Speaking about a SMPS for auto applications (12-13.3 Vin), why do most people use 55V mosfets for this? I built a SMPS prototype using 25V Infineon IPP03N03LA mosfets (2 per side) and seem to have no problems. In fact it seems to be working very well as they have a RDS of only 3 mOhm each.
Second question is why do many designs use a resistor on the gate? I am currently driving the Mosfets with an IXDD414PI 14A Ultrafast driver and am also having no problems. In the final design I plan to use a 4A inverting/non-inverting dual driver.
Second question is why do many designs use a resistor on the gate? I am currently driving the Mosfets with an IXDD414PI 14A Ultrafast driver and am also having no problems. In the final design I plan to use a 4A inverting/non-inverting dual driver.
Assuming that your topology is push-pull, you have to leave room for input voltages as high as 16V and you have to be aware that the leakage inductance of the transformer will cause severe avalanche at turn-off with heavy loads. Your prototype is not likely to survive too long in a car.
Also, a resistor or a ferrite bead are almost mandatory in series with the gate in order to control turn-on and turn-off slopes, to limit gate drive current, and most important, to prevent RF oscillation in the 50 to 250Mhz range due to source lead inductance. Your circuit with a 14A driver attached directly to the gates is probably a nice FM transmitter now.
I have recently found out by experimentation that ferrite beads in series with the gate exhibit very interesting properties and can actually produce faster (ringing-free) turn-on and turn-off transiets than a resistor or nothing, with a dramatical reduction in gate driver dissipation and current consumption. It's all a matter of making the inductance of the bead resonate a bit with gate capacitance 😀😀😀 but I'm too busy now to explain that in detail.
Also, a resistor or a ferrite bead are almost mandatory in series with the gate in order to control turn-on and turn-off slopes, to limit gate drive current, and most important, to prevent RF oscillation in the 50 to 250Mhz range due to source lead inductance. Your circuit with a 14A driver attached directly to the gates is probably a nice FM transmitter now.
I have recently found out by experimentation that ferrite beads in series with the gate exhibit very interesting properties and can actually produce faster (ringing-free) turn-on and turn-off transiets than a resistor or nothing, with a dramatical reduction in gate driver dissipation and current consumption. It's all a matter of making the inductance of the bead resonate a bit with gate capacitance 😀😀😀 but I'm too busy now to explain that in detail.
the automotive mosfet can handle high s-d pulses in both current and voltage.
55 volts seems to handle the peaks from other high current switching circuits and the starter trash that comes through the system. the rds is also ussualy lower the lower the voltage rating of the mosfet.
this is however, under review due to the possibility the automotive voltage systems going to 36 or 48 volts.
55 volts seems to handle the peaks from other high current switching circuits and the starter trash that comes through the system. the rds is also ussualy lower the lower the voltage rating of the mosfet.
this is however, under review due to the possibility the automotive voltage systems going to 36 or 48 volts.
I am using IRF1302 mosfets in my P3A car amp for almost three years now and I haven't replaced anything from it since it was built.
that was when I still have a lot to learn. 🙂 20V 4mOhm mosfets aren't the best choice but they were almost free. to note, I have tested it to work at 15V input and it didn't fail. must be a good example of how much safety margin IR puts in their devices.
since then, I have used 55V volt or greater mosfets but just wanted to show something not designed right could work problem free for quite a while. 😀 not something I would recommend though.🙄
that was when I still have a lot to learn. 🙂 20V 4mOhm mosfets aren't the best choice but they were almost free. to note, I have tested it to work at 15V input and it didn't fail. must be a good example of how much safety margin IR puts in their devices.
since then, I have used 55V volt or greater mosfets but just wanted to show something not designed right could work problem free for quite a while. 😀 not something I would recommend though.🙄
Eva said:
I never thought about the FM transmitter, I think I will be addind a ferrite bead then!
As for the avalanche energy on the Mosftet I can understand that but the particular FET I am using has a rather high avalanch rating, so I hope it will be ok ID=80 A, RGS=25 Ù, Avalanche energy 960 mJ)... ... I have 13 of them here (that is the reason I want to use these).
thefish said:
EMI suppression is something that has to be done "on demand". Haven't you carefully checked gate and drain waveforms (among others) and their switching transients in detail with an oscilloscope having reasonable bandwidth (100Mhz)?
Also, conecting a 1x oscilloscope probe (non-attenuated) to a loop antena with a couple of turns you can get an idea of how much stuff your circuit is radiating, and you can even trace the origin of the ringing...
Eva said:
APT have appnote availlable at their site about ferrite bead stuff...
Selecting best bead size and material is pure guesswork, but luckily these are inexpensive and testing is easy.
I was just checking the drain waveform on my 300mhz scope and I see the squarewave is 24V peak to peak (SMPS connected to 12.0V). Why would this be the case, how is the voltage doubling? Each leg on the primary side only has 12 volts across it right? BTW it is push-pull topology.
Congratulations, you have just discovered the push-pull transformer windings arrangement 😀😀😀
And yes, the switches have to whitstand twice the supply voltage when they are off (plus the inductive spike due to leakage inductance). Do you understand now why 60V MOSFETs are routinely employed?
BTW: When you manage to understand transformer basics you will understand why the voltage is doubled.
And yes, the switches have to whitstand twice the supply voltage when they are off (plus the inductive spike due to leakage inductance). Do you understand now why 60V MOSFETs are routinely employed?
BTW: When you manage to understand transformer basics you will understand why the voltage is doubled.
Eva said:
By trade/schooling I am a mechanical engineer all my electronics knowledge came from only 3 classes (Basic Circuits, Digital Circuits, Experimental Methods and Sensors), so I don’t really have that much of a background in this stuff but I always found it interesting and that is why I am trying to do some more stuff on the side now. All my transformer knowledge is from an “Electronics and Magnetism" physics class I took 4 years ago but it was mostly general theory like the right hand rule, faradays law, flux linkage, ect, but I don’t recall much about it. Could you point me the way to some info about this effect with the push-pull setup in the transformers or just explain the short version?
Seeing as how the load on the MOSFET is 2 times the input voltage then a 40V FET should be ok right? I can get some IRL1404PBF 40V 4mOHM for $1.76 each that is preatty cheap I think.
Eva said:
Fish:
Here is a laundry-list of books I regularly recommend to members asking questions about SMPS design in general, or specific ones about the different sections of a Switcher:
1) High Frequency Switching Power Supply Design - Geo. Chryssis (c) 1989 (Don't know ISBN)
2) Power Supply Cookbook (EDN Series for Design Engineers)
by Marty Brown; 2nd edition May 2001 ISBN: 075067329X
3) Switchmode Power Supply Design - Abraham I. Pressman (Don't know year or ISBN).
These three books are, IMHO, the Holy Graile of power supply design.
The power switch section of any of them gives the trade-offs between MOSFETs and Bipolars and the primary side waveforms for most of the discussed topologies. When Eva congratulates you for discovering windings arrangement of the classic push-pull topology, she means (and I do not presume to speak for her) 😀 is that as the switch (MOSFET) at each end of the primary turns off, there is a spike of at least the supply voltage in the opposite direction, plus the leakage inductance spike.
So, assuming a worst-case scenario of a maximum supply voltage of 16v, you will see a spike of 2 x (16V + leakage inductance spike of, say, 20-40%) plus a guard band of another 15% for a total minimum Vdss of (2x (16 + 6.4)) +6.72) = 51.52V. This is why Eva says MOSFETs w/ 55v Vdss rating is the minimum safe value for a +12V-powered push-pull.
To clarify this point a bit further, if you were to remove the secondary side all together, and just have the two halves of the primary connected as usual, you will see these spikes at the drains of the MOSFETs. If you were to connect high-speed rectifiers at these junctions, and then smooth the resultant waveform with a capacitor, you would have a nice two-phase DC-DC boost converter (look carefully at the drawing) of about twice the input voltage + the leakage inductance spike.
Hope these comments clarify things a bit.
Steve
Hi Fish
Now the answer after leaving you some time to think: The voltage is doubled because the push pull windings are in the same transformer core and thus they are coupled (although not perfectly, that's why there is something called "leakage inductance", but that's a more complex subject). As the center tap is connected to 12V and one of the push-pull sides is pulled to 0V, the other side rises to 24V. All windings in a transformer are coupled (in an ideal transformer each turn shows the exact same voltage drop), you wouldn't be able to get any output from the secondaries otherwise 😀
Concerning the MOSFETs, I tend to dislike exotic or expensive devices unles they are *really* required. Last time I bought 100 inexpensive IRFZ48V and those are the ones that I have been using for building and repairing push-pull 12V converters for a long time.
Now the answer after leaving you some time to think: The voltage is doubled because the push pull windings are in the same transformer core and thus they are coupled (although not perfectly, that's why there is something called "leakage inductance", but that's a more complex subject). As the center tap is connected to 12V and one of the push-pull sides is pulled to 0V, the other side rises to 24V. All windings in a transformer are coupled (in an ideal transformer each turn shows the exact same voltage drop), you wouldn't be able to get any output from the secondaries otherwise 😀
Concerning the MOSFETs, I tend to dislike exotic or expensive devices unles they are *really* required. Last time I bought 100 inexpensive IRFZ48V and those are the ones that I have been using for building and repairing push-pull 12V converters for a long time.
Hi,
it´s not only overvoltage caused within the PSU.
External voltage spikes and dumps are even worse.
Ask LT, why the call the cigarette lighter plug in a "Supply from Hell".
Check also the ST apllication notes AN553 and 554 for behavior and testing of this "only 12 Volt" gear. AN553 is one of the rare documents, that show the shape and voltage of the standard test pulses.
http://www.st.com/stonline/stappl/p...ome.myComboChange&keyword=&ptype=AN&new=false
Regards,
Onra
it´s not only overvoltage caused within the PSU.
External voltage spikes and dumps are even worse.
Ask LT, why the call the cigarette lighter plug in a "Supply from Hell".
Check also the ST apllication notes AN553 and 554 for behavior and testing of this "only 12 Volt" gear. AN553 is one of the rare documents, that show the shape and voltage of the standard test pulses.
http://www.st.com/stonline/stappl/p...ome.myComboChange&keyword=&ptype=AN&new=false
Regards,
Onra
- Status
- Not open for further replies.