• These commercial threads are for private transactions. diyAudio.com provides these forums for the convenience of our members, but makes no warranty nor assumes any responsibility. We do not vet any members, use of this facility is at your own risk. Customers can post any issues in those threads as long as it is done in a civil manner. All diyAudio rules about conduct apply and will be enforced.

LJM Audio

Right, Are you sure you soldered the IC in correct way?
Best wishes,

The 2092 requires quite a few conditions to be right before it wil output pulses to the mosfets. VAA and VSS must be right, check with a scope or multimeter. Mine was about 7v and -7v respectively.

VCC must be very clean, I used a 12 volt regulator but still got spikes on VCC. So i put a 100uf close to the 2092 and a 47uf close to the regulator.
The mosfets need decoupling close by, 220nf from b+/gnd, b-/gnd and b+/b-.
 
Ex-Moderator R.I.P.
Joined 2005
yeah, some of the 55V and 60v are down at 500pf, and very low "miller"
and even more so, the closest N and P devices seems to be fairly well matched
yeah, I vacuumed all the spec sheets a hundred times :eek: :clown:

reason I ask you guys is
some of the interesting higher voltage are recommended for classD("B" devices), and almost all of them have high transconductance
which means high gain, and efficiency, yes ?

so, wheres the hatch, whats the possible problem with such devices ?
instability issue maybe ?
and what may the advantages be ?
 
Right, Are you sure you soldered the IC in correct way? Check the electrolytic caps polarities and be sure there is not short circuit. I also recommend to clean the board from solder paste with isoporpyl alcohol or thinner. After all these, first turn on the aux power to see whether you are getting the reference voltage from the IC and check its temperature as well (without output mosfets).
Best wishes,
That's an insult ;) Yes they are all connected the right way around.
Which MOSFETs are you using ?
It definitely works with irfb4019's.
I have 4 on my setup.

Does the csd pin go up and down ? this would indicate something is wrong with the power supply or the mosfets.
I too am using the irfb4019's.
The 2092 requires quite a few conditions to be right before it wil output pulses to the mosfets. VAA and VSS must be right, check with a scope or multimeter. Mine was about 7v and -7v respectively.

VCC must be very clean, I used a 12 volt regulator but still got spikes on VCC. So i put a 100uf close to the 2092 and a 47uf close to the regulator.
The mosfets need decoupling close by, 220nf from b+/gnd, b-/gnd and b+/b-.
My VCC was very clean and can take the high current spikes if needed. I also added a lot of capacitors (like i said in my last post)
 
yeah, some of the 55V and 60v are down at 500pf, and very low "miller"
and even more so, the closest N and P devices seems to be fairly well matched
yeah, I vacuumed all the spec sheets a hundred times :eek: :clown:

reason I ask you guys is
some of the interesting higher voltage are recommended for classD("B" devices), and almost all of them have high transconductance
which means high gain, and efficiency, yes ?

so, wheres the hatch, whats the possible problem with such devices ?
instability issue maybe ?
and what may the advantages be ?
 
some of the interesting higher voltage are recommended for classD("B" devices), and almost all of them have high transconductance
which means high gain, and efficiency, yes ?

so, wheres the hatch, whats the possible problem with such devices ?
instability issue maybe ?
and what may the advantages be ?

Higher transconductance will lead to a faster change of the MosFet impedance in the moment of switching.
If you have a high transconductance device in a TO-220 package and some reasonable strong driver and reasonable layout, then the resulting di/dt will be dominated by the driver voltage, the voltage of the miller plateau (which sometimes is not a Miller plateau, but and an inductive caused plateau) and parasitic inductance of the TO-220 package.
High di/dt is desirable for low switching losses. On the other hand high di/dt is triggering high frequencies ringing more massive....(Everybody has to find his way of snubbering away resonances. Eva's RLC snubbering appears the most efficient to me so far.)
When playing around with the IRS20957 and IRFB4615, I settled the design at a reasonable di/dt around 700A/us - may be 800A/us.
I especially examined hard switching behavior, means the device which turns ON has to take over the load current and additionally has to deliver the reverse recovery charge of the body diode of the opposite MosFet.
At 31A load current this took slightly less than 70ns.
During these 70ns the MosFet has to handle increasing currents and is still operating at full rail voltage. Voltage sloping does not start before the load current is taken over + the reverse recovery charge has been removed.
==> Under this load conditions you are forcing losses !
At 400kHz already 30A are not really a reliable load for hard switching conditions with the IRFB4615 anymore. At 200-300kHz I could easily experiment with 35A, but at 400kHz 31A things touched the limits of the device. At 450kHz I already got defects at 27A....
Why looking for hard switching behavior? Because it is a typical load condition in class D amps, when driving load currents above the ripple current of the output filter.

Edit:
Rail voltage was +/-55V, well under heavy load dropping slightly below +/-50V....
 
Last edited:
Single channel, IRS2092S driven class D amp, output power up to 600W (with other MOS-FET's and caps) the shown one has 300W. the main application is multi-channel amplifiers. it only requires the main differential supply voltage, and has some ingenious power management scheme to save power as much as possible. Dead-time is set to min. 25ns, and the pk-pk overshoot is about 2V only, due to the tight layout.
I've made them for a friend who integrates them in some amps which he build. Soon i will have a batch for sale on my site too.

Nice!!! :up: :up:
 
hi people,

I have build a prototyp with irs2092 and irfb4212. Two things struck me.

-the amplifier have a strong noise, I can it hear till 50cm from the box.
-the amplifier disturb the the radio(HF-radiation).

Do you have similar experiences?

AMP in the use of noise in this very small, I took him to work in 680K HZ.
Using shielded filter inductance. 22UH, will be about +-40V 680K square wave signal filtering is very thorough.

The remaining approximately 200MV + - 680K HZ sine wave.

Most of the noise because the ground is because the PCB LAYOUT interference caused
Please refer to IR IRAUDAMP7S PCB LAYOUT.


I also have IR official template design products. In use, did not hear a difference.


Thanks everyone
 

Attachments

  • IMG_2690.jpg
    IMG_2690.jpg
    295.7 KB · Views: 2,994
I had a problem with my first 2092 pcb.
I tracked it down to noise on the VCC line.
I fixed it with a 47uf right next to the 12 volt regulator.
I have a question about that. My circuit isn't really functioning well.
I followed the application diagram except i added some decoupling. A BIG capacitor over the drain and source of both fet's, a BIG capacitor over the B+ and B- and a pretty big one from my VCC to B-.

For the VCC i used a 7812 with input +B, reference (gnd) -B and output the VCC line. I found this in the IRAUDAMP document, but that means it has a voltage drop of 40V or something? Isn't that way too high?

THen my last question: Is it possible to test it without connecting the low pass filter and speaker? I want to measure it first before adding a big load..
 
For the VCC i used a 7812 with input +B, reference (gnd) -B and output the VCC line. I found this in the IRAUDAMP document, but that means it has a voltage drop of 40V or something? Isn't that way too high?

THen my last question: Is it possible to test it without connecting the low pass filter and speaker? I want to measure it first before adding a big load..

I use a 200ohm dropper resistor from 45 volts to the regulator.
You will need to decouple the regulator on the input and output as close to the regulator as possible

I was getting something like a 1KHz tone when the regulator wasnt decoupled properly.

I was also getting quite a bit of background noise until I earthed the input ground.
 
Boudemaniak, show the pcb
An externally hosted image should be here but it was not working when we last tested it.

An externally hosted image should be here but it was not working when we last tested it.


I am still unsure about 2 things.
1. Isn't it working because i did add the load?
2. Are my decouple capacitors placed correctly? (1 over each fet from drain to source.. shouldn't it be drain to ground?)
 
1.well load is part of the output filter, so yes you should have it connected (even if few times higher then intended)
2. capacitor from v+ to gnd, and from gnd to v-

Power tracks are narrow, probably too long also, only single gnd (you could find my amp or look for 2092 app note how grounds should be placed)
 
1.well load is part of the output filter, so yes you should have it connected (even if few times higher then intended)
2. capacitor from v+ to gnd, and from gnd to v-

But the feedback is placed before the low pass filter. I thought this shouldnt be a problem.. But ok, i'll add the LPF and a power resistor.

And capacitors should not be placed over the drain and source then? Man, what was i wrong:p I'll try it now :)
 
Last edited:
Well, did all of the above things and added a 100Ohm resistor as load. The current now is 200mA for V- and nothing for V+. It seems like the IC isn't driving the fet's (although we dont have a short now :D).

Is the VCC correct? I used a Lab-supply with the - connected to the V- and +12V connected to the VCC. This means that it's -8V (with +-20V supply) with respect to ground.

Any ideas?

Ps: CSD Isn't resetting, so that's a good thing