1. Home
  2. Electronics Design
  3. Todays's tronix project, LM2596 switcher as a current source

Todays's tronix project, LM2596 switcher as a current source

Todays's tronix project, LM2596 switcher as a current source.

For the big battery loader would it not be nice to have a current source that was simple, and made from stuff I already have in the box.

The big ebay modules are cheap, but a bit of overkill.

Had once (upon a time) made a little test board for the LM2596-3.3V version, that I actually use for any other volts >= 3,3 with the aid of an extra resistor, so, while plopping down on paper some circuits for the battery whatsit balancing act? anyways, why not muddify that LM standard circuit a bit.

This is what such a circuit looks like... 3.3V (or more) and a few amps, no heat.

formatting link

Muddified it looks like this:

formatting link
it will be oblivious, ehhm obvious to the attentive lurker that the 2 resisters in series can be replaced by one resister. And now you may wonder how this wurks, well here it is then, the circuit diagram:
formatting link

That is an extra PNP and an extra NPN, normally (say power on) pin 4 is low, (the feedback pin), and the LM starts outputting all it can. In the output shunt Rs , at least when some load is present, will drop some voltage. Say Rs is 1 Ohm, then the NPN will start conducting at about .7 V or 700 mA. This will cause the PNP to conduct too, and the LM adjust pin is pulled high, and it relaxes and goes watch TeeFee while the current remains at say that 1 A.

None of the resisters are critical [1], the base resistor in the NPN is just to reduce possible peak currents (will those ever happen), the 47 k is probably not needed either in normal temperatures where human life forms dwell, and the other sister is (gets out of chair and measures it) 1.6 k (1600 uncles). that is the 2 in series, so total, OK. Hope this helps, else use SOS.

[1] Just grabbed in the used parts box and use whatever came up, give or take a few grabs This is to support my statement that a good designer makes circuit that also work with 20% components.

So the circuit if guaranteed free from opti-mice.

PS not sure the circuit diagram is right, I soldered the circuit first, the pictures rule.

Oh well.

Reply to
Jan Panteltje
Loading thread data ...

Hey, Jan, if you use the ADJ type of LM2596, you can simply feed its FB pin with your ground sense resistor, and you've got a much more simple current-regulated rather than voltage regulated supply. I mean, since you were willing to lift the load off ground anyway. BTW, if you want the current-sense resistor on the high input side of your load, add an LTC6101, very handy.

--
 Thanks, 
    - Win
Reply to
Winfield Hill

On a sunny day (22 Dec 2015 06:08:56 -0800) it happened Winfield Hill wrote in :

Right, the case was I once needed 1 LM2596 3.3 V, ebay came as 10 for a very low price. so since then I am using these.

Sure enough now you mention that next time I will use the adjustable version.

I also want to mention that if you need to limit voltage as well, (as in the battery case), one of those TL431 from the + output to the base of the NPN should work.

Never ever used TL431, but now 50 are on their way for 1$83 IIRC it was, so _have_ to use those.

Since I actually need to boost too (from 18 V solar) maybe I will fiddle around a bit more. This was actually just a test of part of a bigger thing I painted together today;

formatting link

Anything can still change at this time, some resistors still need to be calculated, not until those TL431 arrive.. And I tested those.

The 'control module' will have RS232 NMEA, and Ethernet using this ebay module:

formatting link
and I _just_ got enough analog inputs on the PIC 18F14K22 to monitor all cells. AN3 according to the errata datasheet has a huge error, but can be used to monitor the shunt voltage, in that case that error is only milli volts, and the lower impedance helps. There are some notes ehh, well to remind myself that the pin-assignment for the ENC28J60 drive signals on the PIC have changed (I have used that chip before and wrote a UDP stack for it:
formatting link
and that works 100 % reliable now for more than a year. So I expect no problems there. I use software SPI, as Microchip errata also mentions broken hardware SPI on that PIC.

Anyways, the original 2N2955 current source in that drawing to be replaced by the thing in the middle. Cell voltages is just subtraction, final calibration 2 bytes per cell in EEPROM, set via the RS232 only once, PIC 1024 mV (or there about) reference.

Thanks for the suggestion, had not thought about that.

Reply to
Jan Panteltje

Cute, but I don't completely understand the biasing of the feedback pin... You;re giving that pin a current, but isn't it voltage-sensing?

There are switchmode circuits intended for current-limit feedback, like L294 (and others, I presume). The wiring is simpler, at least. The intended use, "solenoid driver" doesn't really describe current regulation very well, but that IS its function.

Powering LED lamps ought to be a major application for current regulation.

Reply to
whit3rd

On a sunny day (Tue, 22 Dec 2015 13:04:45 -0800 (PST)) it happened whit3rd wrote in :

There is a resistor to ground _in_ the chip, plus a resistor to the feedback pin, so in the 3.3 V version a voltage divider. In the 'adjustable' version, as Winfield pointed out, you can directly access that comparator input.

There are < 4 $ free shipping step up modules on ebay, I have some:

formatting link
but those have no current limit.

I will have a closer look tomorrow at what chip is used in those, and if I can muddify one for current limit. That would both solve the step up from solar ( Powering LED lamps ought to be a major application for current regulation.

Yep

Reply to
Jan Panteltje

formatting link

Boost (well, probably SEPIC) with a current limit,

Reply to
Jasen Betts

On a sunny day (23 Dec 2015 01:07:57 GMT) it happened Jasen Betts wrote in :

Reply to
Jan Panteltje

On a sunny day (23 Dec 2015 01:07:57 GMT) it happened Jasen Betts wrote in :

The ebay one linked above works, tested it.

formatting link
Now that I am looking at it, the chip is a UC3843A, runs at 500 kHz (explains why so few turns on the toroid) it drives a P75NF758 power MOSFET (75 V 80 A) (on heatsink, I see a mOhm shunt too) and a STP2045CT dual Scottky power diode (45 V 30 A). also on a second heatsink.

PS, of course somebody already dissected it:

formatting link
formatting link
Circuit diagram seems correct. Easy to do the same trick as I did with the LM I'd think.

Reply to
Jan Panteltje

On a sunny day (22 Dec 2015 06:08:56 -0800) it happened Winfield Hill wrote in :

PS, I had a go at that ebay boost converter.

formatting link
Circuit diagram:
formatting link
formatting link

and used exactly the same circuit (and same components :-) as in the LM2596 setup:

formatting link
and it nicely stabilizes the current, test with 11 V in from solar panel, and for test 4 cells at about 14 V.

Of course as it has this diode from input to output, so battery (to be charged) voltage needs to be higher than the input voltage, else huge current flows, but still limited by the solar panel of course.

I pulled the comparator input of the UC3843A towards the 9 volt from the 7809 stabilizer.

No problem.

Simplicity increases.

Reply to
Jan Panteltje

"Muddified", ie: grey on grey as usual. Take a hint from the quality we see from JL.

Reply to
Robert Baer

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.