Does anyone have a source for the inductors needed to build a simple switcher power supply for the LM2576 simple switcher IC. The datasheet indicates the inductors should be an RL-2445 from Renco or PE92108 from Pulse Engineering. These inductors seem to be obsolete products, so maybe there is a better switcher IC someone can recommend to get 5 volts @ 3 amps.
Simple switcher inductors
- Thread starter gbrusseau
- Start date
westaust55
Moderator
inductors
You do not give people much to work with in terms of physical arranagement and value (in mH or other) required.
Have you looked at this thread:
http://www.picaxeforum.co.uk/showthread.php?t=10752&highlight=inductor
You do not give people much to work with in terms of physical arranagement and value (in mH or other) required.
Have you looked at this thread:
http://www.picaxeforum.co.uk/showthread.php?t=10752&highlight=inductor
BeanieBots
Moderator
Have a look for the very popular RM series of cores.
RM6, RM8 & RM10 should all be good for what you want.
The datasheets will tell you how many turns for how many uH.
I've just done a flyback transformer using RM8 + 27 turns = 180uH.
Winding your own is often better. Especially if you use Litz wire.
RM6, RM8 & RM10 should all be good for what you want.
The datasheets will tell you how many turns for how many uH.
I've just done a flyback transformer using RM8 + 27 turns = 180uH.
Winding your own is often better. Especially if you use Litz wire.
premelec
Senior Member
Mouser and Digikey have suitable inductors - keep your eye on the maximum current rating and DC resistance - their listings are extensive so you have to plow through quite a lot of data... You can, of course try making your own and if you can deal with larger physical size it's easier - to get a good high current inductor with small size is trickier.
Thanks for the great replies. So I can build my own and experiment with different wire types and windings to get the best fit. I like that idea.
BeanieBots, that litz wire. Is that the enamel coated wire often used in transformers. I think Radio Shack sells that stuff.
Now I'll have to figure out the best way to monitor coil saturation and thats going to be fun. By fun, I mean I enjoy doing that kind of thing. I would rather build and learn then buy ready made. I'm assuming that breadboarding this will be a little tricky. I'm going to need to keep the jumper wires as short as possible and maybe an RC filter on the output and input.
BeanieBots, that litz wire. Is that the enamel coated wire often used in transformers. I think Radio Shack sells that stuff.
Now I'll have to figure out the best way to monitor coil saturation and thats going to be fun. By fun, I mean I enjoy doing that kind of thing. I would rather build and learn then buy ready made. I'm assuming that breadboarding this will be a little tricky. I'm going to need to keep the jumper wires as short as possible and maybe an RC filter on the output and input.
Maybe not saturate in the way ferrite saturates. But the current can't keep rising indefinitely - eventually you start getting resistive losses and the inductor turns into a (low value) resistor. Then it gets hot, catches fire, and turns into a very high value resistor...
For a given number of turns, an air core will have a lot lower inductance than a ferrite core. So if you want it small, it will have to run at a higher frequency. Or you could make it big with thick wires, like the air core inductors used for speaker crossovers.
The exact brand of inductor would not be critical. It just needs the right value in micro henries, and thick enough wire that it can handle the amps. The DC resistance should be as low as possible. Searching through parts catalogues, they usually list the max current. 3 amps is going to be a chunky sort of inductor maybe an inch cubed with thicker than 1mm wire.
For a given number of turns, an air core will have a lot lower inductance than a ferrite core. So if you want it small, it will have to run at a higher frequency. Or you could make it big with thick wires, like the air core inductors used for speaker crossovers.
The exact brand of inductor would not be critical. It just needs the right value in micro henries, and thick enough wire that it can handle the amps. The DC resistance should be as low as possible. Searching through parts catalogues, they usually list the max current. 3 amps is going to be a chunky sort of inductor maybe an inch cubed with thicker than 1mm wire.
Last edited:
Litz wire
From Wikipedia, the free encyclopedia
Jump to: navigation, search
Litz wire is a special type of wire used in electronics. The wire is designed to reduce the skin effect and proximity effect losses in conductors. It consists of many thin wires, individually coated with an insulating film and twisted or woven together, following a carefully prescribed pattern often involving several levels (groups of twisted wires are twisted together, etc.).
The term litz wire originates from Litzendraht, German for bundled wire.
Today the only place you're likely to encounter real litz wire is for RF transformers and "loopsticks" (AM ferrite bar antennas)
The usual choice is the radio shack stuff - ordinary enameled magnet wire. Some people try to combine large numbers of very fine gauge magnet wire to form litz wire. In the old days it was "shaved" from bar or wire stock - lots of very fine rectangular wires with a lot of surface area, and small cross sectional area, woven with cotton threads, for strength and insulation.
Fascinating Flooby. Thanks for the heads up on litz wire BeanieBots. I just may try some of those twist patterns. I'm going to need to be heavily medicated before I start twisting though. First I need to find a good way of monitoring the efficiency of the coil. Maybe a parasite coil placed near the inductor with maybe a 1 meg load. What do you think?
OK, if I get on the output pin of the SS chip, before the inductor, I should see the current waveform. I should also be able to see the duty cycle of the PWM. I'm thinking by switching out each of the custom built inductors I make, I could weed out the ones with the highest duty cycle. Am I right? Its entirely possible that the LM2576 switcher IC is very forgiving, and the overall efficiency doesn't change much as long as the inductor is around 100 to 150 micro-henries.
BeanieBots
Moderator
The LM2576 only switches at 52kHz. So your inductor will need to be BIG physically. Even with an air gap it will still need to be about 1" cube. It can still saturate with an air gap.
Litz wire only only starts to have an effect above about 20kHz. It will help at 50kHz but the effect will be small compared to other losses.
The actual inductance value will have no effect on efficiency itself but it changes the operating point. This will effect the efficiency of other components. Higher inductance will require more turns and hence higher I^2.R losses. Lower inductance will require higher FET current and hence higher I^2.Ron losses. It's a balancing act.
By far the biggest loss will be any rectification diode volt drop. This can be massive with low voltage outputs. The lower the reverse voltage rating of the diode the better the forward volt loss at any given current. What you can get away with depends on your topology.
Then there's the snubber. A 'stronger' snubber will let you use a lower voltage diode but a strong snubber causes snubber losses. Another balancing act.
Then there's layout. Poor layout requires better snubbing and adds track I^2.R losses and radiated losses.
If you want to experiment with different inductors, then design your board around easy replacement or your experiments won't be accurate. I always fit pins where the inductor goes so that nothing else gets disturbed. Always keep the inductor lead lengths the same. These add to the "leakage" inductance and can be very significant below 1mH.
May as well mention core material as well while we're at it. Forget iron at 50kHz. Just about all ferrite will be fine.
If you want to experiment with air gaps as well, use paper. I often use scotch tape or similar for quick and easy tests but it degrades with time and your inductance could easily double over time as the glue spreads and the plastic thins. Mylar tape works quite well.
Good luck, it's an art rather than a science.
Litz wire only only starts to have an effect above about 20kHz. It will help at 50kHz but the effect will be small compared to other losses.
The actual inductance value will have no effect on efficiency itself but it changes the operating point. This will effect the efficiency of other components. Higher inductance will require more turns and hence higher I^2.R losses. Lower inductance will require higher FET current and hence higher I^2.Ron losses. It's a balancing act.
By far the biggest loss will be any rectification diode volt drop. This can be massive with low voltage outputs. The lower the reverse voltage rating of the diode the better the forward volt loss at any given current. What you can get away with depends on your topology.
Then there's the snubber. A 'stronger' snubber will let you use a lower voltage diode but a strong snubber causes snubber losses. Another balancing act.
Then there's layout. Poor layout requires better snubbing and adds track I^2.R losses and radiated losses.
If you want to experiment with different inductors, then design your board around easy replacement or your experiments won't be accurate. I always fit pins where the inductor goes so that nothing else gets disturbed. Always keep the inductor lead lengths the same. These add to the "leakage" inductance and can be very significant below 1mH.
May as well mention core material as well while we're at it. Forget iron at 50kHz. Just about all ferrite will be fine.
If you want to experiment with air gaps as well, use paper. I often use scotch tape or similar for quick and easy tests but it degrades with time and your inductance could easily double over time as the glue spreads and the plastic thins. Mylar tape works quite well.
Good luck, it's an art rather than a science.
Thanks BeanieBots, Your input is very helpful. I think I'm going to forget about the litz wire. If I understand the term snubber correctly, the SS IC manufacturer recommends a 1N5822 Schottky snubber diode in the circuit example. I'm ordering some of those M8 ferrit cores and some 18 gauge and 20 gauge enameled wire. One more question though. I have some 1" diameter toloroid cores I want to try out. If I wind a second layer, do I continue the second winding in the same direction as the first winding or should the second winding be wrapped in the reverse direction. Reverse direction sounds all wrong, but I thought I would ask.
BeanieBots
Moderator
What matters is the start and end of each winding. That is, the direction of wire turn rather than which way it goes round the circumference of the toroid. If you want to add more inductance then the start of one winding should be connected to the end of the first winding. If you want to make a transformer, you should still make a note of which lead is the start. I always put a bit of coloured insulation over the enamelled wire to denote which is the start.
Be a little careful of toroids in high current designs. There is no air gap (unless you have 'special' ferrite) and they can saturate quite easily with only a few turns and a few amps. They do however offer many uH for very few turns.
Be a little careful of toroids in high current designs. There is no air gap (unless you have 'special' ferrite) and they can saturate quite easily with only a few turns and a few amps. They do however offer many uH for very few turns.