Fusing a Geofex Spyder Power Supply

[davent]

Hello,

Gathering parts together to make a Geofex Spyder-type power supply.  http://www.geofex.com/article_folders/spyder/spyder.htm

For transformers i have five with dual 115v primaries and dual 12v/100ma secondaries. One will be set up with the primaries paralleled and the secondaries paralleled-> rectified->LM317 regulated to 9vdc. The other four will be set up for eight secondaries-> eight isolated 9vdc outputs.

I have no idea how to best fuse the setup, options i see.

One fuse for the whole shebang period (the IEC socket)... do i add one of the following?

1. One fuse for the single transformer and one for the group of four?
-or-
2. One fuse for the single and one for each of the other four transformer set up with the primaries paralleled?
-or-
3. One for the single and one for each of the primaries in the other four?
-or-
4. Fuse all the secondaries rather then the primaries?
-or-
Have i missed the boat entirely?

Then... how do you determine an appropriate value for the fuse(s)?

Thanks guys!
dave

[digi2t]

I built a Spyder 16 (two units in one housing) not too long ago. I have mine fused at the primary, specifically, the main 120vac line coming in. Following the calculation for fusing transformers that I found on the net, I was supposed to be using around a 1/4 amp fuse. Problem was that when I would turn on the unit, the initial surge would blow it, so I`m running a 1/2 amp fuse now. Never had a problem since.

I didn`t want to use a slo-blo, because I would rather err on the side of caution, and have the protection work quickly, instead of cooking before blowing.

I wish I could remember the site where I found the calculation, but there are plenty out there. I think for an 8 output unit, a 1/4 amp fuse should resist the surge, and provide adequate protection. Do the math though, just to be sure.

Read this too; http://www.diystompboxes.com/smfforum/index.php?topic=58651.0

[psychedelicfish]

There's plenty of information on fuses here.

I reckon it would be best to fuse each transformer's primary, because it'll make fixing things easier if anything goes wrong. If you have only one fuse and something fails, you'll have the same problem as with old christmas lights, where they ran lots of 12V bulbs in series across the mains.

You might even want to fuse the outputs too, so if you make a mistake wiring the power in your newest pedal build you won't wreck the regulators in your spyder.

I didn`t want to use a slo-blo, because I would rather err on the side of caution, and have the protection work quickly, instead of cooking before blowing.

It would probably be safer to use a slo-blo fuse that's rated for the right current. If you use a fuse that will blow at twice the calculated current, you'll be able to pull up to twice the calculated current continuosly, which would be far worse than pulling more than the calculated current for a second or so.

[R.G.]

Before we start this, the normal warning: by doing any AC mains power wiring, you are betting that you do it properly and don't endanger yourself, your property, or your life, or those of others. The internet is not a sufficient source of information to get this detailed knowledge.

1. Fuse the AC power entrance.  This is a bare minimum. Use a slow-blow/time delay fuse of a rating that will let it start reliably. This must be larger than you think, because the size of the inrush current depends on exactly where in the AC power cycle the power was turned off last time and where it was turned on this time. I once had a variac that was perfectly normal, but about one time in ten blew a 20A  wall-socket breaker. I found 200A peaks on some power-ons with this. It was all dependent on how much it took to charge up the core and any filter caps it eventually powered. This fuse is a critical part of electrical and fire safety.

2. If you have the time, patience and money, fuse the secondaries. Fusing the primaries won't help, because the only realistic time a primary will over current is when the secondaries are overcurrent, or when the primary is damaged. A secondary fuse may protect the transformer from damaged rectifiers or filter caps. A regulator after the filter cap usually has its own internal current limit, so fusing after is is not all that productive. A fuse per secondary *may*  be useful,   but most professional designs don't bother.

Pick the value of the fuse by picking the smallest size that will both start reliably on input surges and stay on for maximum-normal use.

[davent]

Thank you guys for the info, links and warnings! Looks like i'll go the simple direct route with the one fuse in the IEC socket.

Next question- would something like the Kill a Watt meter be useful/helpful in determining the fuse value needed?

dave

[R.G.]

Let's talk about fuses and current.

A fuse is a resistance wire with a low melting point. The power dissipated in it is P = I²R. The melting temperature is - well, just call it T.

The temperature that a heated object gets to is a composite of the thermal resistance, degrees C per watt. The idea is that if you dissipate 1W in a thing with a combined thermal resistance of 10C/W, its temperature will go up by 10C. Easy, yes?

Unfortunately, no. Thermal resistance is nonlinear. It depends on the fourth power of the temperature difference of the thing about the things around it for radiation, linearly for conductive losses, and highly, highly nonlinearly in a non-computable way on the fluid mechanics of the gas/fluid surrounding it.

To add insult to injury, the resistance of the wire itself is nonlinear, generally being higher as the temperature goes up.

And then there's time dependence. If you heat something up ...very... slowly... you can heat it to its bulk melting temperature. If you heat it fast from an external source, the outside heats up faster, then heat conducts into the interior, so the outside will melt first. If you heat it up VERY fast, like with a picosecond laser pulse, the outside will evaporate before the inside notices any heat at all. If you heat it internally, like with current flow, you can put a pulse of current in that evaporates the whole thing before heat can flow out of it in any manner at all. You've all seen fuses with the insides of the glass tube coated with shiny metal, right? That's what happened.

As you can guess, designing fuses is tricky. And applying them is tricky.  A 1A fuse does NOT melt at 1A, or 1.01A. A 1A fuse is guaranteed NOT to melt at 1A. A 1A fuse may melt in 1 hour at 1.05A, or 30 seconds at 1.5A, or 10mS at 4A. And a 1A slow blow versus a "fast" fuse...

Well, you get the idea.

This is why on the AC input, you put a fuse that will withstand the full operating current, or a little bit more, and make it a slow blow. Slow blow fuses will withstand large pulses for a short time. This is good, because transformers and motors can easily pull 10-20 times their normal max operating current for a half-cycle of the incoming AC power. The time delay lets it get started without choosing a huge fuse to start, then melting down on normal operation without popping the big fuse.

On loads where you want to protect the insides somehow, you can use normal or "fast" fuses. The secondaries of transformers are one example. It may not be worth your time at all to do this for these small transformers because the smaller the transformer, the more it's limited by the resistance of the primary wire.

You can calculate crudely what the max currents will be. Your trannies are rated for 12V/200ma outputs. Actually, they're probably 14-15Vac open circuit, but 200ma RMS outputs. So they put out about 2.8W each at full load. On the primary, that's about 2.8W/120V = 23ma. Actually, it will be greater, because small trannies usually have poorer primary inductances. Call it 30ma each. Five of them are about 150ma. That will be the highest combined primary load you want to see. Fuse the primary at the nearest fuse you can get bigger than 150ma *time delay*/ slow blow rated.

Each output is feeding a full wave bridge rectifier. One of the quirks of full wave bridges is that the RMS/heating current they pull from secondaries is about 1.6 to 1.8 times the DC current that goes out of the filter cap. So if your transformers are rated at 100ma per secondary, and you full wave rectify them, then you can pull out something like 100ma/1.8 = 56ma per secondary on a continuous basis. Notice that the secondary IS conducting 100ma, but you only get 56ma of usable DC from it. So a 100ma fuse would protect the secondary. This may or may not need to be slow blow. The first cycle of AC coming in tries to charge the completely empty first filter cap, so it looks like a short circuit for a half-cycle or two.

I would leave off the secondary fuses and count on the current limiting of the regulators to keep you from sudden disaster.

[davent]

Thanks R.G. for the explanation and advice, very much appreciated! Digikey has a big selection of 160ma 5 x 20mm so've added some to my next-up cart.

Thanks, take care!
dave

[davent]

Finally finished this up tonight, yet to be tested with pedals but powers up fine. Used LM317 regulators i had on hand and other bits and pieces worthy of hiding under a pedal board. First boards i did with the endoscope drill setup.

Thank you for the help!
dave

[digi2t]

If the term «sexy» could apply to a power supply, I would say yeah man, damn sexy!

[tubegeek]

I don't know whether to drool or slobber!

Dude, even your light-bulb limiter is shipshape....

Can I come over and use your toys?

[seedlings]

The only problem I see is the opaque enclosures you're using    Kind of like a supermodel wearing a burka.

CHAD

[J0K3RX]

The only problem I see is the opaque enclosures you're using    Kind of like a supermodel wearing a burka.

CHAD

lmao