Ty Dzony ,
How much watts are you getting??
So you now are using irfp250 or IRFB4227 ?
The inputs are getting 11v , the supply its 12v , the 1 logical its about 9.5v to 12v.
ty
Last edited:
What is the resistor before the 12V zener? You should check it. At your schematic its 4k7, but it should be 6k8 for you supply rail (but that should not be an issue).
If you have a FET with high gate charge you should not decrease the turn on resistor. At higher temperature, the fet usually gets slower, so cross-conduction rises. Ant it can kill not only the FETs, but driver, and every other.
Put some scope probe on the supply rails, and see them at load condictions. You have absolutely no decoupling, they will be extremely noisy. Then anything can happem...
Your gate signals are really messy, but switching signal seems nice.
You must put the oscillation frequency up over 200-250kHz. I looked the spanish forum, and from the schematics (as I cannot understand spanish) nobody tried the normal connection of the comparator, and even raising switching freq. I haven't seen any input stage with differential inputs.
(offset is a tiny problem, and for example differential inputs should decrease it too, and I've seen lots of photos with air-coil inductors... forget them. Most of all the inductors heat up because low switching frequency and the bad filtering)
Rule of tumb: switching freq should be a decade higher then corner freq, to ensure good filtering of carrier signal.
Dzony988: output filter is good as is (for 4-8 Ohm load), with your values, the corner frequency incereases. For 2 Ohm you should redesign the filter: 20uH and 2.2uF would be better.
Well-well-well.
I did the same, tried with cheap IRFP240 first. Then with IRFB4227 for big power...
Now I even made some modules with IRFP250N ("el-cheapo" models), and I was surprised, I expected worse. The disspiation only became a problem at high loads.
new pcb design with dc and current limit
Hi guys,
I think the trouble is they really understand how this amp UCD
and it is likely that the sequence of failure is that the transistor Q5 is set to short, and destroys the IR2110 IC with failure of the other components
as 5401 and the mosfet
I comment on how important this good design is to choose well the first mosfet value
vds max 20% of the value to be used in vcc
and consider it very important Qg value always less than 100 Nc
this depends on which ic not going to burn 2110
I recommend it because it works great is to put a 12 v source independently of negative v vcc pin 3 of IC IR2110 is removed so the transistor Q5 and heater and many possibilities of failure of the amplifier
L1 coil recommend an air core is much more linear at different loads and anyone can do it. how?
16 gauge wire about 56 laps in a diameter of .750 "and diameter of them
the capacitor C4 is very easy to get from any source atx is in series with the main transformer and is proper for high Frequence work
Soon I will put a PCB to Protect DC and current limit
all greetings from MEXICO
Hi guys,
I think the trouble is they really understand how this amp UCD
and it is likely that the sequence of failure is that the transistor Q5 is set to short, and destroys the IR2110 IC with failure of the other components
as 5401 and the mosfet
I comment on how important this good design is to choose well the first mosfet value
vds max 20% of the value to be used in vcc
and consider it very important Qg value always less than 100 Nc
this depends on which ic not going to burn 2110
I recommend it because it works great is to put a 12 v source independently of negative v vcc pin 3 of IC IR2110 is removed so the transistor Q5 and heater and many possibilities of failure of the amplifier
L1 coil recommend an air core is much more linear at different loads and anyone can do it. how?
16 gauge wire about 56 laps in a diameter of .750 "and diameter of them
the capacitor C4 is very easy to get from any source atx is in series with the main transformer and is proper for high Frequence work
Soon I will put a PCB to Protect DC and current limit
all greetings from MEXICO
Air coils are big, ineffective, and radiate noise.
Because they are big they were able to dissipate the loss because of the high current (because of low switching freq).
Since this is an UcD amp, the feedback is after inductor, so every non-linearitiy of the inductor is compensated by the feedback. Gapped ferrites also very linear until 80% of saturation. Iron-powder cores are less linear.
So chosing air-core inductor for linearity is stupid. I think they chose it because any other could not dissipate, mainly because the low switching freq. A bad solution for a bad choice.
Gate signal is closest to mosfet.
Power supply line is direct on mosfet pins.
Capacitors are closest to Drain and Source pins.
What do you think about this new pcb routing?
Is there anything to change ?
I´ll apreciate your comments and tips.
ty
regards!
P.D. I´ll add some test points to pcb.
Power supply line is direct on mosfet pins.
Capacitors are closest to Drain and Source pins.
What do you think about this new pcb routing?
Is there anything to change ?
I´ll apreciate your comments and tips.
ty
regards!
P.D. I´ll add some test points to pcb.
Attachments
Last edited:
2000W/4ohm rms power UCD class D SMPS inteagrated
It is good circuit indeed ,we have developed some amp like this circuit,the concept same circuit not ,works well,powerful and stable,it is not perfect choose to select IRS2110 as driver,it is so hard to proper control the DT time.
2000W-4000W/40hm range behave uncommon.
Prototype pics below:
It is good circuit indeed ,we have developed some amp like this circuit,the concept same circuit not ,works well,powerful and stable,it is not perfect choose to select IRS2110 as driver,it is so hard to proper control the DT time.
2000W-4000W/40hm range behave uncommon.
Prototype pics below:
Attachments
Last edited:
Two methods
ne is IR series DT control Floating Driver(irs20124/irs2092/irs20955 etc.),the other is high speed optocupler + mosfet diver+mosfet body diode unavailable circuit +RC delay circuit.It is only ourselves experiences,maybe friends have good methods on it.
Gate signal is closest to mosfet.
Power supply line is direct on mosfet pins.
Capacitors are closest to Drain and Source pins.
What do you think about this new pcb routing?
Is there anything to change ?
I´ll apreciate your comments and tips.
ty
regards!
P.D. I´ll add some test points to pcb.
PCB seems nice. C9 is not necessary, instead increase C6 to 1u ceramic. Then you can make gate drive path shorter.
What does the opocoupler do? Only MUTE function?
Use the normal connection of the comparator. I reccomend to redesign the input to be balanced. Then the 47k trimmer is not necessary, (if you decresease R5 to 10k then the offset is also smaller, or put a series capacitor with R2 to ground)
This weekend I made an amp with this (with modifications) for sound reinforcement in a church. It is very nice. The 250W @ 4 Ohm amp, with SMPS, and preamplifier with into a 20x10x6 cm box.
I reccomend you do biamping. It is much safer for the tweeter, and you can even push it much harder.