UCD 25 watts to 1200 watts using 2 mosfets

Anyone tried IRFB4127 mosfets? How much output current can you get with a pair of it? (Assuming that the class-d amp is stable)
it's similar to irfb4227 with higher max current (76A vs 65A on the 4227). in real world application it would most definitely not able to sustain max output current continuously, and most definitely not at high temperature.

if i remember correctly max output current drop to 60% at high temperature. do refer to the temp graph inside the datasheet.
but if you run them slow enough, with wide enough dead time, they'll be relatively cool to the touch even during operation with couple of hundred watt of output.

mine (irfb5615) doesn't even heat up with up to ~500W surges in bass duties.
 
D

Deleted member 148505

Thanks, I am targeting to get 22A RMS output into 2 ohms (1KW) for a pair of IRFB4127. Based on datasheet the mosfet can handle 30A at 150deg c case temp.

I can't measure the total deadtime of my amp because my scope is not capable. But I've added additional 67us of deadtime through gate resistor.

Currently with my lowly supply I only get 490W RMS at 2 ohms load. Rail voltage dropped to 36VAC at full load. Next time I'll use higher voltage.

Here's the vid of the measurements.

Measuring voltage drop of EI Core transformer at full load of JLAmp1000D - YouTube
Measuring JLAmp1000D output at 2 ohms using +-65VDC, supply sagged to +-51VDC - YouTube
 
Thanks, I am targeting to get 22A RMS output into 2 ohms (1KW) for a pair of IRFB4127. Based on datasheet the mosfet can handle 30A at 150deg c case temp.

I can't measure the total deadtime of my amp because my scope is not capable. But I've added additional 67us of deadtime through gate resistor.

Currently with my lowly supply I only get 490W RMS at 2 ohms load. Rail voltage dropped to 36VAC at full load. Next time I'll use higher voltage.

Here's the vid of the measurements.

Measuring voltage drop of EI Core transformer at full load of JLAmp1000D - YouTube
Measuring JLAmp1000D output at 2 ohms using +-65VDC, supply sagged to +-51VDC - YouTube
2 ohm is pretty low for a speaker. you're running for sub duties?

if yes then you don't really need to run them at high speed, and with super tight dead time. if the driver chip supports deadtime configuration, then you could configure them with that. else the gate resistor value would do fine.

i do think for 2 ohm, high power duties like for a sub, bridging the amp is a good idea. it removes the rail sag and you don't have to run the amp with high voltage. possibly getting away with a much smaller power supply with it.
 
D

Deleted member 148505

2 ohm is pretty low for a speaker. you're running for sub duties?

if yes then you don't really need to run them at high speed, and with super tight dead time. if the driver chip supports deadtime configuration, then you could configure them with that. else the gate resistor value would do fine.

i do think for 2 ohm, high power duties like for a sub, bridging the amp is a good idea. it removes the rail sag and you don't have to run the amp with high voltage. possibly getting away with a much smaller power supply with it.

Yes it's for sub duty, I used IR2110, it doesn't support deadtime configuration. So I just added 67ns of deadtime through gate resistor.

mine (irfb5615) doesn't even heat up with up to ~500W surges in bass duties.

When you bridge your amp, there is less sag based on your experience but will it overload the transformer more and have heat problems? compared to halfbridge configuration?
 
hi all . here is my amp . i have decided to go the simple way . for more power i will use two channels in bridge mode . your ideas and views are appreciated:)
 

Attachments

  • gtG 1100 schematic.pdf
    35.6 KB · Views: 481
  • gtG 1100 schematic.jpg
    gtG 1100 schematic.jpg
    317.6 KB · Views: 1,621
hi all photos of my project. detex irs900 , thienchay irs900, my smps version of lukas design but with ocp protecton , and my 24db 4way crossover with high pass and low pass on all 4 channels 33hz o 22khz;);):)

>>>after mounting the amps and smps i will resume the new simple amp i posted on #3587
 

Attachments

  • SAM_2651.JPG
    SAM_2651.JPG
    966.5 KB · Views: 1,989
  • SAM_2655.JPG
    SAM_2655.JPG
    963.2 KB · Views: 250
  • SAM_2657.JPG
    SAM_2657.JPG
    938.5 KB · Views: 350
  • SAM_2656.JPG
    SAM_2656.JPG
    961.3 KB · Views: 374
  • SAM_2662.JPG
    SAM_2662.JPG
    948.3 KB · Views: 989
  • SAM_2667.JPG
    SAM_2667.JPG
    966.6 KB · Views: 1,153
  • SAM_2672.JPG
    SAM_2672.JPG
    947.4 KB · Views: 1,298
  • SAM_2670.JPG
    SAM_2670.JPG
    946 KB · Views: 1,619
  • SAM_2658.JPG
    SAM_2658.JPG
    973.8 KB · Views: 263
  • SAM_2665.JPG
    SAM_2665.JPG
    966.7 KB · Views: 316
Last edited:
hi all i have a few questions

>>> is it o.k technically to use one aux 12vlts 2 amps to power the ir2110 and cd4049 for four channels of irs900 or irs4k5. or does it need a 12vlts aux power supply for each channel.?

>>> has anyone else experienced this challenge , when you have irs900 4 channels and same power supply in my case "
(+\-40vts, outputs irfp250n,gate resistor 100ohms, output inductor 22turns eei 35, aux 12voltage each channel with its own 12vlts dropping the voltage from 0-40vlts by a small fet irf640 ) "

if i do not connect a speaker to any channel . it drains a lot of current and cuts the fuse . but if all channels have a load i.e speaker connected everything is o.k and works perfectly provided i do not disconnect a speaker from any of the 4 channels. what is the solution?
 
"is it o.k technically to use one aux 12vlts 2 amps to power the ir2110 and cd4049 for four channels of irs900 or irs4k5. or does it need a 12vlts aux power supply for each channel.?"
Yes,provided you use filtering on each module.

"if i do not connect a speaker to any channel . it drains a lot of current and cuts the fuse "
Try a 1 kohm load and see what it does..

Please post a ground connection diagram (for power and signal paths).
 
When you bridge your amp, there is less sag based on your experience but will it overload the transformer more and have heat problems? compared to halfbridge configuration?
it minimizes the voltage sag to a point that it only drops 1-2 volts during huge bass hits. this is to a sealed 15" sub with only 86 dB sensitivity.

since it's bridged, power supply voltage are kept low. thus allowing for added rail capacitance without using high voltage caps. i have about 40,000uf for each channel.

with bridged configuration for subwoofer use, the sub only handle bass signal which will never be continuous. put enough rail capacitance and transformer load will be reduced as the current will be sourced from the caps. i've put 10A fuse from the trafo line to the cap bank to see if the current draw will surge past 10A and it didn't burn no matter how loud i went.
from the cap bank to the amp modules, current surge is more than 5A during heavy bass hits at high volume. 10A fuse don't burn. 40,000uf in the bank (for each module), another 8,000uf reservoir in the module itself.

my trafo don't even get warm after movies playback. it's only rated for 500VA driving 2 bridged modules with about ~1.2kW each. following the general opinion in here, i need at least 2kW transformer to get full capability.

hi all i have a few questions

>>> is it o.k technically to use one aux 12vlts 2 amps to power the ir2110 and cd4049 for four channels of irs900 or irs4k5. or does it need a 12vlts aux power supply for each channel.?
in theory you could power all the channels with the same 12V line, provided the 12V could handle the current. but in reality, you'll be dealing with lots of heat output which will be hard for the mosfet to manage. if your circuit doesn't draw much current, or if the voltage drop isn't as much, leading to less heat output then it should work.

(feel free to correct me)

>>> has anyone else experienced this challenge , when you have irs900 4 channels and same power supply in my case "
(+\-40vts, outputs irfp250n,gate resistor 100ohms, output inductor 22turns eei 35, aux 12voltage each channel with its own 12vlts dropping the voltage from 0-40vlts by a small fet irf640 ) "

if i do not connect a speaker to any channel . it drains a lot of current and cuts the fuse . but if all channels have a load i.e speaker connected everything is o.k and works perfectly provided i do not disconnect a speaker from any of the 4 channels. what is the solution?
sounds like it's unstable without a load. if you have a scope, check the output to confirm that the amp is still oscillating without a load connected to it.
 
if i do not connect a speaker to any channel . it drains a lot of current and cuts the fuse . but if all channels have a load i.e speaker connected everything is o.k and works perfectly provided i do not disconnect a speaker from any of the 4 channels. what is the solution?

Hi stewin,

Without load(speaker) amp will not oscillate.Provide a dummy load use 1K5/2W or 2K2/2W resistor.

Regards
MANOJ
 
Hi Steve,
your schematic does not show any start up path for the upper gate drive supply.
==> A self oscillating amp without such start up path is likely to lock up, especially during turning on.
You can try a 33k (at least 0.5W) resistor from positive power rail to the VB pin of the IR chip.

Your obseravtion of high current draw and blowing fuse is a little bit making me afraid, that you are also struggling with some parasitic effects which provide some uncontrolled gate drive to the upper MosFets even without driver supply.
This is adding the unpleasant risk of a heavy defect when experimenting with the additional resistor. To reduce this risk, you should put serial light bulbs (i.e 230V/100W) in the pos and neg supply wires before experimenting.


Good luck
Markus




P.S.
Just adding my safety hint from the other thread, because I see the same pictures here.

1. My honest congrats.
2. In the region of the wide isolation barrier, please scratch away the thin copper lines at the sides of the PCB. In the moment this copper line effectively jumpers most of the creepage distance, because it is a conductor which is very close to the primary and also very close to the secondary.
3. Take care with the creepages also in the transformer.
4. Check if the blue cap across the barrier has Y1 rating.
5. Enjoy your build! Great to see your walk of success.

...ohps just seeing now, the tape around the ferrite core should be wider.
Again the same topic - the ferrite comes close the primary and the secondary PCB.
My proposal: Make the tape much wider than the transformer (i.e. 6mm on each side) and have it 3 layers at least. I know that's looking ugly.


Keeping fingers crossed that the inside transformer construction provides sufficient isolation, cannot judge this from the distance.
Fundamental rule for you to check:
Imagine the currents wants to creep from primary to secondary along any surface.
Also indirect path which incorporate conductive parts like unconnected copper or ferrite are critical. I.e. Primary ==> core ==> secondary.
Distances along the surface of conductive material are no creepage, because the conductive material jumpers this distance.
Whereever you find a path which has less than 5-8mm creepage over isolating material, you should think how to enlarge.
[/B]

Disclaimer:
My safety hints are based on my best personal knowledge, but I cannot
guarantee regarding potential errors.
Also there is no way to give full information about safety standards by a few posts. Also there is no way to notice all short comings.
It is just an attempt to reduce potential safety issues, which I noticed while zapping through the forum.
 
D

Deleted member 148505

hi all i have a few questions

>>> is it o.k technically to use one aux 12vlts 2 amps to power the ir2110 and cd4049 for four channels of irs900 or irs4k5. or does it need a 12vlts aux power supply for each channel.?

>>> has anyone else experienced this challenge , when you have irs900 4 channels and same power supply in my case "
(+\-40vts, outputs irfp250n,gate resistor 100ohms, output inductor 22turns eei 35, aux 12voltage each channel with its own 12vlts dropping the voltage from 0-40vlts by a small fet irf640 ) "

if i do not connect a speaker to any channel . it drains a lot of current and cuts the fuse . but if all channels have a load i.e speaker connected everything is o.k and works perfectly provided i do not disconnect a speaker from any of the 4 channels. what is the solution?

You're using IRFP250 which has a high Qg, provide at least 1A aux supply per module.

Looks like the amp is unstable and the mosfets are slightly turning on when the amp is not oscillating. Does each of the irf640 aux supply too far away from the modules?

Quick and dirty solution is to make sure that the amp is oscillating without speakers connected. IRS900 already has a 1.5k 2W dummy load.
Connect a preamp with "turn on pop" so that the amp will oscillate at startup. :D Just use a speaker protect with delay so that you won't hear it.
 
D

Deleted member 148505

it minimizes the voltage sag to a point that it only drops 1-2 volts during huge bass hits. this is to a sealed 15" sub with only 86 dB sensitivity.

since it's bridged, power supply voltage are kept low. thus allowing for added rail capacitance without using high voltage caps. i have about 40,000uf for each channel.

with bridged configuration for subwoofer use, the sub only handle bass signal which will never be continuous. put enough rail capacitance and transformer load will be reduced as the current will be sourced from the caps. i've put 10A fuse from the trafo line to the cap bank to see if the current draw will surge past 10A and it didn't burn no matter how loud i went.
from the cap bank to the amp modules, current surge is more than 5A during heavy bass hits at high volume. 10A fuse don't burn. 40,000uf in the bank (for each module), another 8,000uf reservoir in the module itself.

my trafo don't even get warm after movies playback. it's only rated for 500VA driving 2 bridged modules with about ~1.2kW each. following the general opinion in here, i need at least 2kW transformer to get full capability.

Did you test it with pure sine wave input signal? Music signal's average power is only 1/8 of sine wave, it might be the reason why your fuses didn't burn and your'e only getting 5A current surge.

Anyway as you said since the bass signal will never be continuous, smaller transformer will suffice specially if you put a lot of capacitance.
 
Last edited by a moderator: