Forward converter switching frequency

Thread Starter

Imre Norbert

Joined Feb 24, 2021
24
Okay, I am not familiar with that topology, so I cannot be of any more help. Looks interesting though. The inductors are not performing the function I imagined, so you can ignore what I said about their sizing.

Bob
Well, this is the basic topology for a 2 output forward as far as I know.
Thank you for your time and help, have a good day.
 

Papabravo

Joined Feb 24, 2006
21,159
I thought we were dealing with MOSFET switching losses, but there don.t seem to be any MOSFETS in the half of the complete circuit diagram you provided. For the components in the portion of the diagram that you provided, the losses come from
  1. winding resistance in the inductors and transformer.
  2. ESR (Equivalent Series Resistance) in the capacitors. This one takes datasheet forensics to uncover.
  3. Any resistors in a high current path. (I didn't see any)
If there are switching components on the other side of the transformer we should be looking there.

I'm just curious about why you thought providing half a schematic was going to be helpful?
 

Thread Starter

Imre Norbert

Joined Feb 24, 2021
24
I thought we were dealing with MOSFET switching losses, but there don.t seem to be any MOSFETS in the half of the complete circuit diagram you provided. For the components in the portion of the diagram that you provided, the losses come from
  1. winding resistance in the inductors and transformer.
  2. ESR (Equivalent Series Resistance) in the capacitors. This one takes datasheet forensics to uncover.
  3. Any resistors in a high current path. (I didn't see any)
If there are switching components on the other side of the transformer we should be looking there.

I'm just curious about why you thought providing half a schematic was going to be helpful?
Because Bob asked specifically only for the output side.
Yes that's right I have some losses there but the main problem on the secondary side is the snubber. That gets very hot, at least 70 Celsius.
The second biggest problem on the secondary side are the diodes, they also get very hot in range (70 to 100) C.
I have ordered more powerful diodes so I'm waiting to see those.
The windings don't get warm at all.
On the primary side I have an active clamp configuration but those FET-s are changed and now they are ok.
 

Papabravo

Joined Feb 24, 2006
21,159
Because Bob asked specifically only for the output side.
Yes that's right I have some losses there but the main problem on the secondary side is the snubber. That gets very hot, at least 70 Celsius.
The second biggest problem on the secondary side are the diodes, they also get very hot in range (70 to 100) C.
I have ordered more powerful diodes so I'm waiting to see those.
The windings don't get warm at all.
On the primary side I have an active clamp configuration but those FET-s are changed and now they are ok.
Since your schematic does not have reference designators on the components Being a bit more precise. Is the 470 Ω and 470 pf capacitor in front of the full wave rectifier what you are calling the snubber? You do realize that the impedance of 470 pf at your switching frequency (340 kHz) is about 1K Ω so it is not doing much to snub anything. You might as well get rid of it.
 

Thread Starter

Imre Norbert

Joined Feb 24, 2021
24
Since your schematic does not have reference designators on the components Being a bit more precise. Is the 470 Ω and 470 pf capacitor in front of the full wave rectifier what you are calling the snubber? You do realize that the impedance of 470 pf at your switching frequency (340 kHz) is about 1K Ω so it is not doing much to snub anything. You might as well get rid of it.
I am calling the resistor and capacitor an rc snubber and I drew it in front of the diodes because thats how it is on the board as well. I dont have much experience with snubbers. I mentioned in a comment that i calculated a capacitance of about 27pF,would that be any better?.
 

Thread Starter

Imre Norbert

Joined Feb 24, 2021
24
MOSFETS have an associated behavior called the Miller Effect where the gate voltage on the way up and on the way down is on a relatively flat plateau. This places the device in the linear region where it creates nearly a dead short between the supply and Ground. You can minimize this effect by using a current source to drive the MOSFET gate. There is a datahseet parameter that tells you how much charge you have to provide to turn the gate on or off. It has units of Coulombs
The mosfets have a gate threshold of 4 volts and a maximum rating of 20 volts. I give them 13-15 volts.I think that they are saturated enough.
 

Papabravo

Joined Feb 24, 2006
21,159
The mosfets have a gate threshold of 4 volts and a maximum rating of 20 volts. I give them 13-15 volts.I think that they are saturated enough.
It is not the voltage level that is important. It is the transit time from on to off and off to on where the Miller effect shows up. What you will observe on the scope is a noticeable plateau in the gate voltage between Vth (Gate threshold voltage) and the point where the channel is fully on. Depending on the nature of the gate drive circuit and it's ability to charge and discharge the gate capacitance the amount of time it spends on this plateau can significantly increase the transition time. It is worth taking a look at.
 

Papabravo

Joined Feb 24, 2006
21,159
I am calling the resistor and capacitor an rc snubber and I drew it in front of the diodes because thats how it is on the board as well. I dont have much experience with snubbers. I mentioned in a comment that i calculated a capacitance of about 27pF,would that be any better?.
I don't know what sort of better you might be after. The usual reason for having one is to provide a low impedance path to AC ground for high frequency noise. Since 27 pf is an order of magnitude smaller than 470 pf my it should work for stuff that is higher in frequency than the switching frequency (300-340 kHz IIRC) I still have to ask - what do you think is the actual benefit of having it there in this particular circuit? There is enough downstream inductance to kill most any high frequency hash. Would it kill you to get rid of it since it is causing a problem?
 

Thread Starter

Imre Norbert

Joined Feb 24, 2021
24
It is not the voltage level that is important. It is the transit time from on to off and off to on where the Miller effect shows up. What you will observe on the scope is a noticeable plateau in the gate voltage between Vth (Gate threshold voltage) and the point where the channel is fully on. Depending on the nature of the gate drive circuit and it's ability to charge and discharge the gate capacitance the amount of time it spends on this plateau can significantly increase the transition time. It is worth taking a look at.
Oh, okay. I measured that. It takes about 0.1us from 0V to 10V and about 0.15us to reach 15V.
 

Thread Starter

Imre Norbert

Joined Feb 24, 2021
24
I don't know what sort of better you might be after. The usual reason for having one is to provide a low impedance path to AC ground for high frequency noise. Since 27 pf is an order of magnitude smaller than 470 pf my it should work for stuff that is higher in frequency than the switching frequency (300-340 kHz IIRC) I still have to ask - what do you think is the actual benefit of having it there in this particular circuit? There is enough downstream inductance to kill most any high frequency hash. Would it kill you to get rid of it since it is causing a problem?
Well it was in the original circuit and I figured it is there to limit the high voltage spikes, so you say that there is enough downstream inductance to limit this. What does that mean(downstream inductance)?
 

Papabravo

Joined Feb 24, 2006
21,159
Oh, okay. I measured that. It takes about 0.1us from 0V to 10V and about 0.15us to reach 15V.
OK so is the plateau a significant fraction of the 150 nanosecond turn-on turn-off time. Is it 1/3 to 1/2 of that time period? If so it might bear further investigation. If it is less than 1/3 (50 nsec) then it is probably OK.
 

Thread Starter

Imre Norbert

Joined Feb 24, 2021
24
OK so is the plateau a significant fraction of the 150 nanosecond turn-on turn-off time. Is it 1/3 to 1/2 of that time period? If so it might bear further investigation. If it is less than 1/3 (50 nsec) then it is probably OK.
No it isn't significant, I don't even see it.
 

Papabravo

Joined Feb 24, 2006
21,159
Well it was in the original circuit and I figured it is there to limit the high voltage spikes, so you say that there is enough downstream inductance to limit this. What does that mean(downstream inductance)?
Downstream from the snubber is the full wave bridge rectifier and downstream from that is the common mode choke (inductors) and the filter capacitors. It is the flow from left to right on the schematic.
 

Thread Starter

Imre Norbert

Joined Feb 24, 2021
24
Downstream from the snubber is the full wave bridge rectifier and downstream from that is the common mode choke (inductors) and the filter capacitors. It is the flow from left to right on the schematic.
Yes I understand that as I am moving from left to right things get more stable, this is the point of the inductor capacitor and so on, but the spike still remains across the secondary, or that isn't a problem?
 

Papabravo

Joined Feb 24, 2006
21,159
Yes I understand that as I am moving from left to right the thing get more stable, this is the point of the inductor capacitor and so on, but the spike still remains across the secondary, or that isn't a problem?
Can you see it on the filter capacitors (the ones after the inductors)? What are the rise and fall times with respect to the switching frequency?
 

Thread Starter

Imre Norbert

Joined Feb 24, 2021
24
Can you see it on the filter capacitors? What are the rise and fall times with respect to the switching frequency?
Right now I am not at work but the next thing I'll check will be the output ripple. So if the ripple is okay, then I shouldn't bother with the snubber and spikes?
 

Papabravo

Joined Feb 24, 2006
21,159
Right now I am not at work but the next thing I'll check will be the output ripple. So if the ripple is okay, then I shouldn't bother with the snubber and spikes?
That would be my take. There are also other devices you might wan to consider if high frequency high voltage noise continues to be a problem. I'm just not sure of the frequency range of the devices at the moment. I will check.
 

Thread Starter

Imre Norbert

Joined Feb 24, 2021
24
That would be my take. There are also other devices you might wan to consider if high frequency high voltage noise continues to be a problem. I'm just not sure of the frequency range of the devices at the moment. I will check.
I got rid of the snubber and, here are the results. The first one is with the snubber. Looks like the snubber makes a world of difference. This is the voltage form the first secondary coil's * notation with reference to the ground.
 

Attachments

Top