2
\$\begingroup\$

The image below is a 680uH and 4A inductor:

enter image description here

Can I use 5 of them in parallel to make a boost converter which has input current of 15A and inductance of 136uH. The converter is made of discrete components thus no recommended inductor can be found.

I think to do the same for the pi filter at output which is 10A max and will consist of two parallel 68uH inductors with 1mm wire thickness:

enter image description here

\$\endgroup\$
1
  • \$\begingroup\$ You could try, but each inductor may not be identical and one may take most of the current. \$\endgroup\$
    – Kartman
    Nov 11, 2022 at 3:43

2 Answers 2

2
\$\begingroup\$

Yes, but consider doing a few other things first:

  1. Raise operating frequency. Voltage isn't indicated, but if this is average low voltage stuff (5-30V?), these inductances sound monstrously large for that.

    For current, non-bleeding-edge tech, I would choose values in that range for, oh I don't know, mains voltages or thereabouts?

    Example: 136uH 15A has 20% or 3A peak current ripple, at 100kHz, for ±160V applied, e.g. a buck converter from 320 down to 160V, or a bit slower for other ratios at the same input. That is, V = L dI/dt, for dI = 3A, dt = 2.5µs (quarter a 100kHz wave -- quarter because peak, not p-p which would be half up and half down), and L = 136µH.

  2. Rather than wiring everything in parallel / synchronously, do it out of phase (phase interleaved converter). This is more of a priority with respect to transistor ratings, or these days even moreso just sheer layout limitations, but perhaps it's relevant here, too. The disadvantage is more components: more inverter and filter stages repeated; more control complexity (hopefully solved with a single integrated controller, but if not, it can get quite baroque). The advantage is, input and output ripple goes as 1/N for N stages in phase interleave. Which greatly saves on capacitor ratings, which is actually a rather big deal for compact, high power designs.

  3. Can you really not find bobbin-style inductors in those values? They're fairly common around here at least [western markets]. But maybe I'm missing something from my perspective.

    3a. Does it have to be bobbin-style? There are other semi/fully shielded ferrite cored inductors available, and even composite/molded types with reasonable performance (some with Q over 100 at some frequencies). If you're confining your search to THT, seriously consider expanding your capabilities. It's not that much effort to use SMT once you have the tools/techniques, and product availability is tremendous.

Regarding the toroid pictured: beware, they can have obnoxiously high losses. The yellow core pictured is most likely a Micrometals #26 powdered iron, little more than a resistor above 200kHz -- and quite lossy below there. Its only advantage is low cost; it's only suitable for CCM (continuous current mode, read: low ripple fraction, typically under 10%) converters below maybe 50kHz. Unfortunately, datasheets rarely tell you what material they actually used -- and supplier photos are spotty.

I do know, offhand, Bourns' 2200LL series for example, uses Kool-Mu (black; by Magnetics, Inc.), which is a reasonable powder material. Still most suitable in CCM, but to higher ripple fraction (say 30%?), and at higher frequencies (100kHz, may be a bit more?).

Generally you want ferrite for lower losses (required for high ripple fraction i.e. DCM, such as peak current mode converters), and generally that means cut/molded core shapes with an air gap.

\$\endgroup\$
3
  • \$\begingroup\$ I did my calculations, this converter can boost 7V to 100V with more than 1.2A output, so input current can go 15A. The converter is around 120-150W. I an increase the frequency to take current ripple down but according to my calculations it is only 0.8A in the worst case. I searched for phase interleaved converter in the internet, they look cool but this is my very first boost converter I'll ever make so don't want to confuse yet. I could only find choke inductors at that current rating, my supplier is not that professional. \$\endgroup\$ Nov 11, 2022 at 12:49
  • \$\begingroup\$ I am using a torroid, a choke inductor (cdn03.ciceksepeti.com/cicek/kcm69037033-1/L/…) to make a pi filter at the output (it's too lossy for boost converter L). The pi filter consists of 100uF cap - choke with unknown inductance (recycled item) - 10uF polyester respectively. The 100uF is the cap of boost converter and feedback loop is taken from that capacitors voltage. I was okay for unknow inductance because the cutoff frequency is way below switching frequency even the choke has small L. \$\endgroup\$ Nov 11, 2022 at 12:57
  • \$\begingroup\$ Ok, filter is a good place to use those, yes. The losses reduce filtering a bit, but also help dampen it so you have a reduced transient response without having to add R+C snubbers or whatever. \$\endgroup\$ Nov 11, 2022 at 16:09
1
\$\begingroup\$

Yes, you can.

The small bobbin inductor doesn't appear to be really capable of 4 A. In addition, of you place inductors close together, there may be some magnetic coupling which means the effective inductance will be different from what you calculate (+ or -, depending on the polarity of coupling).

\$\endgroup\$
1
  • \$\begingroup\$ Also wondering how the connection will perform, because of a multilayer or longer return path. \$\endgroup\$
    – RemyHx
    Nov 11, 2022 at 4:32

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.