The TL431 adjustable shunt regulator IC is sometimes used as an inverting amplifier; its datasheet even includes a chart of open loop gain versus frequency. Occasionally you'll see a regulated power supply circuit which embeds a TL431 inside the feedback loop of an IC voltage regulator like an LM317; this cascade of amplifiers increases the regulator's open loop gain at low frequencies (like 60 Hz!) by >50 dB, which improves ripple rejection.
I was curious to find out just how low the TL431 "cathode" terminal could fall, while still maintaining good regulation, i.e., the TL431 Ref terminal remains solidly pegged at 2.50 volts. So I plugged together the little circuit shown below, on my solderless breadboard.
A DC power supply sets VCC=20V. A function generator provides a 12V peak-to-peak triangle wave at 200 Hz, which drives the TL431 by means of an NPN emitter follower. Resistors R1 and R2 form a voltage divider whose midpoint drives the TL431's "Ref" terminal.
Scope picture 2 shows nodes REF and KK (sorry, it's just a 2 channel scope!). When REF levels off, KK starts to fall. Eventually the TL431 drops out. A magnified picture of dropout is shown in picture 3.
I've put horizontal cursor A on the REF node @ 2.500 volts, and horizontal cursor B on the cathode node KK, at the point where dropout occurs, in my opinion. REF is no longer regulated, it begins to rise. KK continues to fall a bit further, and then Boom! it all blows up in ugly oscillation. The dropout voltage on the cathode is 1.96 volts, as shown in the cursor popup-box.
The device measured was an ST Microelectronics TL-431I. This is a lowest-accuracy-bin (2%) part, in a TO-92 AmmoPack package with 0.1" bent leads.
_
I was curious to find out just how low the TL431 "cathode" terminal could fall, while still maintaining good regulation, i.e., the TL431 Ref terminal remains solidly pegged at 2.50 volts. So I plugged together the little circuit shown below, on my solderless breadboard.
A DC power supply sets VCC=20V. A function generator provides a 12V peak-to-peak triangle wave at 200 Hz, which drives the TL431 by means of an NPN emitter follower. Resistors R1 and R2 form a voltage divider whose midpoint drives the TL431's "Ref" terminal.
Case 1: Suppose node "TRIANGLE" (the top of R2) is at 2.7 volts. Node "REF" is 2.5 volts, established by the TL431, and so a current of (2.7-2.5)/330 = 600uA flows in R2. This same current flows in R1, so node KK equals (2.5 - (100*6e-4)) = 2.44 volts. The TL431 does not "drop out" (cease regulating) when its cathode is so very close to 2.50 volts.
Case 2: Now suppose node "TRIANGLE" is at 11 volts. If node "REF" is 2.5 volts, then the current in R2 must be (11-2.5)/330 = 25.7 mA. The voltage dropped across R1 must be (100*2.57e-2) = 2.57 volts, which means the cathode "KK" is below ground! Thus we have a contradiction; node "REF" cannot be 2.5 volts in this case, so the TL431 has "dropped out" (ceased regulating).
I've attached scope photos showing the entire ramp from 0 volts to 12-Vbe volts and back down to 0. Scope picture 1 shows the TRIANGLE node in blue @ 5.0V/div, and the REF node in yellow @ 0.5V/div. There's only a small zone where REF sits at its regulated output voltage, 2.50 volts -- that's the flat spot where REF is a horizontal line. Below that, the input is too low; above that, the TL431 "drops out" -- its cathode voltage falls so low that regulation ceases. As the input voltage rises, the current thru R1 rises, the cathode "KK" of the TL431 falls, and the circuit oscillates (!).Case 2: Now suppose node "TRIANGLE" is at 11 volts. If node "REF" is 2.5 volts, then the current in R2 must be (11-2.5)/330 = 25.7 mA. The voltage dropped across R1 must be (100*2.57e-2) = 2.57 volts, which means the cathode "KK" is below ground! Thus we have a contradiction; node "REF" cannot be 2.5 volts in this case, so the TL431 has "dropped out" (ceased regulating).
Scope picture 2 shows nodes REF and KK (sorry, it's just a 2 channel scope!). When REF levels off, KK starts to fall. Eventually the TL431 drops out. A magnified picture of dropout is shown in picture 3.
I've put horizontal cursor A on the REF node @ 2.500 volts, and horizontal cursor B on the cathode node KK, at the point where dropout occurs, in my opinion. REF is no longer regulated, it begins to rise. KK continues to fall a bit further, and then Boom! it all blows up in ugly oscillation. The dropout voltage on the cathode is 1.96 volts, as shown in the cursor popup-box.
The device measured was an ST Microelectronics TL-431I. This is a lowest-accuracy-bin (2%) part, in a TO-92 AmmoPack package with 0.1" bent leads.
_
Attachments
Officialy the dropout may be no lower then Vref.It shows here that in this case you can go down as low as 2V.Very nice but reliable ?
Mona
Mona
It is a behavior I had already noticed
To be avoided, clearly.
In my experience, it is highly variable, some types behaving almost as darlingtons, whilst others only tolerating a few mV of leeway.
To be avoided, clearly.
When you do, be sure to turn down the function generator's amplitude; the TLV431 is only rated for 7 volts max from cathode to anode. The 11+ volts in the scope photos above, would be fatal.
You may want to redesign the resistor values too, since the TLV device is spec'd for a lower IKmax and a very much lower IKmin, than the TLC device. If you intend to operate just a wee bit above IKmin=100uA in your final circuit, you probably want your sweep testing of dropout to emphasize that region too.
Or just buy a handful and fool around! If you blow one up, change something and start over. If the currents are lower than you'd like, swap in other resistor values. If the currents are higher than you'd like, swap in other other resistor values. If the slopes displease you, change the resistor ratio. Experiment and fiddle and have fun. If you destroy 5 parts during your falderal, that's less than 3 dollars. Whoopy.
Last edited:
- Status
- Not open for further replies.