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

Behringer iNuke NU3000 Measurements
Behringer iNuke NU3000 Measurements
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 17th June 2017, 06:53 AM   #101
ICG is offline ICG  Germany
diyAudio Member
 
ICG's Avatar
 
Join Date: Jul 2007
Location: I had a Déjà Moo - I've seen that BS before!
Quote:
Originally Posted by wg_ski View Post
Speaking of ticking bombs inside Behringer equipment - I found something interesting in the IRS4227 data sheet that I hadn't noticed before. It states that the die attach in the D2pak (RoHS version of course) is subject to wear out after 1000 thermal cycles. By wear out it means a 50% increase in Rth. Must be some sort of epoxy. It will take more than that to die, but it eventually will, just like the old aluminum TO-3. Found out about that from the 78xx regulator data sheets. The TO220 has no such warning but is "not recommended for surface mounting". Crack them open and they are definitely soldered. With a package with that small a thermal mass it's not that hard to rack up several thousand thermal cycles at subwoofer frequencies.
The thermal cycles are not equal to the frequency nor the times it is switched on and off. The thermal cycle in semiconductors is counted from a temperature of ambient temperature (or even the lowest operational temperature spec, which is -10, -20 or even -40°C) to the highest operational temperature, which is around ~150°C. As long as the amplifier got a working cooling system of any kind and no (artificial) heat spots, there are no thermal cycles. These cycles are usually only reached in extreme lab enviroments while testing cycles. So no, that are not 'time bombs'.
__________________
I had a Déjà Moo - I've seen that BS before!
  Reply With Quote
Old 17th June 2017, 06:03 PM   #102
Eva is offline Eva  Spain
diyAudio Member
 
Eva's Avatar
 
Join Date: Oct 2003
Location: Near the sea
Send a message via MSN to Eva
Some right and some wrong information here.

The IRFS4227 datasheet states that thermal resistance junction to case increases from .45 C/W to .65 C/W after 1000 thermal cycles. This is called at "end of life" because themal resistance stabilizes at the latter value until "end of life" of device, not because the device ceases to operate. These values are lower compared to thermal resistance from the device to the air, about a few C/W in this application, so a 10:1 ratio, negligible loss of thermal performance.

On the other hand the idle dissipation in these transistors is low, and they are fan cooled permanently. This results in the transistor dies being close to ambient temperature at idle power and being close to maximum temperature at full power at low impedance load and/or bass duty. There is very little thermal mass. So there are *full* thermal cycles, following music program intensity, taking place in these amplifiers at low load impedances and/or bass duty. Not at 8 ohms full range, of course. A "thermal-cycling life test fixture" could be the optimum sarcasm.

It's well known that wide thermal cycles reduce the life of SMD boards, causing intermittent solder joints requiring rework or in some cases replacement when many parts blow due to an open connection. Fortunately the output stage uses IRS20957 control circuit with built-in short-circuit protection, this will limit the degree of damage in case one FET becomes shorted due to open gate. This protection is not even found in many class D subwoofer power modules from Harman group discussed in some threads here, so the INUKEs are better than that.

In fewer words: There will be intermittent solder joint problems in these amplifiers when operated at lower impedances and/or low frequency duty, repairable in many cases.

The power supply has a similar pair of anti-feature and feature, the gates of the IGBTs are transformer-driven to act as a firewall, because such IGBTs can fail exploding and leaving a hole in PCB if one gate becomes open. Another solution in SMPS is to use power zeners and fusible resistors to contain damage in the unfortunate but common case the gate of a power transistor becomes shorted to the drain, done in the best circuits, often underlooked, just repaired by swapping the transistor and blown gate resistor.

The PCB approach used in more reliable equipment is to have separate SMD boards, of materials and thicknesses optimized for long life of SMD, containing all small signal circuits, and the use a main board of materials optimized for rigidness and power handling, with thru-hole power devices mounted to a heatsink.

Also, the cooling approach used in more reliable equipment is to allow the heatsinks to heat to relatively hot before starting proportionally driven fans, this reduces amplitude of thermal cycles, as the heatsink is allowed to keep hot from one high power passage to the next.

The attached picture shows a *reliable* prototype capable of 2x1000W/8r (BTL) short term (for a few milliseconds, as other products are sometimes rated), >1800W for a few seconds @ 230V, and >600W for indefinitely long period with 50mm fan, 3:1 crest factor. *Reliable* in the sense that it does not exhibit substantially faster aging when operated at maximum ratings, and self-diagnostic too, it gives different fault codes when operated outside ratings, to help locate the problem, even from remote location.

And this other picture shows INUKE: Click the image to open in full size.
Attached Images
File Type: jpg amp_test_painting_room_3.jpg (485.1 KB, 374 views)
__________________
I use to feel like the small child in The Emperor's New Clothes tale

Last edited by Eva; 17th June 2017 at 06:32 PM.
  Reply With Quote
Old 17th June 2017, 06:53 PM   #103
wg_ski is online now wg_ski  United States
diyAudio Member
 
Join Date: Oct 2007
Are those D2pak IGBTs????
  Reply With Quote
Old 17th June 2017, 07:34 PM   #104
Eva is offline Eva  Spain
diyAudio Member
 
Eva's Avatar
 
Join Date: Oct 2003
Location: Near the sea
Send a message via MSN to Eva
Yes, IRGB20B60PD1

Relatively good, but there are newer like IKP20N65H5, one of my preferred ones. With help of 1~10 nF capacitor and gapped core for minimum magnetization current, IGBT can achieve very low loss in amplifier SMPS, also idle loss. INUKE power supply has no resonant switching, and it could have because it is not regulated, duty cycle is fixed at close to maximum, _SHDN pin is the only one driven in the SG3525A.

From the sociological point of view, amplifiers like INUKE are karma checkers, it will last reasonably if there is fair people at the party, parties of friends I mean, including techs in process of learning, and may blow otherwise. The next reliability level does not have that "feature" haha. Karma checks in a sustainable world should be performed only with biodegradable and renewable resources, like fights of throwing surplus tomatoes.

EDIT: I just noticed the symbol shows a MOSFET, but the part number is a fast IGBT with diode from IR.
__________________
I use to feel like the small child in The Emperor's New Clothes tale

Last edited by Eva; 17th June 2017 at 07:59 PM.
  Reply With Quote
Old 18th June 2017, 06:18 PM   #105
Reactance is offline Reactance  South Africa
diyAudio Member
 
Reactance's Avatar
 
Join Date: Jun 2005
Location: Cape Town
Quote:
Originally Posted by Eva View Post
Some right and some wrong information here.

The IRFS4227 datasheet states that thermal resistance junction to case increases from .45 C/W to .65 C/W after 1000 thermal cycles. This is called at "end of life" because themal resistance stabilizes at the latter value until "end of life" of device, not because the device ceases to operate. These values are lower compared to thermal resistance from the device to the air, about a few C/W in this application, so a 10:1 ratio, negligible loss of thermal performance.

On the other hand the idle dissipation in these transistors is low, and they are fan cooled permanently. This results in the transistor dies being close to ambient temperature at idle power and being close to maximum temperature at full power at low impedance load and/or bass duty. There is very little thermal mass. So there are *full* thermal cycles, following music program intensity, taking place in these amplifiers at low load impedances and/or bass duty. Not at 8 ohms full range, of course. A "thermal-cycling life test fixture" could be the optimum sarcasm.

It's well known that wide thermal cycles reduce the life of SMD boards, causing intermittent solder joints requiring rework or in some cases replacement when many parts blow due to an open connection. Fortunately the output stage uses IRS20957 control circuit with built-in short-circuit protection, this will limit the degree of damage in case one FET becomes shorted due to open gate. This protection is not even found in many class D subwoofer power modules from Harman group discussed in some threads here, so the INUKEs are better than that.

In fewer words: There will be intermittent solder joint problems in these amplifiers when operated at lower impedances and/or low frequency duty, repairable in many cases.

The power supply has a similar pair of anti-feature and feature, the gates of the IGBTs are transformer-driven to act as a firewall, because such IGBTs can fail exploding and leaving a hole in PCB if one gate becomes open. Another solution in SMPS is to use power zeners and fusible resistors to contain damage in the unfortunate but common case the gate of a power transistor becomes shorted to the drain, done in the best circuits, often underlooked, just repaired by swapping the transistor and blown gate resistor.

The PCB approach used in more reliable equipment is to have separate SMD boards, of materials and thicknesses optimized for long life of SMD, containing all small signal circuits, and the use a main board of materials optimized for rigidness and power handling, with thru-hole power devices mounted to a heatsink.

Also, the cooling approach used in more reliable equipment is to allow the heatsinks to heat to relatively hot before starting proportionally driven fans, this reduces amplitude of thermal cycles, as the heatsink is allowed to keep hot from one high power passage to the next.

The attached picture shows a *reliable* prototype capable of 2x1000W/8r (BTL) short term (for a few milliseconds, as other products are sometimes rated), >1800W for a few seconds @ 230V, and >600W for indefinitely long period with 50mm fan, 3:1 crest factor. *Reliable* in the sense that it does not exhibit substantially faster aging when operated at maximum ratings, and self-diagnostic too, it gives different fault codes when operated outside ratings, to help locate the problem, even from remote location.

And this other picture shows INUKE: Click the image to open in full size.
Your prototype mounting technique:

How is < 600W of power dissipated using x6 TO-220 packages mounted on the PCB underside, fixed using the PCB as a mount, surely this isn't safe ?
__________________
There’s nothing worse than the smell of burned output transistors!!
  Reply With Quote
Old 18th June 2017, 09:32 PM   #106
Eva is offline Eva  Spain
diyAudio Member
 
Eva's Avatar
 
Join Date: Oct 2003
Location: Near the sea
Send a message via MSN to Eva
Quote:
Originally Posted by Reactance View Post
Your prototype mounting technique:

How is < 600W of power dissipated using x6 TO-220 packages mounted on the PCB underside, fixed using the PCB as a mount, surely this isn't safe ?
Lets assume 85% efficiency at 2000W. This is 300W loss, about 21W * 14 power devices. Thermal grease starts to be effective (low thermal resistance) at low contact pressure, as long as it does not drain/void, contact surfaces are flat, and part to part TO-220 height tolerance is satisfactory. The PCB shown is 2oz 4 layer, so it actually provides some pressure. Also there are 6 screws on each side.

There are also those compressible thermal pads of different thickness, compressibility and thermal conductance, mostly used in computers, but don't think they are better than alumina pads and thermal paste in terms of thermal resistance.
__________________
I use to feel like the small child in The Emperor's New Clothes tale

Last edited by Eva; 18th June 2017 at 09:36 PM.
  Reply With Quote
Old 18th June 2017, 10:49 PM   #107
Reactance is offline Reactance  South Africa
diyAudio Member
 
Reactance's Avatar
 
Join Date: Jun 2005
Location: Cape Town
Quote:
Originally Posted by Eva View Post
Lets assume 85% efficiency at 2000W. This is 300W loss, about 21W * 14 power devices. Thermal grease starts to be effective (low thermal resistance) at low contact pressure, as long as it does not drain/void, contact surfaces are flat, and part to part TO-220 height tolerance is satisfactory. The PCB shown is 2oz 4 layer, so it actually provides some pressure. Also there are 6 screws on each side.

There are also those compressible thermal pads of different thickness, compressibility and thermal conductance, mostly used in computers, but don't think they are better than alumina pads and thermal paste in terms of thermal resistance.
Thanks for the insight. It does appear in your picture that the FR4 board is slightly stressed(curved) and may be fatigued, but then again this is a prototype. What you say is true, a computer CPU shows obvious effects when thermal compound dries, I remember my Intel 3.0GHz CPU alert utility all of a sudden started reaching its default thermal threshold.
__________________
There’s nothing worse than the smell of burned output transistors!!

Last edited by Reactance; 18th June 2017 at 10:51 PM.
  Reply With Quote
Old 19th June 2017, 10:15 AM   #108
Eva is offline Eva  Spain
diyAudio Member
 
Eva's Avatar
 
Join Date: Oct 2003
Location: Near the sea
Send a message via MSN to Eva
The board is intended to be bent, SMD is properly oriented and pads have thermals. Only SMD parts close to inflection points at the center are decoupling caps (1206/1210), snubbers (1206), diodes (SMA/SMB), and sensing NTCs (0603). Bend is across ~75mm length. Have you tried bending a 2oz 4 layer PCB like that? Requires >1 kilogram per pair of TO-220 devices. Material is good fiberglass, 2 copper planes at both sides and 2 in the middle. And you are right that it is a prototype, I'm evaluating it: loosening the screws after a year of operation results in PCB becoming nearly straight again.
__________________
I use to feel like the small child in The Emperor's New Clothes tale
  Reply With Quote
Old 25th May 2018, 09:45 AM   #109
pakoh is offline pakoh  Spain
diyAudio Member
 
Join Date: May 2015
Location: Barcelona
Wink for line 70/100v?

Are these amplifiers valid to feed 70 / 100v lines?
I'm doing an installation with Comunity r5 speakers and I'm assailed by doubts, I've never done it with class D in this type of facility ...
Thank you
  Reply With Quote
Old 27th May 2018, 05:50 PM   #110
Dinithm is offline Dinithm  Sri Lanka
diyAudio Member
 
Join Date: Jul 2010
Location: Gampaha
Quote:
Originally Posted by Eva View Post
Some right and some wrong information here.

The IRFS4227 datasheet states that thermal resistance junction to case increases from .45 C/W to .65 C/W after 1000 thermal cycles. This is called at "end of life" because themal resistance stabilizes at the latter value until "end of life" of device, not because the device ceases to operate. These values are lower compared to thermal resistance from the device to the air, about a few C/W in this application, so a 10:1 ratio, negligible loss of thermal performance.

On the other hand the idle dissipation in these transistors is low, and they are fan cooled permanently. This results in the transistor dies being close to ambient temperature at idle power and being close to maximum temperature at full power at low impedance load and/or bass duty. There is very little thermal mass. So there are *full* thermal cycles, following music program intensity, taking place in these amplifiers at low load impedances and/or bass duty. Not at 8 ohms full range, of course. A "thermal-cycling life test fixture" could be the optimum sarcasm.

It's well known that wide thermal cycles reduce the life of SMD boards, causing intermittent solder joints requiring rework or in some cases replacement when many parts blow due to an open connection. Fortunately the output stage uses IRS20957 control circuit with built-in short-circuit protection, this will limit the degree of damage in case one FET becomes shorted due to open gate. This protection is not even found in many class D subwoofer power modules from Harman group discussed in some threads here, so the INUKEs are better than that.

In fewer words: There will be intermittent solder joint problems in these amplifiers when operated at lower impedances and/or low frequency duty, repairable in many cases.

The power supply has a similar pair of anti-feature and feature, the gates of the IGBTs are transformer-driven to act as a firewall, because such IGBTs can fail exploding and leaving a hole in PCB if one gate becomes open. Another solution in SMPS is to use power zeners and fusible resistors to contain damage in the unfortunate but common case the gate of a power transistor becomes shorted to the drain, done in the best circuits, often underlooked, just repaired by swapping the transistor and blown gate resistor.

The PCB approach used in more reliable equipment is to have separate SMD boards, of materials and thicknesses optimized for long life of SMD, containing all small signal circuits, and the use a main board of materials optimized for rigidness and power handling, with thru-hole power devices mounted to a heatsink.

Also, the cooling approach used in more reliable equipment is to allow the heatsinks to heat to relatively hot before starting proportionally driven fans, this reduces amplitude of thermal cycles, as the heatsink is allowed to keep hot from one high power passage to the next.

The attached picture shows a *reliable* prototype capable of 2x1000W/8r (BTL) short term (for a few milliseconds, as other products are sometimes rated), >1800W for a few seconds @ 230V, and >600W for indefinitely long period with 50mm fan, 3:1 crest factor. *Reliable* in the sense that it does not exhibit substantially faster aging when operated at maximum ratings, and self-diagnostic too, it gives different fault codes when operated outside ratings, to help locate the problem, even from remote location.

And this other picture shows INUKE: Click the image to open in full size.
really good explanation,thank you.
  Reply With Quote

Reply


Behringer iNuke NU3000 MeasurementsHide 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
Behringer iNuke NU3000 w/o any fan...? Soldermizer Class D 159 11th February 2018 03:06 AM
Measurements: When, What, How, Why D OB G Multi-Way 1666 19th January 2014 11:23 AM
OB with vintage Seas and Behringer Inuke DSP zmyrna Full Range 2 23rd April 2012 09:36 AM
Measurements SoNic_real_one Digital Source 26 6th April 2012 12:24 AM
are these measurements okay? eitanwaks Multi-Way 3 23rd March 2010 07:32 PM


New To Site? Need Help?

All times are GMT. The time now is 01:38 AM.


Search Engine Optimisation provided by DragonByte SEO (Pro) - vBulletin Mods & Addons Copyright © 2018 DragonByte Technologies Ltd.
Resources saved on this page: MySQL 14.29%
vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2018 DragonByte Technologies Ltd.
Copyright ©1999-2018 diyAudio
Wiki