LED bargraph ballistics

GroupDIY Audio Forum

Help Support GroupDIY Audio Forum:

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

mobyd

Well-known member
Joined
Jan 27, 2006
Messages
196
Location
Auckland NZ
Have been playing around with a monster VU meter for a public display situation and would like to get a reverse exponential decay if that is possible. A normal RC network on the rectifier output gives an exponential decay, replacing the R with a constant current sink gives a linear decay. What I would like is a slow initial decay followed by a rapid drop off - think of the strongman/hammer games - so that the contestants have an opportunity to see how well they scored. Is there an analog way of doing this ? - I know it can be done in the software, but would rather not at this stage.
Current setup :
Piezo transducer->preamp->rectifier->A/D [PIC] bitstream->UCN5832->192 LEDs
 
Doing it in the software would be most flexible but there are several possibilities using analog circuit tricks.

A capacitor could be discharged by a rate that is proportional to how far the meter is below some peak level. This would decay slowly from high levels and speed up exponentially as it gets further away from the peak level.

A quasi-linear variant on this is discharging with a current source which is the combination of a fast fixed discharge current, minus a current component proportional to the meter's output level above fully discharged. This too would decay slowly at first and speed up as it discharges.

Both approaches involve current mirrors-current sources so might take one opamp, a transistor or two and a few resistors.

Perhaps what you really want is a peak hold that wouldn't decay at all for X duration, then decays more quickly.

Have fun...

JR
 
I use couple of capacitors: small one connected directly and a bigger one connected with resistor in series, so I have 2 time constants, for peaks and averages.

Though it is not what you are asking about, but it satisfies me in the real life.
 
[quote author="mobyd"]A normal RC network on the rectifier output gives an exponential decay, replacing the R with a constant current sink gives a linear decay.[/quote]
So use one capacitor charged by a current source to set a current source which (dis)charges another capacitor. The first cap/source pair sets the velocity, the second cap/source pair determines the position. Have a comparator short out the velocity capacitor when the voltage on the peak position capacitor is tracking the 'actual' signal. This should implement a simple quadratic decay.

Having said that, a software implementation would be much more tweakable.

JDB.
 

Latest posts

Back
Top