Do you use Transient Voltage Suppressors?

Go To Last Post
25 posts / 0 new
Author
Message
#1
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I notice the Atmel Butterfly has TVS (varistors) on the RS232 RXD and TXD inputs. Does anybody else do this?

.

Attachment(s): 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Yes, I used Tranzorbs on an industrial board.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I used them on a board I designed where I used to work, and a client's board I developed software for had them. They were across the regulated supply in both cases. I'd always use them on a system that was used in an industrial location, or in a vehicle. I like AVX TransGuards.

Leon Heller G1HSM

Last Edited: Mon. Jun 13, 2011 - 12:08 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I use TVS, but only in the power supply, not yet on the communications channels..

If you want fast responce the TVS is the guy fou u...

Regards,

Bruno Muswieck

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

It depends a lot on the interface device. If it is ISD-protected/resistant, then I probably won't. If it is not protected, then I generally do.

I put them on the other side of the two resistors. Unless the transistors cannot withstand 2KV or so. That greatly reduces peak voltages and currents.

Jim

 

Until Black Lives Matter, we do not have "All Lives Matter"!

 

 

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I use them on any I/O that is exposed outside the box, as well as on any power supply rails.

Writing code is like having sex.... make one little mistake, and you're supporting it for life.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Thanks guys. It makes sense to me that a PCB that is enclosed in a box could benefit from ESD protection on the connections to the outside world.

But what about PCBs that are not enclosed, like the Xplain board, the Gabotronics Xprotolab, and the Butterfly? I'm beginning to think that putting TVS on RS232 lines, like on the Butterfly, is not worth the effort.

The board I'm working on will probably end up inside a computer case. In the meantime, it will be used for testing.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

It all comes down to how much you want to protect your board. Development boards often leave this circuitry out for cost reasons, and the fact that they are typically not "mission critical". Virtually any field deployed board will have the necessary protection circuitry.

Writing code is like having sex.... make one little mistake, and you're supporting it for life.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I had my second programmer out off job last month.. This time the dragon, I remember some freak with some tutorial about protect your dragon.. Soo, poor protection I/O on those...

Regards,

Bruno Muswieck

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

steve17 wrote:
Thanks guys. It makes sense to me that a PCB that is enclosed in a box could benefit from ESD protection on the connections to the outside world.

But what about PCBs that are not enclosed, like the Xplain board, the Gabotronics Xprotolab, and the Butterfly? I'm beginning to think that putting TVS on RS232 lines, like on the Butterfly, is not worth the effort.

It does not matter how the boards are enclosed or not. I have very rarely broken anything with ESD, if ever.

However connecting a device with floating power supply (not earthed) to a grounded computer will for example kill your serial port if data lines connect before ground on serial connector. There might be huge potential difference between devices, now who says they have never seen sparks in these situations?
Just put an oscilloscope probe to your DVD player or whatever and see nice distorted sine wave of more than 40VAC.. So care must be taken in which order to connect devices and their power supplies together.

Modern display connectors like DVI make damn sure connector shell and multiple ground pins are the first to connect before any other pins.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Jepael wrote:
connecting a device with floating power supply (not earthed) to a grounded computer will for example kill your serial port if data lines connect before ground on serial connector. There might be huge potential difference between devices
That makes sense to me. So Atmel was right to put protection on the RS232 pins. What about protection on JTAG/PDI pins also?
Jepael wrote:

Modern display connectors like DVI make damn sure connector shell and multiple ground pins are the first to connect before any other pins.
I think USB connectors work this way also. Has anyone tried this with standard header pins by having the ground pin stick up higher from the PCB?

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

USB connectors are designed for "hot-swapping". Other hot-swappable connectors are available.

Leon Heller G1HSM

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

For USB there are single chip protection devices available that protect all lines. Like the ST USB6B1; or protect 6 I/O's with the ESDA6V1U1.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

steve17 wrote:
So Atmel was right to put protection on the RS232 pins. What about protection on JTAG/PDI pins also?

Well they used some kind of transistor level shifter for RS232. Real RS232 level translator chips like MAX232 and similar have internal protection and can take up to 15kV ESD pulses, and can handle by design greater voltages up to +/- 30V like it is possible to have on RS232 pins.

JTAG or whatever, there is no difference. Any IO pin will get killed if there is a big potential difference between devices and grounds do not connect first to even out the difference. While the potential difference might be huge, the actual current happening from stray capacitances may be max few tens of microamps, it is just a bigger pulse to discharge it.

Jepael wrote:

Modern display connectors like DVI make damn sure connector shell and multiple ground pins are the first to connect before any other pins.
I think USB connectors work this way also. Has anyone tried this with standard header pins by having the ground pin stick up higher from the PCB?

USB as well. Connector shell first, then GND and 5V, then data pins.

Single GND pin sticking out to connect first won't make much difference. I sometimes use 1x1 jumper wires to connect header pins together between devices, and ground is first.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I have another plan. I will put an anti-static pad on my desk and ground it to my aluminum computer case. I should have done this a long time ago.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

jayjay1974 wrote:
For USB there are single chip protection devices available that protect all lines. Like the ST USB6B1; or protect 6 I/O's with the ESDA6V1U1.

To add to that, there are some like ST's USBUFxxW6 which also provide proper termination of the USB line.

Michael

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
It does not matter how the boards are enclosed or not. I have very rarely broken anything with ESD, if ever.

Me too, but my college said that it happened with some of our components, I don´t remember wich one...

Quote:
However connecting a device with floating power supply (not earthed) to a grounded computer will for example kill your serial port if data lines connect before ground on serial connector. There might be huge potential difference between devices, now who says they have never seen sparks in these situations?

Go point Jepeal, maybe this is the case of my two broken programmer... But I do it with an TTL-RS232 board and over now, all goes fine (until now I´m taking the risk :))..

But, let's go for the power supply side, suppose that your PS is a transformer, the primary side is earthed but the secondary (low voltage side that feed the board) isn't earthed and you got problem... My point is: it's hard to know if the boards connected will not suffer from this problem...

Regards,

Bruno Muswieck

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

brunomusw wrote:
Quote:
It does not matter how the boards are enclosed or not. I have very rarely broken anything with ESD, if ever.

Me too, but my college said that it happened with some of our components, I don´t remember wich one...

How can they mysteriously know what happened to your components when they stopped working? What if there was a short circuit or overvoltage somewhere and the component just burned out? Usually ESD is blamed when nobody has not figured out the real reason or do not bother to figure it out, or there just is no other logical reason.

brunomusw wrote:

Quote:
However connecting a device with floating power supply (not earthed) to a grounded computer will for example kill your serial port if data lines connect before ground on serial connector. There might be huge potential difference between devices, now who says they have never seen sparks in these situations?

Go point Jepeal, maybe this is the case of my two broken programmer... But I do it with an TTL-RS232 board and over now, all goes fine (until now I´m taking the risk :))..

But, let's go for the power supply side, suppose that your PS is a transformer, the primary side is earthed but the secondary (low voltage side that feed the board) isn't earthed and you got problem... My point is: it's hard to know if the boards connected will not suffer from this problem...

The primary side of a transformer is never earthed. The neutral wire might be in the same potential as earth, but the live wire has AC voltage. This is not the issue. What matters is if the secondary side is earthed or not. If it is not earthed, the stray capacitance between transformer primary and secondary will make the output to be capacitively connected to mains.

The issue is that if your transformer has a two-prong mains plug, it cannot connect the secondary side to earth, like most power supplies for consumer devices. Most transformers with three-prong mains plugs do connect the earth to negative lead of secondary side, like laptop power supplies, so the stray capacitance has no chance to cause mains voltage on secondary side.

Even bigger issue are switching mode power supplies, that usually are not grounded (two-prong plug) and some regulations require the secondary side to be capacitively coupled to primary side. So you literally have the otherwise floating output connected to AC mains with a smallish capacitor. The capacitor is just so small that the regulations for leakage current is not exceeded, it might be in the order of few tens of microamperes maximum if you short circuit the output to earth, but it still hurts when you touch earthed things and power supply output, as the voltage on capacitor is about mains voltage.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
How can they mysteriously know what happened to your components when they stopped working? What if there was a short circuit or overvoltage somewhere and the component just burned out? Usually ESD is blamed when nobody has not figured out the real reason or do not bother to figure it out, or there just is no other logical reason.

Good point, maybe this is why I never saw... :) This is an open issue...

Quote:
The primary side of a transformer is never earthed. The neutral wire might be in the same potential as earth, but the live wire has AC voltage. This is not the issue. What matters is if the secondary side is earthed or not. If it is not earthed, the stray capacitance between transformer primary and secondary will make the output to be capacitively connected to mains.

Not never, if you have varistor and capacitor connected from live to earth on the primary side of the transformer to protection, they has some leakage current, soo they are connected to earth, the same case when you have the capacitor on switching mode power supplies.
But if the secondary side isn't connected to earth, it will be floating..

Quote:
The issue is that if your transformer has a two-prong mains plug, it cannot connect the secondary side to earth, like most power supplies for consumer devices. Most transformers with three-prong mains plugs do connect the earth to negative lead of secondary side, like laptop power supplies, so the stray capacitance has no chance to cause mains voltage on secondary side.

Negative lead, you mean the 0V potential, right? Soo, in one part of your circuit, the ground will be connected to earth...

Talking about it Jepeal, in a product, it has power and logic parts, on it's manual, they said that the logic part (digital and analog) is connected to earth thorught a resistor and a capacitor, it's to divide the power part and logic part?

This post are getting really good... :D

Regards,

Bruno Muswieck

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

brunomusw wrote:

Quote:
The primary side of a transformer is never earthed. The neutral wire might be in the same potential as earth, but the live wire has AC voltage. This is not the issue. What matters is if the secondary side is earthed or not. If it is not earthed, the stray capacitance between transformer primary and secondary will make the output to be capacitively connected to mains.

Not never, if you have varistor and capacitor connected from live to earth on the primary side of the transformer to protection, they has some leakage current, soo they are connected to earth, the same case when you have the capacitor on switching mode power supplies.
But if the secondary side isn't connected to earth, it will be floating..

But by definition the neutral wire and the earth are connected at some point together, back in the old days there were just two wires going inside walls and earthed wall sockets just used the neutral as earth. Earthing is something artificial you select - for instance I have used to 3-phase mains coming to a house and distributed as 1-phase around the house, but in US it seems a house gets 2-phase power, 230VAC with neutral wire coming from center tap, so you have two 125VAC phases with opposite phase.

So in any event, if you have a varistor or capacitor going from live to earth, the neutral is still earthed at some point. In fact current going trough earth wire is not normal, unless during brief short circuit before fuse breaks.

Usually things like computer power supplies have filter caps from live to earth and neutral to earth, so if you plug it to a wall socket without earthing, that will make the computer case capacitively coupled to half of mains VAC. And that hurts if you touch computer case and earthed things, kills devices too.

brunomusw wrote:

Negative lead, you mean the 0V potential, right? Soo, in one part of your circuit, the ground will be connected to earth...

Talking about it Jepeal, in a product, it has power and logic parts, on it's manual, they said that the logic part (digital and analog) is connected to earth thorught a resistor and a capacitor, it's to divide the power part and logic part?

This post are getting really good... :D

Yes well usually you earth the negative lead of single output voltage. My lab power supply is floating and I can connect the earth separately to either positive or negative lead, thus I can freely combine two earthed power supplies into +12 and -12 supplies if I want.

Sometimes a device is earthed through resistor and capacitor. The capacitor is there to provide good earth path to high frequency signals, and the resistor is there to somewhat isolate the local signal ground from earth, so that for example signals going in and out of the box to other devices have a better return path elsewhere than through the mains lead earth wire. I don't remember it exactly.

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Quote:
Sometimes a device is earthed through resistor and capacitor. The capacitor is there to provide good earth path to high frequency signals, and the resistor is there to somewhat isolate the local signal ground from earth, so that for example signals going in and out of the box to other devices have a better return path elsewhere than through the mains lead earth wire. I don't remember it exactly.

I was thinking about differents power ground levels, not the same for power and logic side.
But this can cause a problem, suppose that you connect other equipament, since this equipament is connect it's ground to earth (0V level to earth), and connect their ground to make a communication between then, what will happen? One will be at different ground level as it's connected to earth throught a resistor, right?

Regards,

Bruno Muswieck

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

To answer the original question - TVS diodes and often input resistors are a must on industrial/production boards. Mostly to protect the board from damage by idiots - it's better for your reputation to have a board that will tolerate it than try and explain to a company that their employees are inattentive monkeys who connected DC power to the wrong terminal block.

I suppose it also protects from the odd spike from natural causes as well :lol:

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Also to answer the original question, I currently am working on devices to pass DO160 testing and will experience "lightning" surges in the hundreds of volts and amps. I am very thankful for TVS devices otherwise I would have a smoking product after testing :)

Alan

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Only hundreds? Must be baby-lightning then :)

A IEC 61000-4-5 mains surge test can actually subject your device to 4KV and currents up to 2kA; depending on the class these figures can vary.

We have products able to stand up to 10KV on its mains input. But that takes quite some elaborate protection circuitry though ;)

  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I used to work on military radio systems and they had to withstand an EMP event caused by a nuclear weapon. I can't remember the figures, but they were very impressive.

Leon Heller G1HSM