best way to protect ADC from over voltage

Log in or register to post comments
Go To Last Post
41 posts / 0 new
Author
Message
pajuhesh80
pajuhesh80's picture
Level: Rookie
Joined: Fri. Jul 5, 2019
Posts: 38 View posts
#1
Posted by pajuhesh80: Sat. Nov 9, 2019 - 03:39 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Hello.

I'm building a circuit to test linear regulators up to 24V using avr ADC.

There is a voltage divider in output line to reduce voltage to 1/5 (4kOHM and 1kOHM); for example 25 to 5, 20.5 to 4.1 ,...

As the over voltage can damage MCU, I want to protect ADC from high voltages using a 100mA fuse (ADC shouldn't take much current. Should it?!) and a Zener diode to the ground.

I have two options:

  1. to put a 27V Zener diode before voltage divider.
  2. to put a 5.1V Zener diode after voltage divider.

If I use the first method, maximum voltage to ADC pin will be 5.4 and in the second method will be 5.1.

Which method is better and safer?

Note: Regulator has a separate power source and shared ground with MCU power supply.

This topic has a solution.

Tags:

AVR Microcontrollers, megaAVR and tinyAVR, #ADC, solved, Littelfuse, PolyZen, Rugged Circuits
Last Edited: Sun. Nov 10, 2019 - 03:42 AM
Top
ka7ehk
ka7ehk's picture
Level: 10k+ Postman
Joined: Fri. Nov 22, 2002
Posts: 20525 View posts
Location: Tangent, OR, USA
This reply has been marked as the solution. 
Posted by ka7ehk: Sat. Nov 9, 2019 - 04:55 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 1

Since you have the voltage divider, the solution that effects the signal the least is a pair of Schottky diodes. Connect one from ADC input to ground (cathode to ADC input). Connect the other from ADC input to chip Vcc (cathode to Vcc). 

 

Your voltage divider provides the source resistance that will limit current. The diodes keep the ADC input within the spec sheet voltage range. Schottky diodes are preferred over standard silicon junction diodes because of the lower forward drop.

 

Zeners are a poor choice because of tolerance and because the breakdown "knee" is not very sharp (especially on lower voltage devices).

 

Jim

Jim Wagner Oregon Research Electronics, Consulting Div. Tangent, OR, USA http://www.orelectronics.net

Last Edited: Sat. Nov 9, 2019 - 04:56 PM
Top
pajuhesh80
pajuhesh80's picture
Level: Rookie
Joined: Fri. Jul 5, 2019
Posts: 38 View posts
Posted by pajuhesh80: Sat. Nov 9, 2019 - 05:12 PM (Reply to #2)
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Schottky diodes? Does this solution cause fuse breakdown (enough current)?

 
Last Edited: Sat. Nov 9, 2019 - 05:26 PM
Top
ka7ehk
ka7ehk's picture
Level: 10k+ Postman
Joined: Fri. Nov 22, 2002
Posts: 20525 View posts
Location: Tangent, OR, USA
Posted by ka7ehk: Sat. Nov 9, 2019 - 05:23 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Where is it in the circuit? There is no reasonable way that you can get 100mA with the divider resistors you have specified, no matter where you put it. The divider and the diodes provide plenty of protection, so that you don't need any fuse.

 

Jim

Jim Wagner Oregon Research Electronics, Consulting Div. Tangent, OR, USA http://www.orelectronics.net

Last Edited: Sat. Nov 9, 2019 - 05:25 PM
Top
pajuhesh80
pajuhesh80's picture
Level: Rookie
Joined: Fri. Jul 5, 2019
Posts: 38 View posts
Posted by pajuhesh80: Sat. Nov 9, 2019 - 05:44 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Do you mean this? Does it guarantee that MCU will survive an over voltage? (for example 30V to "IN" line)

Last Edited: Sat. Nov 9, 2019 - 05:46 PM
Top
avrcandies
avrcandies's picture
Level: Raving Lunatic
Joined: Wed. Jun 30, 2004
Posts: 3633 View posts
Location: Cincinnati Oh
Posted by avrcandies: Sat. Nov 9, 2019 - 06:03 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

You can add a fuse, it will generally do nothing, unless there is a major short on your PCB (like you drop a bolt onto the board)...so it might help then.

WHERE is your capacitor at the ADC pin? Don't you want a quiet & clean signal to measure?

 

You might want to increase your range to maybe 30V...what if the supply is putting out 28V....your divider only allows up to 25V to be measured!!

 

Be sure to use 1/4 watt (or higher) resistors, so they don't burn up (at least for the 3.9K, the others are proportionally less watts)...1/4 W is the same as a 1206 smd part

 

 

Note: Regulator has a separate power source and shared ground with MCU power supply.

That's fine...they must share gnd to get a measurement  (I've had toooo many students overlook this!)....Ensure your 24V reg gnd is really at gnd (or floating, before being tied to the tester).  If the 24V gnd is at some offset, you have a gnd loop and bad things happen.

For example, if you plugged your 24V supply into an outlet  using a 500 foot extension cord out to the barn, the gnds between it & the AVR would likely differ  (especially if your 24V Vout neg pin shared the AC line gnd pin) .    You can even slightly encounter this sometimes if things are on 2 different workbenches/outlet strips.

You also can also encounter this if your AVR is connected to a desktop PC (or laptop grounded via a 3-pin plug)---that can be remedied using a serial/usb cable isolator. 

These are important considerations if you are building a power supply test fixture

 

 

 

When in the dark remember-the future looks brighter than ever.   I look forward to being able to predict the future!

Last Edited: Sat. Nov 9, 2019 - 06:27 PM
Top
curtvm
curtvm's picture
Level: Raving Lunatic
Joined: Wed. Sep 21, 2005
Posts: 2642 View posts
Posted by curtvm: Sat. Nov 9, 2019 - 06:33 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Your avr most likely already has internal clamp diodes (I don't know every pin of every model, so am saying 'most likely').

 

Then your job is simply to limit current. You can look at an appnote like this-

https://www.microchip.com/wwwAppNotes/AppNotes.aspx?appnote=en591171

where they seem to want you to keep the diode current <1ma (they are not using adc, but I don't think the diodes disappear or are disconnected when you use adc).

Also notice they don't seem to be concerned with voltage in that appnote (they designed up to 1000v by using 1Mohm to limit current to 1ma), so if you take the same precautions to keep the current <1ma, you are probably ok.

 

If you have 4k/1k, you could get about 5ma current through the 'upper' pin protection diode if 48v is applied. So either increase the resistor values, or add a current limit resistor to the adc input. Using a 3.9k resistor (no need for more values) at the adc input would keep 48v to a diode current <1ma.

 

+

|3k9

|_____3k9____adc

|

|1k0

-

 

Nothing wrong with your own set of diodes that 'kick in' a little sooner and take all the punishment, but probably nothing wrong without them if you limit current.

Top
pajuhesh80
pajuhesh80's picture
Level: Rookie
Joined: Fri. Jul 5, 2019
Posts: 38 View posts
Posted by pajuhesh80: Sat. Nov 9, 2019 - 06:40 PM (Reply to #6)
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I just need to read regulator voltage and it is a simple tester; so I think no capacitor is needed.

As I said, this tester is for regulators up to 24V. So divider is OK.

About resistors power (watt), what about using 40kOHM and 10kOHM instead of 4kOHM and 1kOHM?

 

I simulated the circuit I posted above and simulator shows 5.7V output when input is 30V. Is it simulator's fault? If the same thing happens in reality, the MCU will be damaged on over voltage!

 

Note: Thanks for posts, but I just need to limit voltage; nothing more! :)

Last Edited: Sat. Nov 9, 2019 - 06:44 PM
Top
ka7ehk
ka7ehk's picture
Level: 10k+ Postman
Joined: Fri. Nov 22, 2002
Posts: 20525 View posts
Location: Tangent, OR, USA
Posted by ka7ehk: Sat. Nov 9, 2019 - 06:42 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Ohm's Law will tell you all you need here.

 

1. You suggest input voltage of 30V. 

 

2. If diode limits input voltage to 5.3V, then 4.0K effective resistance has 24.7V across it.

 

3. 4.0K with 24.7V across has a current of 6.175mA.

 

4. That current mostly flows through R4. There will be 5.3mA (5.3V/1.0K) through R4.

 

5. So, D2 will really only have 6.175ma - 5.3ma < 1ma.

 

6. Power in combined R43 and R42 is 6.75ma X 24.7V = 167mW. Most of that will be in R42 so you want a 250mW  rating on that resistor.

 

Piece of cake!

Jim Wagner Oregon Research Electronics, Consulting Div. Tangent, OR, USA http://www.orelectronics.net

Last Edited: Sat. Nov 9, 2019 - 06:45 PM
Top
avrcandies
avrcandies's picture
Level: Raving Lunatic
Joined: Wed. Jun 30, 2004
Posts: 3633 View posts
Location: Cincinnati Oh
Posted by avrcandies: Sat. Nov 9, 2019 - 07:19 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I just need to read regulator voltage and it is a simple tester; so I think no capacitor is needed

Then your thinking should have an adjustment...you are making a voltmeter measuring through series resistance , which greatly increases the allowance of noise pickup.  You oppose removing 99% of this noise with a cap right at the ADC pin?  Or you don't have  2 cents for a 0.1uF cap?

 

About resistors power (watt), what about using 40kOHM and 10kOHM instead of 4kOHM and 1kOHM?

 

No, that is creeping  the high end for the ADC limit, if you want best measurements.  20K & 5K would be better.   Higher ohms means more noise as well...get the cap in there!

 

Make sure you have a quiet Avcc.  

 

When in the dark remember-the future looks brighter than ever.   I look forward to being able to predict the future!

Top
ka7ehk
ka7ehk's picture
Level: 10k+ Postman
Joined: Fri. Nov 22, 2002
Posts: 20525 View posts
Location: Tangent, OR, USA
Posted by ka7ehk: Sat. Nov 9, 2019 - 07:33 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

By the way, the  OP's simulator was simply wrong. Ohms Law shows it. The quoted numbers COULD be from standard silicon diodes, not Schottky diodes.

 

Jim

Jim Wagner Oregon Research Electronics, Consulting Div. Tangent, OR, USA http://www.orelectronics.net

Top
pajuhesh80
pajuhesh80's picture
Level: Rookie
Joined: Fri. Jul 5, 2019
Posts: 38 View posts
Posted by pajuhesh80: Sat. Nov 9, 2019 - 07:45 PM (Reply to #10)
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

OK. I'll try to make some space in PCB to place a capacitor there.

About AVcc: Yes, I placed both inductor and capacitor on it. Its as clean as possible! :)

And thanks for resistor values. I'll use 20 and 5.

 

I still didn't get my main answer: Is the circuit I posted in #5 good and "trustable" enough to protect MCU from over voltage? If yes, what is the maximum safe voltage? Should I use a fuse for voltages more than it? (more voltage usually means more current) I don't want over voltage cause damage to any "soldered" part of circuit. (MCU, diode, ...) (Glass fuses are easily changeable)

Top
avrcandies
avrcandies's picture
Level: Raving Lunatic
Joined: Wed. Jun 30, 2004
Posts: 3633 View posts
Location: Cincinnati Oh
Posted by avrcandies: Sat. Nov 9, 2019 - 08:10 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

 

The circuit with your 3amp schottky is fine

The fuse will only protect if a short erupts on your tester (say you drop a screwdriver) & only if the short is in the correct place for the fusing.

 

What if your 24V supply suddenly starts putting out 110V?  Then the resistors will burn up quick & a fuse will prob not help, unless you have power resistors that can wait around severely overloaded for a 100ma  or 50ma fuse to blow .

 

To protect against line voltage you could either:

 

1) Use 200K resistors at the input with a clamp, followed by a buffer (to drive the divider & schottky's & adc)

2) Use a fuse at the input followed by a 33V tranzsorb to gnd (look at 5kp33a device)...the tranzsorb can take a very severe punch to easily blow the fuse.   Smaller 1500W part would prob be fine to blow the fuse.  5KP33a is cheap.   This is the followed by your divider/schottky  setup. 

 

I'll try to make some space in PCB 

    Why would there not be plenty of space?  This is a tester, not a hearing aid.

 

 

 

 

 

When in the dark remember-the future looks brighter than ever.   I look forward to being able to predict the future!

Last Edited: Sat. Nov 9, 2019 - 08:22 PM
Top
gchapman
gchapman's picture
Level: 10k+ Postman
Joined: Tue. Jan 9, 2007
Posts: 11208 View posts
Location: Arlington, Texas, u.S.A.
Posted by gchapman: Sat. Nov 9, 2019 - 09:13 PM (Reply to #2)
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

ka7ehk wrote:
Schottky diodes are preferred over standard silicon junction diodes because of the lower forward drop.
and increased leakage current though there are Schottky diodes with impressively low leakage.

Won't matter for this case due to divider's low source resistance.

ka7ehk wrote:

Zeners are a poor choice because of tolerance and because the breakdown "knee" is not very sharp (especially on lower voltage devices).

Some manufacturers bin zeners to improve tolerance.

Zener breakdown voltage is proportional to temperature; inversely proportional for low voltage zeners.

 

 

"Dare to be naïve." - Buckminster Fuller

Top
ka7ehk
ka7ehk's picture
Level: 10k+ Postman
Joined: Fri. Nov 22, 2002
Posts: 20525 View posts
Location: Tangent, OR, USA
Posted by ka7ehk: Sat. Nov 9, 2019 - 09:25 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The two temperature effects happen to cancel each other out in devices with breakdown voltages near 5.5V

 

Back to #1 and the reported "limiting" (in simulation) with a 27V zener across the input....

 

That can only happen with a voltage source that is current limited, either by plain effective output resistance or by active current limiting. Since the OP has not shared any of that with us, other than a statement of "various regulators", we have no way of knowing whether or not this statement is wishful thinking.

 

I'll bet, however, that the OP has not even considered possible or likely modes of failure. For example, suppose that the regulator ground pin is not adequately soldered and it intermittently looses ground connection. What will the regulator do? What will happen if there is an input-output short, internally or externally. Will it still current limit? (very high probability: no!). There are probably another 4-8 potential failure modes. Have these all been considered? Have any been considered?

 

Seems to me that there is still a bit of work to do.

 

Jim

 

Jim Wagner Oregon Research Electronics, Consulting Div. Tangent, OR, USA http://www.orelectronics.net

Top
gchapman
gchapman's picture
Level: 10k+ Postman
Joined: Tue. Jan 9, 2007
Posts: 11208 View posts
Location: Arlington, Texas, u.S.A.
Posted by gchapman: Sat. Nov 9, 2019 - 09:34 PM (Reply to #5)
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

pajuhesh80 wrote:
Does it guarantee that MCU will survive an over voltage?
Almost (AVR have a rail clamp)

Add an external clamp to keep excess current off the AVR's VCC rail (a relative few regulators are two quadrant, iow, current source and current sink)

A zener of enough size (volume) on AVR's VCC may be correct.

XR21B1420 Data Sheet

[page 59]

Figure 7: VCC Clamp Circuits

via

XR21B1420 - MaxLinear

Enhanced 1-Ch Full-Speed USB UART

 

edit :

Engineer It - How to prevent electrical overstress of analog integrated circuits - YouTube (9m29s, 7:25 for the rail clamp zeners)

Protecting Inputs from Damage— EOS - The Signal - Archives - TI E2E support forums

 

"Dare to be naïve." - Buckminster Fuller

Last Edited: Sat. Nov 9, 2019 - 09:41 PM
Top
avrcandies
avrcandies's picture
Level: Raving Lunatic
Joined: Wed. Jun 30, 2004
Posts: 3633 View posts
Location: Cincinnati Oh
Posted by avrcandies: Sat. Nov 9, 2019 - 11:42 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

As mentioned...add a fuse followed by a 5KP33A at the input...that should hold the fort pretty well.

 

I'm building a circuit to test linear regulators up to 24V using avr ADC.

Does that include a variable, current sink (load)?  What tests?  What monitors the heating, regulation, input rejection, etc?

 ...like saying I'm building an airplane.

When in the dark remember-the future looks brighter than ever.   I look forward to being able to predict the future!

Top
gchapman
gchapman's picture
Level: 10k+ Postman
Joined: Tue. Jan 9, 2007
Posts: 11208 View posts
Location: Arlington, Texas, u.S.A.
Posted by gchapman: Sun. Nov 10, 2019 - 12:38 AM (Reply to #7)
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0


curtvm wrote:
Nothing wrong with your own set of diodes that 'kick in' a little sooner and take all the punishment, but probably nothing wrong without them if you limit current.
Depends on how the ESD suppressors are implemented.

in

Planet Analog - SIGNAL CHAIN BASICS #66: How to interface a 5V transceiver to a 3V controller -

Figure 5. Using a single current-limiting resistor causes damage to the input circuitry.

(LVTSCR reduces leakage)

 

"Dare to be naïve." - Buckminster Fuller

Top
gchapman
gchapman's picture
Level: 10k+ Postman
Joined: Tue. Jan 9, 2007
Posts: 11208 View posts
Location: Arlington, Texas, u.S.A.
Posted by gchapman: Sun. Nov 10, 2019 - 12:56 AM (Reply to #13)
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

avrcandies wrote:
To protect against line voltage you could either:
3) Telecommunications PPTC though the components after the PPTC will be stressed for "a while".

PolySwitch Telecom Resettable PTCs High Voltage Power Protection - Littelfuse

TSM250 PolySwitch® Resettable PPTCs - Littelfuse | Mouser

 

"Dare to be naïve." - Buckminster Fuller

Top
curtvm
curtvm's picture
Level: Raving Lunatic
Joined: Wed. Sep 21, 2005
Posts: 2642 View posts
Posted by curtvm: Sun. Nov 10, 2019 - 02:32 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

>LVTSCR

 

The avr inputs are protected by diodes.

 

>I still didn't get my main answer

 

I think you are making it much more complicated than it is. Take a read of the appnote I linked to. The internal protection diodes already limit the voltage to Vcc+diode drop, or typically 5.5v if using 5v supply. Make sure the diode current stays under 1ma, and you will be just as safe as what is done in the appnote which happens to be working with line voltages. They chose a 1Mohm resistor to handle up to 1000V, at which point the current starts to exceed ~1ma. You can do similar, but you are dealing with lower voltages.

 

If you add a current limit resistor of 10k to the adc pin, you will need 15.5V at the other end of the 10k resistor before you exceed 1ma of diode current- 0.001A*10k=10v,(15.5V - 5.5V). Assuming you are dividing input voltage by 4, that means you under 1ma up to 62V. I would guess it would take much more than 1ma to do any damage, but its probably a nice safe number to stay under.

 

 

 

 

Top
pajuhesh80
pajuhesh80's picture
Level: Rookie
Joined: Fri. Jul 5, 2019
Posts: 38 View posts
Posted by pajuhesh80: Sun. Nov 10, 2019 - 03:42 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The avr inputs are protected by diodes.

I know, but if they take damage, you should change MCU. and usually they aren't very powerful and trustable!

 

Let me explain my board:

User connects it to PC USB port and it uses UART to communicate. Then user selects a regulator in windows software, puts that regulator in a socket on circuit, connects an external power source to board, and presses start test. using a BJT transistor, MCU connects ground to regulator GND pin. Then measures output voltage a few times, and reports average output.

 

I decided to use ka7ehk's solution (circuit posted in #5). Thanks everyone.

Top
Kartman
Kartman's picture
Level: 10k+ Postman
Joined: Thu. Dec 30, 2004
Posts: 22800 View posts
Location: Melbourne,Australia
Posted by Kartman: Sun. Nov 10, 2019 - 04:06 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

You really don't need 3A schottky diodes - apart from leakage (which is less of an issue due to the resistances involved), there's the size and cost. I'd suggest something like BAT54 - you can get these in surface mount and dual. As mentioned many times, you REALLY want a capacitor on the input. The ADC input is not benign - it relies on charging internal capacitors and thus will add an AC component. Adding a capacitor adddresses this. As well, some analog filtering is beneficial and adds a degree of ESD protection.

 

Something else to note is the precision of the AVR adc is pretty average. In some of my current designs I've been using a Microchip PAC1934 chip that has 4 32V inputs to 16 bit adc along with current sense inputs. I think they quote 1% accuracy out of the box for voltage - this is significantly better than the poor AVR adc.

Top
avrcandies
avrcandies's picture
Level: Raving Lunatic
Joined: Wed. Jun 30, 2004
Posts: 3633 View posts
Location: Cincinnati Oh
Posted by avrcandies: Sun. Nov 10, 2019 - 04:24 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I think they quote 1% accuracy out of the box for voltage - this is significantly better than the poor AVR adc

If you can get a tight Vcc, then the ADC is decent (1 in 1000 full scale)...but of course that is not "out of the box"  Too bad Atmel internal Vref is +/-10% ...I wish they'd have an option for +/-1% for an extra 10 cents!!  I'd gladly pay.

Lately, I use a cheap Vcc Voltage regulator with 2% accuracy... I end up doing cal anyhow, but at least I start close pre-cal.  

 

On the other hand I'm using a 24bit adc for something else...yeah sure, I'm reading 0.2uV  ...uh yeah...between where & where!   Too many blindly believe. 

 

When in the dark remember-the future looks brighter than ever.   I look forward to being able to predict the future!

Top
pajuhesh80
pajuhesh80's picture
Level: Rookie
Joined: Fri. Jul 5, 2019
Posts: 38 View posts
Posted by pajuhesh80: Sun. Nov 10, 2019 - 11:23 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

You really don't need 3A schottky diodes

I know. I didn't care about the values because I just designed it to post here.

As mentioned above, I'll use capacitor. Don't worry! ;)

And I think 10 bit resolution is enough to just test if the regulator is damaged or not. (25/1023 = 0.024...)

I want my board to be tiny and low cost.

I also use ADC noise reduction sleep mode to increase noise protection.

Top
DocJC
DocJC's picture
Level: 10k+ Postman
Joined: Tue. Dec 11, 2007
Posts: 12693 View posts
Location: Cleveland, OH
Posted by DocJC: Sun. Nov 10, 2019 - 12:47 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

using a BJT transistor, MCU connects ground to regulator GND pin.

That may not be a great design decision.

 

With the regulator's Gnd not connected, there is still a current path from Vin to Vout and through your voltage divider to Ground, which can partially power the regulator, but it won't be "regulating".

 

It is far better to use a PFet to turn the Vin power to the regulator on and off.

Note that the PFet is often itself driven by an NFet, which is driven by the micro.

Google High Side Driver for more into.

 

BTY, you have mentioned Fuses several times.

Remember that fuses take mSec to Seconds to blow.

They are best considered current protection devices, not transient overvoltage protection devices.

A fuse won't provide any protection against a fast overvoltage spike, but the spike can destroy sensitive electronics in microseconds or less.

 

JC 

 

 

 

Top
pajuhesh80
pajuhesh80's picture
Level: Rookie
Joined: Fri. Jul 5, 2019
Posts: 38 View posts
Posted by pajuhesh80: Sun. Nov 10, 2019 - 02:17 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Controlling Vin instead of GND is a good idea. But about using FET instead of BJT: I found that using FET makes my circuit "a little" complex and I think it's not necessary.

I wanted to use fuse to prevent over current and found that it's not necessary. I know fuses are slow and not for voltage.

A question: is BAV99 good for over voltage protection?

 
Last Edited: Sun. Nov 10, 2019 - 02:40 PM
Top
pajuhesh80
pajuhesh80's picture
Level: Rookie
Joined: Fri. Jul 5, 2019
Posts: 38 View posts
Posted by pajuhesh80: Sun. Nov 10, 2019 - 03:41 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I was searching internet and found that Zener diodes are widely used to protect from over voltage. Are you sure they are untrustable? The board is going to work in room temperature, not industrial.

this is the circuit using Zener:

When voltage goes beyond 5.1V, transistor connects line to GND, causes excessive current and burns 100mA fuse.

I can't find many Schottky diodes in local markets. I live in Iran and because of sanctions sad can't buy from other markets like amazon.

Last Edited: Sun. Nov 10, 2019 - 03:43 PM
Top
gchapman
gchapman's picture
Level: 10k+ Postman
Joined: Tue. Jan 9, 2007
Posts: 11208 View posts
Location: Arlington, Texas, u.S.A.
Posted by gchapman: Sun. Nov 10, 2019 - 04:12 PM (Reply to #27)
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0


pajuhesh80 wrote:
Are you sure they are untrustable?
Trusted; simply be aware of zener characteristics.

pajuhesh80 wrote:
this is the circuit using Zener:
Another one :

TND6093 - Low VCE(sat) BJT's in Automotive Applications

[page 5]

Q1 : 40V 2A SOT-23

Q2 : 40V 800mA SOT-23

pajuhesh80 wrote:
I can't find many Schottky diodes in local markets.
Zener diodes?

431 shunt regulator is an alternate at approximately twice the price.

AS431 (2.5V Reference Voltage Parts) | Diodes Incorporated

follow-on :

NCP431: Voltage Reference, Low Cathode Current, Programmable, Shunt Regulator (ON Semiconductor)

pajuhesh80 wrote:
can't buy from other markets like amazon.
Is AliExpress an available competitor to the western Amazon?

 

"Dare to be naïve." - Buckminster Fuller

Top
pajuhesh80
pajuhesh80's picture
Level: Rookie
Joined: Fri. Jul 5, 2019
Posts: 38 View posts
Posted by pajuhesh80: Sun. Nov 10, 2019 - 04:50 PM (Reply to #28)
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Another one:

Good but I think mine is simpler and good enough.

I finally found a local internet market (javan electronic) that has a wide range of Schottky diodes. (and many other things like resettable fuses! maybe I use onelaugh)

I can't also buy from AliExpress. We can't pay cash to foreign markets because of sanctionsindecision.

Top
avrcandies
avrcandies's picture
Level: Raving Lunatic
Joined: Wed. Jun 30, 2004
Posts: 3633 View posts
Location: Cincinnati Oh
Posted by avrcandies: Sun. Nov 10, 2019 - 05:05 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

The circuit you have in 27 is called a crowbar...I was going to suggest, but you don't need something so complex.

The circuit from Chapman in 28 is a OVP disconnect...that one is not good, since it will ruin your voltage measurement (it might be ok if you replaced its Q1 with a fet)

 

Just take a fuse (or poly switch ,resettable PTC) and run it to a transorb like a 5kp33a or maybe something smaller...that will give all the disaster protection needed against extreme problems, like line voltage  or other extreme amounts somehow coming in.  Then run that signal to  your divider and possible diode clamps at the AVR,...that protects against small overvoltages (say 36V) coming in.

 

You need to switch the high side supplying power to your regulator...use a fet or  a high side driver ---there are many...here is the one you should use to driver your regulators (good for maybe 3 amps with decent layout)

I don't think you mentioned how many amps you will be testing these regulators at (at higher currents the regulators will need heatsinks, unless they are on only for a split second).

https://www.onsemi.com/pub/Collateral/NCV8460-D.PDF  

When in the dark remember-the future looks brighter than ever.   I look forward to being able to predict the future!

Last Edited: Sun. Nov 10, 2019 - 05:10 PM
Top
pajuhesh80
pajuhesh80's picture
Level: Rookie
Joined: Fri. Jul 5, 2019
Posts: 38 View posts
Posted by pajuhesh80: Sun. Nov 10, 2019 - 05:58 PM (Reply to #30)
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

I didn't know TVS diodes (transorb). They look interesting! But aren't they better for ESD protection (as I read in websites)?

I found a wide range of tvs SMBJxyA diodes in market: SMBJ26A, SMBJ24A, 20, 5.0, 6.0 ... perhaps SMBJ26A is good for my purpose. Is this circuit OK?: (D2 is SMBJ26A)

And about regulator, should I use FET exactly like BJT? (Gate to MCU, Drain to V+ and Source to Regulator Vin)

Also, I see two categories in market: JFET and MOSFET. Which should I use?

Top
avrcandies
avrcandies's picture
Level: Raving Lunatic
Joined: Wed. Jun 30, 2004
Posts: 3633 View posts
Location: Cincinnati Oh
Posted by avrcandies: Sun. Nov 10, 2019 - 06:07 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

 

Use a fuse or PTC, with a transorb...the  transorb reacts instantly, shorting out the overvoltage & causing the fuse to blow (or the PTC to open).

What part are you having trouble understanding?   The circuit you show is ok, but will be slower & more complicated  ...it also does not portect against reverse polarity, but the transorb does.

 

What are you using to load down the regulators?  How many amps will you need?

When in the dark remember-the future looks brighter than ever.   I look forward to being able to predict the future!

Top
pajuhesh80
pajuhesh80's picture
Level: Rookie
Joined: Fri. Jul 5, 2019
Posts: 38 View posts
Posted by pajuhesh80: Sun. Nov 10, 2019 - 06:29 PM (Reply to #32)
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

As I said, circuit above uses transorb (D2 is SMBJ26A, a TVS/transorb diode, not Zener) is the circuit correct?
PA0 line is directly connected to MCU; nothing else. Do I need any other load? Divider isn't enough? I tested something like this circuit before and worked well without any extra load.

Last Edited: Sun. Nov 10, 2019 - 06:34 PM
Top
avrcandies
avrcandies's picture
Level: Raving Lunatic
Joined: Wed. Jun 30, 2004
Posts: 3633 View posts
Location: Cincinnati Oh
Posted by avrcandies: Sun. Nov 10, 2019 - 08:26 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

NO   there are TWO parts 

A fuse.

A transorb goes to GND and shorts out (blows ) the fuse (or opens the PTC). There is NO transistor

When in the dark remember-the future looks brighter than ever.   I look forward to being able to predict the future!

Last Edited: Sun. Nov 10, 2019 - 08:27 PM
Top
pajuhesh80
pajuhesh80's picture
Level: Rookie
Joined: Fri. Jul 5, 2019
Posts: 38 View posts
Posted by pajuhesh80: Mon. Nov 11, 2019 - 10:54 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

So I have to delete transistor and everything is OK. Right?

 

Top
avrcandies
avrcandies's picture
Level: Raving Lunatic
Joined: Wed. Jun 30, 2004
Posts: 3633 View posts
Location: Cincinnati Oh
Posted by avrcandies: Mon. Nov 11, 2019 - 07:49 PM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 1

So I have to delete transistor and everything is OK. Right?

 

 there are TWO parts 

A fuse.

A transorb goes to GND and shorts out (blows ) the fuse (or opens the PTC). There is NO transistor

 

Ground belong at the bottom of your drawing NOT at the top  ---why is everything upside down?

When in the dark remember-the future looks brighter than ever.   I look forward to being able to predict the future!

Last Edited: Mon. Nov 11, 2019 - 07:51 PM
Top
pajuhesh80
pajuhesh80's picture
Level: Rookie
Joined: Fri. Jul 5, 2019
Posts: 38 View posts
Posted by pajuhesh80: Tue. Nov 12, 2019 - 03:07 AM (Reply to #36)
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

Sorry but I'm getting confused! Could you please send a schematic?

Top
Kartman
Kartman's picture
Level: 10k+ Postman
Joined: Thu. Dec 30, 2004
Posts: 22800 View posts
Location: Melbourne,Australia
Posted by Kartman: Tue. Nov 12, 2019 - 04:33 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

This is a grab from one of my recent projects. We have Vin which is from an external source. F3 is a PTC or 'polyfuse'.

These are self resetting polymer fuses. Read the datasheet CAREFULLY!

These are seemingly simple but tricky components to specify.

First thing to understand is they are slow to react - especially for small overloads. Gross overloads they react a lot faster.

They are temperature sensitive! This sort of goes without saying as they are a thermal device.

They have a life - don't put them in a situation where they frequently activated else they wear out.

They also don't like extended overloads and have a tendency to go up in smoke.

 

In my application, if someone wires the input backwards, D3 will conduct and F3 limits the current.

If there is an overvoltage, D3 conducts and F3 limits the current - note! Do your thermal calculations! D3 is thermally limited. Having a protection circuit that fails is no better than having none. - all the customer sees is a failure. 

If there is a short on the internal circuit or the external circuit the 12V is routed to, F3 will limit the current.

If there is a minor overvoltage, D3 will likely get hot and die. When diodes die, they normally go short circuit. When that happens F3 should kick in. My main concern was customers making wiring mistakes - shorts and reverse wiring are VERY common, so I've covered 90% of expected faults.

 

Even glass fuses are seemingly simple devices, but are far from it. There is a lot to know. There's many different fuses - for good reason.

 

Fuses , TVS diodes, varistor, in fact all components are not perfect. You need to understand their characteristics.

Eg a 1A fuse doesn't magically open as soon as you exceed 1A. Even when the fuse opens, if there's enough energy, the arc conducts and you get what is called 'let through'.

A 20V TVS diode doesn't fully conduct if you go slightly over 20V - there's a curve! Same with zeners. The zener voltage is specified at 1mA usually. This means you need to be wary of using zeners/TVS diodes where there are series resistors and low currents.

Similarly transistor, mosfet, diodes etc are not perfect switches! With reverse voltage, a diode is not perfectly off.

What this means is you need to understand the component and read the datasheet. 

 

 

When drawing schematic, there's a few conventions:

 

1. input on left, output on right

2. ground usually points towards bottom

3. don't have intersecting wires connect

there's probably more, but this gets you started. Follow these and we'll have an easier time of understanding your schematics. It will also make your schematics look more professional.

 

Last Edited: Tue. Nov 12, 2019 - 04:54 AM
Top
ka7ehk
ka7ehk's picture
Level: 10k+ Postman
Joined: Fri. Nov 22, 2002
Posts: 20525 View posts
Location: Tangent, OR, USA
Posted by ka7ehk: Tue. Nov 12, 2019 - 06:41 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

As Kartman points out, matching fuse ratings with TVS ratings can be quite challenging. On several recent projects, I've spent more time on this than any other part of the hardware. 

 

For the polyfuse, there is the current which, if it stays below, it won't pop. And then there is the current which, if exceeded, it WILL pop. There is the voltage that it will tolerate while open and the forward drop when it is closed. And, time, always time.

 

For the TVS, there is a complementary set of specs. Below, its good, above, it conducts. TVS devices tend to be a lot faster so the fuse dominates the timing. 

 

And, of course, there is always cost. And board area. 

 

'Tis not a trivial task! Which is amazing for a two-component circuit block.

 

Jim

Jim Wagner Oregon Research Electronics, Consulting Div. Tangent, OR, USA http://www.orelectronics.net

Top
avrcandies
avrcandies's picture
Level: Raving Lunatic
Joined: Wed. Jun 30, 2004
Posts: 3633 View posts
Location: Cincinnati Oh
Posted by avrcandies: Tue. Nov 12, 2019 - 08:33 AM
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

There are now (within the past few years) some combo polyswitches that combine the zener/transorb...but I don't have a part number.

 

The 5KP33A transorb/zener/TVS I spec'd is beefy enough that you are generally unlikely to blow it out before the small fuse pops.

Remember that if the zener is a 30V  & it is conducting 2 amps (to blow a small fuse), it will be dissipating 2*30V=60W for a short time....enough to fry a smaller zener/transorb.

 

The mentioned combo parts use this heat to heat up & open the polyswitch faster. 

 

Note the poly switch doesn't "blow"; it goes to high resistance to limit the current & resets  to "zero" after the event goes away. 

When in the dark remember-the future looks brighter than ever.   I look forward to being able to predict the future!

Top
gchapman
gchapman's picture
Level: 10k+ Postman
Joined: Tue. Jan 9, 2007
Posts: 11208 View posts
Location: Arlington, Texas, u.S.A.
Posted by gchapman: Tue. Nov 12, 2019 - 04:57 PM (Reply to #40)
  • 1
  • 2
  • 3
  • 4
  • 5
Total votes: 0

avrcandies wrote:
The mentioned combo parts use this heat to heat up & open the polyswitch faster.
For 15V and less power :

PolyZen Devices for Overvoltage-Overcurrent Protection - Littelfuse

 

P.S.

With thanks to Bob :

10 Ways to Destroy an Arduino — Rugged Circuits

 

"Dare to be naïve." - Buckminster Fuller

Top