Is relay paralel to NTC termistor even worth it?

[MacIntoshCZ]

Hot termistor has resistance 80mR but relay 100mR. Lets assume it will lower resistance two times. But you still need to add control circuitry for relay (time delay for caps to charge).

 

[Nigel Goodwin]

With no context the question is meaningless.

 

[MacIntoshCZ]

With no context the question is meaningless.

What do you mean? I saw design that has relay across NTC termistor, but in datasheet realay has higher resistance than termistor that gets hot. So why someone would bother to add relay and extra circuitry that makes circuit more complex. Two parallel termistors with double nominal resistance could be better solution, but maybe lifespan of hot NTC is lower than lifespan os switching relay? =)

 

[Nigel Goodwin]

What do you mean? I saw design that has relay across NTC termistor, but in datasheet realay has higher resistance than termistor that gets hot. So why someone would bother to add relay and extra circuitry that makes circuit more complex. Two parallel termistors with double nominal resistance could be better solution, but maybe lifespan of hot NTC is lower than lifespan os switching relay? =)

Again, there's no context, we've no idea what the current consumption or surges are. Varistors aren't usually that low a resistance, and by-passsing with a relay is a good idea. Mostly soft-start circuits like this use resistors, not varistors.

 

[MacIntoshCZ]

Again, there's no context, we've no idea what the current consumption or surges are. Varistors aren't usually that low a resistance, and by-passsing with a relay is a good idea. Mostly soft-start circuits like this use resistors, not varistors.

Sorry nigel but i never mentioned varistors. Also current is meaningless since we are talking about change in resistance.

 

[AnalogKid]

I saw design

But we did not. Please post a schematic, or a link to this design. At the ***very*** least, tell us what it is.

Also, please post a link to the relay datasheet.

ak

 

[MacIntoshCZ]

But we did not. Please post a schematic, or a link to this design. At the ***very*** least, tell us what it is.

Also, please post a link to the relay datasheet.

ak

=D I forgot to mention that NTC is used to limit current peaks when PSU is connect it to AC.
sorry nigel

 

[AnalogKid]

Current peaks at the input or the output?

ak

 

[Nigel Goodwin]

Sorry nigel but i never mentioned varistors. Also current is meaningless since we are talking about change in resistance.

Thermistors used in such a way are commonly called varistors - possibly a trade name? - current is critical - as I said you're giving no context, so you question is meaningless.

 

[augustinetez]

Two parallel termistors with double nominal resistance could be better solution,

No, because of tolerances, one will be hotter than another and there will be uneven resistance profile across them.

A relay across a surge limiter (which is what they are) is either bad design or it performs some other function not immediately obvious.

Putting a relay across a surge limiter will also negate another protection function - once energy in exceeds a certain limit, surge limiters will fail - often seen in SMPS units with spectacular results (and they also quite often fail before the fuse).

 

[AnalogKid]

A relay across a surge limiter (which is what they are) is either bad design or it performs some other function not immediately obvious.

Actually, it is good design. I've seen it mostly in high-end, high-power supplies that have more price margin to work with, such as the Westcor Multi-pac.

Inrush current limiting works well only if it works every time. Inrush techniques that rely on the temperature of a device are fine when the supply is cold, but need time to cool off between inrush events. After a brief power outage, the surge limiter might not be cool enough to provide any protection, leading to connector damage and fuse-blowing.

The relay shorts out the surge limiter during normal operation, which has two advantages. Second, it increases system efficiency a small amount, although this is partially offset by the relay coil power.

First, it allows the thermistor to cool down to its high resistance state while the supply is in operation. If input power bounces quickly, the cool thermistor can once again limit the inrush current. Without the relay, the thermistor is in the primary circuit continuously, which keeps it warm, so it cannot protect the input unless power is removed for many minutes, long enough to cool off.

ak

 

[ChrisP58]

Anytime you're looking to compare multiple ways of accomplishing a task, you almost always need to look at more than one factor. So yes, context is usually important.

While not stated here, I'm amusing that controlling inrush current in a power supply is the goal. But how much current? What are the acceptable losses? What is the budget?

NTC thermistors limit inrush current by having a relatively high resistance when cold, and a low resistance when hot. What is that ratio and how hot does it need to be to be low enough to be negligible to the efficiency of the power supply? And how much power (current squared x the temperature dependent resistance) does it need to dissipate to get and stay that hot?

The thermistor is a simple, inexpensive device. A relay and it's associated control circuitry will be more expensive. But in many cases, will be a better choice.

I've never seen a relay used in parallel with a thermistor. But I've often seen a relay in parallel with a fixed resistor. That method is quite common in high power supplies, particularly audio amplifiers. In those cases, the higher cost is justified by the lower power wasted.

 

[crutschow]

It would seem that if you have the relay, then a thermistor/varistor would not have a significant advantage over a small fixed resistor.

 

[Nigel Goodwin]

It would seem that if you have the relay, then a thermistor/varistor would not have a significant advantage over a small fixed resistor.

Well we've still got no context at all, so everything is just wild guess work.

However, the issue with surge limiting resistors in high power amplifiers is that if the bypass circuitry (usually a Triac rather than a relay) fails then the resistor gemerally overheats and fails, presumably a varistor/thermistor will help prevent this?.

 

[MacIntoshCZ]

Well we've still got no context at all, so everything is just wild guess work.

However, the issue with surge limiting resistors in high power amplifiers is that if the bypass circuitry (usually a Triac rather than a relay) fails then the resistor gemerally overheats and fails, presumably a varistor/thermistor will help prevent this?.

nice catch
I met a gentleman who sold in the village on the market used electronics devices and one of these devices was a printed circuit board from a server source. I just looked at it and noticed that there was a component at the entrance with the word "ntc" and relay next to it. I have no documentation. So we can only guess why and if the relay was parallel to the thermistor at all =).
After i got home i looked at random 10A termistor resistance (when it gets hot) and resistance of contacts of random 10A relay.

 

[Nigel Goodwin]

nice catch
I met a gentleman who sold in the village on the market used electronics devices and one of these devices was a printed circuit board from a server source. I just looked at it and noticed that there was a component at the entrance with the word "ntc" and relay next to it. I have no documentation. So we can only guess why and if the relay was parallel to the thermistor at all =).
After i got home i looked at random 10A termistor resistance (when it gets hot) and resistance of contacts of random 10A relay.

Right, so you were only guessing at the thermistor/varistor value, I suspect it was nowhere as low as you imagined.

 

[AnalogKid]

And the relay contact resistance. In a 10 A circuit, a relay contact resistance of 0.1 ohm would dissipate 10 W.

No.

ak

 

[crutschow]

And the relay contact resistance. In a 10 A circuit, a relay contact resistance of 0.1 ohm would dissipate 10 W.

0.1 ohm seems unusually high for power relay contacts.

 

[MacIntoshCZ]

And the relay contact resistance. In a 10 A circuit, a relay contact resistance of 0.1 ohm would dissipate 10 W.

No.

ak

maybe i just picked wrong column
https://www.hadex.cz/spec/l597a.pdf here is datasheet
It looks like contact for relay to switch on. ut i dont know why someone would care about 0,1R when coil resistance is in hundreds ohm range...

 

[crutschow]

here is datasheet
It looks like contact for relay to switch on. ut i dont know why someone would care about 0,1R when coil resistance is in hundreds ohm range...

It does say the contact resistance is 100mΩ, but that would give a contact dissipation of 22.5W at 15A contact current which would fry the relay.
That's why you care about the 0.1R. It has nothing to do with the coil resistance.

So either that resistance value is wrong, or the relay is badly overrated.