0
\$\begingroup\$

I am planning to use LT1959 IC for making a voltage regulator circuit for 3V 1A output. The datasheet does not mention maximum allowed value for input capacitor for the circuit.

In order to meet the ripple current requirements for the input capacitor(I RMS(MAX) =I OUT/2 = 0.5A), I am planning to use 7-10 100 uF electrolytic capacitors (in parallel configuration) with 0.1A max ripple current rating. This is because of time and cost constraints. So, my effective input capacitor value would be 1000 uF with 0.7A-1.0A ripple current rating.

I am again going to put about 10 electrolytic caps of 100 uF in parallel for output capacitor. I will also use few 0.1 uF ceramic caps in parallel with the electrolytic output caps. So my effective output capacitor would >1000 uF.

The evaluation board manual for LT1959 demo circuit designed by Linear Technology uses only 25 uF input capacitor and 100 uF output capacitor.

Questions:

1) Should I be concerned about the relatively high effective capacitance value of input and output capacitors which I am about to use, as compared to the company design?

2) Would there be any other problems in the circuit if I continue with this design?

3) Should I place any extra ceramic input cap in parallel with the input electrolyic caps?

4) Should I place output ceramic cap close to IC or closer to load?

\$\endgroup\$
7
  • \$\begingroup\$ 1. Input no. Output perhaps. Simulate startup to find out. 3. Depends on ESR, ESL and layout. 4. Probably IC, but depends. Can you show your schematic and layout? \$\endgroup\$
    – winny
    Commented Jul 17, 2018 at 15:59
  • \$\begingroup\$ @winny My layout will be same as the schematic diagram on Page.2 ofevaluation kit manual. I will be using a perfboard for this. This is because ,i dont not have much time left to complete the project. \$\endgroup\$
    – Mrchief
    Commented Jul 17, 2018 at 16:12
  • \$\begingroup\$ It seems like excessive start current but all depends on your missing specs for step load regulation , ripple and power dissipation. I would C for SRF at 0.5 MHz and add smaller one for 1.5 MHz where LPF is low Q at 20~50 kHz at max current. \$\endgroup\$
    – D.A.S.
    Commented Jul 17, 2018 at 17:11
  • \$\begingroup\$ @TonyEErocketscientist Load is a dc motor. Do you mean it will take in more current from source at start? \$\endgroup\$
    – Mrchief
    Commented Jul 17, 2018 at 18:10
  • \$\begingroup\$ ”My layout will be same as the schematic diagram on Page.2 ofevaluation kit manual. I will be using a perfboard for this.” I would say perfboard is the exact opposite of what the eval kit is using. \$\endgroup\$
    – winny
    Commented Jul 17, 2018 at 20:05

2 Answers 2

Reset to default
0
\$\begingroup\$

You're in luck. The datasheet for this part goes into extensive detail about the design considerations for the input and output caps. I won't quote it here, because it's several paragraphs. But if you follow those recommendations, and test the design then I'm sure it will be fine.

One problem I'm seeing immediately is that your ESR is going to be too LOW with the ceramic output caps. Most regulators come with a maximum and minimum output ESR. The datasheet goes into more detail about this, but the first thing they say is that for typical applications it will be between 0.05 and 0.2 ohms; you're going to be below that.

Edit: I think you're jumping the gun worrying about this. By default you should:

  1. Design according to the datasheet. Read it carefully and make sure you're within the specs.

  2. Test it.

If you do that, and it still has problems, then you might have to dig into the theory of switching regulators, and what might be going wrong.

\$\endgroup\$
4
  • \$\begingroup\$ I did read the recommendations for the input and output caps .The ripple current rating is the main thing as per the datasheet.I did not find any explicitly mentioned maximum value for input and output capacitance.I would like to know if any problems would arise like lesser efficiency or any other issues due to high output and input capacitance. \$\endgroup\$
    – Mrchief
    Commented Jul 17, 2018 at 16:54
  • \$\begingroup\$ Should i not use the ceramic output cap then(If the ESR goes too low)? I do not have instrument to measure ESR though.Should i only use electrolytic caps for output caps? \$\endgroup\$
    – Mrchief
    Commented Jul 17, 2018 at 16:56
  • \$\begingroup\$ I'd say if the datasheet talks for 3 paragraphs about the output caps and doesn't mention the capacitance, then it probably doesn't matter. That's assuming you read through the datasheet and don't see it spec'd anywhere else. \$\endgroup\$
    – Drew
    Commented Jul 17, 2018 at 20:38
  • \$\begingroup\$ Most capacitors should have a spec in the datasheet for ESR, if they don't you can probably estimate based on the type of cap, voltage rating, etc. And when I say estimate, I mean, find similar caps using google which do specify ESR. \$\endgroup\$
    – Drew
    Commented Jul 17, 2018 at 20:40
0
\$\begingroup\$

Electrolytic capacitors are not the best choice for this application. Typically ceramic capacitors are used due to their very low ESR. Ceramic capacitors can easily handle 0.5A of ripple current on the input. This part seems to caution against using them on the output and recommends a specific type of tantalum capacitor that provides the optimum ESR for this part. Most buck regulators are designed to use ceramic capacitors on the output and this part also requires an external diode since it is not synchronous. If it were me I would choose a different regulator that is designed for ceramic capacitors and is synchronous. I strongly recommend that you simulate this circuit in Ltspice.

\$\endgroup\$
1
  • \$\begingroup\$ The problem is that i do not have ceramic caps with well specified current rating. I have those cheap brown ceramic disc caps with 104 AEC written on them. I have to get this done within a day,So i just have to get this done with what i have now. \$\endgroup\$
    – Mrchief
    Commented Jul 17, 2018 at 17:04

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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