Group Actions
Engagement
  • Author Author: jw0752
  • Date Created: Date Created
  • Views 1170 views
  • Likes 5 likes
  • Comments 8 comments
Related
Recommended

Exploring the Voltage Sense on Small wall wart SMPSs

jw0752

image

 

I am studying the similarities and differences in the circuits  of these Switch Mode Power Supplies from a 5 Volt, 6 Volt, and 12 Volts wall warts. I want to find out how they create their output voltages and how they regulate these voltages. One of the similarities that I noticed right away was that each unit had an Opto-coupler between the primary high voltage side and the secondary low voltage side. In my study of the three boards two had a TL431 Precision Programmable Reference and one had a Zener Diode tied to the LED side of the Opto-coupler. Here are rough approximations of the two variations of the circuit that I found:

 

image

 

The variation of the circuit that uses the Zener diode has a current limiting resistor in series with the Zener which is also in series with the LED side of the Opto-coupler. As the voltage on the output rises and exceeds the value of the Zener plus the forward junction voltage drop of the LED current begins to flow and the LED turns on. This is turn causes the photo transistor in the Opto-coupler to begin to conduct. The conduction of the photo transistor must be tied back to the primary side of the SMPS and provide negative feedback to the circuit generating the primary current.

 

image

 

The SMPS that use the TL431 instead of the zener have basically substituted the TL431 for the Zener. The TL431 is quite a versatile component but in this application it is being used as a Zener which can be programmed by the resistor voltage divider between its Cathode, Reference, and Anode terminals. I decided to look a little more closely at the TL431. Here is a Data Sheet for the component and a link to Newark's listing of the part:

 

http://www.newark.com/on-semiconductor/tl431aclpg/voltage-ref-shunt-2-495v-36v-to/dp/88H5149

 

http://www.farnell.com/datasheets/1904582.pdf?_ga=1.162928828.1418853850.1451543312

 

You will note in my schematic fragment I have drawn the resistor voltage divider as a potentiometer. On the actual boards the resistor are simply small surface mount resistors but since I wanted to experiment with the part I used a 10K potentiometer on the bread board where I began to test one of these cool little components.

 

My first experiment simply involved setting up the circuit as shown in the above schematic. I applied an input voltage of 24 volts and chose my R1 so that the 20 mA tolerance of the Opto-coupler could not be exceeded. The TL431 itself can handle a Max current of 150 mA. I placed my voltmeter between anode and the cathode of the TL431. True to the Data Sheet specs the voltage on the cathode began at 2.5 volts when the reference pin was tied to to the cathode and the voltage continued to rise as the potentiometer moved the reference voltage closer to the anode. Whenever I stopped on a specific regulated voltage I found that I could then adjust the input voltage up or down without any appreciable effect on the voltage across the TL431. It was doing a good job of providing a stable voltage reference. I next put an ohmmeter on the photo transistor terminals of the Opto-coupler and watched as it began to conduct as the input voltage surpassed the set voltage of the TL431 plus 1.2 Volts for the forward junction drop of the coupler's LED.

 

Note: The TL431 has an internal 2.5 volt base reference which keeps it from working below this level. The Max input voltage is listed at 37 volts and the data sheet says that it can provide a reference voltage between 2.5 V and 36 V.

 

The next step in my experiment was to see if I could build a simple linear regulator using the TL431 and the Opto-coupler as feedback. I bread boarded this circuit and began to test its limits.

 


image

 

Here is a picture of the bread board and shots of the meters involve in the test:

 

image

 

image

 

You can see that I am supplying an input voltage of 20.9 volts and my output voltage from the regulator is 6 volts. The 10.7mA input current of the circuit is displayed on the meter on the left. I was able to select output voltages from 4 volts up to 13 volts by adjusting the potentiometer and changing the voltage of the TL431. This experiment provided me with verification of my understanding of the TL431 in this application and proof of concept in how the output of the Opto-coupler could be used to control other components to regulate an output voltage. On the actual wall wart circuit I am sure I will find a small dedicated IC that is providing PWM current to the primary of the transformer. The IC will likely be controllable by the output of the opto-coupler. I will continue to further investigate this to verify the exact mechanism involved but that will be for another day. One question that occurred to me was why would a manufacturer use the TL431 and support resistors when they could just use a zener. It was apparent that the circuits that I am looking at were designed and built with the minimum cost necessary. My only thought is that the TL431 itself costs only pennies and perhaps the ability to fine tune its voltage with resistors and not have to buy an custom voltage zener is actually cost effective. I have ordered a bunch of the TL431s as I can see that they will be useful in a variety of applications in the future.

 

John

  • Hi John!

     

    Great article! The TL431 is really cheap as you say, so is extremely popular. By the way there is also a less-popular part called TLVH431 which is a lower-Vref (1.24V) equivalent (it doesn't come in TO-92 package though).

  • Inspired by I decided to see if I could make some generalizations related to heat dissipation in this simple circuit involving a current limiting resistor and a Zener Diode.

     

     

     

    image

    For the purposes of description of total energy dissipation of R1 and ZD1 I will call this circuit "The System". The energy dissipation of R1 will be called PWR1 and the energy dissipation of ZD1 will be referred to as PWZD1. For simplicity's sake I am going to assume that all components are linear over the voltage ranges involved and that their values are not affected by heating. This false assumption of course puts my results in the category of rough guesses at best.

     

    I have run a series of calculations on this circuit using a fixed voltage for Vin and a fixed resistance for R1. From the results of these calculations I have extrapolated to a general model and I have found that the following generalizations can be made:

     

    No. 1 For a given voltage Vin and a constant current limiting resistor R1 the Zener ZD1 experiences the maximum energy dissipation when its value is one half of Vin.

     

    No. 2 Total energy dissipated in The System increases as the value of ZD1 decreases from Vin to zero volts. When the value of ZD1 = Vin the energy dissipation is zero watts and when the value of ZD1 = 0 the dissipation is (Vin)^2 / R1 watts.

     

    No. 3 Maximum Current in the system is limited to Vin / R1 for all values of ZD1.

     

    No. 4 Max Energy dissipation of R1 is  (Vin^2) / R1 watts.   or   Maximum PWR1  = (Vin^2) / R1 watts

     

    No. 5 Max Energy dissipation of ZD1 is (Vin^2) / (4*R1) watts. or  Maximum PWZD1  = (Vin^2) / (4R1) watts

     

    These of course are all simple results using Ohms law and related power calculation formulas. This exercise was needed to help me better understand some of my own questions after reading what   wrote above in his caution to me above about heat and power considerations. Since I had them on paper I thought I would share them.

     

    John

  • in reply to jc2048

    Hi Jon,

    You are 100% correct on your heat dissipation analysis of this circuit. As is customary for people who aren't quite there, knowledge and experience wise, I have over simplified a complex system. Since most of my experiments would use a zener in a low current situation and seldom over 15 volts I should be OK but I will include your limits into the documentation that I make up for the circuit. Once again thank you for taking the time to clarify for everyone that the circuit has limits and can not be expected to handle 36 volts at 100 mA. Also it was nice to see that the idea in a simpler form has been around for a long time.

    John

  • in reply to jw0752

    Here's an older (discrete) form of the variable zener.

     

    image

     

    It has a long history. This is from 'Handbook of Linear Integrated Electronics for Research' by T.D.S. Hamilton, which was published in 1977, but it wasn't original to him.

  • in reply to jw0752

    "This simple little circuit will allow me to dial in any shunt voltage level between 2.5 volts and 35 volts with a ability to sink 100 mA."

     

    Not quite.

     

    It is a useful circuit, but whilst it can manage up to 35V and 100mA, it can't necessarily do both at the same time.

     

    You aren't considering the thermal management side.

     

    The datasheet says the little TO92 package has a junction-to-ambient thermal resistance of 140C/W. It also says the maximum junction temperature is 150C.

     

    Let's say you are designing just for your lab (with the board, as you show it, open to the elements) and that the operating temperature is always 40C or less (this doesn't apply if the device is in a box that can heat up). 150 - 40 = 110C difference. 110 (C)/140 (C/W) = 0.79W. So that's an absolute max of 0.79W dissipation.

     

    So, for a current of 100mA, the maximum voltage would be 0.79/.1 = 7.8V.

     

    For a voltage of 35V, the maximum current would be 0.79/35 = 22.6mA

     

    For lab use, the best thing might be to draw a derating curve (I vs V), so you could just read off the maximum current for a particular voltage rather than have to calculate it. 100mA up to 7.8V and then decreasing linearly to 22.6mA at 35V.

     

    Bear in mind, though, I've calculated the absolute max and it's not good to run the device at that limit (it will have a major impact on the long-term reliability, even if the device can cope in the short term) so, if your lab really gets up to 40C, perhaps redo the calculation for a junction temperature in the region of 120C or 130C.

     

    Another alternative, of course, if you need the 100mA rating at higher voltages, would be to chose one of the other packages with a lower thermal resistance.