HI all,
I wish to explain some of the latest developments concerning my Universal Switch Mode Power Supply. This has been a non-commercial endeavor of mine to utilize my electronics abilities.
The original circuit came into existence about 3 1/2 years ago. I began a group to share it about 6 months later. Since then, the circuit has gradually evolved.
The first addition to the original circuit which was a break from normalcy was the use of a capacitor in series with the secondary to limit current draw on the output. I have recently found that it works even better to place the capacitor on the primary side where the coupling capacitor normally goes. The reason it is better there is that it reduces the value of capacitance needed to one which is readily available. I had tried using a current limiting capacitor in this location in the past, but I didn't think of adding diodes where it joins to the primary winding. The diodes shunt the primary winding flyback energy to both the common and the primary B+. (more later)
I wish to explain some of the latest developments concerning my Universal Switch Mode Power Supply. This has been a non-commercial endeavor of mine to utilize my electronics abilities.
The original circuit came into existence about 3 1/2 years ago. I began a group to share it about 6 months later. Since then, the circuit has gradually evolved.
The first addition to the original circuit which was a break from normalcy was the use of a capacitor in series with the secondary to limit current draw on the output. I have recently found that it works even better to place the capacitor on the primary side where the coupling capacitor normally goes. The reason it is better there is that it reduces the value of capacitance needed to one which is readily available. I had tried using a current limiting capacitor in this location in the past, but I didn't think of adding diodes where it joins to the primary winding. The diodes shunt the primary winding flyback energy to both the common and the primary B+. (more later)
The original version but with modified gate drive and the improved capacitive current limiting suffers rather much from loop stability dependence on the output load. In the ZVS version, the feedback to the oscillator seems to help keep the output ripple synchronized with the control signal which is used for MOSFET drive. Therefore, loop stability tends to be much better in the ZVS version. The switching frequency does vary during this process, but how much depends largely on the turns ratio of the transformer and the capacitance in parallel with the transformer primary winding.
For the gate drive, I have added three components after each MOSFET gate drive output of the IR211x ICs to act as buffers. I have been building all my bench test circuits this way. (more later)
Please pardon my compact diagram style. The original version but with modified gate drive and the improved capacitive current limiting:
For the gate drive, I have added three components after each MOSFET gate drive output of the IR211x ICs to act as buffers. I have been building all my bench test circuits this way. (more later)
Please pardon my compact diagram style. The original version but with modified gate drive and the improved capacitive current limiting:
An externally hosted image should be here but it was not working when we last tested it.
I have simulated a new feedback circuit for the original type USMPS. The result is improved stability and regulation. The additional transistor increases the gain on the secondary side of the optocoupler. Normally more gain reduces stability. At first I was thinking that the improvement was due to asymetrical charge/discharge of the compensation capacitor. But the slightly simpler way did a similar thing. Now I'm thinking that the real reason is that the extra transistor compensates for the slow response of the optocoupler, as the transistor on the secondary side of the optocoupler does.
Originally I had tried the TL431 in the simulation, but it was not as well-behaved as the transistor and zener diode. I show the circuit with a 5khz sine wave source used as the load on the output. It did well regulating with that load.
I still plan to get back to the topic of my personally favorite rendition of the USMPS idea, the ZVS USMPS.
latest non-ZVS diagram
Originally I had tried the TL431 in the simulation, but it was not as well-behaved as the transistor and zener diode. I show the circuit with a 5khz sine wave source used as the load on the output. It did well regulating with that load.
I still plan to get back to the topic of my personally favorite rendition of the USMPS idea, the ZVS USMPS.
latest non-ZVS diagram
An externally hosted image should be here but it was not working when we last tested it.
I am not showing the ZVS USMPS circuit in a form that I think best represents the real life one on my test bench at this time. The reason is that I have devised two theoretical improvements which I cannot be sure that I will get to incorporate into the bench test circuit but which I would still like to include
in the diagram of the LTspice version.
This LTspice version is done in a way that leakage inductance of the transformer is low although the real-life test version uses a transformer purposely having leakage inductance since it saves needing one extra external inductor. I realize that many people like to use e-core transformers where the windings are on top of one another, generally giving less leakage inductance.
The improvement in the non-ZVS feedback which I mentioned above is one of the two improvements I mentioned for the ZVS version too. On a different note, I have chosen the value of the primary winding coupling capacitor to help limit current in case of an overload condition. Another benefit of its having a smaller value in that way is that under heavy load as it begins to cut back on primary current, the rise time of the kick-back voltage on that primary winding is kept from decreasing as much, theoretically lowering EMI generation.
Latest ZVS USMPS diagram done in LTspice
in the diagram of the LTspice version.
This LTspice version is done in a way that leakage inductance of the transformer is low although the real-life test version uses a transformer purposely having leakage inductance since it saves needing one extra external inductor. I realize that many people like to use e-core transformers where the windings are on top of one another, generally giving less leakage inductance.
The improvement in the non-ZVS feedback which I mentioned above is one of the two improvements I mentioned for the ZVS version too. On a different note, I have chosen the value of the primary winding coupling capacitor to help limit current in case of an overload condition. Another benefit of its having a smaller value in that way is that under heavy load as it begins to cut back on primary current, the rise time of the kick-back voltage on that primary winding is kept from decreasing as much, theoretically lowering EMI generation.
Latest ZVS USMPS diagram done in LTspice
An externally hosted image should be here but it was not working when we last tested it.
- Status
- Not open for further replies.