Back to Basics: ESD diodes – Unidirectional vs Bidirectional

I thought should do a few posts in general and advanced about Electro Static Discharge(ESD). There are a lot of small things which can trip up anyone doing circuit designing for the first time. Firstly, ESD protection at the input is something you should do when designing a product which is meant to be sold anywhere. One is because ESD can definitely ruin your devices and secondly because standards mandate that you provide protection for it.

You provide ESD protection to places where someone can actually touch your device. This is mostly at the input side like cable connectors, and power inputs. When a human body is charged up, it can deliver voltages in the Kilovolt range for a short duration. For those of you who didn’t know, it’s the same shock that you get when you touch a metallic doorknob after walking around on a carpet. This high voltage is more than enough to fry the circuits you can physically touch. So it’s imperative that you protect the circuits from ESD.

The usual method to do it is via TVS diodes. These are diodes which are placed at the input to protect against ESD. In order to simulate ESD events in a test lab, we have a standard called IEC 61000-4-2 which defines the test procedures and various voltage levels of protection. The device under test is applied with contact voltages of up to ±8KV for a tiny interval of time for a Level 4 rating(I will get into the specific test some other day if there is interest). The device should ideally be able to survive this ESD strike. As you see, both positive and negative ESD pulses are given to a circuit as you can be positively or negatively charged in real life.

Now, what are unidirectional and bidirectional TVS diodes? These are diodes which are designed to break down and maintain a voltage when it exceeds a predefined level. This helps it save the circuit downstream from damage. They are connected in parallel to your inputs. Here is where a common confusion occurs, should I be using unidirectional or bidirectional TVS diodes for protection? A wrong way to think of it is, “Because my ESD tests can be in the positive and negative regions, I need to use a bidirectional one for protection”. That is simply not true. If you see the VI chart comparison of the diodes you will clearly see the difference between the two. A unidirectional diode clamps the voltage level in one direction after a threshold but in the other direction, it breaks down immediately. Whereas a bidirectional one has these threshold limits on both sides. Why is this useful? Suppose your normal accepted input voltage can go negative with respect to ground-like audio signals or ±5V . If you don’t use bidirectional ones, it will simply clip the negative range and you lose the signal. Hence choose a diode based on what sort of input voltages you are expecting. Unidirectional ones are cheaper and widely used as most input signals are positive wrt the ground.

If you liked the post, Share it with your friends!

Back to Basics – Watchdog Timers(WDTs)

WDTs are probably one of the most skipped sections in an embedded system because a first-time firmware engineer firmly believes that the code that they wrote is 100% bug-free. Somehow they feel that it’s not needed because they have put the code through tons of testing and still couldn’t crash the firmware. Seasoned folks will tell you that this is absolutely wrong, users on the field will definitely find ways to break your system. WDTs are an absolute necessity for anything which you ship out.

So what are they basically? A WDT is a type of timer that is used to detect and recover from system hangs. It works by continuously monitoring the system and resetting itself after a certain period of time. If the system fails to reset the timer(meaning it has gotten stuck somewhere) before it expires, the WDT will trigger a system reset or other recovery action. It’s analogous to you petting a dog at a constant interval and the moment you stop or miss the petting, the dog will bark/bite to alert you. Hence the name “Watchdog”.

There are diff variations of WDT. Internal WDTs are something which you find commonly in new-age controllers wherein a certain section on the chip is a dedicated timer to perform the WDT. External WDT variant usually utilises an external chip to perform the same. The general consensus is that for anything mission-critical you use an External WDT, just avoid the hardware locking up in some extremely brutal edge cases if you can afford to have that extra part. WDTs in general are not only used for a hard reset, sometimes can be multi-staged wherein you can trigger Action#1 when Timer#1 runs out via interrupts and so on and give the system to correct it and the last one being a hard reset. Some internal WDTs can be used to trigger an interrupt and then give certain processing clock cycles so that some debug state info can be saved before issuing a hard reset.

It’s an understatement to mention that WDTs are essential. NASA Clementine moon spacecraft in 1994 was a casualty of badly implemented WDT in software whereas Mars Pathfinder had a well-built WDT which saved millions of dollars.

If you liked the post, Share it with your friends!
1 54 55 56 57 58 89