BIG setback today/ 60-0-60 SMPS caught fire and smoked big time! is there a known issue driving this type of amplifier with SMPS? or should I go back to using the heavy linear PS? thanks
For current limiting based in voltage drop across Rds-on, the threshold voltage has to be calculated considering that Rds-on increases with MOSFET die temperature (a 2 to 3 factor between 25C and 175C), and that Rds-on increases with drain current. This implies that, to find the optimum Vds point, thermal analysis shall be done, to determine at which current is safe to cut out before the die heats past 175C (requiring both consideration of thermal mass, thermal resistance, and thermal cutout threshold). To fulfill these needs I wrote a little program in BASIC, many calculations get a lot easier writing a program.
A pair of IRFB4227 can do 1600W/2r, in proper circuit conditions.
Power transistors in newest Behringer amplifiers (and others) are exposed to much higher amplitude of thermal cycling to shorten equipment life. Behringer switching power circuits are a joke. When a power transistor gets shorted there is a chain reaction of burning parts (the burned hole in PCB phenomena), to make the gear non-repairable. Power amplifiers are not vegetables, do not grow in trees, are not biodegradable.
NPN/PNP buffers (also both in the same package, by Zetex) are the best approach, fast and simple. Adding a series SMD fuse, or a resistor able to fuse before NPN/PNP get damaged, makes a repairable circuit.
Many gate driver ICs have "not so strong" outputs just because, from the point of view of electrical design, having the buffer separated from the driver IC is the correct approach. On the other hand, the fabrication processes of the gate driver IC and the buffer are not the same. In general, each chip fabrication process is optimized for a purpose (driver is high voltage low current MOSFET, buffer is low voltage bipolar, chemistry and topology is different). This is the best criteria to decide where to put lines between integration sections.
A pair of IRFB4227 can do 1600W/2r, in proper circuit conditions.
Power transistors in newest Behringer amplifiers (and others) are exposed to much higher amplitude of thermal cycling to shorten equipment life. Behringer switching power circuits are a joke. When a power transistor gets shorted there is a chain reaction of burning parts (the burned hole in PCB phenomena), to make the gear non-repairable. Power amplifiers are not vegetables, do not grow in trees, are not biodegradable.
NPN/PNP buffers (also both in the same package, by Zetex) are the best approach, fast and simple. Adding a series SMD fuse, or a resistor able to fuse before NPN/PNP get damaged, makes a repairable circuit.
Many gate driver ICs have "not so strong" outputs just because, from the point of view of electrical design, having the buffer separated from the driver IC is the correct approach. On the other hand, the fabrication processes of the gate driver IC and the buffer are not the same. In general, each chip fabrication process is optimized for a purpose (driver is high voltage low current MOSFET, buffer is low voltage bipolar, chemistry and topology is different). This is the best criteria to decide where to put lines between integration sections.
**************you have to learn somthing**********
1.turn on amplifier first time using low voltage and low current powersupply
(currunt controlled bench power supply is the best option)
OR use:--
*** Dc +18v0 18v 1A transformer to main power supply
*** TO powerup irs2092 and bias mosfet gates (L7812) 12v regularted another tranformer
and (7805 7905) regularted +5 -5 supply to puower up Irs2092 audio circuit
2.
first turn on 12V aux,, and +/-5v then tuch irs2092 input pin by your finger,
and check 2 mosfet gates voltage oscilation.gate voltages are oscillating ,it s ok
3.check irs2092 bootstrap capacitor voltage and Irs2092 floating pin voltage
( bootstrap capacitor voltage >>8v its ok)
4.now turn off all power supplys and turn on Dc +/-18V supply
check any short circuit heating issu and any fault you know
5.now turn on all power supplys (18V,12V,5 V)
6. check- amplifier switching ok,switching frequncy,spike s and heating issus
+++++any other matters
7.operate the amplifier more than 1 houre and check any issue
8.last - power up circuit by any voltage power supply as you wish
Hi hpoche
Sorry to hear that :| ........ but I think we have all burned a number of amplifiers
What SMPS are you using?
Would like to know it is not the same as some of the ones I'm using, bought of of ebay
And no SMPS is normally a good match for class-d
Hi Luka
Sure you're right, that is also what I sctually do, but I think I have had a wrong view; bigger is better ... wanting to use IRFB4227, wherre it is really not needed ....
As Eva rightly points out these should be good for up to 1600W@2ohm ... which is much more than I need in my currrent design.
So I think for now I'll try out IPP320N20N3 as a more powerfull alternative to IRFB5620.
For a future design for larger subwoofers, I'll go for IRFB4227 with a NPN/PNP driver.
Kind regards Baldin
IRFB4227/IRFB4127 have a good trade between body diode related loss and Rds-on loss. Actually these parts have increased Rds-on for faster body diode. In a 250Khz amplifier made with IRFB4227, at 700A/us, switching loss is already 2~4 times higher than conduction loss.
Other lower Rds-on MOSFET have been released since then. I have done the analysis. These have higher body diode related loss (only partially compensated by stronger body diode operated at higher di/dt), for less Rds-on related loss. In other words: intended for soft switching or lower frequency operation.
The generation of MOSFET parts released ~10 years ago is hard to beat for class D. There is probably no better physical way to cook 150V/200V silicon MOSFET for 100~400khz hard switching.
btw: Power electronics without protections = meltdowns. Protections shall be considered part of power stage design.
Other lower Rds-on MOSFET have been released since then. I have done the analysis. These have higher body diode related loss (only partially compensated by stronger body diode operated at higher di/dt), for less Rds-on related loss. In other words: intended for soft switching or lower frequency operation.
The generation of MOSFET parts released ~10 years ago is hard to beat for class D. There is probably no better physical way to cook 150V/200V silicon MOSFET for 100~400khz hard switching.
btw: Power electronics without protections = meltdowns. Protections shall be considered part of power stage design.
Last edited:
IRS2092S and SMPS fire
bought these off eBay from CHINA. 5 are working fine. This one smoked and melted the foil and solder. Lots of fun chasing the flaming balls around on my work bench. Lots of smoke and banging noises....
Hi hpoche
Sorry to hear that :| ........ but I think we have all burned a number of amplifiers
What SMPS are you using?
Would like to know it is not the same as some of the ones I'm using, bought of of ebay
And no SMPS is normally a good match for class-d
bought these off eBay from CHINA. 5 are working fine. This one smoked and melted the foil and solder. Lots of fun chasing the flaming balls around on my work bench. Lots of smoke and banging noises....
thanks for this info. used to do this, but got impatient and we know the result. I think the SMPS was defective. I would like to find a better source for the SMPS if anyone knows OR a good reliable.
@Eva, think you're right in this. And in general there are no magic components, just good design.
@hpoche, could you post picture of the SMPS?
I have a 500W +-70Vdc SMPS PSU here I haven't tried yet. Hope it has some protection built in ... maybe I should test this out before using it ... just see how it does when shorting the output .....
Of course there is the current prtection on the IRS2092, but would rather also have some safety built into the SMPS.
@hpoche, could you post picture of the SMPS?
I have a 500W +-70Vdc SMPS PSU here I haven't tried yet. Hope it has some protection built in ... maybe I should test this out before using it ... just see how it does when shorting the output .....
Of course there is the current prtection on the IRS2092, but would rather also have some safety built into the SMPS.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.