurements, scope, efficiency data... Unless its hard to believe that that amp is good. Just few reasons:
- TO247 FETs whiel can dissipate a lot (however a good class D amp has good effciency, so it does not have to) have higher parasitic inductances. This coupled with Coss and other generates spikes, repsonsible for ringing etc...
- I don't see any snubbers on FETs (maybe there are SMD ones on bottom, if so then its really good, but I doubt it has them), this means that it either has lots of ringing/spikes etc high EMI, or switching is controlled, in this way SLOW.
- Slow switching either increases Integral(U*I) losses, or either comes with high effective dead-time (which gives higher THD and noise)
- Supply is bypassed very far from FETs (or there are SMD bypass capacitors on bottom, same applies as for SMD snubbers) this means that to FETs leg (and internal) parasitic inductance add up with all PCB track inductance. (it must also be kept in mind, that the length for the path of ripple current of the bypass capacitors should be minimised)
- toroids while have relative low leakage field, they have some, since leakage field minimalization needs totally symmetrical winding (ideal world assumption). If mounted this way, they couple to the ground plane (if there is ground plane, if there is no ground plane then its much worse), which acts as a looesly coupled short, decreasing SRF of the inductor (decreasing filter efficiency at high frequencies, so another EMI problem)
- mechanical relay: a relay for speaker potection is a really false hope for powers over 100W. Speaker can be inductive (or any kind of reactive) and a mechanical relay has a hard job disconnecting inductive loads under DC (which appears under proetection condition) due to arc formation. If speaker disconnection needed MOSFET relay must be used, or only speaker protection is needed then crowbar protection

Its MHO but much better amp can be made than (even without much better components). Just some ideas to see that I not olny tell you what not to do, but what to do:
- Forget easy heatsinking, a good class D amp only needs a heatsink 1/10th of this. Use L-profiles if needed. Put the FETs together (less path and also more freedom enables you to eliminate track LOOPs)
- Use TO220 devices, less paraitic inductance. Bypass supply as close to possible. RC snubber as close to possible. Gate driving buffer should be put as close as possible. Usually fast urn off and not so fast turn on is needed. Use 1 PNP transistor per FET pust just next to source and gate) to buffer gate turnoff. Gate turn on curren is smaller, so its path can be longer.
For gate buffer make a list of all transistors you can get. Search the datasheet. Search for current gain at high currents (PEAK is only enough, average current is slow). Don't be afraid of SMD, an MMBT2907A (SMD version of 2N2907A) can boost peak gate turn off current >3A (rated at 0,6A average).
- Draw the current path under all periods of switching cycle (att ny load condition, be aware of the 2 crossover points), see the loops, minimize then, don't put any sensitve part inside it.

If that was not enough I can contione. But if you just ignore my advices (just read application nots, white papers, see hypex.nl, International Rectifier - The Power Management Leader etc) and keep designing that way you will never improve your design.
lorylaci,
Those modules are not done by these who are displaying it, its just a Thailand company goes by the name Detex Audio, who make such crappy things and these guys are just assembling the pcbs supplied by them. For them class-D is just another form of amp out there just like class-AB, nothing to do with skill-set or know how of pcb layout, what soever.

You advice is good for those who want to learn. Sadly, in this thread the case is not that.
  Reply With Quote
Old 11th October 2013, 10:06 AM   #3720
diyAudio Member
 
Join Date: Dec 2011
Default 3000w class d Stereo amplifier kit

3000w class d Stereo amplifier kit
Attached Images
File Type: jpg DSCN4917 (Copy).JPG (788.2 KB, 876 views)
  Reply With Quote

Reply


Hide 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
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
Philips 40/25/15 watts amplifier with Universal preamplifier circuit. renjiish Solid State 55 4th April 2014 07:08 AM
Decibel Watts VS Watts ? fivestring Solid State 11 1st April 2009 01:08 PM
AMP1-B: why only 25 watts @8ohms? theAnonymous1 Class D 10 15th August 2005 06:55 PM
Sound quality of amps at 0.3 watts vs @ typical 5 watts rick57 Solid State 2 26th March 2005 06:25 AM
Need 2 watts Vivek Everything Else 4 17th July 2003 08:29 AM


New To Site? Need Help?

All times are GMT. The time now is 05:47 AM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright ©1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2