Switch-mode power supply cutting out

[Pete Verdon]

I have a Vent-Axia HR25 fan (

). It has a separate switch-mode power supply with both 6v and 24v outputs (for the two fan speeds). Rather than use the pull-cord switch, I've put together an electronic timer/relay module with inputs from a couple of sensors to switch between the speeds as required.

The control module requires 12v. Since I don't want to have another supply stepping down mains, I've arranged it to be powered from the fan's supply, via a voltage regulator chip (item L7812CV on this page:

). This is fed from the 24v supply.

When I built and tested it, this seemed to work fine. However, now that the fan is in place and I'm trying the system out for a bit longer before finally mounting the electronics, I have found a problem. After some period of time (between ten minutes and half an hour - I haven't pinned it down exactly) the fan stops. At first I thought it was a problem with my electronics, but it turns out to be the fan power supply cutting out (no voltage across its outputs). After being turned off for a while, it will work again, until it once more cuts out.

Does anyone have any advice on how to sort this? I am no electronicist; all the non-standard stuff here was put together from kits and a basic reading of component datasheets.

Pete

[NT]

p://

assuming the current draw of your extras is minmal, then its time to replace fan or psu. Hopefully your mods can be undone so nothing shows. You could check average fan current draw.

NT

[Adam Aglionby]

page:

). This is fed from the 24v supply.

L7812 is an amp in a bucket of water, they run hot , its a linear regulator which means its having to dump 12V from your 24V as heat, being a clever integrated regulator its probably entering thermal shut off land. Add a bigger T0 220 heatsink, or bit of bent alloy or insulating wahers to metal case.

Cheers Adam

[Pete Verdon]

That's what I first assumed. But it's not the regulator that's shutting down, it's the boxed PSU for the fan.

Is there something the regulator might do to the supply when thermal shutdown happens, causing an upstream device to get upset?

Pete

[NT]

No. Like most electronic faults, your only way is to find out for yourself. Measure fan power consumption, chieck its in speck. Measure motor resistance as the rotor is turned slowly, chekc there's no short. Check your mod current drain. Then you'lll know which bit is failing.

NT

[Calvin Sambrook]

Yes. My guess is that the 7812 is shutting down because it's too hot and that's causing the relay to turn the fan off (which is what you'd expect) and the SMPS has a minimum load so it shuts down too.

How about re-designing the control stuff to run off 5V. Power it from the

6V but don't use a regulator as there's probably not enough headroom but, having worked out how much current your control stuff takes and assuming it's not too great, consider a simple resistor and zener arrangement. Or do the control at 24V. More, many more, details needed really - like some idea what the control circuit looks like.

[john]

I'd be suspecting your timer/relay module is taking just a bit too much power from the 24V rail and the SMPS is falling over after a while because it's getting to hot. The fan needs 1Amp at 24V at it's "Boost" setting. 'Boost' usually implies only intermittent operation of the fan at this max power, which means the SMPS can be built cheaper, down to maybe a tiny '6W' continuous rating. What I'm getting at, is your control gear may be taking a couple of watts and the 7812 dropper is losing maybe a few watts. Add on a fan

2W load and you could be over the top for the SMPS static power consumption.

[Pete Verdon]

The relay is wired with the fan on common, 6v on NC and 24v on NO. So the fan switches between the two speeds, but is never off.

Not an option unfortunately - it's a pre-built kit rather than my own design. Building it from scratch is probably not *quite* beyond me if done with a microcontroller, but it's a lot more work than I'm prepared to do.

I think it's this one:

Pete

[Pete Verdon]

That sounds plausible. As has already been said, the regulator is having to lose half its input power as heat, plus running a reasonable-size relay at precisely the time that the fan is pulling most power. Some parts of the fan assembly were very lightweight, so the thing is clearly built down to a price.

Sod's law being what it is, the system hasn't cut out since I started this thread. I'm inclined just to leave it and see if it happens enough to be annoying. Or is cutting out like this likely to damage the supply?

Cheers,

Pete

[Calvin Sambrook]

Well that claims to take 55mA tops which, at 12V, is 0.6W. So your 7812, which is also dropping 12V * 55mA, is dissipating 0.6W as heat. Your SMPS is seeing an extra load of 1.2W

From your description of the way you've wired it I have to say it sounds like you're overloading the SMPS, it must be really near the edge in normal operation. A quick test would be to power it from an external power supply and make sure it all works OK.

It's probably not easy but if you could arrange your sensors so that the timer was triggered when the low speed was required rather than the high (and swap the 6V/24V at the relay obviously) then you'd only be adding 0.72W when the fan was running at full power which is when it matters.

[Pete Verdon]

Thanks. I guess I originally thought that a small box of electronics would be trivial set against a great big fan motor - except that in actual fact the fan isn't all that big.

I don't think I can reverse the logic as you suggested, so I guess I'm going to need to power it separately. Any suggestions other than a small wall-wart?

Pete

[Adrian C]

Also, check the 7812 has a couple of nearby 0.1 microfarad caps on both input and output legs of the device. They are known for oscillating without, and drawing excessive current in this fault mode.

[Pete Verdon]

Yep, did that. Couldn't swear to the actual values as they were out of my parts bin, but they would have been reasonably close to what was recommended and I guess the value's not too critical.

I've fully subscribed to the fan-supply-overloaded hypothesis and am now just looking for a suitable small supply for the electronics only. Most packaged supplies I can find seem to be too big, and therefore presumably inefficient at the required load.

Pete