EEVblog Electronics Community Forum

Electronics => Beginners => Topic started by: bryce1 on February 05, 2014, 10:50:09 am

Title: Getting 5-19V to 3.3V
Post by: bryce1 on February 05, 2014, 10:50:09 am
Hi guys,

I am in the process of designing a board that due to the use of the enc28j60 draws around 200ma. Since the Input power should vary depending on the wallwart used, I need to design some solution that gets the 5-19V DC down to the 3.3V used by the enc28j60.

My question here:

What is the most intelligent way to keep the complexity (number of parts) low while still having a reliable solution that is somehow efficient?

In my first try, I used a lm317 (LDO) without proper cooling, which was stupid. My board is fairly simple and I will make "a lot" of them (about 50), so I obviously care about cost as well, but my main concern is my lack of experience in designing buck converters.
Title: Re: Getting 5-19V to 3.3V
Post by: grumpydoc on February 05, 2014, 10:56:48 am
Something like this, perhaps? http://uk.farnell.com/xp-power/tr05s3v3/dc-dc-converter-0-5a-3-3v-sip/dp/2319829 (http://uk.farnell.com/xp-power/tr05s3v3/dc-dc-converter-0-5a-3-3v-sip/dp/2319829)
Title: Re: Getting 5-19V to 3.3V
Post by: fcb on February 05, 2014, 11:39:02 am
The best compromise between efficiency, cost & complexity would be something like a "simple switcher" from National (now TI).  You can pretty much guarantee it will work if you use their online software to design it.

If your super-cost sensitive then something like the MC34063.

Be careful about availability on those all-in-one SIL type regulators.
Title: Re: Getting 5-19V to 3.3V
Post by: bryce1 on February 05, 2014, 02:18:43 pm
@grumpydoc: oh wow, I didn't now these existed! thx
@fcb: I have a problem deciding on inductors. The software is helpful in many ways, but I have to read up some more on inductors.

edit: ok as far as I know, this inductor would be ideal for use with a lm22574-3.3 in this scenario:
http://www.digikey.com/product-detail/en/NS10145T152MNA/587-3120-1-ND/2774735 (http://www.digikey.com/product-detail/en/NS10145T152MNA/587-3120-1-ND/2774735)

Did I get it right?
Title: Re: Getting 5-19V to 3.3V
Post by: rsjsouza on February 05, 2014, 03:13:02 pm
A quick parametric search on the TI site shows six switchers that somewhat fit the bill. Check it here (http://www.ti.com/lsds/ti/power-management/step-down-buck-module-products.page#p238min=4.5;4.5&p238max=20;20&p634min=0.8;2.8&p634max=3.6;6).

You can always tweak the parameters at the top, although this page sometimes shows a strange display depending on the browser I am using.
Title: Re: Getting 5-19V to 3.3V
Post by: bryce1 on February 05, 2014, 03:18:57 pm
thx but this is definetly not the price range for this project.
Title: Re: Getting 5-19V to 3.3V
Post by: qno on February 05, 2014, 04:04:32 pm
Hi Brice1

Try a search on ebay on "buck module"

http://www.ebay.com/itm/1pcs-1-23V-30V-DC-DC-Buck-Converter-Step-Down-Module-LM2596-Power-Supply-Output-/110950894139?pt=LH_DefaultDomain_0&hash=item19d5304a3b (http://www.ebay.com/itm/1pcs-1-23V-30V-DC-DC-Buck-Converter-Step-Down-Module-LM2596-Power-Supply-Output-/110950894139?pt=LH_DefaultDomain_0&hash=item19d5304a3b)
Title: Re: Getting 5-19V to 3.3V
Post by: mariush on February 05, 2014, 07:33:05 pm
You could design something super fast and with minimal components with an AP1509-33 :

http://uk.farnell.com/diodes-inc/ap1509-33sg-13/ic-pwm-reg-buck-150khz-sop-8/dp/1825322RL (http://uk.farnell.com/diodes-inc/ap1509-33sg-13/ic-pwm-reg-buck-150khz-sop-8/dp/1825322RL)
http://www.digikey.com/product-detail/en/AP1509-33SG-13/AP1509-33SGDICT-ND/1301652 (http://www.digikey.com/product-detail/en/AP1509-33SG-13/AP1509-33SGDICT-ND/1301652)

Works with 4.5-22v, 150 kHz,  small pcb space, probably under 2$ each for everything if you get 50pcs.
Title: Re: Getting 5-19V to 3.3V
Post by: bryce1 on February 07, 2014, 06:26:22 pm
Ok I drew up a LM2574M version. Would this be ok (mainly, working for a long long time)?

The LM2574M:
http://www.digikey.com/product-detail/en/LM2574M-3.3%2FNOPB/LM2574M-3.3%2FNOPB-ND/212607 (http://www.digikey.com/product-detail/en/LM2574M-3.3%2FNOPB/LM2574M-3.3%2FNOPB-ND/212607)

With this inductor:
http://www.digikey.com/product-detail/en/NS10145T152MNA/587-3120-1-ND/2774735 (http://www.digikey.com/product-detail/en/NS10145T152MNA/587-3120-1-ND/2774735)

This diode:
http://www.digikey.com/product-detail/en/S1G/S1GFSCT-ND/965720 (http://www.digikey.com/product-detail/en/S1G/S1GFSCT-ND/965720)

Two of these caps:
http://www.digikey.com/product-detail/en/UCW1H221MNL1GS/493-9426-1-ND/3962774 (http://www.digikey.com/product-detail/en/UCW1H221MNL1GS/493-9426-1-ND/3962774)


Title: Re: Getting 5-19V to 3.3V
Post by: mariush on February 07, 2014, 08:21:15 pm
How the hell did you come up with a 1500uH inductor? Did you even look in the datasheet for LM2574 ?

Datasheet : http://www.ti.com.cn/cn/lit/ds/symlink/lm2574.pdf (http://www.ti.com.cn/cn/lit/ds/symlink/lm2574.pdf)   

Inductor selection: See page 14, figure 25.  See Design Procedure, page 13.  I would probably choose 470uH-680uH depeding on what input voltage range is more likely to be used.

The diode you chose is a Standard Recovery 1A one. It's crap. You need a fast recovery / schottky diode. Again, page 13, recommended diode is fast recovery kind with 1.5x the current you want.

Through hole, 1n5817 would work fine: http://www.digikey.com/product-detail/en/1N5817-TP/1N5817-TPCT-ND/950586 (http://www.digikey.com/product-detail/en/1N5817-TP/1N5817-TPCT-ND/950586)
SMD, maybe something like this:

http://www.digikey.com/product-detail/en/DB3J208K0L/DB3J208K0LCT-ND/2704394 (http://www.digikey.com/product-detail/en/DB3J208K0L/DB3J208K0LCT-ND/2704394)
http://www.digikey.com/product-detail/en/SS110-TP/SS110-TPCT-ND/950555 (http://www.digikey.com/product-detail/en/SS110-TP/SS110-TPCT-ND/950555)

Capacitors are probably OK, but you can do better.
You don't need 220uF at the input. Datasheet recommends as little as 22uF. 100-150uF would probably be more than enough. If your input will be a maximum of 19v, then you can choose a maximum of 35v, no need to go with 50v.
For example: http://www.digikey.com/product-detail/en/UCL1V151MNL1GS/493-3955-1-ND/2300382 (http://www.digikey.com/product-detail/en/UCL1V151MNL1GS/493-3955-1-ND/2300382)

You have 3.3v at the output so no point using 35-50v rated capacitors there. Also, the regulator likes low esr capacitors on the output, the one you chose has 0.35 ohm esr which is low, but you can do better. You can go with 220-470uF 10-16v at the output and if you want to, you can add a small 10-22uH inductor and a second capacitor to get output even smoother.

Suggestions:

http://www.digikey.com/product-detail/en/EEE-FP1A331AP/PCE4525CT-ND/1701024 (http://www.digikey.com/product-detail/en/EEE-FP1A331AP/PCE4525CT-ND/1701024)
http://www.digikey.com/product-detail/en/UCL1A331MCL6GS/493-3906-1-ND/2300333 (http://www.digikey.com/product-detail/en/UCL1A331MCL6GS/493-3906-1-ND/2300333)
 

At 3.3v 200-300mA output and about 75% efficiency, the regulator won't heat up much, so you don't need to chose 5000h@105c rated capacitors.

later edit: also read application hints on page 17-20 and do the calculations for the formulas in those pages using your numbers (3.3v and 0.2-0.3a, 5-19v input voltage etc) and you will see why the parts you chose are not so good.

even later edit: as for inductors, something like these would probably work fine:

http://www.digikey.com/product-detail/en/SDR1307-471KL/SDR1307-471KLCT-ND/3043672 (http://www.digikey.com/product-detail/en/SDR1307-471KL/SDR1307-471KLCT-ND/3043672)
http://www.digikey.com/product-detail/en/SRR1280-471K/SRR1280-471KLCT-ND/2127445 (http://www.digikey.com/product-detail/en/SRR1280-471K/SRR1280-471KLCT-ND/2127445)
Title: Re: Getting 5-19V to 3.3V
Post by: Rick60 on February 07, 2014, 09:20:46 pm
I have been using  MICROCHIP - MCP16322T-330E/NG - BUCK, REG, 2A, Max 24VIN, 3.3V, 16VQFN , no need for a diode , a small 4.7uH inductor , just dont tie the en pin to vin :palm:  Makes a very compact design , a bit of a pain to hand solder however due to the thermal pad
Title: Re: Getting 5-19V to 3.3V
Post by: os40la on February 07, 2014, 10:11:02 pm
Take a look at one of these. Price may be a little high. But you can't beat the complexity.  :-+
 
The 7v min may not work for you.

http://www.digikey.com/product-detail/en/OKI-78SR-3.3%2F1.5-W36-C/811-2195-5-ND/2259780 (http://www.digikey.com/product-detail/en/OKI-78SR-3.3%2F1.5-W36-C/811-2195-5-ND/2259780)
Title: Re: Getting 5-19V to 3.3V
Post by: jeremy on February 08, 2014, 01:56:50 am
+1 for qnos link. I use them in everything