3
\$\begingroup\$

I'm in the third year of my electrical engineering major and I want to take things a bit further than the usual analogue electronic circuits I've studied thus far by making a DC-DC step down converter that would convert 12V dc to 5V DC and output 3 amps of current.

So far I've only designed a PCB and I've gathered my components, but before I actually start assembling everything, I would like some advice from you guys especially since I can't test the components since it is my understanding that using a breadboard is not a good idea and perf boards are not really easy to get for me.

The schematic and the design of the PCB are as follows:

enter image description here enter image description here

You can tell from the schematic what components I will be using, however I wasn't able to find those exact resistors. The closest I could get was a 1k resistor and a 3.3k resistor for the voltage divider resulting in 1.163 volts instead of 1.21 or thereabouts. Alternatively, I was able to find 4.7k and 1.5 k resistors which would give me 1.21 volts, however the error for all these components are 5%. Every electronics store I've been to, I was told 1% resistors are considered military grade so they don't sell them or some other bogus. I bought 20 pieces of each of the four resistors and I'll try to pick the ones with the closest value to the theoretical one (using a multi meter) but I don't know how they would fare with changes in temperature and load and what not. So this is one aspect I need help with.

I also wasn't able to find a 22uH inductor that could take 6 amps (apparently it is recommended that I should choose an inductor that could take twice the current the converter will supply) however I was able to find 11uH inductors rated for 6 amps and I was wondering if it would be ok to connect them in series.

And I forgot to mention this the first time I posted the question but I also was not able to get the diode that is in the schematics but I was able to get an SB560 diode which as far as I can tell from the data sheets, it should do the job but in any case if any one would like comment on that, it would be great.

What are your thought on this?

\$\endgroup\$
  • \$\begingroup\$ That link just tells me my browser is out of date and doesn't show anything. Besides that, is all this online after a month? A year? many years? Better attach some images... \$\endgroup\$ – PlasmaHH Feb 12 '16 at 15:29
  • \$\begingroup\$ Actually, I'm looking at the link right now. I can't tell why it doesn't work for you. I'll post some images in 15 minutes. \$\endgroup\$ – user1969903 Feb 12 '16 at 15:30
3
\$\begingroup\$

Regarding components sourcing: just put in an order at Digikey, Mouser, or whatever mail-order supplier you prefer. Your local shop may not have 1% (E96) resistors, but if you order them, you can get whatever you want. And you can get the correct inductor too. Also note that the datasheet has a nice table of suggested part numbers.

Some notes regarding your layout:

Read the datasheet carefully, then read it again. Some things that stand out:

  • the datasheet shows which traces should be kept short. Follow these guidelines!
  • your inductor connection crosses under one of your feedback resistors. The datasheet instructs you to not do this.
  • adjust your layout so that your diode, input cap, and output cap are as close as possible to the regulator. Your cap layout is alright, but your diode could be much closer (the feedback resistors need not be in close proximity to the regulator).
  • a ground pour (if not a ground plane on the bottom layer) would be advantageous. I'm not sure why you're using a single layer design, but there's no cost advantages for a one-off.
  • those big electrolytics may have too high of an ESR, paralleling lower-ESR caps is a good idea. Output ripple voltage is directly proportional to output cap ESR, hence you should do your best to reduce ESR; paralling capacitors is an easy method.

Again, almost all of this is in the datasheet, you simply need to read it carefully. Also, you can make this work with all through-hole parts, but if you're willing to switch some parts to SMT, you will have a more compact layout as well as improved performance.

\$\endgroup\$
  • \$\begingroup\$ Thank you for your time. I did have a read through but I am a superficial reader so I'll take a closer look and keep your advice in mind. The capacitors I have, although they have the exact same specs, they're actually not electrolytics (I chose them specifically with low ESR in mind). I'll rethink my layout and I will try to come up with 2 designs, one with SMD parts and one with through hole parts and I'll post it here once I finish. \$\endgroup\$ – user1969903 Feb 12 '16 at 17:39
  • \$\begingroup\$ Alrighty, that's good you have proper capacitors. I think adding a ground plane and swapping the location of the feedback resistors and diode/inductor would be a big improvement. \$\endgroup\$ – uint128_t Feb 12 '16 at 17:57
  • \$\begingroup\$ I'll do that and I'll post back. \$\endgroup\$ – user1969903 Feb 12 '16 at 18:03
  • \$\begingroup\$ I just sketched a new layout on paper and I don't know why I didn't think of a ground pour. It makes it a lot cleaner. I'm thinking of placing the diode above the controller as close as I can get it. I'll read through the datasheet a couple more times, see what components I'll order and should come up with a new layout in a couple of days or so. \$\endgroup\$ – user1969903 Feb 12 '16 at 22:26
  • \$\begingroup\$ I was looking for parts on one of the websites posted here earlier and I was joggling between different value resistors and I had the idea of using a potentiometer in combination with a resistor, and then I thought, why not use just a potentiometer? Would it be acceptable to place a potentiometer before the output and connect it straight to the feedback pin? And if not, it should be a good idea to place a potentiometer instead of the second resistor of the voltage divider, correct? \$\endgroup\$ – user1969903 Feb 14 '16 at 13:30
1
\$\begingroup\$

For your resistors, you should have little/no problem sourcing them online (depending on your location & possibly with your country's treaty status with suppliers' countries). Here is a link to through-hole resistors from one supplier (DigiKey) who is readily available, and stocks (chip) resistors down to 0.001% tolerance. Whdn purchasing 'standard' values, I order in quantities of >100 from them & get 100 1% tolerance chip resistors for $0.84 (USD).

Alternatively, you can use the 1K resistor you already have, and a 20K linear-taper pot in place of the 3.3K & adjust as required, while watching your metsr across the Vout, to reach you exact desired output voltage.

As for wiring 2 11uH inductors in series; as long as you don't saturate your cores, or physically place your inductors so that mutual inductance would adversely affect your circuit, that should be fine.

\$\endgroup\$
  • \$\begingroup\$ Thanks for the reply. Well I'll be honest. I was hoping to avoid online orders. I admit I like the convenience of going to the shop and picking up what I need right there and then but I'll do it if necessary. How badly would using either pair of the resistors I have already be? Regarding the inductors, I was able to source some enclosed SMD inductors but I don't have the tools to use those. If I order the resistors online, I'll also order an inductor and the diode in the schematic. \$\endgroup\$ – user1969903 Feb 12 '16 at 16:31
  • \$\begingroup\$ Using either pair of resistors you already have would be equal to the deviation from ideal of one resistor + the in-phase deviation of the other resistor + any in-phase deviation contributed by the IC's inherent potential inaccuracy. Best-case scenario: the three could perfectly cancel each other, yielding ideal output. Worst-case: possibly up to ~20% inaccuracy (range of about 4V-6V) in your possible resting output voltage. Measuring the resistors with a good meter can help you match 2 together for closest match to your 1:3.16 'ideal' ratio; the closer you get there, the better your result. \$\endgroup\$ – Robherc KV5ROB Feb 12 '16 at 16:47
  • \$\begingroup\$ Also, definitely read & carefully consider unit128_t's answer. He made several points on design issues whkch are very likely to impact your circuit's success. P.S. Why do you feel you lack the equkpment to solder on a SMT inductor? I very regularly solder SMT inductors, capacitors, resistors & transistors using the same solder & Iron I use for through-hole components. SMDs are a bit 'touchier,' but soldering them with a 'regular' iron is still feasible. \$\endgroup\$ – Robherc KV5ROB Feb 12 '16 at 16:51
  • \$\begingroup\$ A range of 4-6 is a risk not worth taking I suppose. I never had a good experience with soldering irons. The ones I had / have always oxidize the tip (it blackens to the point the solder won't take to it) and this is regardless of how I'm tinning it, I don't keep it on for a long period of time and all the other precautions. I did play with one of my universities soldering station ( a weller brand station) and they were a real joy but they're way over my budget. If I go with SMD I'll probably try using their stations if I can't do it with my iron. \$\endgroup\$ – user1969903 Feb 12 '16 at 17:35
  • \$\begingroup\$ Yes, use their station for now is best idea. As an 'interim' step, idk weller irons very well to recommend a good low-cost model, but if you can afford $100, I LOVE my Hakko FX888D temp-controlled station, and they have some $30-50 pen-type irons that just plug into the wall which suffice for 'general use' & aren't nearly as unpredictable as mkst pen-irons I've failed with in the past ;) \$\endgroup\$ – Robherc KV5ROB Feb 12 '16 at 18:15

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.