Increasing wattage of a SMPS circuit

[maheshkuruganti]

smps circuit

Hi,I want to make a SMPS Circuit with Mains input and 35V/2A Output.I found a Application Note based on MC44608.The application note is AND8034-D
www.onsemi.com/pub_link/Collateral/AND8034-D.PDF. I just want to know what the various outputs are :?:and also what the maximum current is?.Also can the TLC431 be replaced.Thank YouAlso What do you think the PCB Track width should be.

 

[melc]

500w smps 회로도

Read datasheet. There is one usefull output 35V. FBO is a voltage used for a microcontroller AD measurement and REG is the feedback that microcontroller is providing for modifying the output voltage. SBY is a standby input.
It might be a 70W SMPS. About the trace routes you have to dimension that based on the current flow in every trace and copper thickness.TLC431 is a variable zenner, equivalent with LM431 or TL431. Is a cheap component, you can't replace with anything cheaper...

 

[maheshkuruganti]

lm431 circuit

Thank You for the reply.I have another question.How do you know it is a 70W supply.Also can I increase the wattage by simply increasing the gauge of the windings.Thank You.


P.S:The question may be foolish but I am desgning my first SMPS with help ofcourse

 

[DrBob13]

500w smps dadasheet

Hi,
Be very careful about replacing any components in an established SMPS design I totally agree with melc about the TLC431, these are readily available and the only type of component that you should use in the configuration shown.
Regarding output power 70W might be achievable with the circuit shown as melc said. The switching FET will give about 5A continous output current with average heatsinking, this is pulsed so the pulses could be a bit higher as per the device datasheet.
Increasing the power is a bit more complicated than just upping the wire guage, the size of the core is important too and increasing the wire size only works up to a point because of the skin effect. This means that only the outer skin of a wire is carrying current at the high switching frequencies. This is worked around by using several strands of smaller insulated wire in parallel to make up the required current carrying capacity, use wire of about 0.3mm dia or less and as many strands as you need to carry the pulse currents (the books still talk about using Litz wire for this purpose - if you can get it!) design the wire size based on about 3 - 5A per square mm (US designers use 400 to 1000 circular mill per A) (a lower current density means it runs cooler on full load). Designing the magnetic components is OK if you are reasonably up to speed with magnetics but the terminology can seem a little daunting at first, follow the design advice of the manufacturers of the core materials and you will be OK, do not try to be a hero (just yet) and use the material outside its recommended limits, there are cases when this is done but only do it when you know the reasons why. There is a lot of info on this available on the web and books are available, my favourite is Practical Switching Power Supply Design by Marty Brown of Motorola Inc. Use a spreasheet or Mathcad to design the magnetics etc. You can design everything this way and have the whole set of of the design calculations on the sheet. which is very useful for the next design you do, you can just copy it and edit the new version, this will save hours.
Have a look at the website of your chosen magnetics manufacturer, (Ferroxcube, Fair-rite, Magnetics Inc, Siemens etc) most of them have app notes or even design software you can download to enable you to design the transformer / inductors, make sure you can easily get the recommended cores locally in small numbers before committing to a design incidentally). Increasing the output power much above 100W normally requires a diffierent topology to the flyback that is normally used for isolated low power SMPS (Half bridge Forward topology converers are good to about 500W, Full bridge designs for higher powers).
As you probably realise the design of SMPS are tricky but there is no reason why you should not succede if you follow it through logically. The use of design calculations are very important for a SMPS, do not try to design one by guesswork, modifying an existing design unless it it a very minor mod, often even what seems like very minor mods can have a disasterous effect on a design's validity. I have designed a number of these over the last few years and I still often need to teak the design slightly at prototype stage so don't worry if you need to try twice. Be ready to switch off quicky if it smokes, usually though you dont get time because the FET blows first and usually manages to protect the main fuse even if it is of the correct value if anything is seriously wrong with the design / build!

ONE SERIOUS WORD OF CAUTION - THIS IS A MAINS POWERED SMPS WHICH USES DIRECT RECTIFIED UN-ISOLATED MAINS POWER - TREAT IT WITH EXTREME CAUTION WHEN DEBUGGING IT. THE USE OF A 1:1 ISOLATION TRANSFORMER WITH THE SECONDARY GROUNDED AT ONE END MAY BE WORTH CONSIDERING FOR THE DEBUG STAGE. This will not prevent a shock if you touch the wrong place but it may make it less likely to be lethal.
Good luck
Bob.

 

[FvM]

smps transformer design software

Actually, the output power is said in the AN to be 20 W maximum. Apart from component power capabilities, the dimensioning of current sense R11 sets an upper limit.

 

[DrBob13]

smps transformer winding

Thanks for pointing that out FvM, I rambled on and completely forgot to mention about the current sense resistor in my previous post.
For the benefit of maheshkuruganti who is a SMPS newbie, remember to calculate its value based on the peak FET pulse current, not the average current calculated from your input power, the data sheet gives you the calculation for this, it is based on the standard 1V curent sense voltage, its wattage is calculated in the normal way for the average power that it dissipates allowing for normal de-rating (at least 50%) to prevent excessive heat-up of this component. It is easy to forget how much power this component dissipates, so do ensure that it's power rating is large enough and use a 'non-inductive' (i.e. not a wirewound) resistor.
Bob.

 

[maheshkuruganti]

smps 회로도

I get my transformers winded outside so if get a 100W SMPS Transformer with the same Turns Ratio and change the current sense resistor can I get 35V/2A.I dont want build a new Flyback Converter as I have read the Power Supply Cookbook by Marty Brown and the Flyback Voltage seems to be pretty high and I dont want to risk it.I am bascially designing a Programmable Bench Power Supply.

 

[DrBob13]

lm431 smps

The design in your original reference does actually use flyback topology, you can tell this from the transformer winding start marker dots. The flyback topology dumps energy stored in the core inductance into the secondary winding when the FET is non conducting. This topology is in theory good to at least 150W but I am sorry that I will not say for certain, given the information that you have provided that the modifications you propose will produce a working design. The turns ratio of the transformer will be equal or very close to what you have in the reference design but changing the core for a bigger one will change its Al value so the inductance will be different, this will mean a different number of primary turns to produce the same core inductance and hence although the turns ratio will be the same the actual number of turns on both the primary and secondary windings will not.

This is not like designing a simple 50/60 Hz linear PSU, the transformer will have to be designed from first principles by either you or your transformer manufacturer, I suggest you discuss your design requirement with them and they will allocate an application engineer to help you design it.

Changing the current sense resistor for a lower value will enable more power to be drawn provided that the rest of the design is capable of delivering it.

Bob.

 

[maheshkuruganti]

smps circuit transformer design software

I can design a Flyback SMPS but my biggest worry is whether or not my Transformer Supplier can make the required Flyback Transformer.Because if it is not made properly the Voltage Spikes might be too large.One of the consequences of living where I live is that you dont have the cores of Siemens,Magnetics etc.There are local core manufacturers but then thier cores do not have as much of info as that of international makers.Can you point out special considerations for a Flyback Transformer

 

[FvM]

smps circuit mains

I basically think, it's not too complicated. The nominal inductance values can be read from SMPS applications circuits, you could also scale the value given with the present circuit according to your designed power.

As a design criterion, ferrite losses at the present frequency and maximum magnetization current should be acceptable. A probably unknown property is leakage inductance that causes the said voltage spikes. With this design, the free-wheeling spikes are absorbed by the RCD-snubber.

If the peak voltage at the switch transistor respectivly the snubber losses are in a tolerable region, leakage inductance can be regarded acceptable. However, when scaling output power and peak current, the snubber losses are increasing, it may be necessary to redimension this circuit detail to maintain safe operation area for the transistor.

For an off-mains SMPS and its respective isolation requirements, some leakage inductance can't be avoided in transformer design. The circuit has to deal with.