Post by thread:[EE] Wide input voltage step down SMPS IC out = 5V

part 1 2492 bytes content-type:text/plain; charset="iso-8859-1" (decoded quoted-printable) Hi, could anybody please suggest me a good and possibly cheap (it's not for a one-off) small, nice, step down SMPS IC with wide input voltage range (say 6..40 or more V) out = 5V (even better, dual out 5V / 12V ) 100+mA? 40V+ input narrows the range and "cheap" seems to vanish. __________________ LM5085 $1.24/1000 Digikey $1.07/10000 Digikey Reasonably nice IC if $1+ OK. External FET (more $) 4.5 - 75 Vin External FET Freq prog to 1 Mhz. 95% at 500 mA out at 7V in 80%+ at higher Iouts at most Vins. MSOP-8, LLP-8 (3mm x 3mm) http://www.national.com/ds/LM/LM5085.pdf ___________________ Worsish in some ways but aimed at efficiency at lower loads with PFM Graphs on pages 23 + 24 would give you a fright when you first met them being implemented in action. eg Infineon offer eg TLE 629 TLE6389-3G V50CT-ND<http://search.digikey.com/us/en/products/TLE6389-3G%20V50/TLE6389-3G%20V50CT-ND/2231231> However, as that uses an external FET switch a little playing could extend the range indefinitely ___________________ But, there's always the time honoured MC34063 which has a switch abs max voltage of 40V stand alone and as high as you like with external FET. Only 100 kHz ish so larger L's but very minimal parts count up to ~~~ 40V. http://www.st.com/internet/com/TECHNICAL_RESOURCES/TECHNICAL_LITERATURE/DATASHEET/CD00001232.pdf See figs 15+16 for minimum parts buck cct. and PCB layout IC. 3R, 1C, D, L + 2 x filter caps. Not marvellous efficiency at low Vin due to non saturating darlington driver but can add external FET if desired. $0.19 / 1000 Digikey $0.124/50,000 Digikey $US0.05 - 0.10 / China (Motorola licenced copy) _______________________ "My" "God's switching Regulator" [tm] will do this with 3 jellybean transistors 4R, 1D. 1L, 1Z, 2 x filter caps. Attached is someone's apparently parallel reincarnation of my design. That's from here : http://sound.westhost.com//appnotes/an006.htm With a constant current version here http://sound.westhost.com//appnotes/an003.htm I've had versions of this that ran at Vin from 12V to 200V and Vout = 12V.. Below 12V Vout tracked Vin less Q3 saturation. There are 10,000++ of these "out there" so ignore people who say it won't work :-). [It's actually a 'hysteretic converter' but I had not met them when the concept dropped/was dropped fully formed into my mind in time of need :-).] Russell part 2 3174 bytes content-type:image/gif; (decode) part 3 181 bytes content-type:text/plain; name="ATT00001.txt" (decoded base64) -- http://www.piclist.com PIC/SX FAQ & list archive View/change your membership options at mailman.mit.edu/mailman/listinfo/piclist

part 1 1076 bytes content-type:text/plain; charset="iso-8859-1" (decoded quoted-printable) Attached is a cct of my high voltage input version from February 2001 (how the years fly). On line with notes here http://www.piclist.com/techref/postbot.asp?by=thread&id=%5BEE%5D+Low+cost+switching+regulators&w=body&tgt=_top P Channel MOSFET used on high side. Bipolar OK especially at low power. QBUK2 is high voltage transistor (MPSA42) due to voltage input. RBUK3 also has full HV across it. Voltage is somewhat above Vzener. Zener runs on soft knee as shown but end result is amazingly good. Much better Vin and load stability than would be expected from its appearance. (The 2 transistor "Black Regulator" has fewer parts but is much more load sensitive. Oscillation is "chaotic" - nicely spread spectrum noise wise. Ripple on CBUK2 and zener noise both contribute to switching point. Some people have added positive feedback to formalise the feedback. eg (probably) a resistor from hot end of coil to base of QBUK1 would add positive feedback and hysteresis. Russell GSR _GSR part 2 9148 bytes content-type:image/gif; name="GSR 040224.gif" (decode) part 3 181 bytes content-type:text/plain; name="ATT00001.txt" (decoded base64) -- http://www.piclist.com PIC/SX FAQ & list archive View/change your membership options at mailman.mit.edu/mailman/listinfo/piclist

> This circuit has been modified at times to take a much higher input voltage: > http://www.romanblack.com/smps/smps.htm The "Black Converter" is cost efficient and compact but is far less regulated than 'my' 3 transistor GSR. The difference is significant in real world applications. Roman produced his circuit based on a relay driver circuit suggested by Richard Prosser in response to a long ago PICLIst design challenge where I put my design up as straw-man. If the regulation is acceptable Roman's wins out slightly over mine on price. Russell McMaho

On Mon, Oct 31, 2011 at 5:57 PM, RussellMc <apptechnzKILLspamgmail.com> wrote: > The "Black Converter" is cost efficient and compact but is far less > regulated than 'my' 3 transistor GSR. The difference is significant in > real world applications. > > Roman produced Â his circuit based on a relay driver circuit suggested > by Richard Prosser in response to a long ago PICLIst design challenge > where I put my design up as straw-man. If the regulation Â is > acceptable Roman's wins out slightly over mine on price. > Possibly useless anecdote: I designed a buck converter based on the TL432. Regulation was "perfect" but efficiency wasn't so great. Extremely cheap I'd say. -- Martin K.

On Mon, Oct 31, 2011 at 5:03 PM, Dwayne Reid <EraseMEdwaynerspam_OUTTakeThisOuTplanet.eon.net> wrote: > At 04:24 PM 10/31/2011, M.L. wrote: > >>Possibly useless anecdote: I designed a buck converter based on the >>TL432. Regulation was "perfect" but efficiency wasn't so great. >>Extremely cheap I'd say. > > I assume that you meant TL431 rather than TL432? > "The TL432 has exactly the same functionality and electrical specifications as the TL431, but has different pinouts." from http://www.ti.com/product/tl432 -- Regards, Mark markrages@gmai

>>>Possibly useless anecdote: I designed a buck converter based on the >>>TL432. Regulation was "perfect" but efficiency wasn't so great. >>>Extremely cheap I'd say. >> I assume that you meant TL431 rather than TL432? > "The TL432 has exactly the same functionality and electrical > specifications as the TL431, but has different pinouts." from > http://www.ti.com/product/tl432 Worse!!!! The TL431 and TL432 will often be pinout variants but some manufacturers make a TL432 with a TL431 pinout. Manufacturer and distributor must be specified and locked in with any variation from this requiring change control and testing and certification of the alternative. Enforcing that from a distance without "being there" relies totally on integrity and buy-in and competence of the manufacturer. That's an unrealistically large amount to expect of a manufacturer in many cases. Some years ago I saw (afterwards & from a distance) a complete production batch of products made and shipped with a pinout-incorrect variant of the TL431 SOT-23 pkg installed. These had a significantly bad effect on performance. and the factory engineers involved were aware that performance was not as it should be but did not know why and did not report it and afterwards justified it with an explanation re how they thought that was what was wanted. Complete swapout of PCBs or recall of that batch of product was required. Russell McMahon

On Mon, Oct 31, 2011 at 7:03 PM, Dwayne Reid <@spam@dwaynerKILLspamplanet.eon.net> wrote: > At 04:24 PM 10/31/2011, M.L. wrote: > >>Possibly useless anecdote: I designed a buck converter based on the >>TL432. Regulation was "perfect" but efficiency wasn't so great. >>Extremely cheap I'd say. > > I assume that you meant TL431 rather than TL432? > > Do you have the design posted anywhere? > > Thanks! > > dwayne I'm really busy with work on a customer site so I'll have to get back to you. I have it soldered together but I can't seem to find a drawing of the schematic. I think I meant the TL432, if that's the version of the TL431 with lower minimum cathode current. -- Martin K

TLV43x is low voltage reference version = 1.25V. TL43x is 2.5V reference version. TLV has notably lower Imin_reg than TL version (about 80 uA vesis 400 uA from memory. Either would work as a switching regulator. Main limitation is low frequency response. Out of head. Buck regulator. High side P channel MOSFET feeds inductor o output = std Buck. P Channel drain to reverse biased diode to ground = std Buck. TL431 anode grounded. 2 resistor divider across output - tap to TL431 gate.or ref pin. Small pnp emitter to V+, collector to MOSFET gate. resistor base to V+ say 100k. resistor base to TL431 Cathode = ? 10k. MOSFET gate resistor to ground. Say ? 1k. ___ When output high TL431 turns on. TL431 turns on PNP. PNP turns off FET. Inductor delivers energy till I = 0. Output falls. TL431 turns off MOSFET turns on. Voutrises. Cycle repeats. Russell Cycle completes. On 3 November 2011 11:05, M.L. <KILLspammKILLspamlkeng.net> wrote: {Quote hidden}> On Mon, Oct 31, 2011 at 7:03 PM, Dwayne Reid <RemoveMEdwaynerTakeThisOuTplanet.eon.net> wrote: >> At 04:24 PM 10/31/2011, M.L. wrote: >> >>>Possibly useless anecdote: I designed a buck converter based on the >>>TL432. Regulation was "perfect" but efficiency wasn't so great. >>>Extremely cheap I'd say. >> >> I assume that you meant TL431 rather than TL432? >> >> Do you have the design posted anywhere? >> >> Thanks! >> >> dwayne > > I'm really busy with work on a customer site so I'll have to get back > to you. I have it soldered together but I can't seem to find a drawing > of the schematic. I think I meant the TL432, if that's the version of > the TL431 with lower minimum cathode current. > -- > Martin K. >