Connecting two different voltage outputs

[AxleD]

Hi,

I am trying to build the attached circuit which is just a simple "switchable" 3V and 5V  power supply controlled through a rotary switch. If i want to have only a single output, can i connect the output of the 7805 directly to the 3V zener ouput and then to a banana socket? Seems potentially problematic and diodes at the output may be a better option?

Thanks

[German_EE]

There is a lot of duplication there that you can dispose of if you want a 3V/5V supply with a single output. Look carefully at the 3V supply, by putting a second zener diode in series you can generate a higher voltage output.

So, use your existing 4,3V device and put a 2V zener in series, this will give you 5V. Short out the 2V zener and your output will fall to 3V.

Note, you may need to play around with diode voltages until you get exactly what you want.

[AxleD]

That's a very neat solution and i will note that down for future projects.

The circuit is indeed overly complicated and wasteful. A variable LM317 supply would be much cleaner and easier but I am relatively new to electronics so this is more of a learning exercise in "design", modelling in spice, cross checking theory and calculations, bread board setup and building.

I have included only the 3/5V output in the attached circuit but the final supply will have a 6V, 9V and a 12V output as well (so a total of 5 output voltages). Hence my question around having one output rather than many separate

Cascading zener diodes (or regulators) in series and shorting the right junctions is a possibility but any idea if the diode idea at the output will work? Any ideas on another method?

Thanks

[Seekonk]

If you want to experiment, replace that zener with a TL431.  The voltage divider for it will go right to the output and will hold the voltage right at 3.3 as the load changes.  The 431 is a great 3 pin chip you should get to know.   Found in many PC supplies and older wall warts.

[Christopher]

Why not just use an adjustable regulator and a HEF4051 analog mux for the vref resistors and a rotary switch?

Learning is all very good but part of the engineering learning process is choosing the simplest solution that gets the job done.

Also: tinycad sends me loopy 

[IanB]

That's a very neat solution and i will note that down for future projects.

The circuit is indeed overly complicated and wasteful. A variable LM317 supply would be much cleaner and easier but I am relatively new to electronics so this is more of a learning exercise in "design", modelling in spice, cross checking theory and calculations, bread board setup and building.

Engineering design is about simplicity. Don't make things complicated for the sake of it, unless you are doing an art installation.

Cascading zener diodes (or regulators) in series and shorting the right junctions is a possibility but any idea if the diode idea at the output will work? Any ideas on another method?

The simplest method is to use a single voltage regulator and use the switch to adjust the set voltage. Simplest of all would be to use a single LM317 for this with switchable resistance ratios for the voltage setting. This after all is what the LM317 is made for.

[AxleD]

Thanks everyone, I will look into all the suggestions.

I know there are simpler solutions but experimenting seems to provide me with a far better opportunity to learn. I tried electronics a few years ago and took the route of using the most common approach to solve problems which often involved an IC or PIC solution which made things allot easier.

The result was that i learnt slowly and ultimately did not have sufficient knowledge around the basics. I am trying to change that this time around but point taken on the value of simple solutions. This circuit is probably far too complicated for what its supposed to do.

As an aside, the final circuit of mine also includes a seven segment display, controlled through a diode matrix, and a Wilson current mirror to control current through the supply . Completely unnecessary but a very valuable learning exercise.

[Greorge]

What sort of current is it going to deliver?

Interesting to see diagram for all the voltages you have in mind.

[Psi]

+1 for a lm317

[Alex Eisenhut]

Besides all the good comments so far, using a switch like that can lead to two problems
1) If you're switching with full load current, you'll arc, either damaging the switch if not properly rated, or creating havoc for the load
2) Cutting the input to a 7805 regulator could run the risk of Vout > Vin due to the load's input capacitance.
I've always been told that having Vout > Vin on a 78-style regulator is bad for the regulator, maybe even destroying it.
Different datasheets say anything from nothing to outright warning against it.
Since the cure is a reverse diode across the Vin and Vout terminals, I always put one.

and

3) Also, there is no adj pin on a 7805, it's gnd.

I don't know what " a seven segment display, controlled through a diode matrix," means. Like the old frequency displays on desktop PC cases? If all you want is a more or less accurate display of voltage, get a nice ebay voltmeter module.

I whacked one into my super-old college-project power supply.

http://www.ebay.ca/itm/0-99-9V-Panel-Meter-Digital-Voltmeter-RED-LED-DC-One-Decimal-Digit-Display-/330813058644?pt=LH_DefaultDomain_0&hash=item4d05feae54

[AxleD]

The circuit delivers around 260 ma with reasonable supply compliance (3% output voltage change from 0 to 260 ma). Anything above 260 ma and the limiter kicks in to reduce the voltage and ensure that max current in the event of a short is around 300 ma. Haven't built the full circuit though and that was just with limited testing.

Thanks, your explanation is clear on why a reverse diode is required. I opted for no reverse diode on the regulator because i was thinking of having one at the output which would then prevent load current flowing back to the regulator.

I was wondering about the impact of switching between the the different voltage sources and the safety of doing this. I will look up switch arcing for more detail. I suppose that this is an issue with any mechanical switching mechanism and just not this configuration.

Not sure about how those old frequency displays were controlled. In a diode matrix i use diodes at the connection between the supply and each input LED of the display. That's a diode with every connection which results in many, many diodes to display numbers 3,5,6,9. I have just ordered one, thanks!