LM2575 regulator shutting down

[Andrew Cowan]

Apologies for the non-PICAXE question.

I've got a LM2575-12 switching regulator stepping down from 24v to 12v.

I've followed the example circuit exactly, and it worked fine with a previous 12v camera. However, I've upgraded the camera, and come across a problem.

Background:
The camera runs of 12v, and draws 200mA when idle, and up to 500mA when zooming and focussing. The regulator is good for up to 1A continuous.

Situation:
When connected to the camera, the regulator enters current overload mode, where it stays until the load is removed (as per the datasheet).

Looking with my multimeter, the maximum current is 40mA. However, when I use Peak Mode (cutting the detection time to 250uS), inrush current of about 6A is seen (which is causing the regulator to shut off).

So, how do I stop the camera's high inrush current at power-up? I would imagine it has lots of capacitors in it, which charge up when powered. I am doing this for a deadline, so can only use components I have or Maplin sell.

Options:
- Thermistor. However, I doubt the 200mA current it usually draws is enough to heat it significantly to get the resistance down.
- Resistor. I don't mind a small voltage drop - maybe a one or two ohm resistor would help?

Any other ideas? I somehow need to stop this inrush current spike.

Thanks for any ideas!

Andrew

 

[Andrew Cowan]

Hmm - I found I had two high power thermistors.

One didn't solve the problem at all (too low resistance).
One did solve the startup problem, but took over 1 minute before the resistance was low enough for the camera to work.

Due to my time constraints, I've just replaced the switcher with a bog standard 7812.

Andrew

 

[westaust55]

@Andrew,

I presume that camera might have some internal motors for zoom, focus, etc ?
These would draw a higher current on energisation of the motors.

Or as you say maybe just some large caps are the casue of the problem.

Some sort of primative "soft starter" with resistance in the line to camera and cutting the resistance out in steps could be done relatively quickly.
Then of course you need a PICAXE (08M) to control the start-up.

 

[marks]

Hi Andrew,
sounds like your inductor size is to small probaly 330 uh
if you reduce your inputvoltage to less than 20v it will proberly work.

i think you need to try a higher value inductor value ie 470 etc depending on your load,
which should reduce switch current.

ripple will increase with a higher value inductor a higher value output cap may be needed

good luck lol

 

[Andrew Cowan]

I think you might be right - looking at the datasheet a 330uh inductor is really only good for up to about 20v.

I'll try reducing the input and se if that has any effect.

Thanks!

 

[Dippy]

Phew, thats quite a big startup current. Well out of spec. for that switcher.

After you've tried all the other things, I'd be tempted to look at:-
A) Whether your HV side can provide enough oomph.
(I didn't see whether you had actually checked that the LM2575 was really going into shutdown mode OR whether the initial stratup current prevented it from switching properly in the first place).
B) A delayed startup.
C) A switch on the Reg output so that the regulator can get up and running before that big output oomph. Hopefully a low Esr fat cap can handle that.

Sometimes with some stuff, if you limit the startup oomph (e.g. a res on the camera power input) it won't start properly.
Hopefully, this isn't the case and a simple solution will sort it.

 

[BeanieBots]

MUCH larger caps on both input side and output side.
Have look at the National website where they have SMPS design calculators.
Put your new design parameters in and then check the bode plots to make sure you haven't designed an oscillator

 

[Andrew Cowan]

But wouldn't larger caps on the output side increase the startup current? Unless the caps were slowly filled through a resistor...

A

 

[Dippy]

Andrew, instead of copying circuits verbatim it's sometimes a good idea to read and calculate and use any on-line design calculator available.
I wouldn't go too mad on Cin value.

I haven't look at the Nat Semi's WebBench for your requirments - I think you'll struggle getting something representative with the startup transient you have indicated. Anyway, that's something which is useful to have a go with even if not entirely productive.


Check the Data Sheet for Application equations.
Cout:
"The above formula yields capacitor values between 10 μF
and 2000 μF that will satisfy the loop requirements for stable
operation. But to achieve an acceptable output ripple voltage,
(approximately 1% of the output voltage) and transient response,
the output capacitor may need to be several times
larger than the above formula yields. B. The capacitor’s voltage
rating should be at last 1.5 times greater than the output
voltage."
Your crucial parts are also inductor value and TYPE and Cout Value and TYPE.

Personally, before getting carried away, I would.
1. Put a minimal load on the Buck PSU output but leave the camera disconnected via a little toggle.
Power-up the LM2575 circuit and check voltage/noise.
Then flick the toggle.
If the camera powers up OK then this indicates that it's the startup thats the problem.
You can then include a simple R/C based delayed switch into your PSU output.
If it neeeaaarly starts then a larger cap (low Esr) may cure it. (Or 2 of what you have paralleled).
This assumes your other components are sutiable and your construction is good.
Physical aspects can be crucial with S/mode designs.
(I've had instances where 2mm extra track in the loop can make a huge difference).

 

[BeanieBots]

Possibly, which is why I suggested you put your revised design through their FREE simulation software.
Most of the "simple switcher" series devices include soft startup, so for the two minutes it would take to try it would my first test. It has pulse by pulse current limit so I'd be hopeful it can cope with large output caps. The only likely issue is instability but the bode plot will tell you if you are getting close.

 

[Dippy]

But will it bode well for Andrew?

 

[moxhamj]

Seems like a great job for a picaxe. First step - LM2575 5V reg driving a picaxe. Sounds a bit complex, but I've gone for 100% simple switchers over the last few years, and what the inductor and reg costs is usually saved on the heatsink on a standard reg.

1 amp max current. Put the output into whatever the standard circuit says. I think it is 330uF. Run that via a 5 ohm resistor into a much larger capacitor - try tens of thousands of microfarads to start, then work down. At zero charge on that big capacitor, the max current is V=IR, I=V/R 5 ohms, 5V, 1 amp = max for the switcher.

You can work out how long it takes to charge =1A 1F = 1V/S so it depends on the big capacitor.

The picaxe's job is to delay turning on the flash till that big capacitor is charged. Either measure time or voltage on the cap.

Then turn on the camera with a relay or a NPN then PNP transistor 'high side' switch.

 

[Dippy]

Thats roughly what I suggested on the previous page Drac.
Except you don't need anyhting complicated to form a delayed switch.