Disclaimer - I am new to everything in the hardware world. I've been a software engineer for 12 years and have maybe 1 week of tinkering behind me.

I would like to build a fun/retro/useful set of analog indicator gauges that report on real-life events. I can handle network stuff and I will work my way through the microcontroller part but I cannot find suitable gauges. Everything I can find is way over the voltages I am dealing with. The easiest path I can imagine to getting unblocked would be some way to linearly (or at least predictably) map the approximately 0-5vdc range onto the 0-15vdc range so I can use readily available 0-15vdc panel voltmeters as my gauges.

The ways that seem to be possible given my extremely limited exposure:

1. Some simple transformer. I have no idea what the curves look like or if you can get anywhere near linear.

2. Some op-amp setup. I am way out of my league here. I could build it if I had a schematic but even the surrounding nomenclature is intimidating right now.

3. Some simpler transistor setup.

I would love to know if a secondary source of power would be necessary (my op amp reading seems to say they generally require + and - sources at a higher range) and other concerns like that. Alternately perhaps there is some ic that I could buy that does what I need without reinventing the wheel?

Many thanks, I suspect I will be a regular reader of this forum as I build my knowledge.

 

Maybe you can use a suitable ammeter with the appropriate shunt resistance to attenuate the signal to give you full scale for say a 5 volt DC input signal.

Here is the section on ammeters in the AAC ebook. You may get a better idea of what is involved by reviewing the material it contains. Ohms law will get you through most of the calculations you need.

hgmjr

 

hgmjr, you beat me to it

Actually, I was going to suggest starting here:
http://www.allaboutcircuits.com/vol_1/chpt_8/2.html

It shows you how to take a 1mA meter, and adapt it to measuring arbitrary voltages by using known resistances in series with the meter.

Note that a 1mA meter will pass 1mA of current at full-scale deflection. This is of little consequence when measuring such things as computer power supplies. However, a 1mA drain can be quite significant when measuring low power circuits, and the drain of the meter itself needs to be taken into account.

While you're at it, study up on the Wheatstone bridge.

 

Very interesting idea given that I don't really care what type of gauges I use, just that I can use their whole range. I understand the scaling and the effect of the resistor in series as a voltage divider so I'll follow up with one more question before I try to source anything.

Let's say I would like to show 4 distinct "values" on my meter scale. One way would be to design for the full sweep using a particular resistor and then to choose 3 other values that give me the 3 additional meter steady states. I could do that with four microcontroller outputs pretty simply but I end up using more outputs than I would like. If I understand correctly, a digital potentiometer could somehow come to my aid here. Can you give me a quick indication of how hard those components are to use? I would of course have all the math done beforehand. Thanks again.

 

If you use the digital pots with SPI interface you could put them all on the common SPI bus and address them one at a time as you deem appropriate.

This would of course entail a bit of extra complexity in programming to implement this approach. Still it has the advantage of reducing the number of lines used as you add additional meters to your project.

hgmjr

 

One way to do it is to use a D to A output to send voltages to the milliammeter. With a scaling resistor (read voltage to current converter) you can predict the reading for any applied voltage. Use the uC to do the scaling for you by means of the voltage produced by the D to A converter so the meter reads what you want.

 

These sound like excellent ideas. One more to add to the fray - what would happen if I used a pwm output to pulse the meter connection? The Make book I am reading suggests using that technique to get one-wire control over a true voltmeter because the meter responds slowly enough to smooth out the pulses. It is not clear to me what will happen if I am using say a 500uA meter. Any chance this technique would work here? It has the added benefit of not requiring me to buy or learn any new components this particular week.

 

Sure, that would work.

In this case, consider using a small capacitor in parallel with the meter movement; that will smooth the pulses somewhat.