Author Topic: LDO Voltage Regulator Trouble  (Read 4510 times)

0 Members and 1 Guest are viewing this topic.

Offline quiet-techTopic starter

  • Newbie
  • Posts: 9
  • Country: au
LDO Voltage Regulator Trouble
« on: November 13, 2017, 09:44:59 am »
Hi guys,

I’ve been working on a project for a little while now which involves a linear voltage regulator to step down 12v to 3.3v, this works but is problematic because of the excess heat.

My solution is to change the power to 5v in through a USB which powers the circuit as well as getting stepped down through a LDO voltage regulator to 3.3v (current draw will be around 200mA).

My problem is that the LDO regulator I’ve been experimenting with drops the voltage of the whole circuit to 1.9v from 5v and then measures 0.9v at the ‘out’ pin.

My current PCB has two capacitors in parallel with the regulator but I have also tried the output capacitor in series on a proto-board like the data sheet recommends. Both configurations have the same result, the voltage before the regulator drops down to 1.9v instead of 5v like I would like.

I’ve tested this on the fully populated PCB and also on a PCB with only the components in the diagram soldered.
I have measured the voltage of the board as 4.9v, then soldered the LDO regulator and measured the voltage as 1.9v, thus isolating the problem to the regulator.


Here is the regulator:
ROHM BD33KA5FP-E2 ( http://au.rs-online.com/web/p/low-dropout-voltage-regulators/7313873/ ) (D-PAK package).
Data sheet is here: http://docs-asia.electrocomponents.com/webdocs/0f5d/0900766b80f5dc3a.pdf .


I’m happy to use a different LDO regulator but I’m worried I’m making a fatal mistake in the PCB design which is causing the problem.

Any help would be much appreciated,

Oli
« Last Edit: November 13, 2017, 09:47:50 am by quiet-tech »
 

Offline sleemanj

  • Super Contributor
  • ***
  • Posts: 3024
  • Country: nz
  • Professional tightwad.
    • The electronics hobby components I sell.
Re: LDO Voltage Regulator Trouble
« Reply #1 on: November 13, 2017, 09:54:30 am »
Yeah, no, you have that pretty messed up, those caps are all wrong.

You should also be very careful to double check your pinout (of regulator and USB connector especially)

I have modified your diagram, attached.
~~~
EEVBlog Members - get yourself 10% discount off all my electronic components for sale just use the Buy Direct links and use Coupon Code "eevblog" during checkout.  Shipping from New Zealand, international orders welcome :-)
 

Online Ian.M

  • Super Contributor
  • ***
  • Posts: 12852
Re: LDO Voltage Regulator Trouble
« Reply #2 on: November 13, 2017, 10:13:10 am »
There's also a pitfall for the unwary - if you use a ceramic or other very low ESR input decoupling capacitor with a LDO or other type of regulator, and the supply voltage is greater than half the abs. max. input voltage rating, the supply wiring inductance ringing with the input capacitor can cause the max input voltage to be transiently exceeded and blow the regulator.  As your regulator's abs. max. input voltage is only 7V it is at risk of failing this way.   The cure would be to make the input cap an ordinary 1.2uf or 1.5uF aluminum electrolytic in parallel with a 0.1uF ceramic.  The output cap needs to be a ceramic >1uF so use 1.5uF 10V.

See https://www.eevblog.com/forum/projects/protecting-lm317-from-voltage-spikes/msg1316465/#msg1316465 for more background on this issue.
 

Offline Zero999

  • Super Contributor
  • ***
  • Posts: 19491
  • Country: gb
  • 0999
Re: LDO Voltage Regulator Trouble
« Reply #3 on: November 13, 2017, 10:25:01 am »
Yeah, no, you have that pretty messed up, those caps are all wrong.

You should also be very careful to double check your pinout (of regulator and USB connector especially)

I have modified your diagram, attached.
Yes, as the capacitors as they were before, bypassing the regulator, the output voltage would briefly rise to 5V, when the power was applied. If the components on the 3V side can't tolerate 5V, then they might have been damaged.
 

Offline danadak

  • Super Contributor
  • ***
  • Posts: 1875
  • Country: us
  • Reactor Operator SSN-583, Retired EE
Re: LDO Voltage Regulator Trouble
« Reply #4 on: November 13, 2017, 10:51:31 am »
Page 6 of the datasheet shows the considerations for setting
min ESR for the output cap.


http://docs-asia.electrocomponents.com/webdocs/0f5d/0900766b80f5dc3a.pdf


http://www.ti.com/lit/an/slva115/slva115.pdf


In addition page 7 shows diode protection for reverse current
protection of the LDO.

Datasheets and their recommendations are your friend.


Regards, Dana.
Love Cypress PSOC, ATTiny, Bit Slice, OpAmps, Oscilloscopes, and Analog Gurus like Pease, Miller, Widlar, Dobkin, obsessed with being an engineer
 

Offline quiet-techTopic starter

  • Newbie
  • Posts: 9
  • Country: au
Re: LDO Voltage Regulator Trouble
« Reply #5 on: November 14, 2017, 01:52:54 am »
Thanks a lot guys, thats terrific.
The additional info on the low ESR capacitor issue is much appreciated as well.
 

Offline quiet-techTopic starter

  • Newbie
  • Posts: 9
  • Country: au
Re: LDO Voltage Regulator Trouble
« Reply #6 on: November 15, 2017, 09:24:16 am »
I figured out by removing two bits of trace and adding a jumper cable I could emulate the new position of the capacitors on my current PCB.

I've attached a snapshot of the bottom section which has the LDO and power supply (USB).
The adjustments for the capacitors fixes my first problem, the circuit now measures 5v.
However at the LDO regulator output I measure 4.7v not 3.3v.

I re-watched Dave's video on Linear/switching/LDO regulators and he mentions how LDO's can get effected by ground current but I can't work out how this would cause this situation...
Would a diode on the ground rail just after the micro-controller ground solve my problem?

Any help would be much appreciated.

(I will be making the adjustments in regards to ESR on my next PCB, I'm just trying to get the LDO working before I commit to my next PCB).

« Last Edit: November 15, 2017, 09:26:03 am by quiet-tech »
 

Online Ian.M

  • Super Contributor
  • ***
  • Posts: 12852
Re: LDO Voltage Regulator Trouble
« Reply #7 on: November 15, 2017, 09:43:17 am »
With traces that long and looped for its decoupling caps, there's a fair chance of it oscillating.  I see you've got some ground vias next to the regulator's ground tab.  Jumper the  nearest one to the top end of C1 and the same on the other side to C2 to bypass the long looped traces.

Also, without an electrolytic on the input, the regulator is probably blown due to transient input overvoltage, which would explain the high output voltage - change it, and to prevent it blowing again, replace C1 with 0.1uF and fit a 2.2uF 10V or 16V Aluminum Electrolytic (leaded, not low ESR), between the power in track after the fuse and the ground track.  Lay the 2.2uF cap flat and solder its leads directly to and along the tracks.

Finally for testing, patch in a 330R resistor between the regulator output and ground to give it a 10mA load.

When you re-spin the board *PLEASE* discuss the layout here before ordering.  Long skinny traces do NOT make for stable regulator circuits.  You should be using a ground plane and nice fat power traces, with the decoupling as close to the regulator as possible and more decoupling at the power pins of whatever loads its supplying.
« Last Edit: November 15, 2017, 09:47:02 am by Ian.M »
 

Offline jcw0752

  • Contributor
  • Posts: 35
  • Country: us
Re: LDO Voltage Regulator Trouble
« Reply #8 on: November 15, 2017, 04:20:18 pm »
Hi,
Have you ever considered using one of the RECOM R78 Dc to DC converters instead of a Linear Regulator. I use them all the time as they are 90% plus efficient and therefore very little heat.

http://www.newark.com/recom-power/r-78e3-3-1-0/dc-dc-converter-3-3v-1a-sip/dp/51Y4020?ost=R-78E3.3-1.0&scope=partnumberlookahead&exaMfpn=true&searchref=searchlookahead&iscrfnonsku=false&ddkey=http%3Aen-US%2FElement14_US%2Fw%2Fsearch

John
 

Offline mariush

  • Super Contributor
  • ***
  • Posts: 5014
  • Country: ro
  • .
Re: LDO Voltage Regulator Trouble
« Reply #9 on: November 15, 2017, 05:32:07 pm »
If you know the voltage is gonna be 4.75v .. 5.5v from USB , you may want consider adding a diode before the linear regulator. The diode will drop 0.5v to 0.8v so the regulator will have less input voltage to deal with and the heat will be spread on two chips now.  It could also be a sort of reverse voltage protection (though why would you need it with usb connectors..)

It will also be an added protection for the chip, if you're going to go with one that has a maximum voltage of 5.5v - some cheap phone chargers with usb jack will use just a 5.6v zener as reference or to control the maximum output voltage so you may have 5.6v .. 5.7v

A diode will deal with 0.15w of heat without any issues.   

As LDO, you should look into something like MIC5504 , a 300 mA LDO with 0.16v dropout voltage at 300mA and they're super cheap at around 12 cents if you buy 1, less than 10 cents if you buy 25 or more: https://www.digikey.com/product-detail/en/microchip-technology/MIC5504-3.3YM5-TR/576-4764-1-ND/4864028

It's stable with 1 uF ceramic capacitors on input and output (but no reason why you couldn't use 4.7uF or 10uF since ceramic capacitors are super cheap) and can go up to 125c so even with 5v at the input, 0.35w of heat won't get its temperature up to 125c.

Another option would be NCP114 pretty much same specs and same package, same pinout (so interchangeable with the one above) : https://www.digikey.com/product-detail/en/on-semiconductor/NCP114ASN330T1G/NCP114ASN330T1GOSCT-ND/6560649

Or TLV73333 , same sot-23-5 package, same pinout around same dropout voltage yadda yadda : https://www.digikey.com/product-detail/en/texas-instruments/TLV73333PDBVT/296-38492-1-ND/5023884

OR AP2125K , same  everything : https://www.digikey.com/product-detail/en/diodes-incorporated/AP2125K-3.3TRG1/AP2125K-3.3TRG1DIDKR-ND/4505581

There you go , four different chips from four different manufacturers, all with same pinout and relatively same dropout voltage
 

Offline quiet-techTopic starter

  • Newbie
  • Posts: 9
  • Country: au
Re: LDO Voltage Regulator Trouble
« Reply #10 on: November 16, 2017, 08:03:03 am »
Thanks for the responses guys.

Ian.M
I’ve just ordered all the parts you mention in your advice, 2.2uf electrolytic cap, 0.1uf ceramic cap as well as a couple more LDOs.

I’ve just done some reading on ground planes and thicker traces and will absolutely be using them in my next design iteration as well as closer position of anti-oscillation caps and more direct traces.
I will post a new iteration when my next PCB design iteration is done.

jcw0752
This wasn’t on my radar but is definitely interesting. I don’t think I will use it for this project as I need a braided cable (for aesthetic purposes more than anything) and these seem to be cheapest/most easily available for usb connectors, so I need a 5v solution really).

mariush
I love that idea, will definitely be using it to put less heat on the LDO.

I really like the sound of the MIC5504 LDO you recommended, would this still work with a 2.2uf electrolytic cap in parallel as well as a 0.1uf ceramic cap in parallel before the input like Ian.M described?

I originally chose a DPAK LDO because of the good thermal characteristics but I suppose with your diode addition this will be less of a problem.
 

Offline mariush

  • Super Contributor
  • ***
  • Posts: 5014
  • Country: ro
  • .
Re: LDO Voltage Regulator Trouble
« Reply #11 on: November 16, 2017, 11:51:53 am »
I don't see any problems with that.

However, I'd like to make some observation.  If your whole design is mostly surface mount, it would be sort of a shame to use an electrolytic capacitor which would require you to have holes  in the circuit board just for that capacitor.  Of course there are surface mount electrolytic capacitors but what I found is that they're less reliable than capacitors with leads, they're more expensive, they take up more space and they're also quite tall (an electrolytic capacitor with leads you could in theory lay flat on the motherboard like a crystal oscillator or a plain resistor) 
There are polymer capacitors but they're quite expensive, a few times more expensive compared to electrolytic... for example this one is 1$ if you buy 25, or 0.5$ if you buy 1000 : https://www.digikey.com/product-detail/en/panasonic-electronic-components/EEF-CD1C4R7R/PCE3170CT-ND/256710

If the regulator says it's stable with ceramic capacitors, maybe just using a second 2.2uF or 3.3uF or 4.7uF x5r or x7r  and with decent voltage rating (let's say 25v or higher) would be enough, for example this 4.7uF 25v x5r is 5.5 cents each if you buy 100 or 16 cents if you buy just one : https://www.digikey.com/product-detail/en/murata-electronics-north-america/GRM21BR61E475MA12L/490-5422-1-ND/2175229


Some other notes.

It will probably make your board look nicer and also easier to route if you use 45 degree angles in your traces - i see you seem to use only right angle (90 degrees).

It's not wrong and doesn't make much difference, but I generally like to see components like resistors (that 820 ohm resistor and R45 in your case) close to the end of a trace, in your case closer to the headers unless their height would physically block insertion of something in the header, or there could be some risk of mechanical pressure on the resistors.

Also, I'd say in general you should try to aim for as few vias as possible (jumping traces from one layer to another). For simple designs, ideally you'd want to aim for keeping it all on one layer.  You have a few places where you jump on the other layer almost just for the fun of it, basically it can be avoided with a bit more clever routing.

For example, the two traces on top right of the header could be moved all the way to the edge of the board, and go down along the edge of the board this way making the switch easier to connect (and no need to go near the switch to the other layer (btw there's some visual bug there, you go at the switch on bottom layer but don't see the via coming out to connect to yellow trace)
This will also save you going on other layer at the 820 ohm resistor since now you'd have space to put the resistor right there near the header.

Another less common technique to reduce the number of vias is to use 0 ohm resistors as a convenient way to jump over traces.  For example, you have those 3 traces going on the other layer just to cross three other traces.
There are surface mount resistors which you could use to jump over the traces, depending on their length and the trace thickness it would be easy to jump over 2-3 traces.
It's less common because it adds some cost (vias are free but resistors are fractions of a penny) and often it's not used because it increases the completion time for pick and place machines  or the pick and place machines have a limited number of parts they can work with (for example a machine could only pick parts from 64 tapes of components, so adding your 0 ohm resistor would block a tape that could be used for other chips)

However, it's a design with a small number of units made, it could make the layout look nicer and leave you the bottom space for routing traces or could even allow you to use a cheaper single layer pcb

Here's some examples of such resistors, 0805 (probably wide enough for 2 traces to go under it)  1 cents each if you buy 50 : https://www.digikey.com/product-detail/en/yageo/RC0805JR-070RL/311-0.0ARCT-ND/731163
1206 is about the same price : https://www.digikey.com/product-detail/en/yageo/RC1206JR-070RL/311-0.0ERCT-ND/732131
 

Online Ian.M

  • Super Contributor
  • ***
  • Posts: 12852
Re: LDO Voltage Regulator Trouble
« Reply #12 on: November 16, 2017, 11:59:47 am »
Its the long input lead and the USB power connector that's the problem. The lead inductance ringing with a ceramic cap can transiently double the input voltage if the lead is plugged in with the power on.  Either use a regulator with >12V Vin_max rating, or damp it.   If you don't want to use an electrolytic so its ESR damps it, try 2.2uF ceramic with 1.5 ohms in series, then 0.1uF ceramic across the whole lot.
« Last Edit: November 18, 2017, 12:31:59 am by Ian.M »
 

Offline quiet-techTopic starter

  • Newbie
  • Posts: 9
  • Country: au
Re: LDO Voltage Regulator Trouble
« Reply #13 on: November 17, 2017, 11:16:52 pm »
Thanks, I'l put together this new design today and then post it here to make sure I've understood everything you guys have been saying.
 

Offline quiet-techTopic starter

  • Newbie
  • Posts: 9
  • Country: au
Re: LDO Voltage Regulator Trouble
« Reply #14 on: November 19, 2017, 12:16:59 am »
Hi guys,

I re-worked the board and made some changes. I decided to go with the through hole ectrolytic capacitor in the end as my headers are through hole and the mounted micro controller will still be taller than the cap.

I also realised there is a little ridge around the casing that the PCB is going into which would prevent the micro-usb from inserting. To raise the connector I have put in a micro-usb breakout board in its place.

I had decided to use the MIC5504 LDO, but I'm worried that the package is too small for me to hand solder so have left the LDO as the ROHM for now.

Below is a list of the changes I made gathered from this thread followed by a new component list.


Change DC barrel to Micro usb connector female breakout as need extra height to protrude  over 3mm rim of frame.

Make Input cap 1.5uF aluminum electrolytic in parallel with a 0.1uF ceramic.  (will use through-hole electrolytic cap as I have through-hole breakout board and micro which make height of populated pcb atleast 10mm already).


The output cap needs to be a ceramic >1uF so use 1.5uF 10V.


Make traces thicker.

As LDO, you should look into something like MIC5504 , a 300 mA LDO with 0.16v  (SOT-23 package)
http://au.rs-online.com/web/p/low-dropout-voltage-regulators/9113332/
(Worried about too smaller package size to hand solder as well as ‘enable’ pin).
So using http://au.rs-online.com/web/p/low-dropout-voltage-regulators/7313873/ for now

Diode protection for reverse current (parallel from before input to after output)


You should be using a ground plane and nice fat power traces,


Decoupling as close to the regulator as possible.


Add a diode before the linear regulator. (The diode will drop 0.5v to 0.8v)(also protects chip(Omega).


Stable with1 uF ceramic capacitors on input and output
(but no reason why you couldn't use 4.7uF or 10uF since ceramic capacitors are super cheap)


Shorten input lead from USB to LDO.


Route traces at 45 degree angles.


It's not wrong and doesn't make much difference, but I generally like to see components like resistors (that 820 ohm resistor and R45 in your case) close to the end of a trace.


(Ask about fuse, is it necessary and do they fail often?)


Aim for as few vias as possible.


For example, the two traces on top right of the header could be moved all the way to the edge of the board, and go down along the edge of the board this way making the switch easier to connect
(and no need to go near the switch to the other layer
This will also save you going on other layer at the 820 ohm resistor since now you'd have space to put the resistor right there near the header.


Components

Micro usb connector female breakout as need extra height (3mm rim of frame)
https://www.adafruit.com/product/1833

1.5uf Electrolytic capcitor (lead pitch: 3.5mm) (INPUT CAP)
http://au.rs-online.com/web/p/aluminium-capacitors/8454556/

0.1uf ceramic cap SMD 1206 (INPUT CAP)

1.5uF 10V.  Ceramic Cap SMD 1206 (OUTPUT CAP)

Diode before LDO (spent ages looking at SMD diodes but most were switching or tiny packages??)
http://au.rs-online.com/web/p/rectifier-schottky-diodes/8050959/

Diode paralel over input/output on ldo
http://au.rs-online.com/web/p/rectifier-schottky-diodes/8050959/

LDO Voltage Regulator
http://au.rs-online.com/web/p/low-dropout-voltage-regulators/7313873/

 

Online Ian.M

  • Super Contributor
  • ***
  • Posts: 12852
Re: LDO Voltage Regulator Trouble
« Reply #15 on: November 19, 2017, 12:36:13 am »
C1 and C3 need to be in parallel, not series.  You can remove D2, its not needed because of D4.  Use a ground plane, and move the decoupling caps as close as possible to the regulator.

I'm not even going to comment on the rest of the layout!  |O
 

Offline quiet-techTopic starter

  • Newbie
  • Posts: 9
  • Country: au
Re: LDO Voltage Regulator Trouble
« Reply #16 on: November 19, 2017, 02:03:17 am »
Thanks, more like this?
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf