Feed aggregator

The TL431 has been around for nearly 50 years. During those decades, while primarily marketed as a precision adjustable shunt regulator, this legacy device also found its way into alternative applications. These include voltage comparators, audio amplifiers, current sources, overvoltage protectors, etc. Sadly, in almost every example from this mighty menagerie of circuits, the 431’s “anode” pin sinks to the same lowly fate. It gets grounded.

Wow the engineering world with your unique design: Design Ideas Submission Guide

Current sink regulation

The design idea presented here offers that poor persecuted pin a more buoyant role to play, Figure 1.

Figure 1 The floated anode serves as a sense pin for active current sink regulation. 

The Figure 1 block diagram shows how the 431 works at a conceptual level where: 

Sink current = Is = (Vc – 2.5v)/R1 = 0 to 1/R1 as Vc = 2.5v to 3.5v
Vs < 37v, Is < 100mA, Is(Vs – R1Is) < 500mW @ 50oC ambient

Series connection adds an internal 2.5-V precision reference to external voltage input on the ANODE pin. The op-amp subtracts this sum from the voltage input on the REF pin, then amplifies and applies the difference to the pass transistor. If the difference is positive (sum < REF), the transistor turns on and shunts current from CATHODE to ANODE. Otherwise (sum > REF), it turns off.

If the 431 is connected in the traditional fashion (REF connected to CATHODE and ANODE grounded). In that case, the scheme works like a shunt voltage regulator should, forcing CATHODE to a resistor-string-programmed multiple of the internal 2.5-V reference voltage. But what will happen if the REF pin is connected to a constant control voltage (Vc > 2.5 V); and the ANODE, instead of being grounded, floats freely on current sensing resistor R1?

What happens is the current gets regulated instead of the voltage. Because Vc is fixed and can’t be pulled down to make REF = ANODE + 2.5, ANODE must be pulled up until equality is achieved. For this to happen:

Is = (Vc – 2.5v)/R1 

Constant current sink regulation of 1/R1

Figure 2 illustrates how a fixed voltage divider might be used (assuming a 5-V rail that’s accurate enough) to use a floated-anode Z1 to regulate a constant sink current of:

Is = (3.5v – 2.5v)/R1 = 1/R1

It also illustrates adding a booster transistor Q1 to accommodate applications needing current or power beyond Z1’s modest TO92ish limits. Notice that Z1’s accuracy will be unimpaired because whatever fraction of Is that Q1 causes to bypass Z1 is summed back in before passing through R1.

Figure 2 Booster transistor Q1 can handle current and voltage beyond 431 max Ic and dissipation limits, while the 3.5-V voltage divider programs a constant Is.

Programming sink current with DAC

Figure 3 shows how Is might be digitally programmed with a 2.5-V DAC signal. Note the DAC signal is inverted (Is = max when Vx = 0) while Z2 provides the necessary level shift:

Is = (2.5v – Vx)/(2.5R1) = 0 to 1/R1 as Vx = 2.5v to 0

 Figure 3 DAC control of Is, the DAC signal is inverted, while Z2 provides the necessary level shift.

Programming sink current to Df/R1 with DAC

Figure 4 shows an alternate programming method using PWM with Is = Df /R1 where Df equals the 0 to 1 (0% to 100%) PWM duty factor:

Is = (2.5R2/(R3/Df))/R1 as Df = 0 to 1
Df = IsR1R3/(2.5R2)
Df = Is R1

Figure 4 PWM control of Is, where Is is the ratio of the PWM duty factor and R1.

The 8-bit PWM resolution and 10-kHz frequency are assumed. The R2C1 single-pole ripple filter has a time constant of approximately 64x the PWM period (10 kHz = 100 µs assumed) for 1-lsb peak-to-peak max ripple and 38-ms max settling time.

Speeding up settling time

One shortcoming of Figure 4 is the long settling time (~40 ms to 8 bits) imposed by the single-pole R1C1\2 ripple filter. If an extra resistor and capacitor won’t break the bank, that can be sped up by a factor of about 5 (~8 ms) with Figure 5’s R5C2 providing 2nd-order analog-subtraction filtration.

Figure 5 The addition of R5 and C2 provides faster settling times with a 2nd-order ripple filter.

Programmable current sink application circuit

Finally, Figure 6 shows the Figure 4 circuit combined with an inexpensive 24-W AC adapter and a 5-V regulator to power a small digital testing system. Be sure to adequately heatsink Q1.

Figure 6 The combined current sink and small system power supply where the max Is is 1 A, Max Vs is 20 V, and Is = Df.

Thanks for the implicit suggestion, Ashutosh!

Stephen Woodward’s relationship with EDN’s DI column goes back quite a long way. Over 100 submissions have been accepted since his first contribution back in 1974.

Related Content

• Cancel PWM DAC ripple with analog subtraction
• TL431 Model
• PWM-programmed LM317 constant current source
• A negative current source with PWM input and LM337 output
• A high-performance current source
• VCO using the TL431 reference

The post Precision programmable current sink appeared first on EDN.

Dear EDN Readers,

We’re thrilled to announce the successful expansion of our Design Ideas section. Thanks to your support, we now publish two new DIs every week!

We’re also excited to revitalize our Tales from the Cube column. This platform allows engineers to share their unique experiences in solving challenging design issues—whether they encountered a product failure, dealt with troublesome equipment, or tackled a persistent problem on a personal project.

We aim to regularly update this column with fresh content. With your contributions, we hope to gradually breathe new life into Tales from the Cube with new articles to feature in our Fun Friday newsletter.

Here are some FAQs to get you started:

What are Tales from the Cube articles?

Tales from the Cube are generally brief, focused narratives where engineers outline how they arrive at a solution to a specific design challenge or an innovative approach to a design task. This can relate to a personal project, a contract, or a corporate design dilemma. Here are some basic guidelines that might help you as you write out your article:

• 600-1000 words 
• 1-2 images
• One-sentence summary of your story that goes along with the title
• A short author bio

What technology areas are allowed?

We’re open to a wide range of technology areas, including (but not limited to) analog and digital circuits, RF and microwave, programmable logic, hardware-design languages, systems, programming tips, utilities, test equipment and techniques, power, and more. If you’re not sure about your topic, just email us at editors@aspencore.com.

Do I get paid for a Tales from the Cube article?

Yes! Monetary compensation for each Tales from the Cube article is $200 USD, not enough to keep the lights on, but it does offer you an avenue to tell your unique engineering story to tens of thousands of engineers globally and engage in some interesting conversations about your engineering remedy. 

How can I submit a Tales from the Cube article?

Feel free to email us at editors@aspencore.com with your questions, thoughts, or a completed article. So, Tell us your Tale!

The post Tell us your Tale! appeared first on EDN.