LM3914 and sound output!?

[jhovel]

Hi all,
I need to record changes of voltage in a circuit over time. I need to record this for days at a time in an analysable format.
The voltage will only change in known jumps (probably no more than 6-8 values) and usually for short times only (milli seconds to seconds) between 0 and +12V.
So I figured maybe an LM 3914 would dothe job, if I could drive a tone generator and I pipe the output into the microphone input of a laptop. I could then simply record the "music" in a low resolution wave file format. as an aside, has someone got a suggestion for a particularly efficient codec?
Anyway, I'm no electronics expert at all, but handy with a circuit diagram, a multi meter and a soldering iron
How would I change the pitch of a tone generator (and what kind to use) from the output of the LED Voltmeter using the LM 3914 in dot mode?
Maybe there is a way of making a tone generator voltage dependent directly?
Is there a better way to go?
Any help will be greatly appreciated!

 

[chemelec]

Why not use a Voltage to Frequency Converter?

An "EXAR" XR4151 is an Excellent Example of these devices.

 

[Nigel Goodwin]

I think no matter how you do it, you are probably going to get some massive files!.

Assuming you use an 8 bit A2D and sample at 1mS intervals, which you will probably have to do to meet your short term variations, you're looking at 86MB per day!.

Converting it to an audio frequency and recording as WAV files is only going to make the files larger.

What exactly is the application?, there may be an easier way - for example, only storing changes outside a certain range, and storing a timestamp along with the data when they occur.

Assuming you require millisecond resolution, you should be aware that Windows isn't very good at accurate timings.

 

[jhovel]

driving small relays with LM3914?

OK, I hear what you are saying. I have since talked some more to my partner about the issues and he suggests that we should have the facility to record for a week or two! That makes sound recordings way too big!

I've gone back to my original idea of recording voltages via LM3914 and relays - using a security system....

What kind of voltage and current can the LM3914 "supply" to drive small relays? or do I need to give that task to a switching transistor for each output?

Any ideas where I can find a circuit to do this?

Someone asked for application information: I work part-time in the security system industry helping a mate troubleshoot systems. We come across a few systems where there are 6 IR motion detectors wired in series (all of them normally closed). When things go wrong, they alarm (open their relays) lots of times at random times and for random periods.
The security guard goes there and finds absolutely nothing - all the while the alarms keep going (or start again later or the next night).
Of course because of the way they were wired, we can't tell which one is doing it!
These things are expensive and we can't just replace them all without good reason, and often there are environmental reasons for it - eg warm air circulating in weird patterns, even mice having a party on a curtain rod near a sensor or whatever - so that replacing them does not solve the issue!

Our solution is of course to rewire the building when it can be justified at all, but mostly these are schools with low budgets who can't afford that.

To track where the hits come form, I want to put a resistor across the alarm terminals of each sensor - a different value for each sensor. That way I could simply measure the voltage drop when one sensor opens and identify it by its "voltage drop signature".
It is relatively cheap (here, if you know the right people) to have an alarm panel monitored without getting any guard responses. So we now want to set up a testing panel/dialler with six inputs - provided by the relays pulled by the LM3914 in response to six different voltages. The friendly security firm then e-mails us a log every night or week or whatever and we will know which sensor went into repeated alarms (called 'continous single hits' or 'multiple alarms' in the industry). We can then either replace the sensor that plays up or relocate it away from a heat source or move the mouse dance floor away from the sensor.
The system would be powered on along with the existing system, which also passes on the alarms on the "whole circuit" for responses as usual.

I hope that helps a little in understanding the issues. Sorry it is so lengthy....

All we need now is a circuit to drive 6 relays in response to six voltages

Cheers,

 

[Roff]

I think this will work for you. The LM3914 is programmed for 10ma per output, so you can use either 5 volt or 12 volt relays. The outputs will saturate if you use 12 volt relays, but should still operate OK.

 

[jhovel]

Thank you very muich, Ron!
That looks exactly like what we need!
I'll build one and let you know how it works.
Cheers,

 

[Roff]

Thank you very muich, Ron!
That looks exactly like what we need!
I'll build one and let you know how it works.
Cheers,

Great! I would appreciate the feedback. I hope it works for you.

Ron

 

[gerty]

Very good Ron!! I have a few friends in the alarm business that'll kill for this circuit.. There is nothing worse than a alarm that looses it's credibility!!
This will make finding the offending detector simple to find..