Take a good luck
the first SMPS I had to fix was in an HP portable test set, then a Genrad test set -- in the early days things blew up with regularity. I have this view of SMPS designers in the 1970's being akin to "Doc" in the "Back to the Future" series.<p> Some of the components bear loads which were probably not properly analyzed two decades ago, and a little fatigue and whammo, there goes a cap and one of the switching transistors.
the first SMPS I had to fix was in an HP portable test set, then a Genrad test set -- in the early days things blew up with regularity. I have this view of SMPS designers in the 1970's being akin to "Doc" in the "Back to the Future" series.<p> Some of the components bear loads which were probably not properly analyzed two decades ago, and a little fatigue and whammo, there goes a cap and one of the switching transistors.
Switch speed vs. Slew/Transition speed...
I see there are a lot of good points being brought up here.
The comment about a lot of output capacitance is a good one. Most SPS designers are looking at mass production, and like to limit the number of expensive parts (i.e. filter caps). Board space is also a consideration. More caps on the output is always good. If you're homebrewing it, spend a litle extra for caps. Also, you'll have to research which caps to use, and pick some rated for the high frequency current. High frequency lytics are usually not cheap...
The point about slew rate is also a good. However, ther are some things to consider there. You can only get rid of so much inductance, there will always be some in the circuit. Looking at the SMPS controller, If the maximum duty cycle of the PWM chip is 90%, you'll need to turn off the switching device in 10% of your duty cycle (Often referred to as dead time). That means, there is a limit to how low the slew rate can be. If there is a lot of inductance in the circuit, the only option is to lower the switching freqeuncy to buy more time to turn off the transistors. The other thing to consider with slew rate is that as the switch transitions become slower, switching losses increase. That is to say there is a practical limit to how slow the slew rate can be to minimize heating in the transistors.
Probably the most important thing will be to reduce leakage inductance. Focus on a good transformer design, and keep all lead lengths as short as possible, If you need a long leads/traces, try to keep them close togher and parallel.
As for the comment about the car audio amps... ...they all run Push-Pull power supplies. There's no other practical way to get that much power from a low voltage source. Most of them run around 50Khz or so, and usually a TL494 or similar SMPS controller (My personal favorite is the LM3524/CA3524/SG3524...). With the ammount of current being switched leakege inductance becomes a major problem, and reducing switch speed is the only way to keep the transients from taking out the switching transistors.
-Dan
I see there are a lot of good points being brought up here.
The comment about a lot of output capacitance is a good one. Most SPS designers are looking at mass production, and like to limit the number of expensive parts (i.e. filter caps). Board space is also a consideration. More caps on the output is always good. If you're homebrewing it, spend a litle extra for caps. Also, you'll have to research which caps to use, and pick some rated for the high frequency current. High frequency lytics are usually not cheap...
The point about slew rate is also a good. However, ther are some things to consider there. You can only get rid of so much inductance, there will always be some in the circuit. Looking at the SMPS controller, If the maximum duty cycle of the PWM chip is 90%, you'll need to turn off the switching device in 10% of your duty cycle (Often referred to as dead time). That means, there is a limit to how low the slew rate can be. If there is a lot of inductance in the circuit, the only option is to lower the switching freqeuncy to buy more time to turn off the transistors. The other thing to consider with slew rate is that as the switch transitions become slower, switching losses increase. That is to say there is a practical limit to how slow the slew rate can be to minimize heating in the transistors.
Probably the most important thing will be to reduce leakage inductance. Focus on a good transformer design, and keep all lead lengths as short as possible, If you need a long leads/traces, try to keep them close togher and parallel.
As for the comment about the car audio amps... ...they all run Push-Pull power supplies. There's no other practical way to get that much power from a low voltage source. Most of them run around 50Khz or so, and usually a TL494 or similar SMPS controller (My personal favorite is the LM3524/CA3524/SG3524...). With the ammount of current being switched leakege inductance becomes a major problem, and reducing switch speed is the only way to keep the transients from taking out the switching transistors.
-Dan
Brett,
For low voltage input <20 V, I like push pull foreword mode converters. The basic advantage is the number of switching devices needed. In full bridge, you end up with two switches in series with the primary at all times. For power levels above a couple of hundred watts the losses start to become a concern. When you get to a 2000W, the current is very high. (I've recorded 250+ Amp peaks). At 200+ amps, even .01 ohms in the power switches is a lot of voltage drop (=2Volts!). If you were to use a full bridge, that voltage drop would double to 4 volts or so. Half bridge, just isn't practical because of capacitor requirements. Flybacks and the other foreword mode converters aren't practical because of flux swing in the core.
Toroids are nice. It's easier to wind lots of the big wire for the primary on them...
I've spent a good bit of time doing SMPS supply repair in Car Audio amps for a friend of mine. He owns a car audio dealership and his customers break a lot of stuff... ...Everything I've seen so far is push-pull built around a toroid.
-Dan
P.S. I tried emailing you, and it says you don't have email enabled...
For low voltage input <20 V, I like push pull foreword mode converters. The basic advantage is the number of switching devices needed. In full bridge, you end up with two switches in series with the primary at all times. For power levels above a couple of hundred watts the losses start to become a concern. When you get to a 2000W, the current is very high. (I've recorded 250+ Amp peaks). At 200+ amps, even .01 ohms in the power switches is a lot of voltage drop (=2Volts!). If you were to use a full bridge, that voltage drop would double to 4 volts or so. Half bridge, just isn't practical because of capacitor requirements. Flybacks and the other foreword mode converters aren't practical because of flux swing in the core.
Toroids are nice. It's easier to wind lots of the big wire for the primary on them...
I've spent a good bit of time doing SMPS supply repair in Car Audio amps for a friend of mine. He owns a car audio dealership and his customers break a lot of stuff... ...Everything I've seen so far is push-pull built around a toroid.
-Dan
P.S. I tried emailing you, and it says you don't have email enabled...
Sorry Email Enabled Now.
Looks Like Toriod is the way to go. What Model do you recommend. Ive been looking for days at variouse brands etc.
I am having trouble referencing the manufactures design software for inductors to a Push Pull Power conversion design.
I need 2KW, 12v Input.
Thanks heaps
Brett
Looks Like Toriod is the way to go. What Model do you recommend. Ive been looking for days at variouse brands etc.
I am having trouble referencing the manufactures design software for inductors to a Push Pull Power conversion design.
I need 2KW, 12v Input.
Thanks heaps
Brett
Brett,
I like these guys, but all of the manufacturers are about the same.
http://www.mag-inc.com/
You could probably look at P or R type material.
Good applications notes on designing transformers/inductors.
http://www.mag-inc.com/library.asp
I would recommend: Switching Power Supply Design (Hardcover)
by Abraham I. Pressman (Check it out at amazon.com).
The book it a little weak on modern mosfet technology (Mosfets have advanced by leaps and bounds since this book was written), but has a bunch of good information. It's also a little weak on low voltage input supplies, but so are most other books.
As for the software goes, the transformer calculations for are really not very complicated. I always do these by hand, or in a spreadsheet. A few hints to remember. Reduce input voltage used in your calculations by the level dropped across the mosfets and in circuit resistance (traces, primary, and fuses). Also include output diode drop when calculating the secondary.
-Dan
I like these guys, but all of the manufacturers are about the same.
http://www.mag-inc.com/
You could probably look at P or R type material.
Good applications notes on designing transformers/inductors.
http://www.mag-inc.com/library.asp
I would recommend: Switching Power Supply Design (Hardcover)
by Abraham I. Pressman (Check it out at amazon.com).
The book it a little weak on modern mosfet technology (Mosfets have advanced by leaps and bounds since this book was written), but has a bunch of good information. It's also a little weak on low voltage input supplies, but so are most other books.
As for the software goes, the transformer calculations for are really not very complicated. I always do these by hand, or in a spreadsheet. A few hints to remember. Reduce input voltage used in your calculations by the level dropped across the mosfets and in circuit resistance (traces, primary, and fuses). Also include output diode drop when calculating the secondary.
-Dan
Have any of you used a toroid for mains operation?
I've been wanting to make a few different 120V SMPS designs.
1st one, I'm wanting to make 18V output or so to charge batteries in a 14.4V portable amp, so I have the option of either battery, or mains power
2nd one, is to make a +/- 25 or 30V railed 800W PSU for a small HT system.
For high powers, how powerful do the input transistors have to be? I'm thinking of using 2-3 pairs of 400V transistors, and make similar to computer PSU..
I've been wanting to make a few different 120V SMPS designs.
1st one, I'm wanting to make 18V output or so to charge batteries in a 14.4V portable amp, so I have the option of either battery, or mains power
2nd one, is to make a +/- 25 or 30V railed 800W PSU for a small HT system.
For high powers, how powerful do the input transistors have to be? I'm thinking of using 2-3 pairs of 400V transistors, and make similar to computer PSU..
One can also achieve near unity PFC directly with Phase shifted full bridge also by alternate ways.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.