getting multiple voltages out of a single wallwart?

[dentaku]

Someone gave me an old (early 90's) InterTAN/RadioShack 300 in 1 Electronic Project Lab. I never had one of those as a kid so it looks like fun and the breadboard on it is rather nice.

Anyway... it was stored for many years with six AA batteries in it an of course they leaked everywhere and even ate away some of the metal inside the battery compartment.

I don't like batteries. I'd like the put a DC socket on the back of the enclosure and power it from a DC adapter BUT the way the positive rail on the breadboard is made is interesting.
It's split into 1.5, 3, 4.5, 6, 7.5, and 9V. I'd like the keep it that way instead of just plugging a battery into the + rail like you would on a regular breadboard.

QUESTION
Can I simply use an old 13.8V 750mA DC adapter through an LM317 adjusted to 9V and use 6 resistors in series to make a long voltage divider that will give me all those voltages?
I tried 6 resistors of equal value with a 9V battery and it gave me all the correct voltages at the intersection of the resistors but I still don't know if this is really a practical way of doing it.
Will I just end up with a bunch of warm resistors if I try to power anything with this?
Will it be able to only supply tiny amounts of current?
If it IS a reasonable way of doing this, roughly what size/type resistors should I use?

[Corporate666]

QUESTION
Can I simply use an old 13.8V 750mA DC adapter through an LM317 adjusted to 9V and use 6 resistors in series to make a long voltage divider that will give me all those voltages?
I tried 6 resistors of equal value with a 9V battery and it gave me all the correct voltages at the intersection of the resistors but I still don't know if this is really a practical way of doing it.
Will I just end up with a bunch of warm resistors if I try to power anything with this?
Will it be able to only supply tiny amounts of current?
If it IS a reasonable way of doing this, roughly what size/type resistors should I use?

I wouldn't do it that way... the voltage will change with the load, and because it's a breadboard, you would likely have various different loads and therefore various different outputs from your resistive dividers.

Regulators aren't expensive, so I'd just create your voltages with a regulator for each voltage you want, if you are set on using wall power vs. batteries.

[alm]

Voltage dividers work, but only for low currents, for example biasing a small-signal transistor. The problem is that the current draw by the load affects the voltage if it gets in the same order of magnitude as the current flowing through the divider without load (quiescent current). If you have a divider of three 300 Ohm resistors connected to a 9 V source, so 10 mA flows through the divider, and you connect another 300 Ohm resistor that should draw 3 V / 10 mA based on Ohms law, it will drag down the voltage across that part of the divider to only 1.8 V (because it will make it effectively a 150 Ohm resistor).

To avoid this, you want the quiescent current to be at least an order of magnitude greater than the current drawn by the load. So for a 3 V / 10 mA load you'd design a divider with 30 Ohm resistors for a 100 mA quiescent current, which would be down to about 2.8 V with 10 mA load (still fairly close). But you're wasting 9 V * 100 mA = 900 mW for a 30 mW load (3.3% efficiency), and would need 1/2 W resistors in the divider to prevent them from overheating. For a 3 V / 100 mA load you might design a divider with 1 A quiescent current. This would be wasting almost 9 W, which would require some serious power resistors and a bigger power supply, just for that 100 mA load.

In practice voltage dividers are mostly used when the current draw is negligible, so the quiescent current can also be small. In general these are not considered power supplies because they don't supply substantial currents. One example would be your DMM that likely has a 9.9 MOhm + 100 kOhm voltage divider to divide the voltage on the 200 V range down to the 2 V that the ADC can handle. The current flowing through this divider is tiny, and the current flowing to the ADC is even less. For anything anything drawing more current you'd use a real voltage regulator, for example an LM317 or 78xx type regulator. You could even make the voltage from the LM317 adjustable using a pot.

[lapm]

Better solution would be get heatsink to that lm317 and use it to regulate voltages you need. As other have already pointed out resistive division docent work on variable loads in way you describe it.

But then again i'm sure you can figure out how to calculate correct resistor values for lm317 to give you output voltage you want. It is after all adjustable voltage regulator. Get one of those rotary switches or something to choose resistor you want.

Also recommend reading datasheet of LM317, good info and example circuits there.

Personal note: LM317 is nice little device and i have used it in quite few projects myself. Not the most efficient, but usually gets the job done nicely when i'm not running off battery power.

[amyk]

Keep in mind that linear regulators dissipate the voltage difference as heat, so with 13.8V@750mA you will get far less than the ~10W the adapter can provide. An array of buck converters might be a better, cooler solution.

[John Coloccia]

Personally, I would get a 6 batter AA holder, solder it in and get on with it.  If you really want to do this, though, to do it right you really want to do it with switching regulators.  Voltage dividers work best when you're drawing negligible current.  For example, it's a great solution for biasing a transistor, or maybe an op-amp input.

[John Coloccia]

You know, something occurs to me.  They have those wallwarts that provide a selectable 1.5V, 3V, 4.5V etc.  How do they work?  I've not taken one apart, but I would have to think that the simplest way to make one is with a multi-tap transformer.  See where I'm going with this?

If you get your hands on one of those and take it apart, and assuming it works the way I'm guessing, you can probably modify it pretty easily to bring all of the taps out, and you'll have your voltages right from the transformer.  Now you do have to be a little careful about exceeding the current ratings when you're using multiple taps at the same time.  For example, if the thing is rated for 100W, for example, you need to realize that it's the TOTAL draw, not 100W for each individual tap.  Other than that, that would be a nice, simple solution.

And try not to electrocute yourself, OK?

FYI:  you will have to rectify and filter each tap.  No big deal, but I wanted to mention that.

[Psi]

i always assumed those adjustable ones with the little voltage setting switch used a LM317 or similar adjustable regulator inside.

Multiple taps on the transformer would probably cost more in copper than a cheap regulator IC

[madires]

i always assumed those adjustable ones with the little voltage setting switch used a LM317 or similar adjustable regulator inside.

Multiple taps on the transformer would probably cost more in copper than a cheap regulator IC

It's not just the regulator, you also need a heatsink and a larger box with venting holes. Those universal wallwarts I've seen all got multiple transformer taps. But there are also universal SMPSUs where the voltage is set by a resistor in a small plug (like a jumper).

[sleemanj]

I haven't seen a selectable voltage PSU with an actual transformer in a long time, 20 years or more. 

What is really needed here is some very small buck converters.  I'm sure I've seen somewhere such devices constructed as 78xx TO220 direct replacements and the same size, but not very common or cheap that I recall. 

I wish they were more common, because they'd be damn handy to just be able to slap one in place of a 78xx (as long as you didn't care about the switching noise in exchange for the efficiency gains).

[Zero999]

The power dissipation problem can be solved by putting the LM317s in series but you need to make sure there's enough headroom for the drop-out voltage.

Connect: the 9V and 7.5V regulators to 13.8V, the 6V to the output of the 9V, the 3V to the output of the 6V, the 4.5V to the output of the 7.5V and the 1.5V to the output of the 4.5V.

[John Coloccia]

I haven't seen a selectable voltage PSU with an actual transformer in a long time, 20 years or more. 

What is really needed here is some very small buck converters.  I'm sure I've seen somewhere such devices constructed as 78xx TO220 direct replacements and the same size, but not very common or cheap that I recall. 

I wish they were more common, because they'd be damn handy to just be able to slap one in place of a 78xx (as long as you didn't care about the switching noise in exchange for the efficiency gains).

They're still quite common in the US, and I'm assuming Canada too.  You can get a cheap one for less than $10.  A lot of them don't go down to 1.5V (though some do), but if you REALLY needed 1.5V for a circuit, at that point you could just build a regulator on the breadboard.

[mariush]

You know, something occurs to me.  They have those wallwarts that provide a selectable 1.5V, 3V, 4.5V etc.  How do they work?  I've not taken one apart, but I would have to think that the simplest way to make one is with a multi-tap transformer.  See where I'm going with this?

Old power adapters did have multiple taps and then a simple bridge rectifier would convert the selected tap to DC and smooth it out a bit with a capacitor. Here's a power adapter from around 1990-1995 (ignore the crap solder on the AC power cable and the transparent scotch tape, this adapter was played with to hell and back):






There are still really shitty power adapters sold that still use multiple taps, for example here's one that was bough from a "everything for 1$" type of store, but this one is actually so bad it doesn't even work right (taps soldered in wrong places, typos on label etc):








Most new power supplies that let you select output voltage are the switching type, like this good one I bought maybe 2 years ago for about $10. It's quite good:









(will edit to inline the pictures)


 

[dentaku]

I have one of those ancient adapters with a switch on it.
It seems pretty horrible

Here's what it calls 6V DC (scope set to 5V/CM, DC)

It's a Recoton like allot of the stuff from that time.


The 13.8V adapter I mentioned in my first post gives me a perfectly flat line so I guess there's a big difference in quality.

[amyk]

It could also be that it's just the filter caps have dried out over time and have basically no capacitance left.

I doubt a tapped transformer needs much more copper than an untapped one - the taps are just short stubs inserted periodically during the winding process and are negligible compared to the wire used for the whole coil (which would be the same for an untapped).

[mariush]

In the case of a classic power adapter, yes, that would be true.

However, in the case of switching power adapters like the last adapter I posted images for, there's much less copper and much less weight in the overall transformer so you not only save money on the copper itself, but also on shipping costs, and you can also offer much higher current in a smaller weight (you can't do 3-12v 1A without a 15-20va transformer, which is large and heavy)

[John Coloccia]

Are you sure you're not thinking of a multiple coil transformer, where you have separate coils?  A multi-tap transformer just pulls out a tap to a partial coil after the appropriate number of winds...and then it keeps winding.  It's the same amount of copper.