5V to 48V Switch-Mode Power Supply using MC34063

I needed to build a 5V to 48V DC to DC converter. After scouring the web for sample circuits, I found one that was based on the SG2534N. Well, the circuit had problems and I spent entirely too much time trying to get it to work. I guess the SG2534N has been around quite a while, but the amount of hobbyist understandable information is very small.

But, while flailing with with the SG2534N, I found quite a few references to the MC34063 IC. After a week of little luck with the SG2534N, I ordered the MC34063 from mouser.com.

While waiting for the parts, I spent time just experimenting with SMPS in general.

Besides being less expensive ($0.50 vs. $2.00) I was able to get this circuit up and running without too much trouble.


First, here are some references:

Determining the Overall Circuit Layout

The way the passive components are connected to the MC34063 depend on what type of power supply you are building: boost, buck,  boost-buck, or inverting. I am building a boost circuit, so the overall circuit layout will look like this:


There are a few components that have static values:

  • Cin – This capacitor is next to Vin and should be 100uF.
  • R – This resistor should be 180 ohms.
  • D – This is a schottky diode of sufficient power rating. I used an MBR1100RLG.

I need to calculate L, Rsc, Cr, C0, R1, and R2. This is where the calculator comes in. To use the calculator, you are going to need to know

  • Vin – this is the minimum input voltage. 5V for my calculations.
  • Vout – this is the required output voltage. 48V.
  • Iout – this is the output current. I want 25mA (note that the input current will be much higher than the output current when stepping up this far – in my final tests, it takes 220mA @5V to generate 25mA @ 48V).
  • Vripple – the voltage ripple in the output. I’ve been using 1mV for this.
  • Fmin – the minimum frequency. The chip runs at a max of 100MHz, so that is what I use.

Using these requirements, according to the calculator, the passive values should be:

Ct=366 pF
Ipk=593 mA
Rsc=0.506 Ohm
Lmin=62 uH
Co=2060 uF
R=180 Ohm
R1=1.5k R2=56k (47.92V)

The value for Co is really big. When I do the calculations by hand, I get 1/100th of that. I don’t know if that is exactly how far it is off by, but if you just use the calculator, I’d question the Co value if it is large.

Using the worksheet found in the TI datasheet, here are my calculations. Note that I take the Vf and Vsat values from the “How to Build a Switch Mode Power Supply Circuit with an MC34063” tutorial.


This gives me:

  • Ct – 362pf. Closest I have is 330pf.
  • Ipk – the peak input current should be 527mA. So my input power supply and inductor need to be able to support that much current.
  • Rsc is .569 ohm. I have a .5 ohm, so I’ll use that (reducing resistance increases current).
  • L(min) is 72uH. The next size inductor up I have is 100uH.
  • Co is 20uF. The only 100V capacitors I happen to have are 47uF, so that is what I will use (I believe C0 is a minimum value).

I’m just missing R1 and R2. I arbitrarily decide to use a 100K ohm resistor for R2 so I just need to do some algebra to determine what R1 must be:


So if I set R2 to 100K, R1 needs to be 2674 Ohms, or 2.7K which I have.

This worked amazingly well. I got about 48.5V and could draw 23mA. Here is my final circuit:


Jan 22 2016 update:

I found this post the MC34063 on EEVBlog. Wish I had found this before. It explains everything, better than I can.

Oct 2016 Update:

One ‘tip’ I discovered later but forgot to document here: keep all of your wires as short as possible.

This entry was posted in c-electronics and tagged . Bookmark the permalink.

9 Responses to 5V to 48V Switch-Mode Power Supply using MC34063

  1. Ahmad says:

    Hi Dan
    Thanks for sharing.
    MC34063 has 40V maximum rating. There is many newer products for Boost converters.
    Such as TPS family of TI. We use TPS40210 and TPS61040.

    • Dan TheMan says:

      Thanks for the comment. When I looked at the datasheet, I interpreted it as a maximum of 40V in. Now that I read it again, what you say makes sense and the internal transistors support only 40V.

      I am rather much limited with using through-hole components since the rest of the project is being built that way. Looking at the TPS40210 and TPS61040, neither have a through-hole variant.

      Can you recommend anything that is through-hole?

      • Ahmad says:

        Finding through-hole packages in nowadays is very hard. But if you are rich-man🙂 my selection is:
        LT1172CN8 in P-DIP8 package for output voltage up to 65V
        LT1172CT in 5-LEAD TO220 package for output voltage up to 65V
        LT1172HVCT in 5-LEAD TO220 package for output voltage up to 75V

  2. Sean Straw says:

    I find that for prototyping on a breadboard or perfboard, it is handy to have a selection of SMD-to-TH adapter boards on hand for different footprints. It’d be possible to use one in a final design if necessary, but even if the final design were to be SMD, having the exact same IC to use when breadboarding is good – it’s the same spec, you can order in quantity, etc.

    If you find some useful part and order a quantity, wouldn’t you like to expect that it’ll still be useful for designs in a year or two when you shift to SMD?

    While I don’t personally use either of the TI boost converters Ahmad suggests, I very much like the TI TPS62160, which is an awesome SMPS buck converter (opposite of what you need for your application here), capable of 1A output and being driven to Vin=Vout (unlike even an LDO linear regulator which will need a couple tenths of a volt). So, if the TPS boost series has some of the same DNA as their bucks, they should be great parts too.

    All that said, MC34063 is much less expensive than the TI TPS series. The TPS61040 has a max output of 28V, which means it isn’t applicable to your specific requirements (much as the MC34063 isn’t ). If the MC34063 were suitable for the voltage and current you require, an added advantage of it is that the one part is capable of being utilized in buck, boost, or inverting circuits. I’ve bought them for as little as $0.044 cents apiece via AliExpress, and the ST version can be had for $0.22 domestically in single unit quantities.

    • Dan TheMan says:

      I was VERY disappointed when I realized I hadn’t read the MC34063 datasheet properly (lesson painfully learned) and I shouldn’t be using it for 48V. Because I’m (at least for now) limited to PTH, I looked at all my options and they weren’t good. Even the SG3425 that I had initially tried to use is not acceptable either and, if I recall, everything else I looked at was ridiculously expensive if I ever hope to do something commercial with my project.

      While stewing in disappointment, it occurred to me that I probably didn’t really need 48V and 40V may be just fine. I’m monitoring a telephone handset to detect if it goes offhook. As soon as it does, I simply connect it to the telco line via SSRs. I don’t want the phone to ring until I say it can ring, but I still need to know if the user picks it up to place an outbound call.

      After some testing 40V works just fine even if it is a bit below the ‘minimum spec’ of 48V. In fact, for every phone I can find to test, they will work down to 24V, but I’d rather stay as close to 48 as possible just in case there are some handsets out there that are really picky.

  3. Sean Straw says:

    Have you considered masking the ring signal? US POTS ring signal is something like 75-90V AC at 20Hz (on top of the DC line). Because the ring signal is AC superimposed on the DC, masking the ring should be a matter of introducing a diode bridge between the phone line and the handset. This should serve to disable the ring signal for the connected handset without disabling the line for making calls or picking one up because you can hear a ring elsewhere (or if you have ESP). A single pole double-throw switch (or an equivalent relay device) would serve to switch one phone lead between the bridge or directly to the line to re-enable the ring. or omit the switch and just have the ringer permanently disabled at that jack. No 5v supply or boost necessary, and it should have no impact on other phones not clients to the bridge setup. Your phone is also never actually disconnected from the phone line for making calls.

    I’ll admit I haven’t rushed out and tinkered with this myself. It is applicable to POTS and SHOULD work for VoIP which locally mimics POTS for use with old handsets.

    Also, the DC voltage should be *NEGATIVE* 48V. That’s all a matter of where you reference the ground from your boost.

    Ring disable may or may not work with a half-wave rectification.

    • Dan TheMan says:

      The rectifier sounds like a neat trick, but I’m doing more than just blocking the ring. My box interrogates caller ID while the handset is disconnected. Once my box determines the call can pass, it regenerates caller id for the handset, then connects it back to the telco so I can hear the phone and still see who is calling.

      I am trying my best to keep the telco side extremely simple to decrease my chances of problems getting FCC certification if I ever manage to get that far. The telco side has only 2 SSRs transferring control of the handset between myself and the telco and a device already certified to provide CID.

      I spent a ridiculous amount of time trying to get an inverting boost power supply working to provide -48V. For reasons I still don’t understand, I constantly had heating issues. A straight boost power supply was, comparatively, a piece of cake. Then out of the blue it occurred to me to just switch the polarity for the handset. As far as it is concerned, it is getting -48 (well, -40V anyway).

      One positive side-effect of using 40V rather than 48V that I have noticed is that when I improperly connect the telco line into the handset jack (I’m testing using an old PBX, not a POTS line) the MC63035 sees the voltage is now > 40 and cuts off. I’ve left it in this state for up to 10 minutes at a time and found no ill effects.

  4. Sean Straw says:

    So, where is the 5V coming from? Does the device have to run off of batteries or a wall-wart? What happens if the battery runs down or there’s a mains power outage – does the circuit fail to a functioning (albeit, ring-able) phone, or is the hot Co-Ed in Halloween XXIII cursing that they have no dialtone?

  5. Dan TheMan says:

    5V comes from a wall wart. That drives a Teensy MCU which then provides 3.3V to everything else except the 5V to 40V boost circuitry. The SSRs are connected such that if there is no signal from the Teensy, the handset is connected to the telco (the N.O. side).

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s