Author Topic: Has anyone connected a linear regulator to the output of a switching reg ?  (Read 7102 times)

0 Members and 1 Guest are viewing this topic.

Offline ez24Topic starter

  • Super Contributor
  • ***
  • Posts: 3082
  • Country: us
  • L.D.A.
The Art of Electronics suggests to use a linear regulator to reduce the ripple from a buck but it gives no suggestions.

Some linear datasheets give some ripple reduction information and some do not.   If there is any info in a datasheet then the info is different for another device.  Some application notes state that the linear can be used to reduce the ripple from a switcher but give no data.

In general, I believe low dropout voltage linear regulators have less ripple reduction than a linear with higher voltage dropout.  In other words more watts lost more ripple is lost.

Has anyone done any experiments using a linear to reduce the ripple from a switcher?

thanks
YouTube and Website Electronic Resources ------>  https://www.eevblog.com/forum/other-blog-specific/a/msg1341166/#msg1341166
 

Offline swingbyte

  • Newbie
  • Posts: 3
Yes it works quite well at reducing the noise. Used to increase the number of bits of the ADC that I could use above the noise floor
 

Offline Chris C

  • Frequent Contributor
  • **
  • Posts: 259
  • Country: us
I think you'll find Linear Technology Application Note 101, "Minimizing Switching Regulator Residue in Linear Regulator Outputs", quite useful:

http://cds.linear.com/docs/en/application-note/an101f.pdf
 

Offline ez24Topic starter

  • Super Contributor
  • ***
  • Posts: 3082
  • Country: us
  • L.D.A.
I think you'll find Linear Technology Application Note 101, "Minimizing Switching Regulator Residue in Linear Regulator Outputs", quite useful:

http://cds.linear.com/docs/en/application-note/an101f.pdf

Good info - thanks   Especially the test hookups.  Cannot buy a HP-461A right now, I put it on my watch list.  BUT maybe I will be able to use this

http://www.ti.com/tool/TI-PLABS-AMP-EVM


I added the link to

https://www.eevblog.com/forum/beginners/electronics-primers-course-material-and-books/120/

To another subject - what do you think about making your answer about "parts" a sticky ? 
YouTube and Website Electronic Resources ------>  https://www.eevblog.com/forum/other-blog-specific/a/msg1341166/#msg1341166
 

Offline Chris C

  • Frequent Contributor
  • **
  • Posts: 259
  • Country: us
Now that I'm had some sleep, I remember some more info. :)

If you're not familiar with the general differences in internal construction between standard and LDO linear regulators, this is a good reference:

http://focus.ti.com/download/trng/docs/seminar/Topic%209%20-%20Understanding%20LDO%20dropout.pdf

See Figure 1, and imagine the effect of a high frequency input spike on the output transistor of each topology.

For the standard reg, the output current is controlled by the base-to-emitter difference.  But that spike directly hits the collector, and is mostly ignored.  There are still some other means by which the spike can be coupled through, but the tendency is towards attenuation of spikes.

But for the LDO, the output transistor is configured in a way where it directly amplifies the spike!  Until the gate driver (feedback loop) can correct it, and as it's fighting the MOSFET gate capacitance to do so, its speed is limited; short transients it simply never has a chance to fix.

You said, "In other words more watts lost more ripple is lost."  So in general, that would appear to be true.  But not always.

For example, let's say you have a high current 5V supply from a switching regulator, which you want to regulate down to 3.3V via linear reg to reduce ripple.  The dropout voltage is small enough that the linear reg must be an LDO.  But assume you also have a lower current 12V rail around.  It seems theoretically possible you could put a standard linear reg on the 12V supply, set to produce 3.3V.  This wouldn't provide the 3.3V rail directly, but rather drive an external bipolar transistor to do so, just like you would to provide higher current than the linear reg's internal limit.  As the linear reg output can swing up to 9-10V, dropout voltage for your added output transistor ceases to be an issue!  And that 5V rail could be reduced to 4V, for even higher efficiency.

I can sympathize with the lack of adequate test equipment.  Often I just have to design defensively, and hope for the best.  Whenever switching noise may be an issue, I put a Pi filter on the switcher's output, with a ferrite bead as the inductor.  Be aware that there are different ferrite materials, each working best at different frequency ranges, and choose your material appropriately.  Also try to choose a part with low parasitic capacitance, which tends to bypass the ferrite.  I chose Fair-Rite #2512067007Y3 based on this info as best as I understand it, and it halved my ADC errors compared to the best bead formerly in my parts collection.

As for the "parts" thread.  The last forum I was active on, I posted something helpful that's up to 103K views, 449 replies, and still going after several years.  But it's not a sticky, the mods there have a very restrictive set of criteria.  I'm new here, not sure what the criteria are.  And prefer not to endorse my own posts for stickification as a point of etiquette anyway.  If others feel it's sufficiently helpful and wish to endorse it, I have no issue with that.
 

Offline ez24Topic starter

  • Super Contributor
  • ***
  • Posts: 3082
  • Country: us
  • L.D.A.
Chris C thanks for all the info.  By any chance do you have any YT videos or any online training classes?  I also do not know about the "sticky", I will post a question later under General and see if we can get some answers.

When I got my equipment, I looked at it and realized I could not use it because I did not have anything.  So I think the parts post is valuable for newbies like me.

Right now I am trying to get this

https://www.eevblog.com/forum/projects/ti-have-a-great-offer-on-an-opamp-evaluation-board/msg734841/#msg734841

I think TI is investigating me  :-\
YouTube and Website Electronic Resources ------>  https://www.eevblog.com/forum/other-blog-specific/a/msg1341166/#msg1341166
 

Offline T3sl4co1l

  • Super Contributor
  • ***
  • Posts: 21687
  • Country: us
  • Expert, Analog Electronics, PCB Layout, EMC
    • Seven Transistor Labs
Seven Transistor Labs, LLC
Electronic design, from concept to prototype.
Bringing a project to life?  Send me a message!
 

Offline Chris C

  • Frequent Contributor
  • **
  • Posts: 259
  • Country: us
My training has been slow and very fragmented, so I'm afraid there's no particular videos or classes/books that I can personally recommend.

I described what's been helping me most on another thread, final post:

https://www.eevblog.com/forum/beginners/frequency-counter-measurements-and-noise/15/

To put that in the current context:  I can design a standard or LDO linear regulator in SPICE.  And a switching regulator as well.  Then put the two together and see how they interact.  Yes, I actually did this at one point.  It was only after observing this interaction and looking for verification, that I found and understood the previously linked AN101.  If I'd seen AN101 prior to my simulated experiment, I wouldn't have grasped its significance.  A similar pattern played out on the individual regulators - read how they work with minimal understanding, design and simulate, observe, then re-read with much deeper understanding.  I guess it's a weird way of learning, but it seems to work best for me.
 

Offline ez24Topic starter

  • Super Contributor
  • ***
  • Posts: 3082
  • Country: us
  • L.D.A.
My training has been slow and very fragmented, so I'm afraid there's no particular videos or classes/books that I can personally recommend.

I meant do you have any YT videos that you made?  How about books, have you written any?

Can I ask - do you have a PHd ?   I am guessing you have 2 - one in electronics and one in English.  I have never seen anyone that can write to your level of skill of electronics and English.  Someone who can proofread and correct MCU datasheets.  :clap:

So I guess you are a retired MIT professor.  Am I right?

By any chance have you played around with a PIC32 Microstick II ?

One more thing - members like to see what country you are from by the flag icon in your Profile.  It helps when giving help so members can see if the resources are in their country.  Some members say they do not want to show their flag by showing a black flag.  I hope you do not do this.

Roger in San Diego, Calif.
YouTube and Website Electronic Resources ------>  https://www.eevblog.com/forum/other-blog-specific/a/msg1341166/#msg1341166
 

Online Zero999

  • Super Contributor
  • ***
  • Posts: 19525
  • Country: gb
  • 0999
A very low drop-out regulator can be made from a MOSFET and the TL431 if it's biased from a higher supply voltage.

The output of buck converter could be set just a couple of hundred of mV above the TL431.

The difference between the input of the buck and output of the LDO needs to be high enough to turn on the MOSFET though.
 

Offline Chris C

  • Frequent Contributor
  • **
  • Posts: 259
  • Country: us
Can I ask - do you have a PHd ?   I am guessing you have 2 - one in electronics and one in English.  I have never seen anyone that can write to your level of skill of electronics and English.  Someone who can proofread and correct MCU datasheets.  :clap:

So I guess you are a retired MIT professor.  Am I right?

Hope you didn't make any side bets on that.  I'm a programmer by profession, electronics hobbyist, and didn't make it through college.

I meant do you have any YT videos that you made?  How about books, have you written any?

No electronics-related YT videos.  Or books.  Though I've been toying with the idea of a Web-based series of tutorials.

By any chance have you played around with a PIC32 Microstick II ?

I just started using the PIC32 six weeks ago.   But no Microstick II.  I avoid pre-built dev boards, that's passing up an opportunity to practice and learn.  And I wanted something that was easily deployable in a variety of actual projects as well.  So I made a 1.95" x 1.95" board, that accommodates any PIC32 in TQFP44 package.  All MCU pins broken out to headers, plus separate ICSP header, headers for multiple connections to the primary SPI/I2C busses, and header for plugging in a nRF24L01+ RF transceiver.  Built-in switching power regulator with up to 21V input.  Soft-power circuit and measurement of the input voltage, so it can shut itself completely off with true zero drain, should it be powered by a battery that drops too low.  Holes for standoff mounting, often used with some magnetic standoffs I made that let it snap securely onto any ferrous surface.  Total cost per fully populated board including RF is a bit under $10.  It's the most complex board I've made.  As an additional challenge I chose to design it from the PIC32 datasheet, and order PCBs, before having or trying a PIC32.  Think I expanded my collection of gray hairs with that choice.  But considering I only need two jumper wires to fix mistakes and make it work, it came out better than I expected. ;D

One more thing - members like to see what country you are from by the flag icon in your Profile.  It helps when giving help so members can see if the resources are in their country.  Some members say they do not want to show their flag by showing a black flag.  I hope you do not do this.

Good point, there are several times I've used others' flags to tailor my responses.  My flag has been added.
 

Offline ez24Topic starter

  • Super Contributor
  • ***
  • Posts: 3082
  • Country: us
  • L.D.A.
Quote
Hope you didn't make any side bets on that.  I'm a programmer by profession, electronics hobbyist, and didn't make it through college.

I am really surprised.   But I guessed the US and was right.  Now I guess CA or Wash 

Quote
No electronics-related YT videos.  Or books.  Though I've been toying with the idea of a Web-based series of tutorials.

What subject ?  Are you thinking of doing it with videos or text?  I sure hope you do. 

Would it be free or for profit?
I can help if you like, by proof reading, beta testing, video editing, etc.  My name is Roger, I am 67, and was a civil engineer.  I live in San Diego and I have a lot of free time.  Also I can contribute some money to your cause.  Even if your project is for profit,  I will volunteer my time.  Because I could learn a lot from you just by editing your work.

My electronic and language skills are light years below yours but who knows, I might be able to do something that can help you.

Quote
Holes for standoff mounting, often used with some magnetic standoffs I made that let it snap securely onto any ferrous surface.

I thought of this once (that is as far as I went).  Are you planning on teaching a lab and using the metal standoffs and your PIC32 board for the class?  Are you going to sell your board?  There is a "For Sale" section on this forum.  I believe there are several members selling boards to each other.

I bought a Microstick II thinking I could use if for an online course from Cornell.  It turned out that the course is not online nor posted on YT. 


YouTube and Website Electronic Resources ------>  https://www.eevblog.com/forum/other-blog-specific/a/msg1341166/#msg1341166
 

Offline JoeN

  • Frequent Contributor
  • **
  • Posts: 991
  • Country: us
  • We Buy Trannies By The Truckload
I think you'll find Linear Technology Application Note 101, "Minimizing Switching Regulator Residue in Linear Regulator Outputs", quite useful:

http://cds.linear.com/docs/en/application-note/an101f.pdf

I just started reading this and as soon as I hit the second line "Banishing Those Accursed Spikes" I knew that was going to be Jim Williams and the next line confirms it.    :-DD  He had a way with words.

Quote from: Jim Williams RIP
Large numbers of capacitors and aspirin have been expended in attempts to eliminate these undesired signals and their resultant effects.
« Last Edit: August 24, 2015, 04:12:10 am by JoeN »
Have You Been Triggered Today?
 

Offline Chris C

  • Frequent Contributor
  • **
  • Posts: 259
  • Country: us
Now I guess CA or Wash

I'm one a dem Cajuns from Louisiana, cher.

What subject ?  Are you thinking of doing it with videos or text?  I sure hope you do. 

Would it be free or for profit?
I can help if you like, by proof reading, beta testing, video editing, etc.  My name is Roger, I am 67, and was a civil engineer.  I live in San Diego and I have a lot of free time.  Also I can contribute some money to your cause.  Even if your project is for profit,  I will volunteer my time.  Because I could learn a lot from you just by editing your work.

My electronic and language skills are light years below yours but who knows, I might be able to do something that can help you.

Subject would be discrete and analog circuits, as that's what I've had the most trouble with.  Probably no video, as it forces the viewer to learn at the pace the creator of the video sets; I prefer text.  And it would be free. 

I find there's a gap in educational material.  Plenty of bare-basics beginner material written in a friendly, conversational style.  Plenty of expert level material bogged down in terminology and math.  Very little in the middle.

For example, let's say Mr. John Doe wants to use a bipolar transistor as a switch.  He can easily look up an easily understandable basic tutorial on that.  But what if the switch operates too slowly?  That tutorial will typically cover one method of improving it.  Two if he's lucky.  But off the top of my head, I know four methods.  Plus which cheap and common "jellybean" transistors he probably already has, or can easily obtain, that will operate faster.  And which datasheet specs he'll need to examine if he needs to find something faster still.  As it stands, John will have to look elsewhere to find that scattered info, likely finding himself in documents that are way over his head.

Another example, Class AB transistor amps.  Any basic tutorial will explain what crossover distortion is, and show you one or two biasing schemes to eliminate it.  But they can be touchy, and involve some trade-offs.  The expert level resources mainly show the same schemes, though with a much better (and technical) explanation of the trade-offs.  I built some, and though the end result was adequate, I never really felt comfortable that I had it right.  Then, two years back, I wandered into an audiophile forum, for some reason forgotten to me (I love music but am not an audiophile).  There I discovered a variation of the Class AB plus biasing called the "diamond buffer", apparently originating from the internals of an old op-amp IC.  It's simple, yet works better and with less fuss, than any other biasing scheme I've seen.  I would have loved to have found out about it decades ago.  Yet it seems unknown outside audiophile circles.  I have not seen it elsewhere.  And despite mentioning multiple places, I have not found a single non-audiophile who recognized it.

There is no reason why additional info like this can't be written into a tutorial written for beginners, in a way that it's still easily understood.  I think it would be helpful and appreciated.

The main barrier, as always, is time.  I figure I'll need two virtually uninterrupted weeks of free time to focus and lay the groundwork, after that I can expand it in increments.  Just too many other things I want to accomplish right now to give it serious consideration.  In the meantime, a little forum activity helps others, and helps me polish my language and explanation skills.  Your offer of help is quite gracious.  Should I start moving ahead on this, I'll let you know.  I'd love to have someone vetting my work, before putting it out there for the world to see.

I thought of this once (that is as far as I went).  Are you planning on teaching a lab and using the metal standoffs and your PIC32 board for the class?  Are you going to sell your board?  There is a "For Sale" section on this forum.  I believe there are several members selling boards to each other.

No lab/classes.  Again, time constraints.  But I've also been a serious programmer for 35 years now, have forgotten what it's even like to be a beginner.  Even moving to MCUs a couple years back didn't faze me, felt just like coding on my old Atari 800.  I'm good at helping with specific problems, but in the broader context of a class?  I know from experience I move too fast and soon lose everyone.  It's just too second nature to me, I am the WRONG person to teach that particular topic!

As for the board, when I revise it to fix the known errors, I'm going to post the design here to solicit suggestions from folks more experienced than me.  Then roll them into the design, and do another fab run.  Assuming there are no new issues, I'll update my post with the final design.  If anyone expresses interest, I can sell some bare boards as a convenience to others.

I bought a Microstick II thinking I could use if for an online course from Cornell.  It turned out that the course is not online nor posted on YT.

Don't let a little thing like that stop you. ;)  Find a starting point, anything at all, and start doing stuff.  I think there's a book called "Feel the Fear and Do It Anyway".  For a somewhat cheesy maxim, it applies here perfectly.
 

Offline ez24Topic starter

  • Super Contributor
  • ***
  • Posts: 3082
  • Country: us
  • L.D.A.
Quote
I'm one a dem Cajuns from Louisiana, cher.
I would have never guess this  :)  I was thinking you were a chip designer in Silicon Valley.
Quote
Subject would be discrete and analog circuits, as that's what I've had the most trouble with.  Probably no video, as it forces the viewer to learn at the pace the creator of the video sets; I prefer text.  And it would be free. 

Would this be through a free blog service or your own web site (do you have one ?)  Give some thought about immortality.  If you could do YT they might be immortal lessons.  If they are "hosted" then somebody has to pay to keep it up.  I am not sure how free blogging services work (ie could they be immortal ?)  (idea for new topic!)
Maybe there is a way of doing both.  As a wild idea maybe you could do written and I could make some videos of your lessons ?  I have basic equipment including a Rigol DS-1054Z.  No exotic stuff.


Take a look at all the links in this post

https://www.eevblog.com/forum/beginners/electronics-primers-course-material-and-books/

There are many links to online tutorials that you can use as examples of presentation.  Also if you come across a good educational link, please post it there.
Quote
The main barrier, as always, is time.
You need to stop reading MCU datasheets trying to find errors in them  :-DD
Quote
helps me polish my language and explanation skills
Oh brother, any more polish and I wouldn't be able to read.  I already thought you had a PHd in English.
Quote
It's just too second nature to me, I am the WRONG person to teach that particular topic!
I figured this by reading your first post where you explained why Microchip made errors  :-DD  To be honest I could not understand much of it.
Quote
Should I start moving ahead on this, I'll let you know.  I'd love to have someone vetting my work, before putting it out there for the world to see.
 
I will send you a PM with my email address.  Check out my immortality post.








YouTube and Website Electronic Resources ------>  https://www.eevblog.com/forum/other-blog-specific/a/msg1341166/#msg1341166
 

Offline Chris C

  • Frequent Contributor
  • **
  • Posts: 259
  • Country: us
Would this be through a free blog service or your own web site (do you have one ?)  Give some thought about immortality.  If you could do YT they might be immortal lessons.  If they are "hosted" then somebody has to pay to keep it up.  I am not sure how free blogging services work (ie could they be immortal ?)  (idea for new topic!)

I was thinking my own website (I run a few but none explicitly for myself).  I want it heavily hyperlinked, much like a typical Wikipedia article.  There will be a linear progression of articles, each building on the last.  But many people are going to jump to a specific article of interest.  Any time I mention a concept explained in an earlier article, I want it to be clickable, so they can quickly jump to anything unfamiliar.  Though I did not consider immortality.

I have basic equipment including a Rigol DS-1054Z.  No exotic stuff.

You're set up better than me in that department.  I have a Hitachi V-302B.  It apparently didn't sell in an estate sale, and was tossed to curbside in a thunderstorm.  I did my best with it, but it's still cranky on a good day. ;)

You need to stop reading MCU datasheets trying to find errors in them  :-DD

LOL!  I read datasheets to find errors in my own code.  Finding errors in datasheets is a natural consequence.  I'd rather do that than what I spent all yesterday doing, tracking down a Heisenbug (the kind that disappears when you try to observe it) that was causing crashes at random times/locations in 20,000 lines of C code (my current PIC32 project).
 

Offline Chris C

  • Frequent Contributor
  • **
  • Posts: 259
  • Country: us
I just started reading this and as soon as I hit the second line "Banishing Those Accursed Spikes" I knew that was going to be Jim Williams and the next line confirms it.    :-DD  He had a way with words.

Somehow I missed your post.  And I'd forgotten all about that quote from the appnote!  Thanks for reminding me.  I think I've seen some humor in a few other LT appnotes too, must have been Jim's handiwork.

I love it when engineers have a sense of humor.  RIP, Bob Pease.
 

Offline ez24Topic starter

  • Super Contributor
  • ***
  • Posts: 3082
  • Country: us
  • L.D.A.
Back to LDOs

Some have agreed with me that generally a low voltage dropout is going to have low ripple rejection from a buck.  But also someone said this is not always the case.

I have been using Digikey and looking for one and so far no success.  I am looking at datasheets and getting confused.  Since I have more detailed questions I am going to start a new post on PSRR and dB calculations.

thanks
YouTube and Website Electronic Resources ------>  https://www.eevblog.com/forum/other-blog-specific/a/msg1341166/#msg1341166
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf