Home Theater Forum and Systems banner
1 - 20 of 23 Posts

·
Registered
Joined
·
212 Posts
Discussion Starter · #1 ·
Geoff, the VU meter circuit on your website is exactly what I've been looking for.
I have some Modutec meters and wonder if/how I should modify your circuit for these meters. I haven't found any specs on these, but the DC resistance across the terminals is 1585 Ohms and the needle deflects to around -0.5 VU during the reading. The needle can act erratic during the multimeter reading. Does that mean there is an internal diode? Should I turn the level all the way down on a 1kHZ signal and gradually increase it until the meter reads 0 to determine the current rating? :scratch:
 

·
Registered
Joined
·
212 Posts
Discussion Starter · #2 ·
I neglected to add the web address before:
http://www.axtsystems.com/index.php...icle&id=69:vumeter&catid=42:vumeter&Itemid=74
And... there it says that the capacitor on the base of the transistor can adjust the ballistic 'a little'.
How simple would it be to add a switch from VU to Peak meter?

It's not so funny how difficult a simple circuit can be for me. I got a simple technical degree in electronics years ago. I need to review op-amp design and fancy transistor bias. I went into stage lighting after that schooling and didn't get a chance to ingrain those skills.
 

·
Registered
Joined
·
132 Posts
Glad to see you guys are using Vu meters! it seems like everything has gone peak reading with the advent od digital.

I got a couple of the hugh Simpson Vu meters that I'm gonna tack on the back end of this circuit.

Thanks for the schemes.

Pep
 

·
Registered
Joined
·
63 Posts
There are VU meters and there are VU meters, ideally you should never stick a meter directly across the signal you want to monitor since the internal diodes will add a small amount of distortion to the waveform. Hence the active circuit to drive the meter, which presents very little load to the signal being monitored. The ability to easliy calibrate the VU meter FSD is also important, since you need to calibrate it for 0VU to the levels you use during mixing or mastering.

It helps if you know the FSD current for the meter and the impedance. To determine this you need to use your bench power supply as a current source (ie use a large value series resistor) and measure the current at FSD. It is usually 50 or 100uA, so be careful or you will smoke your meter. Use a 100K resistor in series with your PSU as a starting point. Wind it up until the meter is at FSD and measure the current (you can use your voltmenter to measure the voltage across the 100K resistor and Ohm's law will give you the current). Now measure the voltage across the meter terminals. This voltage, divided by the current you just calculated will give you the impedance of the meter. It may be around 600 ohms, depending on the meter.

You also need to know whether the meter has an internal bridge, or if it is just a DC meter and requires on an external bridge to drive it. You can either test this with an AC signal, if the meter just sits there as you wind the signal up then it's a DC meter and requires an external bridge. Another way to test it is with your 100K series resistor and your bench PSU. Wind up the voltage until the meter is just deflecting up a little and reverse the terminals. If the meter goes backwards (ie reads a negative value) then it's a DC meter.

A simple bridge can be constructed using 4 schottky diodes (or germanium if you can get them) and attached to the terminals at the back of the meter.
 

·
Registered
Joined
·
63 Posts
How simple would it be to add a switch from VU to Peak meter?
I would not recommend this. Analogue meters make horrible peak meters, their mechanics are way too slow. Use it as a VU meter and get some LED peak meters, which are far more accurate for peak metering. I usually use the DAW metering for peak metering anyway, so no need for an outboard peak meter. The VU meter gives me a good indication of the mix characteristics, particularly at the low end where a mix can be broken very easily with too much gain on the low frequencies.
 

·
Registered
Joined
·
63 Posts
I've got a couple of nice LED peak meter DIY circuits and PCB layouts already done. I'll try and get them onto the site soon as a reference for interested folks.

I've got a few other DIY projects as well, some in prototyping, some complete that are not on my site either. These include monitor switching and calibration (analogue signals only), a mic preamp and DOA, VCA based compressor and various mods to bits of outboard gear. Will let the forum know as they appear.
 

·
Registered
Joined
·
132 Posts
Glad to see you guys getting in the science of things and pulling out the voltmeters. Yes, I can remember studying some AES papers regarding the proper "ballast" of Vu meters. There's a certain time constant (R/C network) that was the AES standard. This was all geared to voice recognition and again the "way back" standard of 600 ohm load.

Never use Vu meters as a FSD meter. You will be fooled depending on what freq spectrum your sound presents. Try playing the sound of a triangle and monitor with Vu and FSD peak meters. The peaks will be over and the Vu will be lucky to give -10db. If your trying to record this... good luck. I remember trying to convince a college music professor that his synth's level was fine even though he insisted on pushing the gain to get the Vu meters to respond using his triangle patch. It was a funny situation!:bigsmile:

PS I'd love to see your peak meter circuit, if you can get it together to post.

Keep it coming.

Pep
 

·
Registered
Joined
·
63 Posts
Never use Vu meters as a FSD meter.
FSD = 'Full Scale Deflection'
It applies equally to any meter, they all have a full scale deflection. :)

I think you mean not to use them as indicators of peak or digital overloads.
They are not fast enough for that, which is why the 0VU point has to be calibrated against a known Full Scale Digital (also mistakenly called FSD), for example 0VU=-14dBFSD.
 

·
Banned
Joined
·
66 Posts
The VU (actually a VI, Volume Indicator) meter was developed as a joint effort between Bell Labs and two broadcast companies in 1939. It is tightly standardized so that one meter will exactly agree with another reading the same program material. The calibration was for an exact power level of 1mw into a 600 ohm line, plus a fixed attenuator setting. The ballistics are also tightly standardized for 300ms rise and fall time, and maximum allowable overshoot is also specified. The ballistics were chosen so that speech could be consistently metered.

The things to remember when using a VU meter in this century are:

1. We have abandoned 600 ohm transmission lines in studios, so power calibration makes no sense. We’ve adapted to a voltage calibration. This means we have to be aware of what “0 VU” means. It no longer means 1mw into 600 ohms.

2. The meter is NOT a loudness meter, a power meter (anymore), an SPL meter, or a peak meter, though it is possible to calibrate so that “0” corresponds to some reference of any of these, excepting loudness. However, doing so creates a new, non-standard application, and takes you away from the next point.

3. The only point of using a VU/VI meter is consistent level interchange between other facilities or users with standard VU meters. It means that in reality the only thing a VU meter shows you is how that program material will look on another standard VU meter. Any alteration to scale or especially ballistics makes this meaningless.

If you’re going to use a VU meter, it should be just that, a VU meter, and correspond to the standard in every way possible. If you’re going to re-calibrate “0”, that calibration should be carefully noted. But ballistic changes, especially if arbitrary, make the meter much less useful, and one of a kind.

Loudness is a complex psychoacoustic quantity, and a simple averaging voltmeter can’t do it. A peak meter is even worse as a loudness meter. But, there’s a really good (and free!) solution. And you won’t have to build anything.

http://www.orban.com/meter/

Download the free public Beta. You’ll get several meters, one of which has VU ballistics, though not the standard meter face and scale. Another has PPM ballistics, and there are two standardized, for-real, loudness meters in there too. Can’t beat the price and performance. You’ll have a calibration task, but it can certainly be done. Combine this with K meters, and you’re basically loaded for bear, and ready to meter for the 21st century.

Final comment, if you’re dead set on using a pair of old VU meters, drive them with a buffer amp followed by a 3900 ohm resistor to maintain ballistics. Don’t externally rectify, peak detect, or mess with ballistics. Just feed the meter through the resistor from an opamp buffer. You can change the amp gain to adjust calibration, but don’t change the meter resistor.

Jim
 

·
Banned
Joined
·
66 Posts
The jlmaudio.com page links to Rob Squire's DIY VU article in Audio Technology (the have it in a .pdf). Rob makes a good point that while real VU meters (your Simpsons, and Modutecs) include internal full-wave rectifiers, some cheapie "VU Meters" are DC meters, so his driver circuit provides rectification using silicon diodes so those cheap meters can work with an AC audio signal.

Well... there are two problems. First, there's a fixed voltage drop across a silicon diode of about .7 volts, so perfect rectification at audio levels is sketchy, especially with the 3K9 resistor there. It would be better to use a precision rectifier circuit, which compensates for the diode drop, like this:
http://www.discovercircuits.com/DJ-Circuits/fullrect3.htm
Follow the precision rectifier with the 3K9 built-out resistor, of course.

But that's really minor. The bigger issue is, if it says "VU", but it's a DC movement, it's really unlikely the manufacture followed any of the VU specs at all, ballistics included. So what you have is an unknown DC meter with a VU scale on it.

To test if you have a real VU meter, here's two signals you can use. This one is for zero cal. Play it, and set it so your meter reads "0".

http://www.movingbitsproductions.com/download/0VUref.wav

This next one is a ballistics test. It's a 300ms tone burst with 300ms of silence in between. That will deflect a real VU meter to 95-99% of the reference reading, with no more than 1% overshoot or undershoot. It should also fall back to 5% of scale during the silence sections.

http://www.movingbitsproductions.com/download/VUtest.wav

(everyone now runs off to see if his VU meters follow spec or not...) :T

Jim
 

·
Registered
Joined
·
41 Posts
Yes but unless your using my avatar a full wave rectifier is also going to drop .7v if its silicon. Im not familiar with a precision rectifier. And i agree ballistics are another matter entirely.
 

·
Registered
Joined
·
63 Posts
Yes but unless your using my avatar a full wave rectifier is also going to drop .7v if its silicon. Im not familiar with a precision rectifier. And i agree ballistics are another matter entirely.
A silicon full-wave rectifier will drop 1.4V typically. That's why germanium and schottky rectifiers are used, which drop about half this.

The precision rectifier uses and active element (usually an op amp) to overcome this problem.
 

·
Registered
Joined
·
132 Posts
Yea... I was talkin' digital FSD! It's nice you know about "full scale deflection". Sounds like your an ol timer too. Ever hear of the "Rm equation"?? My electronics professor in college, Bob Koval, had the patent/copyright on this... worked for RCA for 20 years. One of the most brilliant men I know.

Course we had a million tests on this... which I forget all of it!

Back in the 80's I built my first "State of the Art" mixer and wanted to try the faster LED peak meters using the "new" National Semiconductor LED driver chips.
The most coveted book was the "Audio Handbook" and I believe I used a precision rectifier from that book. Based on a few opamps and ... I think Germanium signal diodes.....It's been so long. it worked really well with variable decay of the LED stream. Still got the board stuck away.

I agree w/u Jaddie, change the ballistics and your making your own meter. Thanks for those test files and downloads. I'll check them out.

Pep
 

·
Registered
Joined
·
63 Posts
Not sure what an 'old-timer' is. Conjures up images old blokes with bushy beards! I'm still under 50 :)

Can't say that I've heard of the Rm equation before, maybe it went by another name. Then again my electronics professor was Richard Small, so we got to learn a thing or two about loudspeaker modelling!
 

·
Registered
Joined
·
132 Posts
I'd prefer to learn about the modeling also. HA HA

I remember taking a test for a full time job and they had the "Rm" (resistance of the meter movement..relates to the accuracy of the meter) problem for you to solve. I was like.... if you guys use analog meters, I don't want to work here.:neener:

Some guys still love the analog meter guitar tuners... so I guess meters are still in.
They look pretty on those old desks too.

P.S. I'm not that much ahead in age.... but I have been thinking about a beard.

Pep
 

·
Banned
Joined
·
66 Posts
Yes, a silicon full wave bridge rectifier would have a 1.4 volt drop, making it pretty much unusable for a directly connected VU meter. The original meter specs call for a copper oxide rectifier and takes not only its voltage drop into account but its series resistance as well. Germanium diodes would be ok, but they aren’t as common as silicon diodes.

But, neither Rob Squire’s circuit nor the atxsystems.com circuit use a full wave bridge, just a full wave rectifier, so only one diode is conducting at a time, so just the single diode drop of .7 (approximately!) volts. Both would benefit from using the active precision rectifier circuit, as the diode drops are servo-ed out, and you get real rectification down to several millivolts. Of course, you could just use a real VU meter and forget the external rectifiers...

You actually can make a mechanical meter display peaks. That’s what a PPM is. You start with a precision rectifier, follow with a log amp (a real PPM is a log display) add a quasi-peak detector (a PPM isn’t true-peak) with sample/hold and you got it. You can make it a true-peak meter rather than a PPM by making the peak detector fast enough to grab 1/2 cycle of a 20KHz wave. If you don't need a log scale, you can forget the log amp and go linear, with of course, the limitations in range. You can make the sample/hold freeze that value long enough for a mechanical meter to swing up and display it. In fact, you can electronically accelerate the needle so that it even moves fast enough to look like a peak meter. You can do this by putting a network corresponding to the mechanical ballistics in the feedback loop of a meter driver, and really speed the pointer up a lot. It’s even possible to over-damp the meter and remove a lot of overshoot. But it’s a lot to do when a display made up of lights doesn’t need any of that. However, a mechanical meter is often preferred by operators for watching fast changing signal levels, so it might be worth the effort. I’ve done it, it makes for a really impressive, fast, if somewhat nervous looking meter. If you slow the fall-back time, it's easy to read, and actually one of my favorite level displays.

Yes, I probably qualify as an old-timer, whatever that is. No beard, it would be gray if it were there. But I hate to see anyone head down the wrong road, or the right road for the wrong reason. A VU meter is only good if you want to see what your stuff looks like on another VU meter, and that’s only true if you stick with a meter that matches the specs. Otherwise, there are better things for level measurement, loudness measurement, recording level set, etc. Lots better.

‘Course, there isn’t much that’s cooler looking than a 4” Simpson VU meter lighted by under-voltage #47 lamps in a darkened room...
 

·
Registered
Joined
·
212 Posts
Discussion Starter · #18 ·
A meter is easy to read and not just another bright light in the eyes. I mean if you want to keep track of general levels. It shows a continuous metering of what the signal is doing, instead of a sequence of LED dots. If you are running through a lot of analog equipment, or even recording to magnetic tape it is a good indicator. Peak meters are good for digital, ie., dB peak to full scale dBFS.

I don't know what I'm saying. I got 6 Modutec meters for $60 and want to build them out for the front middle and end of my signal chain. I intend to use the K system of calibration, since it seems to be an objective measure.

And I talk a lot, since I'm putting together my first professional studio. I just remember how well the meters served me doing those recordings to tape. And I have a Revox and Sony [tube] decks now, to test out.

Maybe people forgot or never knew what real analog sounds like; the phase distortion of FM radio when driving in and out of reception... It's all a part of the human sound, technology included. The real grit and grime of physics, not simulated.

My 0.002 cent
 

·
Registered
Joined
·
132 Posts
You actually can make a mechanical meter display peaks. That’s what a PPM is. You start with a precision rectifier, follow with a log amp (a real PPM is a log display) add a quasi-peak detector (a PPM isn’t true-peak) with sample/hold and you got it. You can make it a true-peak meter rather than a PPM by making the peak detector fast enough to grab 1/2 cycle of a 20KHz wave. If you don't need a log scale, you can forget the log amp and go linear, with of course, the limitations in range. You can make the sample/hold freeze that value long enough for a mechanical meter to swing up and display it. In fact, you can electronically accelerate the needle so that it even moves fast enough to look like a peak meter. You can do this by putting a network corresponding to the mechanical ballistics in the feedback loop of a meter driver, and really speed the pointer up a lot. It’s even possible to over-damp the meter and remove a lot of overshoot. But it’s a lot to do when a display made up of lights doesn’t need any of that. However, a mechanical meter is often preferred by operators for watching fast changing signal levels, so it might be worth the effort. I’ve done it, it makes for a really impressive, fast, if somewhat nervous looking meter. If you slow the fall-back time, it's easy to read, and actually one of my favorite level displays.
Yea, my precision full wave detector was modeled after the circuit diagrams that were found in the LM3914 LED driver data sheet. Didn't look the circuit above though. Two 914's or 4148 diodes w/dividing resistors all in feedback loop.

You guys have really lit a match to me getting these Simpsons wired for your above description. It may help with my K-system problem over in the Mastering forum. I think it would be a nice addition.. and yea.. they do look sweet!

Pep
 
1 - 20 of 23 Posts
This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top