Author Topic: Advice on multiple MC34063 circuits  (Read 10167 times)

0 Members and 1 Guest are viewing this topic.

Offline Electron_1Topic starter

  • Contributor
  • Posts: 18
Advice on multiple MC34063 circuits
« on: January 19, 2015, 04:51:53 pm »
Hi all,
This is my first time posting on EEVBlog so I hope I am posting in the right forum.
Just wondering if I could get advice on a circuit. I am planning on building a Dc-Dc convertor  for my CNC machine control box using the MC34063 (based on Dave's video blog #110)
I have 48v running to my motor controllers. My plan is to take that 48v and bring it to about 39v (maximum 40v input for the MC34063). I then hope to build three independent circuits on one pcb board, giving me three outputs: 24v, 12v and 9v
Basically my question is, are there any pitfalls I should look out for when placing the 3 circuits on the one board eg. Having the inductors too close to each other etc. or perhaps there is a reason why I shouldn't place them on one board at all ?
Any and all advice is appreciated.
Many Thanks
 

Online mariush

  • Super Contributor
  • ***
  • Posts: 4983
  • Country: ro
  • .
Re: Advice on multiple MC34063 circuits
« Reply #1 on: January 19, 2015, 05:44:29 pm »
How much current do you need to have on 24, 12 and 9v ?

DC-DC converters based on MC34063 will only be able to give you about 0.75A.

In addition, the chip is inefficient at about 60-75%, so if you only need.. let's say 50-100 mA on 9v and you won't need the maximum current on 12v, it would be cheaper and about the same to just use a linear regulator to get 9v from 12v.

There are other chips that can handle up to 60-75v and will also be more efficient. Lots of them are available in surface mount and fewer in through-hole packages but just on a quick search here's some suggestions :

http://uk.farnell.com/texas-instruments/lm2591hvt-adj/switching-reg-1a-adj-2591-to2205/dp/9489541
http://uk.farnell.com/texas-instruments/tl2575hv-adjikv/ic-step-dn-sw-reg-adj-1a-to220/dp/1755275
http://uk.farnell.com/texas-instruments/lm2594hvn-adj-nopb/ic-step-down-regulator-8-dip/dp/1693988  (only about 0.5A maximum output current)
http://uk.farnell.com/linear-technology/lt1072cn8-pbf/ic-switching-reg-1-25a-1072-dip8/dp/9559574

Check the datasheets, you'll find some example circuits inside and recommendations on how to calculate the parameters for components you need. The chips may be a bit more expensive but the other components you need (inductor, diodes, resistors) may be a bit cheaper than the ones you'd use with mc34063.
 

Online ajb

  • Super Contributor
  • ***
  • Posts: 2582
  • Country: us
Re: Advice on multiple MC34063 circuits
« Reply #2 on: January 19, 2015, 06:21:21 pm »
No reason you can't put as many switching converters as you want on one PCB.  The most important part is that each converter is laid out properly.  The datasheets of many parts will have an example layout to help you out, but the bottom line is that you want to keep a continuous groundplane under the IC and keep the critical paths (between input/output caps and the switching node, primarily) as short and wide as possible--polygon pours are your friend here. 

If you've never done a switching converter before, you might start with TI's Webench tool.  It's a bit cheating, but given your current and voltage requirements it'll spit out a whole selection of suitable parts with schematics and BOMs, and so eliminates much of the tedium of selecting a solution.  For a more in depth understanding of what's going on and why various component choices matter, download LTSpice and play around with some designs and see what happens when you change component values.
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 21606
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Re: Advice on multiple MC34063 circuits
« Reply #3 on: January 19, 2015, 06:35:28 pm »
Best suggestion: don't!  MC34063 is old, crappy, and best left to history.  There are better, more efficient, and easier to use parts out there.

For such a high input voltage, I would suggest browsing TI's offerings of switching converters; they'll have some that are geared towards higher voltage (automotive and telecoms) applications, that will only need a few support components to use.

Tim
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline Electron_1Topic starter

  • Contributor
  • Posts: 18
Re: Advice on multiple MC34063 circuits
« Reply #4 on: January 19, 2015, 06:51:36 pm »
Many thanks for the answers guys. You've given me a lot to think about.
I did suspect that there might be better ways to go than the MC34063 but I don't have enough experience, so often get bogged down in the multiple options out there until I don't know how to proceed.
As far as current draw on each output, it's quite small (between 200ma and 300ma) with the exception of the 24v output. It's purpose is to provide power to proximity switches. There are five of them and each one pulls about 300ma but the chances of all switches being activated at the same time are almost zero.
I'll start checking out alternative chips and see if I can come up with a better suited circuit.
Thanks again.
 
 

Online ajb

  • Super Contributor
  • ***
  • Posts: 2582
  • Country: us
Re: Advice on multiple MC34063 circuits
« Reply #5 on: January 19, 2015, 07:06:24 pm »
It's purpose is to provide power to proximity switches. There are five of them and each one pulls about 300ma but the chances of all switches being activated at the same time are almost zero.
???

Each switch draws 7.2W, but only when activated?  What kind of switches are these?  Even an active proximity sensor should draw only a fraction of that.  Are you perhaps looking at the load current rating for the switch (which could certainly be on the order of 300mA)?  If it's driving an input on a controller or drive, load current should be almost nothing.
« Last Edit: January 19, 2015, 07:08:09 pm by ajb »
 

Offline Electron_1Topic starter

  • Contributor
  • Posts: 18
Re: Advice on multiple MC34063 circuits
« Reply #6 on: January 19, 2015, 08:11:04 pm »
Each switch draws 7.2W, but only when activated?  What kind of switches are these?  Even an active proximity sensor should draw only a fraction of that.  Are you perhaps looking at the load current rating for the switch (which could certainly be on the order of 300mA)?  If it's driving an input on a controller or drive, load current should be almost nothing.
Hi. My understanding is that 300mA is the maximum load current of these switches. I would expect no load current to be in the region of 10mA or so. Was basing my design on the switches maximum load current.
 

Offline dannyf

  • Super Contributor
  • ***
  • Posts: 8221
  • Country: 00
Re: Advice on multiple MC34063 circuits
« Reply #7 on: January 19, 2015, 08:48:05 pm »
Quote
There are five of them and each one pulls about 300ma but the chances of all switches being activated at the same time are almost zero.

It is probably not prudent to design a system that will behave unexpectedly when all of such switches activate at the same time. So I would think you want to design for a draw of 5x300ma.

If I were you, I would use a standalone 24v smps - they are tones of them and they are quite inexpensive.

From there, I would use a 12v or 9v regulator, either linear (for simplicity) or smps (34063 or otherwise, for efficiency).
================================
https://dannyelectronics.wordpress.com/
 

Offline Electron_1Topic starter

  • Contributor
  • Posts: 18
Re: Advice on multiple MC34063 circuits
« Reply #8 on: January 19, 2015, 09:01:47 pm »
It is probably not prudent to design a system that will behave unexpectedly when all of such switches activate at the same time. So I would think you want to design for a draw of 5x300ma.
Thanks. I had initially based my design on this as I thought the MC34063 could deliver 1.5amps. I don't think I was correct in that though.

Quote
If I were you, I would use a standalone 24v smps - they are tones of them and they are quite inexpensive.
My New Years resolution is to learn more ! So whichever route I take I want to build it from scratch to further my own understanding.

Quote
From there, I would use a 12v or 9v regulator, either linear (for simplicity) or smps (34063 or otherwise, for efficiency).
That sounds like good advice thank you. I'll certainly take it into account
« Last Edit: January 19, 2015, 09:05:23 pm by Electron_1 »
 

Offline dannyf

  • Super Contributor
  • ***
  • Posts: 8221
  • Country: 00
Re: Advice on multiple MC34063 circuits
« Reply #9 on: January 19, 2015, 09:13:33 pm »
Quote
I don't think I was correct in that though.

It can be made to deliver a lot more than that. It is just that if you are starting new, it is a lot easier to go with a ready-made solution.
================================
https://dannyelectronics.wordpress.com/
 

Online mariush

  • Super Contributor
  • ***
  • Posts: 4983
  • Country: ro
  • .
Re: Advice on multiple MC34063 circuits
« Reply #10 on: January 19, 2015, 09:37:36 pm »
Linear regulators work by dissipating the difference between input and output as heat, so you can't really use them like a "magic pill".

For example, most can only dissipate up to about 15 watts, with a very good heatsink... in practice, with normal regular heatsink you'd want to NOT go over about 5-7 watts.

So, let's say you want 9v and 1A of current and you want to use a 7809 linear regulator to achieve that ... this means the chip will produce  (12v input - 9v output )  x 1A current = 3 watts as heat. This is reasonable and a normal heatsink and some air around the regulator will be enough.
You can't however have a 24v power supply and use a 7812 linear regulator to get 12v at 1A ... the chip would dissipate 12 watts, which is kind of hard to move away from the chip even with a large heatsink, a fan would probably be required.

Another thing you should understand. 
With linear regulators,  what current goes in, that much current goes out ... so if you want 9v and 1A using a 7809 linear regulator, whatever gives power to 7809 must be able to give 1A of current.
With switching regulators, it's not the same..  the current at the input is smaller than the current at the output, in the case of step-down regulators.  As an example, if you use a switching regulator chip to produce 9v at 1A from a 12v power supply, that means  the circuit consumes  about  9v x 1A = 9w plus about 1w due to circuit being about 80-85% efficient only so in total about 10w.  With a 12v power supply, this means the current used from 12v is about 10w / 12v = 0.84 A  ...  With 24v input it's about half that.
 

Offline digsys

  • Supporter
  • ****
  • Posts: 2209
  • Country: au
    • DIGSYS
Re: Advice on multiple MC34063 circuits
« Reply #11 on: January 19, 2015, 10:15:58 pm »
Unless you REALLY want to learn about DC-DC conversion, I'd just get a module off the shelf -
eg  au.element14.com/dc-dc-converters-isolated-board-mount-single-output   or the non-isolated versions. Only cost from $4 upwards,
some with 92++% efficiencies, put as many as you want next to each other, just add 1,000uF low ESR cap on the supply as an example.
No other parts required. AND EMI is taken care of ! 2-5W, IF you need it, cost a bit more ~$10-15, but you don't have to wind inductors.
Hello <tap> <tap> .. is this thing on?
 

Offline rob77

  • Super Contributor
  • ***
  • Posts: 2085
  • Country: sk
Re: Advice on multiple MC34063 circuits
« Reply #12 on: January 19, 2015, 10:32:13 pm »
As an example, if you use a switching regulator chip to produce 9v at 1A from a 12v power supply, that means  the circuit consumes  about  9v x 1A = 9w plus about 1w due to circuit being about 80-85% efficient only so in total about 10w.  With a 12v power supply, this means the current used from 12v is about 10w / 12v = 0.84 A  ...  With 24v input it's about half that.

sorry but 10W input and 9W output is 90% efficiency ;) with 9W output and 80% efficiency as you mentioning above you would need 11,25W input - and that's almost the same as the 12W with  linear regulator ;) so in cases where the voltage difference is low (12V->9V, 5V->3.3V etc...) the linear regulator would be the better choice (no switching noise and less costy).
 

Offline Electron_1Topic starter

  • Contributor
  • Posts: 18
Re: Advice on multiple MC34063 circuits
« Reply #13 on: January 23, 2015, 11:51:40 pm »
Hi again all.
Update on this. I spent this evening drawing up a rough design. I'm going to use an LM2576HVT for both the 24v and 12v outputs and then an MC7809 linear regulator to get my 9v output from my 12v.
Anytime I've used linear regulators in the past I've included 1N400* diodes for polarity protection and would like to do so this time but I'm unsure about correct placement with the LM2576.
I usually place one between Vin and Vout on the regulator and one between Adj and Vout.
Could somebody advise on the safest way to do this with the LM2576 ?
Many Thanks !     
 

Online mariush

  • Super Contributor
  • ***
  • Posts: 4983
  • Country: ro
  • .
Re: Advice on multiple MC34063 circuits
« Reply #14 on: January 24, 2015, 01:18:20 am »
A single diode in series with each regulator would do enough protection, but would also waste a lot o power and produce heat.

My advice would be to do better and use a p-channel mosfet for reverse voltage protection, it's really not that much complicated:



Make sure you're sizing the inductors and schottky diodes correctly in those LM2576HVT regulators. Pay attention to currents etc.
 

Offline Electron_1Topic starter

  • Contributor
  • Posts: 18
Re: Advice on multiple MC34063 circuits
« Reply #15 on: January 25, 2015, 07:34:27 pm »
Thanks for that mariush. Very informative. I'll definitely be using the p-channel mosfet.
Have spent quite a bit of time matching the inductors and schottky diodes so hopefully I'm good to go. Have ordered the parts so will find out during the week when I build the circuit.
 

Offline Electron_1Topic starter

  • Contributor
  • Posts: 18
Re: Advice on multiple MC34063 circuits
« Reply #16 on: March 05, 2015, 10:47:44 pm »
Hi guys,
I've had very little time since starting this project and have only gotten around to drawing up a schematic now. I was wondering if somebody might take a look and let me know if there is anything glaringly wrong with it before I commit to laying it out and milling the board.
I'm not very experienced so any help is much appreciated.

Thanks
« Last Edit: March 05, 2015, 10:50:38 pm by Electron_1 »
 

Offline MrAl

  • Super Contributor
  • ***
  • Posts: 1395
Re: Advice on multiple MC34063 circuits
« Reply #17 on: March 06, 2015, 05:29:16 pm »
Hi,

Couple quick notes:

1.  There is a somewhat better chip, the LM2596, which has a higher switching frequency.
2.  Is that a 47k potentiometer in the feedback path?  Max resistance to ground for the feedback network is given in the data sheet for either chip.  I think it is 4.7k if i remember right.  You will have to lower the impedance of that network or you might get too much noise in the feedback signal.
3.  What current do you need on the output?  If you are going for 2 amps max (with no overcurrent expected) then the Schottky diode shown might be ok, but if you are going for the full 3 amps output you should use a 4 amp or better Schottky.  A 5 amp device would be much better.

 

Offline Electron_1Topic starter

  • Contributor
  • Posts: 18
Re: Advice on multiple MC34063 circuits
« Reply #18 on: March 06, 2015, 05:51:35 pm »
Hi.
Thanks for taking a look for me.
That's good to know about the LM2596. It can be time consuming for me to get parts where I live so I'll probably press ahead with the LM2576HVT for the moment as I have a few in front of me, but I'll certainly remember it for the future.
That is a 47K potentiometer. I'll go and investigate that again. This would be where my inexperience shows. Is there a recommended way to calculate the best values of the components for this part of the circuit ?
In regard to current, the absolute max that should ever be pulled is 1.5 amps and to be honest something would have to have gone catastrophically wrong to lead to that scenario. Otherwise about 300mA would be the norm. 
Much appreciate the feedback !
 

Offline MrAl

  • Super Contributor
  • ***
  • Posts: 1395
Re: Advice on multiple MC34063 circuits
« Reply #19 on: March 06, 2015, 09:55:52 pm »
Hi.
Thanks for taking a look for me.
That's good to know about the LM2596. It can be time consuming for me to get parts where I live so I'll probably press ahead with the LM2576HVT for the moment as I have a few in front of me, but I'll certainly remember it for the future.
That is a 47K potentiometer. I'll go and investigate that again. This would be where my inexperience shows. Is there a recommended way to calculate the best values of the components for this part of the circuit ?
In regard to current, the absolute max that should ever be pulled is 1.5 amps and to be honest something would have to have gone catastrophically wrong to lead to that scenario. Otherwise about 300mA would be the norm. 
Much appreciate the feedback !

Hi,

I checked the data sheet and they state 5k max for that lower resistor, which we'll call R1.  The upper resistor is then called R2.  These two form the voltage divider for the feedback, with no pot.
The equation is then:
R2=R1*(Vout/Vref-1)

where Vref is taken to be 1.23 volts.

What would be good is say a 2k resistor for R1, and then calculate R2 as needed for whatever Vout you need.
If you really need to adjust, put a trim pot in series with R1 and that pot becomes R2, or else use R2 lower than needed and a small value pot in series to 'trim' the output to the desired level.  The smaller the pot value relative to R2 the better the temperature characteristics will be, so if the trim pot is say 10 percent of R2 that should be good.
Have the wiring in the feedback circuit as close to the chip as possible, and make sure it comes from the output cap just after the inductor, not from the very output after some secondary filtering.

The LM2576 ADJ is a pretty nice chip...stable and reliable, and fairly easy to use.

« Last Edit: March 06, 2015, 09:58:08 pm by MrAl »
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf