VU meter on an amp

Status
Not open for further replies.
Jump to Latest
L

laki101

Member
Joined 2009
#1
so I am building my first amp and cause I love disco lights I want to install an vu meter into my amp but cause I'm sure how to make it work in the amp and not to interfier with the signal cause all of them do draw a lot of the signall

oh yes and I forgot I making the vu meter with leds so the ones that turn leds on and off depending on the music
 
#2
The usual place to tap off audio is directly at the speaker outputs.

LED VU meters do not draw any appreciable current from the audio source.
They may however draw significant current from whatever power supply they require. But some of the modern ultra-bright LEDS give very good brightness from currents as low as 5mA.

If you want to build it from scratch, search for the datasheet and appliation notes for National's "LM3915 bargraph driver" IC.
 
#5
A clipping indicator is a good idea too. Elektor published a thorough one a while ago - it used a comparator and a monostable so the slightest clipping would be visible. It might be included in their "Elektor Audio Collection" CDROM, if you can find that someplace.

It's worth looking through the service manuals for various power amps to see how they do it. One (old) amp I own has a string of comparators rather than a bar graph chip. That allowed them to choose 5 dB steps, allowing 10 LEDs to cover a very wide range. On the other hand, a lot of pro amps just have a "activity" and "clipping" indicators which is all you really need in a working amp.

DJ mixers often have attractive level displays, and some good brands like RANE have schematics online.

There are relatively cheap wide-range level displays from brands like "American DJ". Velleman has several audio level kits, K4304 to K4307. The K4307 is powered directly from the audio output, so it's easy to add to an amp or speaker. The manual even has the schematic and board overlay, but doesn't show the PCB layout. The kit seems overpriced, compared with the American DJ "DB Display" which gets you two channels and more LEDs. There's one on eBay now with buy-it-now price less than the K4307, if you move fast.
Velleman Inc.
There has to be some cheap displays on eBay if you can find them.
Radio Shack carried a few different models of add-on power meter, some analog, some LED. The APM-300 looks suitable, if you find one cheap enough.

It may be feasible to salvage the display from some gaudy car stereo or home stereo component. Or check the catalog of a chip maker that makes stuff for consumer applications.

If you want a really impressive, though not necessarily useful, display, multiplex a bunch of LEDs so each column is driven by the bar graph chip in sequence. Persistence of vision makes it look like an oscilloscope display.

And, finally, there's the retro option: a magic eye tube display.
 
Last edited:
#6
so I just connect it in parallel with the speaker output on the amp and it will work

and also how do I overcome the problem that if I make it run at low power when I turn the volume up it will be always on top and if I make it run on max volume it wont work on low
 
#7
Basically yes, you connect across the output but with an attenuator network so that max output of the amp drives the meter to 0DB

The second problem needs the meter to have a logarithmic scale, not a linear one.
 
#8
Hi,
make the voltage steps for the LEDs logarithmic, i.e. based on -dB ref max output voltage.Look at the three LED drivers, VU3916, linear3914, logarithmic3915. They can all be cascaded to increase the range that can be displayed.
 
#11
Good audio-activated light displays include an automatic gain control or AGC to compensate for varying levels. There was a "color organ" kit from SWTPC that had that. I couldn't find much evidence of it on-line, but someplace I've got a photocopy of the manual with schematic and layout. The SWTPC kit would drive several incandescent bulbs (or strings of small lights), with each responding to a different frequency range of music. Google "color organ" and you may find a kit or plans that you like. It's not a very meaningful display, but it's lively.

Another mode is to just respond to the beat. I came across a design for the ultimate beat follower, that had been used in some disco lighting effects. It also used an AGC to make it respond to the beat at varying volume levels. I've seen a simpler beat detector in a mixer that just used a comparator. The reference input to the comparator was derived from the average signal level, so it should respond to peaks (or beats) over a wide volume range. Maybe a beat detector is even less practical on a power amp, but if you want a lively display, use the beat detector to trigger a digital counter and drive a display that changes every beat. A 4017 could drive a sequence of LEDs. A binary counter would give a more random looking display, depending on how you wire the binary outputs to LEDs. Or, like a project in ELV a while back, feed the binary count to an EPROM that's programmed with a bunch of different output sequences.
 
#14
katana2112 said:
How would I wire an analog VU meter? What is the difference between a VU meter and a power(watt) meter?
Click to expand...

Wire it to what ?
It depends on the sensitivity of the meter, what you are driving it from etc. You need to rectify the audio signal to feed a meter, probably using an active rectifier is best to eliminate the inevitable diode volts drops (germanium is best by the way for a "passive" rectifier)

The terms are very specific... VU etc
VU meter - Wikipedia, the free encyclopedia

A power meter reads power... except most don't, they are just calibrated voltmeters to show what the power would be into a fixed load.

For example a 100 watts into 8 ohm is approx 28 volts RMS. So a power meter on an amp is really nothing more than a voltmeter, calibrated in this example so that 28 volts reads as 100 watts. The problem is if you disconnect the load the power meter still reads "100" because there is still 28 volts present... but no power is dissipated (no load). A true power meter is much much more complex and gives a true figure based on voltage and current actually supplied to the load... and it becomes even more complex when you want it to show "true" power caused by a reactive load... ie a capacitor.
 
#16
Mooly said:
The problem is if you disconnect the load the power meter still reads "100" because there is still 28 volts present... but no power is dissipated (no load).
Click to expand...

One way around this would be for the system to operate from the voltage dropped across the RE resistors. Apply an opamp to add gain and then add another opamp after this to invert the signal.

Take the inverted and non inverted signals and pass them through a diode to remove the -ve component of each signal.

Combine the outputs of the two diodes and use this as your signal to drive your indicating system.

Depending on the biasing in the amp you might need to cap couple the signal from the RE resistors.
 
#17
5th element said:
One way around this would be for the system to operate from the voltage dropped across the RE resistors. Apply an opamp to add gain and then add another opamp after this to invert the signal.

Take the inverted and non inverted signals and pass them through a diode to remove the -ve component of each signal.

Combine the outputs of the two diodes and use this as your signal to drive your indicating system.

Depending on the biasing in the amp you might need to cap couple the signal from the RE resistors.
Click to expand...

That's a good idea and would work well for a resistive load (one that varies as well), as the voltage across RE is proportional to the power dissipated in the load. So if you changed from an 8 ohm to a 4 ohm load the meter would register the corresponding increase... and tbh that's probably good enough without making it to complex.

The "true power" problem I mentioned is showing power in a reactive load. A capacitor across the output is the extreme example. The current and voltage are 90 degrees out of phase, so the net power produced in the load (the cap) is zero, despite the high current flowing and the high voltage across the load.

It is do able... perhaps deriving RMS values for the current (RE) and the voltage across the load and using a four quadrant multiplier to obtain an indication of "apparent power" in the load... the power meter would end up more complex than the amp 😉
 
#18
If anyone still cares about the SWTPC color organ I mentioned, I found the manual (with barely legible schematic, and parts overlay which shows the pads and traces). The model is QX-III (or QXIII or QX III or QX3 depending on what page you look at). There's no separate image of the PCB artwork. Someone could conceivably drill a bare board using the overlay, then freehand the traces and pads with a pen or something. Or use the overlay as the background for a PCB CAD program that has that feature, then trace it.

I've scanned the pages, but not touched them up and PDF'd them yet. If I can get a scan or backlit picture of the mostly-stuffed board that I have (not an original SWTPC board - it looks like someone copied the layout using tape and donuts), I'll include that. And maybe data sheets for the more obscure semiconductors if I have them.
 
Status
Not open for further replies.
Top Bottom