• Welcome to our site! Electro Tech is an online community (with over 170,000 members) who enjoy talking about and building electronic circuits, projects and gadgets. To participate you need to register. Registration is free. Click here to register now.

DC to DC Conversion

Status
Not open for further replies.

Scarr

Member
Hi All,

I have a 28vDC supply and want a 5vDC 10mA + 3.3vDC 100mA deriving from it. I have used tried a few LDO regulators just wired up by hand to test the outputs and I have tried two configurations.

First configuration was with the 28vDC feeding both regulators, both regulators can handle 30v in, but for some reason the 3.3v regulator popped and and blew up the HC-O5 I had attached to it, this happened twice after approx 30-40 seconds!! I did it a second time because I thought it might have been a faulty regulator. When I did not have the HC-05 connected the output showed 3.3v without issue.

The second configuration has the 5v regulator output feeding the 3.3v regulator input, but in this configuration the 5v regulator got hot (didn't blow just got hot to the touch) .

So my questions if what is the best way to get 5vDC 10mA + 3.3vDC 100mA from 28vDC supply. both the 5v and 3.3v supply IC's

Thanks

Steve


P.S. I am thinking about using two Zener + resistors but am not sure this is best approach, also if any part of my email does not sense don't flame just ask me to clarify.
 

MikeMl

Well-Known Member
Most Helpful Member
You are not putting the regulators on a proper heat sink, are you?

If you cascade the regulators, the 5V one is dissipating P=I*V =0.11* (28-5) = 2.53W which requires it to be bolted onto a large heatsink...

The 3.3V regulator will be dissipating 0.1* (5-3.3) = 170mW, which depending on the package, and how it is mounted, probably doesn't require a heatsink.

Post a link to the regulators that you actually used.
 
Last edited:

MikeMl

Well-Known Member
Most Helpful Member
Use switch-mode regulators,...
and then pray that the 100mV of 150kHz noise crap that the switcher puts out doesn't screw up the low-level analog sensor/opamp you were planning to run from the 5V power supply...

The alternative is to put the switcher in a shielded box with lots of filtering on both the power input and power output leads. Maybe use a (shielded) switcher to drop 28V to ~9V, and then use linear regulators to get the 5V and 3.3V.

(kinarfi , are you listening?)
 
Last edited:

Scarr

Member
It is so easy to forget to mention things :oops: I have very little room as this fits in a slot that is out on my control! so it look like I will just have to connect the GND pins of the 5v reg to a GND pour and dump the heat from the LDO 5v regulator, then feed the 3.3v from the 5v as MikeMI says?

5v

http://www.infineon.com/dgdl/Infine...n.pdf?fileId=db3a304314dca38901150402f8920c01

3.3v

http://ww1.microchip.com/downloads/en/DeviceDoc/20001826C.pdf

Anyone any idea of size of the pour, as I say there are physical limits I have that I cannot change. (these are 48mm X 10mm approx)
Steve
 
Last edited:

MikeMl

Well-Known Member
Most Helpful Member
Here is the relevant part of the '4264 data sheet:

4264.png

Starting from a 50degC ambient, and keeping the junction temp at 125degC or less, you would need a thermal resistance of (125-50)degC/2.53W ≈ 30degC/W = 30K/W. Notice that per the data sheet, even with 300mm^2 of copper pour (17mmx17mm), the thermal resistance junction-to-ambient is 81K/W, so I do not see how you can get there with 10mmX48mm.

Getting rid of 2.5W of heat is usually handled with one of these:



Clamped to one of these...

You can also put a power resistor upstream from the 5V reg to drop the 28V down to something lower, say 9V. R = E/I = (28-9)/0.11 = 173Ω, say a standard value of 168Ω. That would move P=R* I^2 = 168*0.11*0.11 ≈ 2W of the dissipation into the resistor, leaving ~0.5W in the 5V reg...
 
Last edited:

Scarr

Member
Dam Dam and Dam, switcher it is then :) having said that I quickly soldered a scalpel blade (yes I was surprised it would solder!) to the GND tab and it did not get too hot or even warm, I think this is due to the fact the HC-05 does not take maximum current all the time. I had already ordered the PCB's with a large GND area (as large as possible) so I will try this first, but I think Nigel is correct for the long term with a switcher.

Steve
 

tomizett

Active Member
If the maximum current is reasonably well-defined, you might be able to take some of the dissipation away from the regulator by connecting an appropriately-sized resistor in series with the input. Or even a power zener.
Of course the dissipation will still be in the same geographic area, so the heat from your resistor/zener will tend to raise the ambient for your regulator.
 

Nigel Goodwin

Super Moderator
Most Helpful Member

Scarr

Member
Hi again all,

OK luckily I have been able to stop production of the PCB's without penalty, so I think Nigel's suggestion of a switching regulator is best as all my worries about heat will go away, however this brings two issues for me, PCB design for a switching regulator due to increased components and component values, also noise.

So any suggestions on the simplest "switching regulator" you know, when I say simple I mean as simple a LDO / Linear regulator 1 main component and a couple of caps would be fantastic. and what if any issues do you think I will have with noise?

[EDIT] I looked at the AMS1117 you pointed out Nigel, this looked ideal but then I spotted its max 15v input I need max 30v input or am I mising something? [END EDIT]

Thanks everyone so far and meanwhile I will go hunting for a switching reg.

Steve

P.S. I am still thinking of a 5v switching regulator into a 3.3v regulator, so please tell me if this is a good / OK./ bad idea.
 
Last edited:

MikeMl

Well-Known Member
Most Helpful Member
Switching regulators usually take a lot more board area than linear ones. You usually need board space for an inductor, a Schottky catch diode, input bypass cap, output filter cap, programming resistors, and sometimes the FET power transistor.

How critical (for noise) is what you are powering from the 5V tap? Is it used for analog signal conversion/conditioning? If so, I wouldn't commit to pcb production until after I breadboard the entire system...
 
Last edited:

Scarr

Member
MikeMI

Whilst you were posting your reply I found the LT3470 it looks small and whilst there are more components I think I can get small versions of them all e.g. 0402 the only one I am unsure about is the inductor does this look suitable here?

Thanks

P.S. to answer your question, noise is probably an issue for the 5v as it''s feeding a ADUM1201 digital isolator
 

MikeMl

Well-Known Member
Most Helpful Member
You must use an inductor which will not saturate with ~120mA dc current in it. Look at the data for this one...

ind.png

ind2.png

I say again:
"How critical (for noise) is what you are powering from the 5V tap? "
 
Last edited:

Nigel Goodwin

Super Moderator
Most Helpful Member
So any suggestions on the simplest "switching regulator" you know, when I say simple I mean as simple a LDO / Linear regulator 1 main component and a couple of caps would be fantastic. and what if any issues do you think I will have with noise?
It depends on the exact application, but Mike is being rather alarmist - it's VERY unlikely to be a problem, in the vast majority of applications.

[EDIT] I looked at the AMS1117 you pointed out Nigel, this looked ideal but then I spotted its max 15v input I need max 30v input or am I mising something? [END EDIT]
Nope, not missed anything - which was why I said to use that to get the 3.3V from a previously regulated 5V source.

One of those, fed from the module I linked to in post #3, should do the job - IF it would fit, and assuming no great sensitivity to slight supply ripple.

There are these, which are nice and small:

http://www.banggood.com/10Pcs-Mini-...e-Power-Supply-Step-Down-Module-p-951163.html

But maximum input voltage is rather tight :D

'Possibly' feeding through a suitable zener diode 'might' help drop the input voltage?
 

Scarr

Member
sorry MikeMI I thought I have answered your question

P.S. to answer your question, noise is probably an issue for the 5v as it''s feeding a ADUM1201 digital isolator
As this is a going to be used to isolate 232 comms at 5v levels I would think it's noise critical but hell what do I know!

Steve
 

MikeMl

Well-Known Member
Most Helpful Member
...As this is a going to be used to isolate 232 comms at 5v levels I would think it's noise critical...
Actually, since this a "digital" use for the 5V, then it is not particularly noise sensitive...

Using the switcher to make 5V, and the linear reg to make 3.3V out of 5V is the way to go...

Have you found an inductor capable of passing 120mA that will fit in your available space? That might be your next big problem.

Ps, I found that Linear makes available a "jig" for the 3470. I modified it to run with a 28V input while delivering 125mA. Note that the peak current in the inductor is actually more like 175mA, so you should be looking for an inductor that wont saturate at 175mA...

3470.png
 
Last edited:

ChrisP58

Well-Known Member
Most Helpful Member
Texas Instrument's Webench Designer is a nice tool for selecting and designing switching supplies. http://www.ti.com/

Put in your parameters and it gives you choices and suggestions. You can optimize your results for cost, efficiency or board area.

And it gives recommendations for suitable components for all of the other parts you need, and it already knows all of the parasitic parameters for the parts in it's library.
 

kinarfi

Well-Known Member
Mike, I'm listening, You'll appreciate this!!
5+3.JPG
If your power supply has the wattage to handle this, same wattage, just used where you don't pop the LDOs and your project can get rid of the heat.
 

Attachments

Status
Not open for further replies.

Latest threads

EE World Online Articles

Loading
Top