EEVblog Electronics Community Forum
Electronics => Projects, Designs, and Technical Stuff => Topic started by: coppercone2 on August 31, 2018, 07:14:25 pm
-
So I have always strayed away from dealing with anything switchmode.
Have you ever actually been baffled or ran into really odd ball problems making a switching supply?
Are there like, certain designs that use more (optional) parts and design choices to avoid certain problems which are not typically followed in mass production (muntzing) for cost reasons but increase complexity significantly?
I don't want to kill myself designing something for a one off with design techniques that become financially relevant when you make more then a few thousand a year. Are there any things to look out for when implementing designs that can make it easier for me?
Since I would be designing it myself time is most imporant so I don't care if you need to spend a few extra bucks to make it more stable or something.
Any really obvious offenders? I am just wondering about general choices that can be made to weigh things in my favor.
-
Bang your head against the wall and come back...
The question is too broad to answer, you seem going into it but fearing something. Make ine experience, come back with the weak points or specific problems you found and we'll help you.
One think to be careful before hand, LAYOUT. Start with something low power and low frequency. Low power means cheaper, less likely to blow, ovearall easier. Low freq means easier with layout traps, but could lead to big inductors, low power makes those inductors smallers.
JS
-
Hi,
This post is rather mysterious, and you sound a bit like an oracle ;D. I understand you've got a problem with switching power supplies, but I quite don't understand what it is exactly, or what you're trying to say to us.
What is your problem with SMPS ?
While the ripple, inductance spikes etc... can be quite a pain in the ass in analog circuits and other sensible designs, no linear regulator will ever be as efficient and versatile as modern SM controllers.
Like anyone, I stay away from benchs SMPS, but in a complex electronic design that needs to be efficient, I can't see how you could stay away from SMPS.
Please explain your issue a little further.
-
Hi,
This post is rather mysterious, and you sound a bit like an oracle ;D. I understand you've got a problem with switching power supplies, but I quite don't understand what it is exactly, or what you're trying to say to us.
What is your problem with SMPS ?
While the ripple, inductance spikes etc... can be quite a pain in the ass in analog circuits and other sensible designs, no linear regulator will ever be as efficient and versatile as modern SM controllers.
Like anyone, I stay away from benchs SMPS, but in a complex electronic design that needs to be efficient, I can't see how you could stay away from SMPS.
Please explain your issue a little further.
Linear regulators can be more efficient than SMPS in some applications, would you like to think it for yourself?
For what I understood he didn't have any problems with them, he hasn't build anything, he is just trying to let other people tell him what problems he will have when he start... I think he should find them on his own and come back asking about them.
JS
-
Horror stories. Like when something really ended up being bizarre, worst case engineering fuck up that goes completely against conventional wisdom.
I honestly can't think of one, electronics has been pretty good to me. Maybe when my compiler malfunctioned and the PIC were not obeying clock directives, took me forever to figure out why serial was not working.
I heard a two stories about inductor/ferrite bead literary melting for some RF reason. (situation 1 claim was that the guy choose some kind of inductor that had bad properties at some frequency in order to get his way with the boss, because he wanted a PCB redesign but people just wanted some kind of filter inductor or something, situation 2 was someone not knowing wtf they are doing, but I don't know the specifics).
I was hoping maybe exploding transistors at least?
And about muntzing, I mean like... the signal level equivalent would be.. i don't know, trying to avoid using buffers in analog circuits? Like for most things I would advise people to go for the improved howland current pump with a feedback buffer, rather then using the regular howland current pump (but its appealing to do so because you save on an opamp, or save a part with the unbuffered improved HCP). Digital equivalent would be like, not using schmitt triggers on off board I/O signals? (usually it works but its bad practice), or maybe trying to work a switch without some kind of debounce (maybe you can do it in code acceptably, and people might not see it as a problem, but its a potential source of alot of grief because you want to save between 5 cents on passives to a few bucks on a debounce IC)
Or maybe even something as simple as not killing yourself with old single transistor amplifier designs at low frequencies unless you want something really exotic because op-amps are cheap now?
I just don't wanna run into some kind of cost related land mine with the design. The instinct is to go for chips from LT, but even then I don't know if there some kind of particular product family or something to avoid.. and it gets even more murky with high power designs. Like maybe some designs are just known to be unstable, noisy, etc in comparison to others for some nuanced reason I don't see because I don't know the lay of the land (like cheaper parts are specified), or the circuit is physically smaller.. and its just one of those things that't obvious to skilled designers.
-
To give you a meaningful answer, please specify
- Input voltage range
- output voltage (range)
- min and max output current
- maximum allowable output ripple voltage
this is about the minimum spec you have to deliver to start a design from.
-
To give you a meaningful answer, please specify
- Input voltage range
- output voltage (range)
- min and max output current
- maximum allowable output ripple voltage
this is about the minimum spec you have to deliver to start a design from.
I am looking for things that can be treated as outliers, just in case. Something that psychologically stands out (because your blood pressure raises when you start thinking about that issue that you had)
The video game equivalent might be a sewer level, jump level or insanely disproportional boss fight. :Like the cradle level in thief 3, or the fight sequences in the 2013 tomb raider being impossible on slow processors. Like there is a 'impedance mismatch' in how you deal with the problem or what you are used to. Or realizing you live in a high crime area and that most people don't consider your way of life normal (dead bolting your back door during daylight hours in nice suburbs after you move from the city).
I am guessing since this did not pop into people's heads its not that bad.
A chemistry equivalent might be starting grignard reactions. Small step can turn into a huge debacle. But you use argon, dry everything excessively, put a dessicant tube on your glassware, and it works first time (I did not even know they are hard because I followed all the recommended procedures that were not even considered necessary for the process by certain literature, then I see later people complaining heavily about wasted reagents because they don't want to add ~30$ of equipment to a 500$ setup). Like I would always do this, as a general precaution, and just not try to save the few bucks/time because I knew stupid stuff can happen even though that reaction has a 'good' reputation in a area of chemistry that generally has a hard reputation. In my head its a blanket policy that seems to make everything work right, I don't see the point of removing parts to ease my cleaning/setup requirements, even when it usually seems to work for people.
Or in a machine shop, lubricate everything but wood for drilling.. good habit IMO.
-
what did you smoke recently ? :scared:
-
?
i want to avoid smoking parts
-
OK, then I suggest that you start with a small but useful DIY project like a switcher creating a regulated 5Volts, 50mA from an AA battery.
Chances everything explodes are minimal here.
Of course, there are thousands of options available. I tried it with the following designs:
- an LT1073 with a "low noise" circuit from an LT appnote.
- some royer converters using different transistors
- flyback converters (did not work very well)
If you got confident with this, you could gradually move on to higher currents and voltages.
Example: make a 5V/2A logic supply from a car battery (10-18V).
Line powered switchers come last, definitively.
-
?
i want to avoid smoking parts
No you don't, its fun and you can learn from it. If you blow a lot of small parts you wi blow less bigger parts and it's much safer that way...
JS
-
... just to add a few arguments as the devils advocate:
Simulation is not a sin. really. >:D
-
... just to add a few arguments as the devils advocate:
Simulation is not a sin. really. >:D
It seems like it would be misleading unless you can connect the inductor to a impedance analyzer. I think the inductors are my biggest issue.
Is there like a switching inductor kit or family I can buy that are generally good for switching supplies that come with models?
I for instance own resistor sets in e96 by decade (yes its alot), this gets more difficult and less useful with capacitors (maybe you can do it good with foil and some ceramic), but (cored) inductors just intimidate me, the HF stuff seems kinda easier because you can wind the inductors yourself easily and the air always the same, and they seem to follow the helical simulator online very well.
Electrolytics are kinda bad too, because you can get so much variation in ESR with different product lines, but it typically does not matter too much. Have yet to put together some comprehensive kit of capacitors. I do have decade inductors from ebay but they are low power signal type, manufacturer unknown, so I am not thrilled about them.
There are too many transistors but I am thinking that accumulating a general switching diode set might be feasible.
-
Try coilcraft. They have inductor kits for you.
They also have demo circuits and proven designs.
-
Try coilcraft. They have inductor kits for you.
They also have demo circuits and proven designs.
Would you say there their inductor kits are in-line with the expectations of engineers from LT in general for lower power designs?
Yo, there are like 100 kits. Thats not really helpful. That's what I mean. I would need to drop like several thousand dollars to establish a prototype capability. :(
0603PS
0805PS
1008PS
1812PS
DO1605T
Coupled Essentials
DO1606T
DO1607B
DO1608C
DO1813H
DO2010
DO3308P
DO3314
DO3316H
DO3316P
DO3316T
DO3340P
DO5010H
DO5022P
DS1608B
EPL2010
EPL2014
EPL3010
EPL3012
EPL3015
LPD5030V / 8035V
LPO2506
LPO3010
LPO3310
LPO4812
LPO6013
LPO6610
LPO4815
LPS30xx
LPS3314
LPS40xx
LPS high L
LPS4414
LPS5010
LPS5015
LPS5030
LPS6225
LPS6235
ME3215
ME3220
MSS1048
MSS1246
MSS1246T
MLC12xx / 15xx
MOS6020
MSD1278
MSS1038
MSS1260
MSS1260T
MSS1278
MSS1278T
MSS4020
MSS5121
MSS5131
MSS6122
MSS6132
MSS7341
PFL1609 / 2010
PFL2510 / 2512
PFL4514 / 4517
SER1052
SER1400
SER1360
SER1590
SER2000
SLC & SLR
SLC7530
SRT8045
XAL, XFL Essentials
XAL1010 / 1060
XAL40xx XAL50xx
XAL60xx
XAL7020
XAL7030
XAL7070
XEL35xx
XEL40xx
XEL50xx
XEL60xx
XFL2005
XFL2006
XFL3010 / 3012
XFL4012 / 4015
XFL4020
XPL2010
XTL7030
-
Would you generally at least got with torroidal or other, and shielded or non shielded? That would narrow it down.
Am I going to get screwed by committing to a certain criterion here with a bunch of edge cases?
-
Another hint:
- Work at moderate switching frequencies. Components are much less critical there.
- A switch IS an RF circuit. So PCB design is important. Short leads, please, a groundplane, good, low ESR electrolytics, ...
will make your life easier.
- You can (and should) include parasitic effects like stray inductance, cap ESR, ... into your simulations
That works quite well, I tried some.
Switchers are bread and butter nowadays. At smaller power levels, they are not very hard to design if you rely
on the standard solutions provided by the manufacturers of your switcher IC. Just try !
Have fun :)
-
Also in addition to shape of core and shield/nonshield, is there some kind of particular geometry I should try to stick to building a switching inductor kit?
Like trying to stick to a certain aspect ratio in the magnetic cores dimensions (length/width/height) that will work well? There seem to be generally agreed on aspect ratios for optimal air inductors (unless you are looking to utilize a parasitic for some purpose that can probobly be handled with more parts), but the magnetic material is weird to me.
-
Why on earth would you like to *build* an inductor kit ? You can buy a nice collection of inductors from several manufacturers,
the come with datasheets and sample applications.
even eBay has such stuff:
https://www.ebay.com/itm/145Pcs-Inductor-Choke-Assortment-Kit-Plastic-Box-12Values-10uH-10mH-/183263074238 (https://www.ebay.com/itm/145Pcs-Inductor-Choke-Assortment-Kit-Plastic-Box-12Values-10uH-10mH-/183263074238)
-
I was hoping maybe exploding transistors at least?
OK I have recipe for that. Start then like every second n00b: Offline SMPS, at least 2kW to start with, 100 amps minimum. Or maybe even better, start by building your own motor inverter drive, regenerative braking back-fed to mains. >:D
Large IGBT packs spew sticky gooey snot around the workshop when they blow up...
Teaches you also really fast where NOT to stick your scope ground probe on live circuits >:D
Incendiary Tantalum caps are probably soon banned by some Geneva convention of non-conventional weapons but those are not limited just to SMPS.. :-DD
-
Why on earth would you like to *build* an inductor kit ? You can buy a nice collection of inductors from several manufacturers,
the come with datasheets and sample applications.
even eBay has such stuff:
https://www.ebay.com/itm/145Pcs-Inductor-Choke-Assortment-Kit-Plastic-Box-12Values-10uH-10mH-/183263074238 (https://www.ebay.com/itm/145Pcs-Inductor-Choke-Assortment-Kit-Plastic-Box-12Values-10uH-10mH-/183263074238)
Coilcraft has like 50 kits, you can buy the parts seperate in quantity, or try to put together a kit from various sources.
The kit you linked has no data sheets, its mystery meat. I wanted spec sheets at least. I already have something like that but its useless for a simulator
-
The eBay kit is dirt cheap, thats why it has no proper datasheet. It could still be measured :)
The coilcraft kits got everything, but are not that cheap :(
-
Messiest I saw: a heavily cookbook based design, for a module with a few more requirements than a cookbook design method can reasonably support.
When we got the prototypes in, I did a seemingly simple test: sweep DC input voltage over the rated range (8 to 80V -- this was an automotive related project).
It went through no less than three hysteresis loops (with respect to input current or output voltage, I forget what). I was impressed. Nevermind unintended behavior, this was emergent behavior above and beyond what the designer, or reviewer (me), expected!
(Those quirks were easily fixed, but the final design was still way more expensive, and under performing, versus what the customer needed. :( )
To be more specific, the block diagram was this: PMOS for reverse polarity protection on input; overvoltage detector + switched pre-regulator (80V max input, 20V output); uModule 8-30V input, 12V output, main regulator; overcurrent detect and disable; and I forget if there were some voltage monitor, temperature limit, enable or other logic around. Input and output also had MOVs attached, which I don't think actually did anything because MOV clamping voltages are stupid to begin with, but particularly bad at low voltages.
The hysteresis arose from the switched preregulator and various quirky logic. It was very much an illustration of what goes wrong when you focus on individual points in an operating space (limiting your thinking to "I need this function at this voltage"), without making any consideration whatsoever for how it will go between points, both in the transfer function and in the time domain. (You can put a MOSFET bypass across a preregulator, sure, but how long is it going to take to turn on and off? Is it fast enough to deal with a surge or swell condition? Much of the logic was resistors, zeners and logic level MOSFETs -- probably not!)
The uModule alone was an easy enough design choice (80V maximum input requirement aside), but it's one of those decisions that has consequences. That one part was half the customer's desired BOM cost already. It was extremely noisy, throwing off low ~ns pulses from both sides (it was an H-bridge "flying inductor" or buck-boost topology), and that was after my best efforts in the layout.
And by sticking to that decision early in the design process, it dictated a two stage conversion, and all the other ugly logic. The logic at least would've been avoidable with some insight (a buck regulator that can deliver 100% duty cycle is its own bypass, no need for logic), but that wasn't done.
For my part, I suggested a single stage SEPIC, as well as various improvements and tweaks to the design as it was. Mostly left on the table in the name of schedule. :-\
The moral is, go for the easy solution (often, a cookbook approach) when the problem is easy. If your early attempts are turning up lots of conflicting issues, don't try to force fit it. Take a step back, think wider about the problem, read about similar problems and their solutions, then go back in for some solutions. Read datasheets and see if you can find chips that will handle part or all of your functional blocks. If worse comes to worse, a discrete design is still perfectly doable today, but it may miss cost or size targets, and definitely requires more expertise to synthesize. Don't shy away from diagramming an RTL-equivalent* diagram, it's helpful towards understanding the system!
*In digital logic, RTL = Register Transfer Level, i.e., a schematic which consists of registers and function blocks (gates, muxes, adders, multipliers, state machines, RAM..), connected with wires and buses. The analog equivalent is using op-amps and functions and such. Fully implementing op-amps wouldn't be a fruitful exercise, but implementing a system with op-amps and such, is.
I suppose the other moral is to work quickly, on a high level. Understand the problem in the abstract. Use the simplest possible solution, no simpler (to paraphrase Einstein). Then dig down into each block, layer by layer, understanding the implementation of your abstraction. If you can't see the trees for the forest -- if you can't take a look up from the low level and see the abstraction -- you're going to spend far too much time resolving those minutiae, trapping yourself in a local minima that isn't a very good minima globally.
Maybe that's too much to ask for, I have no idea. I honestly have no idea. You can't simply ask someone to "just think smarter!" A person only has what logic and inference they do. At least, I've seen little evidence that it's something one can learn, and asking that of someone just makes them frustrated, or angry. It would fit the data just as well that they could, but don't want to, because asking is seen as offensive. One can certainly practice thought, or be given the tools necessary to harness it, more quickly and powerfully. But moving into a whole other order of magnitude, no, that doesn't seem to happen. Where is this going, other than rambling? I guess it leads to a contradiction against the cultural staple of "you can be anything you set your mind to", or "The American Dream(TM)", or however you like to put it. You should be ambitious, yes, but getting yourself into a position well above your capability is probably not a good idea in the long run; and in some fields, people die that way. So definitely don't do that.
Well, anyway.
Tim
-
Messiest I saw: a heavily cookbook based design, for a module with a few more requirements than a cookbook design method can reasonably support.
When we got the prototypes in, I did a seemingly simple test: sweep DC input voltage over the rated range (8 to 80V -- this was an automotive related project).
It went through no less than three hysteresis loops (with respect to input current or output voltage, I forget what). I was impressed. Nevermind unintended behavior, this was emergent behavior above and beyond what the designer, or reviewer (me), expected!
(Those quirks were easily fixed, but the final design was still way more expensive, and under performing, versus what the customer needed. :( )
To be more specific, the block diagram was this: PMOS for reverse polarity protection on input; overvoltage detector + switched pre-regulator (80V max input, 20V output); uModule 8-30V input, 12V output, main regulator; overcurrent detect and disable; and I forget if there were some voltage monitor, temperature limit, enable or other logic around. Input and output also had MOVs attached, which I don't think actually did anything because MOV clamping voltages are stupid to begin with, but particularly bad at low voltages.
The hysteresis arose from the switched preregulator and various quirky logic. It was very much an illustration of what goes wrong when you focus on individual points in an operating space (limiting your thinking to "I need this function at this voltage"), without making any consideration whatsoever for how it will go between points, both in the transfer function and in the time domain. (You can put a MOSFET bypass across a preregulator, sure, but how long is it going to take to turn on and off? Is it fast enough to deal with a surge or swell condition? Much of the logic was resistors, zeners and logic level MOSFETs -- probably not!)
The uModule alone was an easy enough design choice (80V maximum input requirement aside), but it's one of those decisions that has consequences. That one part was half the customer's desired BOM cost already. It was extremely noisy, throwing off low ~ns pulses from both sides (it was an H-bridge "flying inductor" or buck-boost topology), and that was after my best efforts in the layout.
And by sticking to that decision early in the design process, it dictated a two stage conversion, and all the other ugly logic. The logic at least would've been avoidable with some insight (a buck regulator that can deliver 100% duty cycle is its own bypass, no need for logic), but that wasn't done.
For my part, I suggested a single stage SEPIC, as well as various improvements and tweaks to the design as it was. Mostly left on the table in the name of schedule. :-\
The moral is, go for the easy solution (often, a cookbook approach) when the problem is easy. If your early attempts are turning up lots of conflicting issues, don't try to force fit it. Take a step back, think wider about the problem, read about similar problems and their solutions, then go back in for some solutions. Read datasheets and see if you can find chips that will handle part or all of your functional blocks. If worse comes to worse, a discrete design is still perfectly doable today, but it may miss cost or size targets, and definitely requires more expertise to synthesize. Don't shy away from diagramming an RTL-equivalent* diagram, it's helpful towards understanding the system!
*In digital logic, RTL = Register Transfer Level, i.e., a schematic which consists of registers and function blocks (gates, muxes, adders, multipliers, state machines, RAM..), connected with wires and buses. The analog equivalent is using op-amps and functions and such. Fully implementing op-amps wouldn't be a fruitful exercise, but implementing a system with op-amps and such, is.
I suppose the other moral is to work quickly, on a high level. Understand the problem in the abstract. Use the simplest possible solution, no simpler (to paraphrase Einstein). Then dig down into each block, layer by layer, understanding the implementation of your abstraction. If you can't see the trees for the forest -- if you can't take a look up from the low level and see the abstraction -- you're going to spend far too much time resolving those minutiae, trapping yourself in a local minima that isn't a very good minima globally.
Maybe that's too much to ask for, I have no idea. I honestly have no idea. You can't simply ask someone to "just think smarter!" A person only has what logic and inference they do. At least, I've seen little evidence that it's something one can learn, and asking that of someone just makes them frustrated, or angry. It would fit the data just as well that they could, but don't want to, because asking is seen as offensive. One can certainly practice thought, or be given the tools necessary to harness it, more quickly and powerfully. But moving into a whole other order of magnitude, no, that doesn't seem to happen. Where is this going, other than rambling? I guess it leads to a contradiction against the cultural staple of "you can be anything you set your mind to", or "The American Dream(TM)", or however you like to put it. You should be ambitious, yes, but getting yourself into a position well above your capability is probably not a good idea in the long run; and in some fields, people die that way. So definitely don't do that.
Well, anyway.
Tim
Do you mean the Umodules that AD sells now? The little BGA things that have the inductor built in the chip that cost 20$ each and are advertised as low noise?
Did they need paramedic to come in when someone saw the BOM cost of that part?
-
Yes, AD (formerly LT) uModule. This one was $30 in modest quantity.
Tim
-
Did they need paramedic to come in when someone saw the BOM cost of that part?
I tried to justify it once, on paper, but I believe I ended up deleting that document. :-DD
Maybe 15$ unless your building the smallest fucking thing ever. 10$ because its a BGA and no one wants to inspect it.
-
At first I was like ;D
Then I saw the price, compared the solutions ???
Then I thought about inspection and repairability :o
-
Was it one of the LOWEMI ones? Like this one
http://www.analog.com/en/products/ltm4613.html#product-overview (http://www.analog.com/en/products/ltm4613.html#product-overview)
Did it match (https://incompliancemag.com/wp-content/uploads/2011/10/1110_F5_fig5.png)
My main draw to it was that it would not be a noisy thing that requires careful layout since its all integrated.
I do find it appealing for my own use, but not for sale
-
I don't remember what the spectrum was like. I remember it needed a metal box (also for cooling, because efficiency as I mentioned) and a nice big ferrite bead looped on the cables, which sucked that much more because of the feed-thrus, and the wires being soldered in. It was all rather tight and fiddly.
Package size was still as expected, though.
Tim
-
Once I got ultrasonic coupling from corner to corner of a PCB, coming from an inverter charge pump, as not every switchmode power supply requires inductors.
The design was RF related and the component side height-constrained due to some heatsinking and standard card format requirements.
There was an very low phase noise RF synthesizer, requiring a fairly high (uF's) capacitor for its loop filter, and the only capacitors meeting height constraints were class 2 ceramics, which are piezoelectric and a no-no for this application unless you want to make a very sensitive sismograph (as we discovered later). Area constraints also didn't allow for paralleling too many caps so in the end we just chose the lowest capacitance density class2 ceramic cap that met requirements.
There was an inverter charge pump on the other side of the board which again required ceramic caps due to height constraints. It wasn't strictly required but using it slightly improved (2dB or so) linearity of the ADC preamp and the overall SFDR in consequence for some gain settings.
We power up boards and the output from the synthesizer shows -80 dBc spurs at the switching frequency of the charge pump. Touching the board at some places or installing/removing the heatsink varied the spur level, so we suspected mechanical coupling (the charge pump was separated 20cm from the synth, and shared no power rails with it). We changed the capacitor to something non-piezoelectric on one of the boards and the spurs dropped to -95dBc, but were still there. Turns out that the interference was also sneaking through the decoupling capacitors of the synth's charge pump power rail.
In the end we just removed the charge pump and it was OK.
-
thats amazing
-
How do you dampen ultrasonic vibrations anyway? Like say you have a DC DC converter box in the same enclosure?
It sounds like you can't get away from it. Does it transfer on wires too if you broke the PCB in half with tiny cable or maybe flat flex (so you can still have a ~solid ground plane? Join the two PCB halfs together with a tight but flexible braid plane?
How about isolating it from the floor, even if you manage to break it up some how? All those springs and stuff are meant for low frequencies, will they help? Or transmit through?
Does anyone have any figures or graphs relating to ultrasonic energy transmission through a PCB so spacing guidelines can be established?
-
In my early days of repairing stuff, switchmode PSUs were somewhat mysterious and hard to diagnose. Once I had an ESR meter and knew how to use it, I started having much better success. By far the most common problems I see are bad electrolytic capacitors. The most tricky to diagnose types I've encountered are some of the older self-oscillating designs that lack any sort of controller. The one Tek used in the TDS400 series scopes is also a real bear, half the problem is the lack of good documentation and the stupid ceramic hybrid module.