Hi, before you actually build anything, I recommend intense simulation with (for example) SwitcherCAD from Linear Technology. If you have access to a book Fundamentals of Power Electronics by Erickson et al, you should have enough basic information about the design constraints.
Personally I tend to start with the power devices and transformers. If you have somewhat clear specifications for your project, you can estimate the maximum voltage & current stresses and choose component based on the calculations. Once you have selected components for the power section, you can proceed with filters and control circuits.
I guess there are no single best starting point, but since the power section is perhaps the most crucial, you may want to check that first. My knowledge on this subject is rather limited, at least compared to experts like Eva, for example - so don't believe only me
Good luck on your project(s)!
Personally I tend to start with the power devices and transformers. If you have somewhat clear specifications for your project, you can estimate the maximum voltage & current stresses and choose component based on the calculations. Once you have selected components for the power section, you can proceed with filters and control circuits.
I guess there are no single best starting point, but since the power section is perhaps the most crucial, you may want to check that first. My knowledge on this subject is rather limited, at least compared to experts like Eva, for example - so don't believe only me
Good luck on your project(s)!
Hi,
The best, & most accurate simulator for SMPS is prototype board only
Trust me, that even with best in class software, you will not be able to simulate the whole design if you don't ABSOLUTLEY familiar with SPICE, physics of magnetics, RF etc. Because the performance of SMPS is strongly depending on a lot of things and conditions like PCB layout, temperature of your magnetics, quality of components and many others, that very difficult to simulate. What you can do - to simulate important parts of your project: feedback networks, filters, etc.
The best, & most accurate simulator for SMPS is prototype board only
Trust me, that even with best in class software, you will not be able to simulate the whole design if you don't ABSOLUTLEY familiar with SPICE, physics of magnetics, RF etc. Because the performance of SMPS is strongly depending on a lot of things and conditions like PCB layout, temperature of your magnetics, quality of components and many others, that very difficult to simulate. What you can do - to simulate important parts of your project: feedback networks, filters, etc.
Dear jegandren,
Since you are stepping into the enormous world of power electronics, you can start from any where as you have to learn, try,mistake, learn,try........... But follow the safety rule in toto. Since you have done some simulation, you have gathered some idea, though the simulation may not be very useful in practical field. Try to gather working knowledge regarding the parts, circuits etc. In power electronics, knowledge means 100% knowledge regarding the circuit you try to make. Even 1% mistake is fatal here.
If you decide to step into, don't hesitate. Even the tallest building in the world starts from the first brick.
First, you can try with the proven circuits given in cookbook/internet.
Gather knowledge, follow safety rules, and try again and again. With the passage of time you will be able to do, what you thing impossible now. Don't listen to those so called experts, who discourage. Nothing is impossible but can take some time. Man can do what men have done.
Matsusita, the founder of National Panasonic Company of Japan didn't attend any school and started his career as a street light bulb boy whose duty was to remove the damaged bulbs from street of Tokyo. But with the passage of time he invented many electrical items and created the big company single handedly.
------------------------------------------------------------------------------------
IDEAS ARE MONEY, IF YOU CAN USE IT ON GROUND, NOT IN FORUM OR CLASSROOM ONLY.
-----------------------------------------------------------------------------------
Since you are stepping into the enormous world of power electronics, you can start from any where as you have to learn, try,mistake, learn,try........... But follow the safety rule in toto. Since you have done some simulation, you have gathered some idea, though the simulation may not be very useful in practical field. Try to gather working knowledge regarding the parts, circuits etc. In power electronics, knowledge means 100% knowledge regarding the circuit you try to make. Even 1% mistake is fatal here.
If you decide to step into, don't hesitate. Even the tallest building in the world starts from the first brick.
First, you can try with the proven circuits given in cookbook/internet.
Gather knowledge, follow safety rules, and try again and again. With the passage of time you will be able to do, what you thing impossible now. Don't listen to those so called experts, who discourage. Nothing is impossible but can take some time. Man can do what men have done.
Matsusita, the founder of National Panasonic Company of Japan didn't attend any school and started his career as a street light bulb boy whose duty was to remove the damaged bulbs from street of Tokyo. But with the passage of time he invented many electrical items and created the big company single handedly.
------------------------------------------------------------------------------------
IDEAS ARE MONEY, IF YOU CAN USE IT ON GROUND, NOT IN FORUM OR CLASSROOM ONLY.
-----------------------------------------------------------------------------------
jegandren said:
Try searching this forum for threads on half-bridge PSUs. Perhaps the most commonly implemented smps is found in every AT & ATX pc smps: The TL494-controlled half-bridge. Also, try this link. This should serve as a good starting point. Voltage-mode w/feedback, soft-start, current-limiting, all the basic stuff you need. No current-mode (cycle-by-cycle current-limiting), but you can't have everything. This circuit has no PFC, but there's quite a bit of info on the web (and this forum) on PFC.
Before you start anything, please be advised: Voltages in these ckts are LETHAL, and you are advised to exercise to appropriate caution.
Hope this helps.
Cheers,
Steve
THE FIRST STEP IN MAINTAINING SAFETY IN SMPS
Hi jegandren,
While designing and working on smps, provide a discharge path for the reservoir/filter capacitors. These capacitors can store a huge amount of energy and if touched can release that energy on you, which may be fatal.
The discharge path could be a high value resistor or a resistor in series with LED in both input and output section of your circuit. I personally prefer the later as after input supply is removed, the LEDs will indicate whether the Capacitors have discharged or not. Wait till the LEDs go off and after that you will have full freedom to put your hands anywhere in the circuit without any sort of risk of electric shock. Electric shock in smps can be fatal and please don't allow yourself to have that experience.
_____________________________________________________________________________
IDEAS ARE MONEY IF THEY CAN BE USED ON GROUND.
বাস্তবে ব্যবহৃত চিন্তাধারাই মানুষের প্রকৃত সম্পদ।
Hi jegandren,
While designing and working on smps, provide a discharge path for the reservoir/filter capacitors. These capacitors can store a huge amount of energy and if touched can release that energy on you, which may be fatal.
The discharge path could be a high value resistor or a resistor in series with LED in both input and output section of your circuit. I personally prefer the later as after input supply is removed, the LEDs will indicate whether the Capacitors have discharged or not. Wait till the LEDs go off and after that you will have full freedom to put your hands anywhere in the circuit without any sort of risk of electric shock. Electric shock in smps can be fatal and please don't allow yourself to have that experience.
_____________________________________________________________________________
IDEAS ARE MONEY IF THEY CAN BE USED ON GROUND.
বাস্তবে ব্যবহৃত চিন্তাধারাই মানুষের প্রকৃত সম্পদ।
If you want to get your feet wet with switch mode supplies, may I suggest first starting with something relatively harmless - like a DC/DC upconverter. Something that starts with a (current limited) lab supply and goes to something that won't kill you if you touch something. Once you can reliably overcome the inevitable newbie hurdles, and the pile of vaporized HEXFETs you'll get along the way, then move on to something off-line.
Tolik said:
I don't want to rain on anybody's parade, but SMPS design is not to be undertaken lightly.
Before you embark on such a project, you should have a clear idea of what your objective is. Then think whether you can achieve the same result with a conventional power supply. These are much better suited to home construction.
SMPS are often problematic in terms of emc. EMC.
EMC regulates the permitted emissions of equipment.
It is the designers responsibility under European law to ensure that devices do not emit excessively. This legislation is part of the criminal code, and successful prosecutions can result in jail sentences.
I am not suggesting that you run the risk of prosecution, but you should be aware of the potential antisocial consequences of your designs.
Fullscale EMC testing is impractical (uneconomic) for amateurs.
Unless you anticipate a career in SMPS design, it is a field better avoided.
w
OK, you may not be in Europe...
Tolik said:
It's not a good idea to trivialise this subject.
Before EMC legislation background noise levels were rising globally to the point where, if unrestrained, the normal functioning of many devices we take for granted such as mobile phones might nowadays be difficult or impossible.
The legislation was successful in reversing this rise.
If you operate an untested SMPS you run the risk of minor or even severe interference to your neighbours and even, in rare instances, to safety critical services.
In my book, this constitutes antisocial behaviour, however minor.
Radio noise is environmental pollution.
Do you deny that this is the case?
w
Tolik, I'm not taking a dig at you. One of the things we most need prototypes for is EMC test.
I agree...
Generally we are talking about power supplies for audio equipment. Excessive EMI, will first interrupt functioning of whole audio system... So building of LOW noise PSU is main target here. Next, almost 80% SMPS discussed here, are low power - less than 300...500W, and almost neither of these projects not going to be produced in quantities more than 1-2 units. I don't really think, that several, even bad designed, low/ medium power supplies are containing a threat to ecology, or communications.
You better tell us how to measure or avoid EMI/RFE . It will be much more intresting here.
Typical EMI probes can be seen here
Uncalibrated probes can be fabricated quite readily from rigid co-ax, although the sheild must be perforated to permit correct connection in the case of the loops. You can find the connection details in e.g. the ARRL Antenna Book. The straight probes simply have a small stub of the inner conductor exposed.
It is unfortunately not possible to quantitatively test equipment in the absence of a quiet environment, which usually means a shielded room.
It is an unusual SMPS that passes EMC test on first submission, be it ever so small.
w
Most SMPS fail on conducted emissions. These can be detected by coupling to the supply cable. Oh, keep your current loops small. Watch your risetimes. IIRC CISPR B testing goes up to 2GHz. You ain't gonna hear that on no audio system...
Uncalibrated probes can be fabricated quite readily from rigid co-ax, although the sheild must be perforated to permit correct connection in the case of the loops. You can find the connection details in e.g. the ARRL Antenna Book. The straight probes simply have a small stub of the inner conductor exposed.
It is unfortunately not possible to quantitatively test equipment in the absence of a quiet environment, which usually means a shielded room.
It is an unusual SMPS that passes EMC test on first submission, be it ever so small.
w
Most SMPS fail on conducted emissions. These can be detected by coupling to the supply cable. Oh, keep your current loops small. Watch your risetimes. IIRC CISPR B testing goes up to 2GHz. You ain't gonna hear that on no audio system...
I have seen hugely excessive EMI levels (20-30 dB over the B-requirement level 30/37 dBµV) even with very small non-isolated power supplies. I don't think that such excess will affect the audio system in any other way than FM/AM-reception. Most problematic frequencies are generally around 50-200 MHz.
Personally, for me, low EMI levels indicate carefully designed circuit/PCB and gives also other performance benefits. I feel that excessive EMI-levels are a clear indication that something is very wrong and must be fixed.
Most of the cases I have seen, it is rarely the PCB itself which radiates, but the cables attached to the PCB form an effective antenna for common-mode noise. Differential mode is seldom a problem, due to small loop areas. Common mode radiation is much harder to suppress.
A very good cheap test for EMI (requires reasonably wideband oscilloscope, though) is to short circuit a oscilloscope probe right at the tip to the ground collar. Throw that ground clip away though, it is pretty much useless for SMPS and digital measurements due to excessive loop area it creates, thus it picks all the magnetic noise.
Then connect this short circuited probe to the ground of the PCB. Now, do you see any high-frequency noise at mV levels? If you do, then don't bother to try EMC-testing until you have fixed what you see using this test. Remember, it only takes few µA of common mode RF-current to create high EMI levels. Higher the frequency you see, then more serious the issue is.
Regards,
Janne
Personally, for me, low EMI levels indicate carefully designed circuit/PCB and gives also other performance benefits. I feel that excessive EMI-levels are a clear indication that something is very wrong and must be fixed.
Most of the cases I have seen, it is rarely the PCB itself which radiates, but the cables attached to the PCB form an effective antenna for common-mode noise. Differential mode is seldom a problem, due to small loop areas. Common mode radiation is much harder to suppress.
A very good cheap test for EMI (requires reasonably wideband oscilloscope, though) is to short circuit a oscilloscope probe right at the tip to the ground collar. Throw that ground clip away though, it is pretty much useless for SMPS and digital measurements due to excessive loop area it creates, thus it picks all the magnetic noise.
Then connect this short circuited probe to the ground of the PCB. Now, do you see any high-frequency noise at mV levels? If you do, then don't bother to try EMC-testing until you have fixed what you see using this test. Remember, it only takes few µA of common mode RF-current to create high EMI levels. Higher the frequency you see, then more serious the issue is.
Regards,
Janne
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.