LM317 as switcher

I needed 12V to 3.3 V, and lower power losses than the LM317 analog I am using now. Various small switcher configurations I tried were not that good. I was specifically afraid of a switch transistor failing, and killing a wole lot of chips at once by putting a high power 12V on those, or just frying those at low load with an unexpected voltage rise.

Then I remembered the LM317 datasheet had a switcher example in it. So I build a small test circuit, and that was only so-so, but it worked. Not a very good waveform, and erratic start-up. Also it needed a BIG inductor (600 uH). Are all LM317 the same????? Did some tinkering with that circuit, to see what effected what, nice change from spice to use a soldering iron and a scope. Was actually looking to make it slow-start, to fix the startup (rush in current) by adding caps, but that accidently significantly improved the whole thing! Here is the diagram (do not look if you have low contrast eyes, or are allergic to pencil and paper): ftp://panteltje.com/pub/LM317_switcher/LM317_switcher_diagram_img_2316.jpg The original values from the datasheet are between brackets. The 100 nF cap was added by me. This is the waveform without that cap, with 230 uH (versus theirs 600 uH): ftp://panteltje.com/pub/LM317_switcher/LM317_switcher_no_100nF_img_2320.jpg This is the waveform from my circuit, nice: ftp://panteltje.com/pub/LM317_switcher/LM317_switcher_final_img_2312.jpg This with the .25 Ohm resistor shorted, you really need it: ftp://panteltje.com/pub/LM317_switcher/LM317_switcher_no_.25_Ohm_img_2319.jpg The setup looks like this: ftp://panteltje.com/pub/LM317_switcher/LM317_switcher_test_setup_img_2318.jpg

The power PNP is a BD??? found it somewhere in the junk box, no idea what exactly it is. The nice thing about this circuit is that at zero, or very low load, it stops switching, and the LM317 just works as a linear regulator. As soon as you draw power it starts switching, nothing gets hot, I measured 63 % efficiency at 160 mA load, what I will use it for. Efficiency should go way up at higher loads, the data sheet mentions 3 A.

Reply to
Jan Panteltje
Loading thread data ...

On a sunny day (Wed, 18 Aug 2010 18:50:18 GMT) it happened Jan Panteltje wrote in :

PS period time is 12uS, say 80 kHz.

Reply to
Jan Panteltje

Would one of those other designs be out of the question if you threw a crowbar circuit on the output with a pico fuse as the fail safe?

Reply to
WangoTango

On a sunny day (Wed, 18 Aug 2010 15:41:02 -0400) it happened WangoTango wrote in :

That would be possible. But this was all in the junk box :-)

Reply to
Jan Panteltje

Ah, yep, I somehow end up with a lot of LM2596s stuck on things because I ended up with a bunch of 5V, 12V and adj versions and a reel of 68uH shielded inductors. Funny how it works out that way.

Reply to
WangoTango

ISTR a version of the 723 datasheet that gives a switcher among the example applications:

formatting link

The circuit inside this chip includes an error op-amp and you can tailor the design to iron out the bugs you describe.

It also readily accomodates an external power transistor that you can make as big or small as you need - or just use the one in the chip.

Reply to
ian field

Did you change ftp server config? Hung on Google Chrome, so I tried wget:

grant@deltree:/home/common/zzz$ wget ftp://panteltje.com/pub/LM317_switcher/LM317_switcher_final_img_2312.jpg

--07:41:41-- ftp://panteltje.com/pub/LM317_switcher/LM317_switcher_final_img_2312.jpg => `LM317_switcher_final_img_2312.jpg' Resolving panteltje.com... 81.207.135.196 Connecting to panteltje.com|81.207.135.196|:21... connected. Logging in as anonymous ... ^C grant@deltree:/home/common/zzz$ wget -p ftp://panteltje.com/pub/LM317_switcher/LM317_switcher_final_img_2312.jpg

--07:42:57-- ftp://panteltje.com/pub/LM317_switcher/LM317_switcher_final_img_2312.jpg =>

`panteltje.com/pub/LM317_switcher/LM317_switcher_final_img_2312.jpg' Resolving panteltje.com... 81.207.135.196 Connecting to panteltje.com|81.207.135.196|:21... connected. Logging in as anonymous ... Logged in! ==> SYST ... done. ==> PWD ... done. ==> TYPE I ... done. ==> CWD /pub/LM317_switcher ... done. ==> PASV ... done. ==> RETR LM317_switcher_final_img_2312.jpg ... done. Length: 11,070 (11K) (unauthoritative)

100%[====================================================>] 11,070 12.91K/s

07:43:05 (12.90 KB/s) - `panteltje.com/pub/LM317_switcher/LM317_switcher_final_img_2312.jpg' saved [11070]

FINISHED --07:43:05-- Downloaded: 11,070 bytes in 1 files

Cheers, Grant.

Reply to
Grant

723 is quite ancient; I've never been tempted to buy one. That said, if you're going to buy one, you might as well make it an MC34063 or something like that. Just as poor performance (BJT output), but at least it's made for switching.

Tim

--
Deep Friar: a very philosophical monk.
Website: http://webpages.charter.net/dawill/tmoranwms
Reply to
Tim Williams

now.

low

current) by adding caps,

to pencil and paper):

Then your got really upset: grant@deltree:/home/common/zzz$ wget -p ftp://panteltje.com/pub/LM317_switcher/LM317_switcher_no_100nF_img_2320.jpg

--07:46:13-- ftp://panteltje.com/pub/LM317_switcher/LM317_switcher_no_100nF_img_2320.jpg =>

`panteltje.com/pub/LM317_switcher/LM317_switcher_no_100nF_img_2320.jpg' Resolving panteltje.com... 81.207.135.196 Connecting to panteltje.com|81.207.135.196|:21... connected. Logging in as anonymous ... Error in server response, closing control connection. Retrying.

--07:46:15-- ftp://panteltje.com/pub/LM317_switcher/LM317_switcher_no_100nF_img_2320.jpg (try: 2) =>

`panteltje.com/pub/LM317_switcher/LM317_switcher_no_100nF_img_2320.jpg' Connecting to panteltje.com|81.207.135.196|:21... connected. Logging in as anonymous ... Error in server response, closing control connection. Retrying.

--07:46:17-- ftp://panteltje.com/pub/LM317_switcher/LM317_switcher_no_100nF_img_2320.jpg (try: 3) =>

`panteltje.com/pub/LM317_switcher/LM317_switcher_no_100nF_img_2320.jpg' Connecting to panteltje.com|81.207.135.196|:21... connected. Logging in as anonymous ... Error in server response, closing control connection. Retrying.

^C Not worth looking at.

exactly it is.

switching,

% efficiency at 160 mA

Reply to
Grant

On a sunny day (Wed, 18 Aug 2010 22:06:46 +0100) it happened "ian field" wrote in :

Ah yes, I see they have a switching example on page 9. It has more pins than the LM317 though :-) I think this is the Joerg chip?

Reply to
Jan Panteltje

It certainly is - but I recently checked for someone else, Farnell carry stock in practically every case style - including the*ancient* metal can.

There's no harm suggesting a chip that was *designed* for switch mode, but its often the case that an OP has set on a particular topology.

Reply to
ian field

On a sunny day (Thu, 19 Aug 2010 07:44:39 +1000) it happened Grant wrote in :

ftp://panteltje.com/pub/LM317_switcher/LM317_switcher_final_img_2312.jpg

ftp://panteltje.com/pub/LM317_switcher/LM317_switcher_final_img_2312.jpg

ftp://panteltje.com/pub/LM317_switcher/LM317_switcher_final_img_2312.jpg

ftp://panteltje.com/pub/LM317_switcher/LM317_switcher_final_img_2312.jpg

`panteltje.com/pub/LM317_switcher/LM317_switcher_final_img_2312.jpg'

12.91K/s

`panteltje.com/pub/LM317_switcher/LM317_switcher_final_img_2312.jpg' saved [11070]

I dunno, nothing changed here, but I noticed some network outages today, could not reach google, well did eventually, but took a long time. A ping 88.8.8.8 had only 17.8 ms. I did accidently have a ping at that IP running all the time.... Cannot imagine it ate all the bandwidth (only one per second). Probably read tomorrow that google had a problem... Just looking at the ftp log, you could be 123.2.77.8, as only that one only downloaded that file, the rest all the others too :-) Wed Aug 18 23:42:56 2010 1 123.2.77.8 11070 /home/ftp/pub/LM317_switcher/LM317_switcher_final_img_2312.jpg b _ o a -wget@ ftp 0 * c Yes wget, its you: grml: ~ # ip_to_country -i 123.2.77.8 ip=123.2.77.8 (2063748360) "AU" "AUSTRALIA" BTW I wrote ip_to_country :

formatting link

Reply to
Jan Panteltje

Just for grins does anyone know who designed the 732? If that person is still around i would like to ask some questions about the design goals.

Reply to
JosephKK

Didn't you mean 723? 732 is a stereo separator.

I think the 723 was designed by Widlar... long ago deceased

The 723 works well, though cantankerous with regard to stability. ...Jim Thompson

--
| James E.Thompson, CTO                            |    mens     |
| Analog Innovations, Inc.                         |     et      |
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    |
| Phoenix, Arizona  85048    Skype: Contacts Only  |             |
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  |
| E-mail Icon at http://www.analog-innovations.com |    1962     |

                   Spice is like a sports car... 
     Performance only as good as the person behind the wheel.
Reply to
Jim Thompson

"Jim Thompson" wrote in message news: snipped-for-privacy@4ax.com...

Never had any great difficulty with servicing gear that used a 723, for home brews I went through a L200 phase (mainly because I had a ready supply) the rest of the time I've used the 78xx types or modified small open frame switchers to do what I want.

Reply to
ian field

I used it well into the '90s, as it made it simple to design very low noise supplies for applications such as reference oscillators.

The on-die reference voltage was relatively high (a zener) and appeared on a pin which allowed it to be filtered. (C.F. the plethora of regs available at the time which used a ~1.2V bandgap and were as noisy as all hell.)

There was also a problem with a typical application for the 723 that used an external PNP pass transistor. I don't recall the exact details now, but under certain conditions, one of the transistors in the differential pair inside the 723 could become C-B forward biased, and this would latch up the circuit with the external PNP in saturation (possibly frying whatever device was being fed by this regulator). The fix was simple (but again, I don't recall exactly what it was).

Regards, Allan

Reply to
Allan Herriman

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.