relay's versus power mosfets for "contact" resistance

[Nigel Goodwin]

so basically you say that relay contact resistance is far lower than wht they often tell you ? I was hoping to use a relay as it is very robust but I also want a solution that makes as little heat as possible, I expect experimentation is the only way of find out a few facts, although won't the contact resistance increase with age ?

As I've said all along, relays are the way to go - they have loads of advantages (the most important been reliability) and very few disadvanatges. You're really over-thinking the entire thing.

 

[Thunderchild]

well if they tell me that the relay contacts have a resistance of 50 mohms then yes i will be asking questions for if it were the case that would be a dissipation of 20 watts, of course people never like to be specific so I'm left guessing less than 50 mohms can mean anything from 1-50 mohms although I get the impression that 50 mohm is an industry figure for putting relays into categories, relays are by far the easiest and I'll consider them first. plan is to stick with current diodes that currently have a dissipation of up to 8 W, then try relays, if thats a non starter i will have to go for power mosfets.

 

[Nigel Goodwin]

Your car is full of relays - have you ever found one that dissipates 20W?.

 

[Thunderchild]

Your car is full of relays - have you ever found one that dissipates 20W?.

true, I'll be experimenting

 

[Thunderchild]

well I've bought a 12 V 40 A relay with a current consumtion of 175 mA at 13.8 volts and put it in the case I intend to use and powered it with a temp sensor in there (meat thermometers can be real useful - mine may see meet some day ).

Well with an ambient of 16/17 C it has already hit 34 C, thats before I put any other circuitry in there ! thats a power consumption of 2.415 W (so not all of it is heat) and a temp rise of 17 C, thats 7C/W temp coeficient. The diodes are obviously out at this point and I've not finished my temp test yet it is rigged up now and may well be rising as I type.

Now something else has occured to me: what about a relay activated by a sort of motor mechanism or in practive it is a sort of sprocket system and the relay is switched by simply pulsing it so once the state has changed it needs no current to keep it on, I know these ar used in light circuits in houses so that any of a number of pushbuttons can activate one light but those were for 240 volts, do they exist at 12 V ?

 

[kinarfi]

although won't the contact resistance increase with age ?

Are familiar with the concept of contact wiping? When the contacts first make contact, they are moving, duh, but as they come to rest, they have slid against each other and thus wiped each other clean. On some larger relays this is quite visible as it happens when relay is closed manually.

 

[nigelwright7557]

I would go for a relay as the MOSFTE will need a hefty heatsink and drive circuitry.

 

[Ubergeek63]

I would go for a relay as the MOSFTE will need a hefty heatsink and drive circuitry.

not the ones I had suggested:

http://www.irf.com/product-info/data...3004-7ppbf.pdf for <1.5W @40V

http://www.irf.com/product-info/data...fp4310zpbf.pdf x4 for 100V

at 1.5W each, either solution only requires a simple heat sink or 2sqin copper on the pcb... not even that if you want to consider the unattended TO220 sufficient (typically they CAN handle 2W in free air)

Dan

 

[Thunderchild]

Are familiar with the concept of contact wiping? When the contacts first make contact, they are moving, duh, but as they come to rest, they have slid against each other and thus wiped each other clean. On some larger relays this is quite visible as it happens when relay is closed manually.

yes I see your point

 

[Thunderchild]

not the ones I had suggested:

http://www.irf.com/product-info/data...3004-7ppbf.pdf for <1.5W @40V

http://www.irf.com/product-info/data...fp4310zpbf.pdf x4 for 100V

at 1.5W each, either solution only requires a simple heat sink or 2sqin copper on the pcb... not even that if you want to consider the unattended TO220 sufficient (typically they CAN handle 2W in free air)

Dan

I'm starting to tend that way myself, ok I rechecked figures and the temp rise for the relay was more like 14 C but that is still 5.79 C/W and I've just run an unloaded relay, I've yet to account for drive circuitry and contact dissipation, mosfets may prove to be the way, honestly its a case of just testing it with all three solutions

A latch / pulse relay would be really good though infact perfect but I can't find any at a reasonable price AND size/format

 

[crutschow]

P-channel MOSFETs always have a higher on resistance than N MOSFETs. It's something to do with the charge carriers, I think electrons are better than holes, I can't remember in great detail, someone else will probably fill in the gaps.

Yes, the majority carriers are electrons in N-doped devices, and holes in P-doped devices. Electrons have higher mobility than holes which affects the frequency response and on-resistance of the devices. The difference in mobility is related to electrons being faster then holes (which are really the absence of electrons in the semiconductor).

 

[k7elp60]

so basically you say that relay contact resistance is far lower than wht they often tell you ? I was hoping to use a relay as it is very robust but I also want a solution that makes as little heat as possible, I expect experimentation is the only way of find out a few facts, although won't the contact resistance increase with age ?

Some manufactures list spec's for contact resistance and some don't. I'm sure that the contact material affects the contact resistance and I think were people get into trouble with relays is using them for switching other than their intended purpose. For example if one were to use a relay rated for AC and use it on DC, I would expect the contact life would be much less. I did some more research on automotive relays and was impressed with the results. Some for example that are used for resistive loads like lamps, there are some that have coil resistances near 200 ohms. They even have dual coil ones with DPDT contacts that are designed for h-bridge configuration.
So I agree completly with Nigel, and I am going stick with automotive relays, especially for random switching of DC current.

 

[Thunderchild]

well this application is a bit more complicated than a relay sitting somewhere under the bonet, this is to carry the full power of the cars dynamo (old car) and be placed in a box that dissipates heat badly (as any box would) like 10 C/W