LM2596 help

[flagZ]

Hi all,

I am a newbie in electronics and this year I took some time to focus on it and learn the basics. I am trying to build possibly the most common raspberry pi bot… a battery, a couple of wheels + motors, a buck converter, a motor controller. I want to use a 9V battery with a buck converter (LM2596) and bring it down to 5V (to use it with the raspberry). I bought a few but turning the potentiometers do not seem to vary the output voltage. It is stuck at 8.3/8.5.
Can you give me any pointers about what could i try to make it work? Any help appreciated :-)

Thanks

[Ian.M]

If you are using a typical cheap 'LM2596' module, the potentiometer for output voltage is part of a potential divider that sets the 'target' output voltage in the range 3V to 40V.  However its only a buck converter not a buck-boost so if you set a voltage greater than the input, it cant actually provide it so with only 9V in, more than 3/4 of the potentiometer range will have no effect on the output voltage, which will be stuck at the input voltage.   Its also likely to be a multi-turn preset, so there's lots of turns with no effect.  Keep turning it one way till either the voltage drops or it starts clicking gently as you  turn it.   If it starts clicking, you've reached one end so reverse the direction and keep turning.  It should start controlling the voltage *before* it reaches the other end and starts clicking again.  If it doesn't either you are turning the wrong preset or the module is faulty.

N.B. I wouldn't trust a cheap adjustable 'LM2596' module to power a Pi in a robot.  It only needs enough of an impact or vibration for the potentiometer wiper to loose contact with the track and the output voltage will shoot up, destroying the Pi.   A crowbar circuit set to 5.6V to protect it would be advisable if you are using anything more expensive than a Pi Zero.  Also, although a genuine LM2596, properly heatsinked, with the right caps and an inductor with a high enough saturation current (>6A)  can actually deliver 3A.  the current ratings of 'chinesium' LM2596 modules are purely imaginary.  A 3A module is unlikely to be good for more than 1.5A and for long term use, is unlikely to be reliable over 1A, so don't try to power any model of PI B with a full compliment of USB peripherals from one!   I've put quotes round the LM2596 above because the cheap modules out of china often have fake chips.  One easy way to tell, is to check its operating frequency.  If its significantly lower than the 150KHz specified in the LM2596 datasheet, you've definitely got a fake.  Unfortunately if it is 150KHz, that doesn't guarantee its genuine, as it may be a better grade of fake. 

[jmelson]

N.B. I wouldn't trust a cheap adjustable 'LM2596' module to power a Pi in a robot.  It only needs enough of an impart or vibration for the potentiometer wiper to loose contact with the track and the output voltage will shoot up, destroying the Pi.

Right, if there's no need for adjustment, the pot can just cause problems.  You could use an LM2575-5.0 which is a fixed 5 V switching regulator.  (LM2595-5.0 is a newer version of the same basic function.)

Also, these need a little bit of a load to regulate, so if you are testing it with just a DVM for a load, it may well behave as you see.  Put a 100 Ohm resistor on as a load and see if it goes down to the right voltage.

Jon

[thinkfat]

Also, these need a little bit of a load to regulate, so if you are testing it with just a DVM for a load, it may well behave as you see.  Put a 100 Ohm resistor on as a load and see if it goes down to the right voltage.

Jon

Usually the feedback voltage divider provides enough load so that the regulation will work with no other load connected.

Gesendet von meinem Nokia 6.1 mit Tapatalk

[Ian.M]

@Thinkfat,
Thanks for the comment about the feedback divider providing sufficient minimum load, but *PLEASE* fix the mis-attributed quote.  Jmelson said what you quoted, not me, and is is generally preferable to put your reply after the closing /quote tag! Edit: Done

[thinkfat]

@Thinkfat,
Thanks for the comment about the feedback divider providing sufficient minimum load, but *PLEASE* fix the mis-attributed quote.  Jmelson said what you quoted, not me, and is is generally preferable to put your reply after the closing /quote tag!

Corrected. Sorry, tapatalk might be easier to browse a forum, but replying is a pain in the neck.

Gesendet von meinem Nokia 6.1 mit Tapatalk

[kolbep]

Those cheap ALI LM2596 modules have a Multiturn Pot.
You need to turn it several times in one direction to see a change in output voltage.

[radiolistener]

These Chinese LM2596 modules are made with fake chips. Actually this is not LM2596.

This is remarked XL2579 which works at 52 kHz 

[flagZ]

Thanks all for the responses... wow lots of things i have to learn :-)

It turned out it was a multi turn... 15/20 turns and i had the voltage conversion...

I did not know about crowbar circuits. I am using a Raspberry PI 2 B that i had in the house, but I rather not fry it because the vibration changes the resistance of the potentiometer... can i buy a crowbar circuit? or build it easily?

[thinkfat]

These Chinese LM2596 modules are made with fake chips. Actually this is not LM2596.

This is remarked XL2579 which works at 52 kHz 

Ah, interesting. Yet the dimensioning of the capacitors and the inductor is usually taken from the TI design documents, meaning that the inductor will be way off the mark for 52kHz. In fact, 5 to 10 times off. That definitely explains a lot of the problems people are seeing with these modules!

[malagas_on_fire]

Another aspect if looking also in the onsemi datasheet are the capacitors and its recomended ratings... Higher esr capacitors for cheapness could also  cause troubles

[sleemanj]

These Chinese LM2596 modules are made with fake chips. Actually this is not LM2596.

This is remarked XL2579 which works at 52 kHz 

Some are, not all.

[radiolistener]

Some are, not all.

I didn't seen these with real LM2596. Which one you're talking about? Could you give me the link on product page?

[sleemanj]

I didn't seen these with real LM2596. Which one you're talking about? Could you give me the link on product page?

Well, I'm sure I have checked and had real ones in the past and posted a photo of a scope trace here once some years ago but I can't find that post maybe I'm imagining it.

I just checked a random sample from my current lot and indeed ~50kHz.  I can't say I make a habit of checking these, so, who knows how long I've been getting 50kHz ones.

Trouble of course with any Chinese supply, you never know if you're always going to get the same as you got before.