Author Topic: Switching boost converters vs linear...  (Read 6567 times)

0 Members and 1 Guest are viewing this topic.

Offline AtheusTopic starter

  • Contributor
  • Posts: 47
  • Country: gb
Switching boost converters vs linear...
« on: March 24, 2017, 05:57:21 am »
Hi,

I'm planning on turning an ATX PSU into a bench power supply (as you do) but I don't want to be limited to 12V or any standard voltages. So. My idea is to get all the 12V lines together, hook them up to some kind of voltage converter/regulator, and then to the outputs and a pot on the front.

Searching for ways to do tis I obviously came across those modern switching "boost"/"buck" boards which appear to do the whole lot for you - but - the problem with that is I don't get to take care of my own heatsinking, fusing, I'd have to replace the pot from a trimmer to a front knob etc... basically I want to choose my own components for this not modify something.

The regulation part is sorted. I can easily built an LM317-type (but higher current and LDO) circuit out of my bits box and heatsink/pot/fuse it as needed. But what about the conversion???

Stupid Beginner Question: How do I turn 12V into 24V (or 30-ish) before feeding it into the regulator? Is there a simple linear way to do this, rather than paying £10 for one of those switching converters; half of which components I'd want to change? How would you go about increasing the voltage?

Cheers!

~Atheus
 

Offline ebclr

  • Super Contributor
  • ***
  • Posts: 2328
  • Country: 00
Re: Switching boost converters vs linear...
« Reply #1 on: March 24, 2017, 06:34:16 am »
A typical ATX power supply have a lot of rails

Only with wires you can a get of lot of voltages higher than 12 V

 

Offline AtheusTopic starter

  • Contributor
  • Posts: 47
  • Country: gb
Re: Switching boost converters vs linear...
« Reply #2 on: March 24, 2017, 06:49:48 am »
Damn :/

Thanks though! I'm not gonna be producing a PCB for this so I think it'll have to be the bought switching DC-DC boost controller. The problem is, if I use ONLY this device (with mods... possibly... there does not seem much room for a heatsink on these things and they are only rated up to 2-3A usually.

Another question: if I convert the power with the DC-DC converter to say 30v, and then run that in series with the linear voltage regulator (to add my own pot etc), would most of the heat be dissipated by my linear circuit (which I can heatsink) and relieve the DC-DC module? Or would I be creating needless extra heat by adding more resistance and voltage drop over the regulator? My assumption is the latter...

I guess this leaves me with 2 choices. Buy a boost converter which I can modify, or build one myself... considering both are difficult (for me! I write code this is just a hobby), and I don't know how I would get a decent heatsink on that converter chip with those little boards being so tightly laid out... what would everyone recommend?

Maybe I'm missing something and I can offload some of the heat onto MOSFETS inserted somewhere in the regulator circuit but I don't see how...

... Re: ebclr: Thank you, but that schematic doesn't help much! Are you saying if I have 2 separate 12V rails I can connect them for 24V? Don't they always have the same ground? Or are you suggesting I pull a higher DC voltage put of somewhere in the PSU? This sounds highly dodgy, but hey, what do I know.

Cheers!

~Atheus


 


Offline joseph nicholas

  • Frequent Contributor
  • **
  • Posts: 408
  • Country: mx
Re: Switching boost converters vs linear...
« Reply #4 on: March 24, 2017, 10:38:31 am »
Go look at Alfred on youtube.  He shows how to step up the output on an ATX power supply using a simple inverter (I think).  The video is not in English, but its got some English subs, like "yes" and "no".  PS There are no danger warnings, very refreshing.
 

Offline ebclr

  • Super Contributor
  • ***
  • Posts: 2328
  • Country: 00
Re: Switching boost converters vs linear...
« Reply #5 on: March 24, 2017, 11:35:35 am »
Switching converters have high efficiency and very low heat, the transistor is on or off only in a switching power supply, and heating will be generated for a very low  collector-emitter voltage + transition time energy, On the other side, Linear regulator will take very high voltage on the VCE, Power = V. I , lower V means lower heat.
 

Offline b_force

  • Super Contributor
  • ***
  • Posts: 1381
  • Country: 00
    • One World Concepts
Re: Switching boost converters vs linear...
« Reply #6 on: March 24, 2017, 02:15:34 pm »
If it's just for testing purposes, you can also use a variac (variable autotransformer) in front of your linear circuit (with a 1 to 1 isolation transformer if the variac isn't isolated)
That way the dissipation also will be less + it can be nice to have variable AC mains voltage.

Offline james_s

  • Super Contributor
  • ***
  • Posts: 21611
  • Country: us
Re: Switching boost converters vs linear...
« Reply #7 on: March 24, 2017, 09:16:43 pm »
There is no such thing as a linear boost converter, to boost voltage by nature requires switching to produce AC. Linear can only drop voltage across a resistor or active element.

You can often wire power supplies in series though, or you can find a higher voltage power supply. I have a little bench supply I built a while back where I used a 32VDC power brick from one of those printer/scanner gadgets and then a $5 buck regulator module and $3 LED panel meter.
 

Offline AtheusTopic starter

  • Contributor
  • Posts: 47
  • Country: gb
Re: Switching boost converters vs linear...
« Reply #8 on: March 25, 2017, 02:27:21 am »
Thanks all. I've got my answer. There is no such thing as a linear boost converter. Cheers james_s. Switching boost converter is the way to go. Can anyone recommend a specific one (preferably available in the UK) which has good thermal design and/or headers/holes for the pot rather than a trimmer built in? Compact is also good...

Re: ebclr: "What are you goals ?"

Well, in short I'm a programmer by trade, getting back into electronics in preparation for a move into the so called IoT market. I'm trying not to break the bank on the lab. I just picked up a fully working (I think! Testing...) PM3055 scope off Ebay for £50 for example. My goal with the PSU specifically is to a) get something for nothing or as little as possible, b) make use of a good quality 500W PSU I happen to have and c) just to learn about power supplies and to do a hardware hack for the sake of it.

Re: "Why are doing this, instead of buy a cheap Dc power supply ? Here one as a reference"

Well apart from for fun and trying to learn something.. that's basically just a wall wart isn't it? I'm getting by on those right now, but I'd like a bench supply with a nice 10-turn pot rather than a DIP switch or several warts or whatever. I'd also like to use the high power (and smooth) DC provided by this PSU with all the internal safety stuff already there for the protection of my reckless self while I learn. Even if I put extra circuitry in there, as long as it's after the DC output line, anything I plug in should be protected by the PSU to some extent. Thermal shutdown and fan control etc etc will remain intact. I think it's worth it. Besides - I already own pretty much every component I need for this except the DC-DC converter, and even without that I could make a 3.3, 5, 7, 12, and variable 1-10V supply, literally for nothing.

Not that money is a big problem - I'm buying quality where I need to. It's just we're trying to save for a deposit on a house and my girlfriend will kill me if I casually spend thousands on an EE lab right now :D

~Atheus
 

Offline anishkgt

  • Frequent Contributor
  • **
  • Posts: 769
  • Country: qa
    • George Hobby
Re: Switching boost converters vs linear...
« Reply #9 on: March 25, 2017, 03:28:23 am »
Well basic power supplies this is a good one for beginners if your not very much into logic devices. https://www.ebay.com/itm/291966048354 and you'll need this use for its max potential https://www.ebay.com/itm/182174554806

A cheaper version would be https://www.ebay.com/itm/201843032157


Sent from my iPhone using Tapatalk
 

Offline ebclr

  • Super Contributor
  • ***
  • Posts: 2328
  • Country: 00
Re: Switching boost converters vs linear...
« Reply #10 on: March 25, 2017, 05:49:39 am »
If you want only power up your iot circuits the easy, cheap and convenient way is to buy, instead of make

Buying you will have a nice case, voltage and current limitation, convenient connections, Don't reinvent the wheel this is available for near nothing

https://pt.aliexpress.com/item/0-30V-5A-Mini-DC-Power-Supply-Practical-Switching-Power-Supply-LED-Display-Digits-Variable-Adjustable/32653133769.html?spm=2114.02010208.3.19.MZKhQT&ws_ab_test=searchweb0_0,searchweb201602_2_10065_10068_433_434_10136_10137_10138_10060_10062_10141_10056_10055_10054_301_10059_124_10531_10099_10530_10103_10102_10096_10052_10144_10053_10050_10107_10142_10051_10106_10143_10526_10529_10084_10083_10080_10082_10081_10110_10111_10112_10113_10114_10037_10078_10079_10077_10073_10070_10122_10123_10124-10531,searchweb201603_3,afswitch_1,ppcSwitch_5,single_sort_0_default&btsid=72f92684-6b2b-45d5-b83e-8bcd1840005d&algo_expid=6b51b7a9-8a25-4484-a52f-e5110e4c52d7-5&algo_pvid=6b51b7a9-8a25-4484-a52f-e5110e4c52d7

https://pt.aliexpress.com/item/Upgrade-DPS-305BM-keypad-Digital-Programmable-Adjustable-DC-Power-Supply-30V-5A-0-1V-0-001A/32697817791.html?spm=2114.02010208.3.2.c0CIIw&ws_ab_test=searchweb0_0,searchweb201602_2_10065_10068_433_434_10136_10137_10138_10060_10062_10141_10056_10055_10054_301_10059_124_10531_10099_10530_10103_10102_10096_10052_10144_10053_10050_10107_10142_10051_10106_10143_10526_10529_10084_10083_10080_10082_10081_10110_10111_10112_10113_10114_10037_10078_10079_10077_10073_10070_10122_10123_10124,searchweb201603_3,afswitch_1,ppcSwitch_5,single_sort_0_default&btsid=9674c158-8c2b-458c-b425-5161b433037e&algo_expid=9b16e03e-e7f9-4e5f-8f69-aba5424a2f41-0&algo_pvid=9b16e03e-e7f9-4e5f-8f69-aba5424a2f41
 

Offline Nusa

  • Super Contributor
  • ***
  • Posts: 2416
  • Country: us
Re: Switching boost converters vs linear...
« Reply #11 on: March 25, 2017, 06:09:55 am »
Ask yourself what voltages and currents you actually need for your IoT development. Try to think of actual situations where you would need more than 12V or 2A.
 

Offline TheUnnamedNewbie

  • Super Contributor
  • ***
  • Posts: 1208
  • Country: 00
  • mmwave RFIC/antenna designer
Re: Switching boost converters vs linear...
« Reply #12 on: March 25, 2017, 09:23:55 am »
I feel like building a step down supply is a typical project for beginners. What you thus can do is just use an off-the-shelf supply to boost, and then build your own discrete LDO type circuit. There is a few topologies you can play with, and it's always fun since if you are a beginner and you underestimate capacitance, you /will/ have built an oscillator. Great way to learn about compensation capacitance and such.

Linear is also, I think, a bit easier to begin with.
The best part about magic is when it stops being magic and becomes science instead

"There was no road, but the people walked on it, and the road came to be, and the people followed it, for the road took the path of least resistance"
 

Offline ebclr

  • Super Contributor
  • ***
  • Posts: 2328
  • Country: 00
Re: Switching boost converters vs linear...
« Reply #13 on: March 25, 2017, 09:30:06 am »
A good development power supply, will not only supply your circuits, will also measure voltage and current applied, and limit the applied  power, avoiding you to burn things due to mistakes
 

Offline AtheusTopic starter

  • Contributor
  • Posts: 47
  • Country: gb
Re: Switching boost converters vs linear...
« Reply #14 on: March 26, 2017, 04:48:18 am »
Re: "if your not very much into logic devices.":

On the contrary, I am very much into logic devices - I'd be using such things almost exclusively - for work anyway...

Re: "If you want only power up your iot circuits":

Well... at the end of the day this is going to be the direction I head, but I just started this stuff...

Re: "Ask yourself what voltages and currents you actually need for your IoT development. Try to think of actual situations where you would need more than 12V or 2A.":

I do know quite a bit about the voltage/power requirements needed in this field. I work in software and had a job recently at a IoT company where we were fairly closely integrated with hardware - hanging around the EE lab there inspired all this - and we were working with TINY TINY amounts of power. Our devices were meant to last 5 years and had all sorts of special kit I do not understand to attempt to achieve this. I had to be aware of anything in my code communicating with the device itself, what it would cause the device to do in hardware, and the power cost of it. That's not the point.

Guys, I am just starting to learn low level electronics - in my CompSci degree we started at logic gate level - and practicality is way down the list of priorities. I'm hacking around in a slowly clearing fog. Don't bother asking me what the point is. I literally don't need more than 9V right at this moment but a) I'd like to have it on hand when I do need it, b) I'd like to build it for the hell of it, and c) if I'm going to make an adjustable supply I'm going to need a buck converter at least (to get ~12V down to 1V out) so why not boost while I'm at it?

I don't NEED the old oscilloscope I just bought off Ebay for £50 (after a little tinkering I have a trace on the screen but the probes have no ground clips!!! Arrggg... separate thread) either but I sure as hell want it :)

Re: "I feel like building a step down supply is a typical project for beginners. What you thus can do is just use an off-the-shelf supply to boost, and then build your own discrete LDO type circuit.":

I did wire one of these up on a breadboard yesterday and it worked first time. Just a couple of resistors (I think forming a voltage divider providing some kind of reference???), a pot, and a couple of caps. Plus the regulator chip itself of course. I don't know if I accidentally built a R-C oscillator! As soon as I get my scope working I'll find out.

The thing is, the bought boost/buck converters inevitably have variable step down built into them. Not using it and hooking up my own linear circuit seems like a waste of power. Especially given the bought converter is supposed to be 98% efficient.

Re: "A good development power supply, will not only supply your circuits, will also measure voltage and current applied, and limit the applied  power, avoiding you to burn things due to mistakes":

Good point. I have a little LCD voltmeter ready to slot in the front, but had not considered current much, except in terms of fusing and 'how much can I draw from this particular wire'. How might I provide a constant and/or limited current? This might be a rabbit hole of a question but any recommendations on tutorials etc to read would be much appreciated.

FYI the converter I bought is this:

https://www.amazon.co.uk/gp/product/B0171JFSJ0/

Has 3 pots built in - presumably in series and of decreasing capacity to provide fine adjustments. I will probably replace them with a single 10-turn wire wound pot on the front of the device.
« Last Edit: March 26, 2017, 04:50:38 am by Atheus »
 

Offline Kleinstein

  • Super Contributor
  • ***
  • Posts: 14076
  • Country: de
Re: Switching boost converters vs linear...
« Reply #15 on: March 26, 2017, 08:40:38 am »
The buck / Boost converter board could work together with the ATX power supply. It might still need a fuse, just to make sure that in case of a fault nothing bad (like a fire) happens. An ATX power supply can give a lot of power - more like to much power. The better starting point for a self made bench supply would be a laptop supply or similar.

This switched mode regulator will give a reasonable power, but also have quite some residual ripple / noise. So it will not work well for sensitive circuits. Adding current readings is a little tricky as this adds a shunt resistance. However these cheap switched mode regulators are usually not that accurate anyway. For a bench supply reading the current might be the more useful part as the voltage usually is where it should be.

For just powering digital circuits one could use ready made wall warts with adjustable voltage - there are models for something like  1.5 - 3 - 4.5 - ... 12 V and 500 mA. This is usually enough power.
 

Offline janoc

  • Super Contributor
  • ***
  • Posts: 3781
  • Country: de
Re: Switching boost converters vs linear...
« Reply #16 on: March 26, 2017, 11:06:35 am »
Honestly, if you are just starting with electronics, forget about an ATX supply, boost converters and what not. For all that trouble that you would need to make that ATX supply into something semi-usable, with noisy, dirty output full of switching noise and happy to burn your house down with the 20A+ an ATX supply is able to deliver, you could spend hundred bucks and buy a proper lab supply. With current limiting, both current and voltage indications, maybe even programmable. And it will still probably cost less than the hacked up crappy ATX supply.

ATX supplies are good only for powering motors and similar high current stuff in the lab, that doesn't care about clean, low ripple output or accurate voltages. The current limiting feature every proper lab supply has will alone pay for itself in the components that would have blown up due to mistakes while bringing up new circuits.

For what you want to do you don't need more than a 15V-30V/3A supply and these are cheap as chips these days. Get a linear, not switching one, though - the last thing you need is switching noise from a poorly filtered PSU screwing up your circuit while you are still working on it.

If you are starting up the last thing you need is fighting with your tools. Building a PSU is a good project for a beginner - but not rebuilding an ATX supply. That adage is from the era when a PSU meant a transformer, rectifier bridge and maybe a linear regulator at the output.

« Last Edit: March 26, 2017, 11:12:06 am by janoc »
 
The following users thanked this post: ebclr

Offline b_force

  • Super Contributor
  • ***
  • Posts: 1381
  • Country: 00
    • One World Concepts
Re: Switching boost converters vs linear...
« Reply #17 on: March 26, 2017, 01:08:49 pm »
Although I agree with the fact that an ATX is probably asking for trouble, nowadays a lot of them actually have pretty good decent noise levels.
the early 2000' noisy crappy ones are history.

Offline TheUnnamedNewbie

  • Super Contributor
  • ***
  • Posts: 1208
  • Country: 00
  • mmwave RFIC/antenna designer
Re: Switching boost converters vs linear...
« Reply #18 on: March 26, 2017, 04:39:55 pm »
Re: "if your not very much into logic devices.":

On the contrary, I am very much into logic devices - I'd be using such things almost exclusively - for work anyway...

Re: "If you want only power up your iot circuits":

Well... at the end of the day this is going to be the direction I head, but I just started this stuff...


Re: "I feel like building a step down supply is a typical project for beginners. What you thus can do is just use an off-the-shelf supply to boost, and then build your own discrete LDO type circuit.":

I did wire one of these up on a breadboard yesterday and it worked first time. Just a couple of resistors (I think forming a voltage divider providing some kind of reference???), a pot, and a couple of caps. Plus the regulator chip itself of course. I don't know if I accidentally built a R-C oscillator! As soon as I get my scope working I'll find out.

The thing is, the bought boost/buck converters inevitably have variable step down built into them. Not using it and hooking up my own linear circuit seems like a waste of power. Especially given the bought converter is supposed to be 98% efficient.


I meant with this not using a off-the-shelf voltage regulator but build one using some opamps and such. The basic circuit is very simple, but what I like about it is how much you can learn in going from a very basic zener diode with a transistor to a decent supply can be a path with a lot of "gotchas". Ofcourse, if digital is what you want to go into, I can understand it not being a fun thing.
The best part about magic is when it stops being magic and becomes science instead

"There was no road, but the people walked on it, and the road came to be, and the people followed it, for the road took the path of least resistance"
 

Offline janoc

  • Super Contributor
  • ***
  • Posts: 3781
  • Country: de
Re: Switching boost converters vs linear...
« Reply #19 on: March 27, 2017, 09:36:24 am »
Although I agree with the fact that an ATX is probably asking for trouble, nowadays a lot of them actually have pretty good decent noise levels.
the early 2000' noisy crappy ones are history.

The one someone like the OP is going to use is most likely going to be one of the "recycled" ones, not a brand new high quality one. So likely old design, dried out capacitors, etc.

Even the brand new ones are not a guarantee, ATX spec allows up to 120mV ripple. Just look here:
http://www.tomshardware.com/reviews/seasonic-prime-titanium-650w-psu,4690-9.html

And those are brand name supplies, a cheapo street special will be most likely much worse.

 

Offline b_force

  • Super Contributor
  • ***
  • Posts: 1381
  • Country: 00
    • One World Concepts
Re: Switching boost converters vs linear...
« Reply #20 on: March 27, 2017, 02:01:29 pm »
Although I agree with the fact that an ATX is probably asking for trouble, nowadays a lot of them actually have pretty good decent noise levels.
the early 2000' noisy crappy ones are history.

The one someone like the OP is going to use is most likely going to be one of the "recycled" ones, not a brand new high quality one. So likely old design, dried out capacitors, etc.

Even the brand new ones are not a guarantee, ATX spec allows up to 120mV ripple. Just look here:
http://www.tomshardware.com/reviews/seasonic-prime-titanium-650w-psu,4690-9.html

And those are brand name supplies, a cheapo street special will be most likely much worse.
The limits can be 120mV, put you see typical <20-30mV or so on full load!
That's not great, but not bad AT ALL for general testing purposes.
There are also a lot of solutions to make these figures a lot better!
« Last Edit: March 27, 2017, 02:09:25 pm by b_force »
 

Offline janoc

  • Super Contributor
  • ***
  • Posts: 3781
  • Country: de
Re: Switching boost converters vs linear...
« Reply #21 on: March 27, 2017, 06:48:55 pm »
The limits can be 120mV, put you see typical <20-30mV or so on full load!
That's not great, but not bad AT ALL for general testing purposes.
There are also a lot of solutions to make these figures a lot better!

Or simply spend $100 and buy a cheap linear lab supply that will have ripple in the millivolts range without you having to do anything and come adjustable, having both current limiting and voltage/current meters.

A decent ATX supply is like $60-80 (good luck with crappy cheapies or an old supply recycled from some computer carcass).  And you still need the filter to clean up the ripple and still lack anything like a current limiting or adjustable outputs, which are essential features for a lab supply.

So what is the point of such exercise? To prove that it can be done? :-//
 

Offline MrAl

  • Super Contributor
  • ***
  • Posts: 1395
Re: Switching boost converters vs linear...
« Reply #22 on: March 27, 2017, 07:49:16 pm »
Hello there,

That boost converter you picked out on Amazon looks decent, if it has enough power for you.
it looks like it already has adjustments for both voltage and current, so you can adjust both.

The key to a clean output with reasonable efficiency is a switch mode power supply on the front end and a linear regulator on the back end.  Normally the switcher would be adjusted automatically when you adjust the linear by using a tracking circuit of some kind.  But you can also do this manually if you dont expect the output voltage to have to change too much with any given experiment or test.  The idea is to use an LDO and adjust the switcher output to an amount just over the required LDO output voltage (usually 1 or 2 volts) and this keeps power dissipation in the linear very low while providing the cleaner output and better regulation of a linear.  This is if you want the cleanest output that is.
 

Offline james_s

  • Super Contributor
  • ***
  • Posts: 21611
  • Country: us
Re: Switching boost converters vs linear...
« Reply #23 on: March 27, 2017, 08:20:47 pm »
I've actually had remarkably good luck with old salvaged ATX PSUs, nobody is going to buy a nice new one to use as a bench supply but I used an old one for years to power various things. Not saying it's the most sensible bench PSU out there but they actually work reasonably well and they are easy to find for little or nothing.
 

Offline b_force

  • Super Contributor
  • ***
  • Posts: 1381
  • Country: 00
    • One World Concepts
Re: Switching boost converters vs linear...
« Reply #24 on: March 27, 2017, 08:36:12 pm »
The limits can be 120mV, put you see typical <20-30mV or so on full load!
That's not great, but not bad AT ALL for general testing purposes.
There are also a lot of solutions to make these figures a lot better!

Or simply spend $100 and buy a cheap linear lab supply that will have ripple in the millivolts range without you having to do anything and come adjustable, having both current limiting and voltage/current meters.

A decent ATX supply is like $60-80 (good luck with crappy cheapies or an old supply recycled from some computer carcass).  And you still need the filter to clean up the ripple and still lack anything like a current limiting or adjustable outputs, which are essential features for a lab supply.

So what is the point of such exercise? To prove that it can be done? :-//
I agree totally.
Doesn't mean it can't be done ;)  8)


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf