Hi.
Can you please post a copy of your circuit, in CAD or a picture of a hand drawn circuit in jpg, png?
Can you post a picture of your project so we can see your component layout?
What are you using as your 24Vdc supply, is it filtered properly?
Have you fitted a fuse in the circuit?
Have you got the 100uF Cap as close to the input pin as possible, and are you using a good wide gnd for all the components that use the gnd connection.
Have you built a custom circuit board? Have you followed the layout guidelines from the manufacturer? Switching DC-DC converters are highly sensitive to layout (due to the relatively high current being switched at high speeds). Breadboards are kryptonite for switching converters, for example.
DrAzzy:
Have you built a custom circuit board? Have you followed the layout guidelines from the manufacturer? Switching DC-DC converters are highly sensitive to layout (due to the relatively high current being switched at high speeds). Breadboards are kryptonite for switching converters, for example.
I am using a matrix board. Are they really that fickle??? I.E. They fail completely as opposed to just provide sub-standard regulation?
Maybe DIY with these devices just aint worth the hassle and I should instead just purchase the pre-soldered LM2576 based power boards available on ebay.
Post a picture of your circuit as you have wired up. Include the load you are trying to power, as well as the input power (Volts and amps). Perhaps it's just a simple mistake you keep overlooking.
Also, are you sure your inductor is 100uH, and the voltage divider is correct? On the input side, what capacitors are you using to smooth the voltage? Diode?
boylesg:
I am using a matrix board. Are they really that fickle??? I.E. They fail completely as opposed to just provide sub-standard regulation?
Maybe DIY with these devices just aint worth the hassle and I should instead just purchase the pre-soldered LM2576 based power boards available on ebay.
How long are the high current paths? They need to be super-short, and that's the problem, usually on generic protoboard, they end up not being so. I've made a switcher work on one of these: (they come from my store), but I'm sure it didn't perform as well as it should have even on that.
I have used LM2576T for a 6 Volt 3 Amp. I etched an on purpose printed circuit board. As stated above switched must have a very short ground paths. My choose of the inductor was based on 10 volt input. You would need something like i50 microhenrys. See the attachment
Matrix board is NOT suitable for swmode psu chips. Use a pcb and follow the manufacturer's recommended layout and components.
The switches in these devices are extremely fast (a few nS) , and inductances of even a nH or two in the wrong place will wreak havoc.
Suppose it switches 1A in 10nS => that's 100A/uS. Perfectly feasible
Such circuits can radiate into the 100's of MHz which is why RF designers hate them.
And given V= L dI/dT even 1nH will give 100mV error - quite enough to upset a control loop. A typical via is 0.2nH. and track can be >0.5nH/mm. An inch of wire will give > 1volt!
be careful! You have to treat these ic's as if they were high power rf devices.
allanhurst:
Matrix board is NOT suitable for swmode psu chips. Use a pcb and follow the manufacturer's recommended layout and components.
The switches in these devices are extremely fast (a few nS) , and inductances of even a nH or two in the wrong place will wreak havoc.
Suppose it switches 1A in 10nS => that's 100A/uS. Perfectly feasible
Such circuits can radiate into the 100's of MHz which is why RF designers hate them.
And given V= L dI/dT even 1nH will give 100mV error - quite enough to upset a control loop. A typical via is 0.2nH. and track can be >0.5nH/mm. An inch of wire will give > 1volt!
be careful! You have to treat these ic's as if they were high power rf devices.
Allan
I think this confirms I should bother trying to solder these up myself, just purchase the already soldered modules which are only a couple of dollars each on ebay anyway.
By the time I purchase the low esr caps, etched board and inductors (in order to solder them myself) I would probably paying more than the cost of the pre-soldered modules anyway.
boylesg:
I have tried many times to get one of these pricks of things to properly reduce rectified 24VAC to 5VDC and every time I have failed.
I simply don't know what I am doing wrong. The circuit diagram in the datasheet is fairly simple and I am certain I have followed it correctly.
With may latest attempt the LM2576 just heated up severely but failed to regulate the voltage.
Question: do all the manufactures have the same pinout with their TO-220 devices or can the pinout vary?
That is the only thing that I can think of that I am doing wrong consistently.
They can't be this difficult to use surely????
Hi there,
They dont call it a "Simple Switcher" for nothing, it's one of the easiest chips on earth to use if you need a switcher and not a linear.
You do have to use the right parts however besides the chip itself.
#1: The inductor.
The inductor must be able to stand the peak current without going into saturation. #2: The diode:
The diode must be a high speed type like a Schottky. If you try to use a regular rectifier diode it will fail every time as something will heat up badly.
Those are the main two rules for component selection.
There is also a chance you got fake chips. These chips are so easy to use i would have to think something is really wrong if it doesnt work. I've used them for several things including battery chargers and just plain power supplies for things like a flat bed scanner for example. They work up to 3 amps (with proper inductor of course) which is pretty good for a lot of things. I use one up to 1 amp only but it does not even need a heatsink with that lower current.
If you post a list of parts we can go over it and see if anything isnt right.
If the inductor saturates even for a short time it could cause heating.
If the diode is not the right type it can cause over heating.
The caps should be somewhat quality caps.
The lower feedback resistor is limited to a max of something like 5k (check data sheet here though for verification).
(they come from