which P channel mosfet

[Horwig]

My diy dew heater controller at the moment is low side switching, which is not ideal, I'd prefer a zero volt rail and switched volts.

So, any advice on choice of p channel mosfet, googled about a bit, and the IRF4905 seems like it would do everything I need and more, but has anybody got real world experience of doing high side switching on 12v?

Huw

[Gina]

I've done it.  Wrote a thread about dew heater control.

[wasyoungonce]

AUIR3315 or common name IR3315.  I played with a lot high side DEW heater control ICs and this one is pretty much the bee's knees.

Its a P channel FET, self protected, rev polarity protected and current limit set protected (it shuts down if over current...aka short cct etc).  Basically it was made to drive automotive lamps, injectors etc so is well suited to DEW heater systems.  There are others in the family IR3313, IR3317, but the 3315 can be limited to a 3Amp output (via a resistor) which is ideal.  Pretty much if you need more than 3A per channel then ..."wow"!  I built a 6 channel PWM dew heater based on this FET.

All this said, PWM for dew control is probably not the best idea mainly due to power line noise, so, ensure you use a separate PSU to power the dew heaters than that used to power cameras etc.

[Gina]

Details of my high side switching are in this thread from here on.

[perfrej]

I suggest you use a smart-swicth. You get protection and very reliable operation. Look at Infineon's ProFET series.

/per

[Ajohn]

Just in case some  one wants to do this sort of thing thoroughly.

The maximum current level is needed in order to answer the question. Taking the IR3315 for instance that will be sort of ok with 2.8 amps continuous without a heat sink. You need to consider when the unit is first turned on. Mount it on a 60C / watt heatsink and even at automotive ratings it will be happy with 3.9 amps. Automotive ratings are at 85C ambient. Yours will probably be lower and there should be enough info in the data sheer to work this area out if needed. It's a simple calculation , say the 60C / watt was changed to 120C. The temperature rise would double and the device would go pop some where around when it reached 125C. If changed to 30C/watt the temperature increase would halve. Sums are done by adding the numbers up starting from the junction to tab figure and adding a bit for an insulator, then the value for the heat sink.

2 other aspects complicate this area. As the device gets hotter the resistance increases which makes the dissipation higher. This can result in thermal run away. The answer to this is to look at the curves and make some assumptions before doing the above sums. The other aspect is switching on and off. This often dissipates more power than when the are on continuously if the rate is high enough.  This is why IR mention a 1msec min pulse at a rate of 200hz under automotive conditions. They show a graph of the transient thermal resistance for sorting that aspect out if needed but it doesn't seem to account for repetition rates. I've been away from this sort of thing for some time now. There is probably application data around and also I would suspect a Spice model for sorting this sort of thing out under simulation. There is an interesting simulation package here called electronic work bench.

http://sine.ni.com/nips/cds/view/p/lang/en/nid/202311

It has oscilloscopes, power supplies and meters etc so is pretty intuitive to use. No idea of prices or if it has a thermometer but sometimes copies of it and similar packages crop up on ebay, even official ones. It really should have some form of temperature measurement as it's a rather thorny area when power circuits are designed. There may also be an open source etc simulation package around.

John

-

[Ajohn]

Personally I would use low side switching for this anyway. 2 reasons really. Lower on resistance is readily available and easier to switch rapidly from one state to another and also the fact that all of us have 3 core cable all over the place carrying earth, neutral and live 240v AC which seldom causes problems, if cable frays we see it and replace or throw it away.  If I was concerned I would just fit a fuse. If I was using a battery source there would most definitely be a fuse associated with that anyway, right at the battery. The gear on a telescope is usually run at something like 12v dc which has no where near as many problems as the mains does and actual earth doesn't have to have anything to do with it. 0 volt is not earth. I would be more concerned about the cable to the dew heater being up to the job current wise. For short runs like that a mains cable rating should be more than adequate and have a lot more insulation than it really needs. If that type of cable can't be found with a suitable rating cheaper heavy duty speaker cable could be a suitable alternative, no need for gold plating.

Automotive applications are a bit different because the chassis of the car forms one of the conductors and there is plenty of opportunity for cable running through bulk heads etc to chaff and cause a short. They fit fuses anyway and also use high side drivers because the part being driven is often some way from the switch itself which is likely to be in some ECU somewhere on the vehicle.

John

-

[Mick J]

AUIR3315 or common name IR3315.  I played with a lot high side DEW heater control ICs and this one is pretty much the bee's knees.

Its a P channel FET, self protected, rev polarity protected and current limit set protected (it shuts down if over current...aka short cct etc).  Basically it was made to drive automotive lamps, injectors etc so is well suited to DEW heater systems.  There are others in the family IR3313, IR3317, but the 3315 can be limited to a 3Amp output (via a resistor) which is ideal.  Pretty much if you need more than 3A per channel then ..."wow"!  I built a 6 channel PWM dew heater based on this FET.

All this said, PWM for dew control is probably not the best idea mainly due to power line noise, so, ensure you use a separate PSU to power the dew heaters than that used to power cameras etc.

Nice to see you here Brendan, first post - welcome to SGL

[perfrej]

I checked my schematic. The one I used first, and that was by far hte simplest to use, was the BTS 432. Datasheet attached. It has logic high for on, which may be convenient.

/per

Infineon-BTS432E2-DS-v01_01-en.pdf

[perfrej]

One thing that most designers miss when it comes to PWM dew control is to use a low frequency. Given the thermal properties of the object being heated, choosing 0.1 Hz or even lower is fine. I PWM my dew band in my new design at that frequency. A 50% setting gives 5 seconds on, five seconds off. Slew limit the switching and you something that disturbs nothing.

How's that

/per

[Gina]

I checked my schematic. The one I used first, and that was by far hte simplest to use, was the BTS 432. Datasheet attached. It has logic high for on, which may be convenient.

/per

Thank you Per - that seems a very useful device

[pete_l]

One thing that most designers miss when it comes to PWM dew control is to use a low frequency.

The other way to do it would be to use a constant voltage source to power a dew strip made from discrete resistors (i.e. a "normal" power supply) and switch more resistors in / out as the temperature changes. That could be done either with a simple manual switch or an Arduino and some sensors.

[Gina]

Yes, I thought of doing that.  With Peltier TEC cooling too.  PWM switching is inefficient as the power dissipation due to resistance depends on the square of the current (or voltage) and the Peltier effect just on the current so running on a constant voltage or current is best.  When using PWM the extra power dissipation has to be handled by the cooler on the hot side.

[Horwig]

One thing that most designers miss when it comes to PWM dew control is to use a low frequency. Given the thermal properties of the object being heated, choosing 0.1 Hz or even lower is fine. I PWM my dew band in my new design at that frequency. A 50% setting gives 5 seconds on, five seconds off. Slew limit the switching and you something that disturbs nothing.

How's that

/per

Yes, very true Per, I'm always afraid of splats from the switching getting into the camera data somehow, I've run my dew heaters in screened cable, and the camera is fed off a different supply to the mount, focuser and dew heaters.

Can I ask if you use a snubber network across the load?

Huw

[Horwig]

Sorry Per, just re read your post, yes, slew limiting the switching is the way to go.

Huw

[hughgilhespie]

Hi All - and particulary Per,

I am about to use some Infineon ITS436L2 profets for my dew controller. These are very similar to the one that Per sent details of. These will be used in a PWM Arduino circuit. My question is, do I need to tie the IN pin of the profet to ground with a - say 1 Mohm resistor - or can you connect the IN pin direct to the Arduino port output? 

I have searched for the answer on line but without any success.

Thanks,

Regards, Hugh

[Gina]

Arduino use a totem pole output on all digital pins set to OUTPUT.  So you can connect MOSFET inputs directly to Arduino outputs without any resistor.

[bitnick]

My diy dew heater controller at the moment is low side switching, which is not ideal, I'd prefer a zero volt rail and switched volts.

Just curious - why do you feel that low side switching is not ideal?

[hughgilhespie]

Hi Gina,

Thanks for that. Errr - what's a totem pole output?

I just had a vague memory that a high value pull-down resistor was needed with MOSFET input gates as if not pulled low, they could accumulate enough charge from static or whatever to turn themselves on. I don't know if that applies to the Profets as they have signal input conditioning , so probably not, but I don't know enough electronics to glen the information from the data sheets .

Regards, Hugh

[Gina]

A totem pole output has devices to both pull down and pull up the output as required (not at the same time of course ).  This terminology goes back to the TTL logic devices.  (TTL = Transistor Transistor Logic).  Also known in the audio field as push-pull.

[hughgilhespie]

Hi Gina,

Thanks again. My knowledge of electronics was OK with things like EF91's but is now a tad out of date!

Regards, Hugh

[Horwig]

Just curious - why do you feel that low side switching is not ideal?

My dew heaters are fed from phono plugs and sockets. With a low side, I have to isolate the socket body from the panel its mounted on. With high side switching, the phono body would always be at ground potential. A much more elegant solution.

Huw

[pete_l]

Hi Gina,

Thanks again. My knowledge of electronics was OK with things like EF91's but is now a tad out of date!

Regards, Hugh

No w*that* would make  great dew heater

[hughgilhespie]

Bring back 6.3 volts!!

[Gina]

Ha ha... but I'd rather have my Arduino that a room full of heated glass   I don't mind heated glass of it's a telescope or lens though