PDA

View Full Version : Sorry, yet another power supply question.



DICKEYBIRD
10-07-2012, 02:47 PM
I know, I know go ahead & say it...so stop with the noob electronics questions already but I have to ask.:o

Got a 36V transformer on the way to feed the 36V KB speed control for the nice Pittman motor Jim sent me. I needed an enclosure for it so I stripped everything out of a dead PC power supply except the fan, the AC plug and switch. The fan is 12VDC and I could've just stuck a wal wart in there to power it but decided to build a little 36VAC to 12VDC 500MA supply. I had everything on hand except the cap so I gave it a whirl using an LM317T regulator.

Here's what I came up with. Without the 36V xfmr on hand to test with, I preset the my Variac to 36V. When I powered it up the breaker on that circuit in the shop popped instantly. No smoke, no flash, just a couple lights in the shop and SWMBO's flat screen shut down. (Which reminds me I'm due for some circuit revamping in the shop but that's another subject.)

I know the initial current draw on that Variac is pretty high but I've used it recently plugged into the same power strip when I was testing the Pittman motor with far more power draw with no ill effects. For this test, I connected everything up and switched it on with the power strip switch. Should I have powered up the Variac 1st and then switched on the power supply or do I have something screwed up in my circuit? (schematic attached) You'd think that little thing wouldn't add enough load to trip a breaker unless it's dead-shorted somehow but if that's the case the magic smoke should have erupted. I can try the Variac on 1st then supply on test but decided to wait and let somebody else take a gander & see where I screwed up.

http://i57.photobucket.com/albums/g227/DBAviation/FanPwrSuppy2.jpg

http://i57.photobucket.com/albums/g227/DBAviation/FanPwrSuppy.jpg

ikdor
10-07-2012, 02:59 PM
36V rectified is around 50V DC. The 317T won't take that.
Also, depending on where in the line period you switched off, the transformer retains magnetic field. This can lead to huge inrush surge current.
I'll let someone else with a real keyboard come up with the solution for that :-)

Igor
On his tablet...

DICKEYBIRD
10-07-2012, 03:24 PM
36V rectified is around 50V DC. The 317T won't take that.
Oops, I thought I read it would tolerate up to 40V diff in/out?

If not, is there a cheap chip that will take 50V?

halac
10-07-2012, 03:32 PM
We used Variacs in the calibration lab to do tests on power supplies. I found that if you are using one on a circuit that already has loads on it, i.e. lights & TV, it will trip a breaker if you have it turned up any amount. I usually start out with the Variac at zero, then apply power to it. I then turn it up to the required setting.

J Tiers
10-07-2012, 03:56 PM
It almost does not matter if turned up or not... unless the output is shorted somehow. The inrush for a 15A variac will pop "home type" or other breakers about 30% to 50% of the time, depending on timing relative to the power sinewave. That is true especially with dual breakers, with two handles in one "single width" case.

If you have Q0 breakers, you can get HID lighting breakers that will deal with inrush.

Elsewise, use single-per-slot type breakers and take your chances.

Yes, DO look at the voltage ratings..... regulators have maximums from 25V to 50 V or so.

For a fan, I would usually just use a resistor... fairly foolproof, and won't waste any more power than a "linear" regulator like that.

ikdor
10-07-2012, 04:12 PM
The LM317HVT can tolerate up to 60V, so that should stay functional with your other hardware.

I wonder if you can put a capacitor over the primary to "degauss" the variac on shut-off. That way it shouldn't hit you with the high inrush current. I'll let J Tiers comment on the feasibility of that.

Igor

J Tiers
10-07-2012, 04:22 PM
Not sure it matters.....

The "detail" is at what point on the magnetizing curve the core is left at when power is off. I have no clue if the cap would act to degauss or not, my suspicion is not, unless it had a strong oscillation at a frequency the core could respond to.

And even then, you still have to establish a field in the core... whether you have to remove one first, or not.

If the core is not "frozen" at a bad point vs incoming power phase, it will draw a "smaller" surge.... no clue if it would be "enough" smaller.

better breaker has other benefits..........

Still like a resistor for teh fan...... set for the right voltage drop at a hair under the fan typical current.

DICKEYBIRD
10-07-2012, 04:43 PM
Thanks for the chip tip Igor.

The Variac was just a temporary method of testing before my ebay xfmr gets here. I have a paying job almost due so the I'll just wait 'til the other one gets here to try again and put that big 'ol thing back in its hole.

DICKEYBIRD
10-07-2012, 04:53 PM
Still like a resistor for teh fan...... set for the right voltage drop at a hair under the fan typical current.Thanks jerry; simple's good too.

So if I've got 50.4VDC from the bridge/cap and the fan draws 200ma, I'd need a 250 ohm resistor?

darryl
10-07-2012, 04:59 PM
The current pulse to charge that 1000 uf cap is a small fraction of that drawn by the variac when it's plugged in. It wouldn't matter if you had the setting up to 36 volts already or not. Usually you would first power up the variac, then while monitoring the dc voltage on the cap, you would slowly bring up the dial. The voltage reading should rise pretty much linearly with the rotation of the variac dial. If it stops rising before you have reached the correct setting, then you have a circuit problem.

I would suspect though in this case that there are enough loads already on that circuit in your house, and switching on the variac (probably at a moment when the ac input voltage is near a peak) cause enough extra current draw to trip the breaker.

One of my variacs can be switched on once and all is well, but if it's turned off then on again within a second or two, a breaker will pop. It does have to charge a good sized cap when it comes on, and if you don't use the load side on-off switch, the load discharges the cap when the variac is turned off. So a turn on, turn off, turn on again has the variac delivering the cap surge current twice in short order. Breaker pops. In this case it's the power supply for my lathe motor, so I don't usually ramp the dial down- I'll power the variac, then use the motor switch when I'm ready. I adjust the dial only when I need to change the speed.

On your math- you have 50v to start with, and the fan needs 12v. That leaves 38 volts to be 'lost' across the resistor. R= 38/.2 which means 190 ohms. You could use a 180 ohm and have about 13 v to the fan, or use a 220 ohm and have about 11v to the fan. It will work either way, though if you don't need the full volume of airflow I'd go with the 220 ohm. Slightly quieter running fan as well. The power rating should be about 10 watts, which is about the same as the 317 would have to dissipate on a continuous basis. That's a significant level of heat being developed by the regulator, probably requiring a larger heatsink than you are using. I'd use the resistor instead of the regulator myself, since I'm not fond of running semiconductors at high heat.

Barrington
10-07-2012, 04:59 PM
Oops, I thought I read it would tolerate up to 40V diff in/out?

If not, is there a cheap chip that will take 50V?

You were right the first time. The 317 is doesn't 'know' at what voltage it is sitting above 0V - the spec is, as you say, the voltage dropped across it.

Some questions you might have considered in an effort to diagnose the problem:-

Q1: Does the variac trip the breaker with no circuitry connected ?

Q2: Does the variac trip the breaker with just the rectifier and capacitor connected ?

Q3: Is it an earth/ground leakage trip and do you have some point in the circuit connected to ground ?

Q4: Or even - Is the variac wired the right way round - i.e. the neutral straight through to the output ?

In general I would not advise the inexperienced to use a variac in this way - it's very easy to forget that the low voltage circuitry you're tinkering with isn't isolated...

Cheers

.

DICKEYBIRD
10-07-2012, 05:42 PM
Hmmm, interesting. I didn't have a resister near that value to try but I wanted to try my supply again based on Barrington's statement about the 317. This time I powered up the Variac first (breaker didn't trip) then switched on the supply (I put a 2A fuse inline this time), the fan started to run and the fuse blew. Not sure if the 317 didn't like the high voltage or the fuse (which is what I had on hand already in a holder) is just too small.

I don't really have enough stuff in stock to mess with it any more so I think I'll go with the wall wart. I know it'll work and it'll be safe.

I did learn some things though and that's a good thing. I cut over 300 pink foam model airplane wings using the same Variac on the same outlet 10 yrs. ago and always wondered why sometimes when I turned it on the lights would go dim for a split sec. and sometimes not. Now I know.:)

darryl
10-07-2012, 05:55 PM
Good point, Barrington, about the variac output not being isolated from the line. On my lathe power supply, the variac I used was purchased damaged, and I re-wound it with an isolated secondary. It's not a true variac anymore, since I made several taps on the secondary and I just switch between them for my various voltages.

For a safer way to use an unmodified variac, find yourself a working transformer from an old amplifier, preferably one which was quite powerful. You can power the primary of that from the variac, then take your voltage from one of the secondary windings. Besides now having an isolated voltage, you also often have a grounded shield between the primary and secondary windings. If you ground this with your 'test setup', you have even more isolation from live wiring.

You are going to find various voltages on the secondary windings, so you would probably want to mark them. The setting on the variac no longer gives a direct readout of the voltage, but you can either make up a chart or simply use a voltmeter whenever you are dialling up a custom voltage.

The amplifier transformer may be fully encased, but it's easy to figure out the secondaries. If it's not cased, you might also find that there's room around the core for perhaps one layer of extra wire. You can wind something on there yourself if you're ever needing a very low voltage, high current source, such as for hot-wire cutting of foam. You still have control with the variac, which makes it easy to set the exact operating parameter for any particular length of hot-wire.

J Tiers
10-07-2012, 06:30 PM
Isolated type "variacs" do exist, but are rare.

About 200 ohms... 12V is for the fan.... 0.2 x 200=40 Will dissipate 8W... so will the regulator.

The Artful Bodger
10-07-2012, 07:43 PM
I would go with the wall wart as I am not too sure what would happen with a resistor to drop the voltage to an electronic fan motor. My fear would be that the motor electronics would see the full 36V at startup and before current has began to flow.

jkilroy
10-07-2012, 07:58 PM
"The inrush for a 15A variac will pop "home type" or other breakers about 30% to 50% of the time, depending on timing relative to the power sinewave."

Thats not my experience. I have a variac running into a rectifier and a large cap for a variable speed power supply for a HUGE shop fan. Never blown a breaker yet, the circuit is a 15A GE style breaker, nothing fancy.

J Tiers
10-07-2012, 09:09 PM
I would go with the wall wart as I am not too sure what would happen with a resistor to drop the voltage to an electronic fan motor. My fear would be that the motor electronics would see the full 36V at startup and before current has began to flow.

it actually works quite well, I have done it in a number of commercially successful products. If you are nervous, a zener at somewhat over the fan voltage will limit spike/surge voltage, and won't be "on" for long if at all. I suggest a couple hundred uF of capacitor across the fan as well, which both limits the rate of rise of voltage, and provides some stability to the fan voltage while running. In at least one case, the fan manufacturer actually had the resistor approach shown in an app note (don't recall who, now).


"The inrush for a 15A variac will pop "home type" or other breakers about 30% to 50% of the time, depending on timing relative to the power sinewave."

Thats not my experience. I have a variac running into a rectifier and a large cap for a variable speed power supply for a HUGE shop fan. Never blown a breaker yet, the circuit is a 15A GE style breaker, nothing fancy.

We specified high inrush for the lab where I used to work. the contractor actually put in boxes that didn't have that option available, and installed a LOT of the dual breakers. Must have been cheaper...... but they were a big hassle.

We had a lot of inrush-related breaker opening, and we had them on and off a lot. Ditto in several other locations. Where I am now, much less trouble, but it does happen.

It depends on the breaker type, breaker brand, etc...... actually it depends on the time-current response curve of the specific breaker in question. if you don't have problems, great..... I have seen it rather a lot, however, and often where the contractor coaxed the building owner into allowing a "home type" breaker box instead of a good industrial type, or if the "two in one" type breakers were installed.