TVS - Vclamp to hight

Thread Starter

Jordinonell

Joined Dec 9, 2018
6
Hello,
Thanks for your support in this forum.
Currently I have a couple of PCB that I need to provide a double input power supply. The specs for this couple of PCB are Vin= 9 to 17.5V. (12V nominal), consumption is about 18-24W/each (working mode dependant).
Long cable runs and hot plugging could induce transient voltage spikes which may exceed the input voltage limit and I need to protect it.
By other hand, the system will have 2 VDC inputs, and will be capable to have a hot-swap between them. As you can see in the attach schematic, the 2 Scotty diodes allow to do this.
finally, the input battery used could provide up to 16.8V (max).
I am looking for a TVS Diode (see it in the attach file).The problem I am facing is about the specs in TVS diodes. Vclamp should be 17.5V, but then, Vr and Vbr are too low.
Any idea how to solve this short margin between Vr/Vbr and Vclamp? Could I place a Zener to clamp it to 17.5V?
Any tip will be welcome.
 

Attachments

dl324

Joined Mar 30, 2015
16,845
Welcome to AAC!

An easier to read version of your schematic:
clipimage.jpg
What is the maximum current to the loads and voltage that they will withstand?
 

ronsimpson

Joined Oct 7, 2019
2,986
Vclamp should be 17.5V
Part numbers will really help.
Most "over voltage limiting devices" have a specification that is not clear. When a part is labeled "17V" it just turns on (uA) at that voltage. You want to know from the data sheet what voltage at, say, 1A. It might be 25V. So give us a part number and lets see what it will do.
 

Thread Starter

Jordinonell

Joined Dec 9, 2018
6
Thanks for your replies.
The only data I have from the PCB datasheet (Load Devices) is the Vin range, Vin= 9 to 17.5V and I need to use it with a batteries that offers maximum 16.8V, so I have low margin for the TVS. I want to protect it from the transient voltage spikes as recomended by the manufacturer, specially for the hot pluggin. The consumption is about 18-24W.

The TVS I found is SMF16VTR, from Rohm Semiconductor. The Vr is 16V, and Vbr is 17.2, the current will be 1mA@Vbr, then it starts to protect the load. I assume the Schottky diode will have about forward voltage Vf=0.4V , and the battery will be at 16.8, so, I expect no many current will flow thru the TVS during the period that the battery is draining 0.4V.
The problem is the Vclamp, it is at 26V..
I would like to clamp the maximum voltage at 17.5V..
Thanks in advance.
 

ronsimpson

Joined Oct 7, 2019
2,986
The SMF16VTR will draw 20mA at 17.2V which is not much current. And 7.7A at 26 volts. I don't have any data at other voltages. My guess is that at your 17.5V it is pulling 25mA or not enough to clamp.
Why do you want to limit the voltage to 17.5V? Is the next thing down stream the reason?

It is interesting that this graph does not match the words in the data sheet.
If this graph is true, and you are drawing 1.5A then the inductive kick would be about 1.5A and the part will limit the voltage to 19.2 volts.
1616877118040.png
 

Thread Starter

Jordinonell

Joined Dec 9, 2018
6
Thanks both,
Yes, the 2 schoktty diodes are basically redundancy switch, and it works pretty well. This is because the equipment could be used with battery or external power supply, and I use the diodes to avoid any issues.

By other hand, @ronsimpson, thanks as well for your answers. I apologise because I am not used to work/calculate this kind of devices, and probably I am misunderstanding the TVS parameters.

I was looking for a VClamp=17.5 because it is the maximum input range for the electronic PCB I am integrating. The batteries I need to use works between 10-16.8V. So, if I want to protect the device, I should to clamp the transient voltage spike to 17.5V, this is why I am looking for a TVS with Vclamp=17.5V.
 

ronsimpson

Joined Oct 7, 2019
2,986
There is not much room from 16.8 and 17.5. So the clamp needs to draw 0 current at 16.8V and x amps at 17.5V. What happens when the battery is cold and the charging voltage is above 16.8? Or the charger has a +/-5% error and charges above 16.8V?

What in your project needs to not see more than 17.5V?
I think there is an answer that does not require 0.5% parts.
 

du00000001

Joined Nov 10, 2020
117
#1 Use 1 TVS diode for each input - "before" the resp Shottkx diode! (Otherwise an overvoltage event will break the Shottky).
This gives you some additional headroom as the output voltage will be the clamping voltage minus 1 Vf.
#2 Provided the loads cannot stand a voltage beyond 17.5 V, the TVS approach is just not feasible.

Why not use dual e-fuses like the TPS2663 - incorporating reverse polarity protection, overvoltage protection, current limiting and more? Pair these with the TVSses of your liking...
 

RPLaJeunesse

Joined Jul 29, 2018
252
Its pretty simple. You MUST design your electronics to run up to the peak voltage seen at the TVS. The actual rated input maximum is more of a "tell the customer" number, not really for design other than making the TVS selection. I've done many production designs for "12V" motor vehicle operation, and I consider 36V rated circuitry almost marginal, with 40V preferred.
 

Thread Starter

Jordinonell

Joined Dec 9, 2018
6
Thanks to everyone! And sorry for my late reply.

This is really annoying, because I have a 3rd party PCB I want to protect (it is expensive), and I want to protect from surge peak due hot plugins and cable lengths. I have no more information that this Input Rate: 9-17.4VDC.

I am just trying to understand the Ipp as @ronsimpson mentioned about 1.5A it clamps to 19.2V. What is the Ipp value? My device rated current? I read in a lot of papers the Ipp is the surge current, but the current is V /Rdyn dependant.... So, I am worried because I think I don't really understand about this 1.5A than @ronsimpson mentioned.

Thanks again
 

du00000001

Joined Nov 10, 2020
117
Thanks to everyone! And sorry for my late reply.

This is really annoying, because I have a 3rd party PCB I want to protect (it is expensive), and I want to protect from surge peak due hot plugins and cable lengths. I have no more information that this Input Rate: 9-17.4VDC.

I am just trying to understand the Ipp as @ronsimpson mentioned about 1.5A it clamps to 19.2V. What is the Ipp value? My device rated current? I read in a lot of papers the Ipp is the surge current, but the current is V /Rdyn dependant.... So, I am worried because I think I don't really understand about this 1.5A than @ronsimpson mentioned.

Thanks again
"Ipp" translates to "Peak Pulse Current" (see the resp. diagram in post #5).
Regarding your basic problem: you haven't got many choices as your headroom (difference between max. input voltage and max. rated voltage of the device) is marginal.
Maybe you should consider to use a special "LDO" - made to survive voltages up to maybe 40 V, but limiting its output to some 17 V - low enough as not to ruin the device. Tricky, but not impossible. If you can afford the power dissipated . . .
 

Thread Starter

Jordinonell

Joined Dec 9, 2018
6
"Ipp" translates to "Peak Pulse Current" (see the resp. diagram in post #5).
Regarding your basic problem: you haven't got many choices as your headroom (difference between max. input voltage and max. rated voltage of the device) is marginal.
Maybe you should consider to use a special "LDO" - made to survive voltages up to maybe 40 V, but limiting its output to some 17 V - low enough as not to ruin the device. Tricky, but not impossible. If you can afford the power dissipated . . .
Thanks for this answer. Yes, I see everywere Ipp is Peak Pulse current, but how can I determine it for this kind of peak surges? As I read it's generally accepted that the worst case scenario will be when I activate the power supply at the source end of the cable. Due to transmission line effects, I could see a peak of up to double the applied source voltage, but What about the current? It depends of the source?
 

du00000001

Joined Nov 10, 2020
117
Ipp is a function of the input voltage (and the TVS's properties). The input voltage in turn is affected by Ipp as there is some source impedance that results in some voltage drop.

Your "worst case scenario" is more from signal theory, where voltage doubling might really occur - provided the line is open on the receiving end. Asssuming your circuit is connected when the power supply is attached, this extreme case shouldn't occur. But a bit of settling might occur. What can be done: add a reasonably large capacitor at the input of your board: the capacitor is a "shorted end" (when discharged) - resulting in some sort of settling without significant voltage overshoot.
 
TVS diodes are not precise. Look at analog.com surge stopper devices and ideal diode controllers.

Hot swap should have connectors that support it. i.e. some shorter/longer pins.
 

crutschow

Joined Mar 14, 2008
34,281
You could use the TLV431 programmable shunt reference voltage (a more ideal Zener equivalent) for more precise control of the clamp voltage.
Below is the LTspice simulation of an example circuit with an added P-MOSFET to clamp high currents.
It shows an accurate clamp at about 17.2V [V[out)] for transient over-voltages [V(bat)] of 17.5V, 20V, and 25V, giving a peak clamp current [Is(M1)] of over 80A for the circuit values shown.
R_Bat simulates the stray circuit resistances that will limit the peak current.

The value of R1 can be tweaked to change the clamp voltage.
The clamp occurs when the voltage at R2 reaches 2.5V, at which point the TL431 starts to conduct, turning on the MOSFET and maintaining the voltage at the clamp value.

C2 speeds the clamp operation to minimize any voltage overshoot for fast transients when it clamps.

1618286426627.png
 
Last edited:

Thread Starter

Jordinonell

Joined Dec 9, 2018
6
Thanks for all your replies. Really appreciated.

I see some solutions here. The programable shunt, efuse, even a power input filter with capacitor could works to avoid the overshoot when I plug in the power in this PCB (see this ti app note) ad mentioned by @du00000001

Thanks again
Jordi
 
Top