VU meter driver assembly

[peter purpose]

[quote author="JPrisus"]just to clarify, VR4 is the trimpot, and U1 could be half of a TL072?[/quote]

VR4 is indeed a trimpot and U1 on mine was a 5534 single op amp. Overkill, but that's what I had to hand. There is additional circuitry to glue on (compensation caps and such). Just look at the specs of the op amp you're going to use.

As PRR mentioned, op amps are so cheap that running two signals into one dual op amp is a false economy and requires additional gizmos.

As NYD says, this circuit is really only good for a gain of more than one.
I was under the impression that you were running unbalanced, as was I.

ayethengyaw
pooter

 

[NewYorkDave]

Seth: Assuming infinite open-loop gain (which doesn't happen in reality, but the assumption is usually made with IC opamps since their o.l.g. is very high), the gain of the stage is expressed:

Rf/Ri for inverters, and
(Rf/Ri)+1 for noninverting stages

where Rf is the feedback resistor--the resistor from the opamp's output to the inverting input--and Ri is the input resistor on the inverting input. On Peter's schematic, R3 is Rf, and VR4 is Ri.

As you can see from the second formula, the gain of the noninverting stage cannot go below 1, even with 100% feedback.

 

[JPrisus]

[quote author="peter purpose"]


As NYD says, this circuit is really only good for a gain of more than one.
I was under the impression that you were running unbalanced, as was I.

ayethengyaw
pooter[/quote]

Actually Peter, i am running an unbalanced monitoring chain, which this will be fed from. I'm monitoring several balanced +4 units, but taking them into my monitor single-ended for a 6dB loss.

 

[jrmintz]

As you can see from the second formula, the gain of the noninverting stage cannot go below 1, even with 100% feedback.

Thanks Dave.

 

[nwsoundman]

Hello,

Does anyone have a schematic for a VU meter buffer that run balanced or unbalanced via a switch. I have some Dixson meters with a buffer PCBA already on them. I just can't seem to figure out the pinout on em.

I want to be able to switch the meter to look at pre or post EQ in some strips I am making. I will be able to bypass the pre or bypass the EQ so I need to be able to meter off either as needed. Also the pre is balancd out and the EQ is unbalanced out for now. They are some old QE parts.

Thanks for any help on this. My brain hurts.

NWSM

 

[NewYorkDave]

I have an idea for a fairly "universal" VU buffer which I will draw up and post as soon as I get a chance, possibly tonight.

 

[PRR]

> buffer that run balanced or unbalanced via a switch.

A proper floating input will work balanced or unbalanced.

Several 1-amp designs work, sort-of.

2-amp and 3-amp designs work very well.

A transformer works, but costs money and loads the source more than a couple 19 cent chips.

 

[nwsoundman]

Thanks!!

I've been stuck in the digital world for the last bunch of years and I have gotten weak in some basic theory. That is very frustrating to me.

I look forward to finishing this box so I get get to finishing the other half built stuff.

NWSM

 

[PRR]

This is a simple but generally adequate VU meter buffer, for when you really have "line levels":

Here it is with optional gain, for use at lower levels:

You need two op-amps to get proper impedance for both differential and common-mode voltages. At today's prices for utility opamps, there is no reason to use the one-amp solution.

The input impedance is 200K differential (driving both inputs), 100K single-ended (if you have an unbalanced source, tie the extra input to ground). That is usually negligible.

It relies on the fact that a VU meter is really floating, neither end has to be grounded. As-shown, it assumes the meter has a rectifier; if not, put a big silicon bridge in front of it or use four Ge or Schotkey diodes.

TL072/TL074 is eminently suitable.

The power supply really does have to be up above +/-9V. +4dBm is only 1.23 volts, 1.8V peak, but a VU has to see those transient peaks 10dB-15dB higher to respond as expected on speech/music. If you used +/-5V supplies, it would read fine on test tones, but read low on loud speech/music, which is dangerous.

 

[NewYorkDave]

I didn't have time to draw out a complete circuit tonight, but I did sketch my idea in outline form. PRR already posted a good working buffer circuit, but since I already took the trouble to scan this...

The reason for fixed gain in the amplifier and an attenuator on the input is to reduce the chance of overload and signal rectification at high levels. It's assumed that the operator will switch the attenuator to the next lossier position if the needle is slammin' :wink:

The instrumentation amp could be either a purpose-built IC, or a standard I.A. circuit assembled around a quad opamp. The input can be run either balanced or unbalanced.

The idea of the I.A. and the odd-looking input circuit is to give a high common-mode impedance so it looks nearly "floating" to the signal line across which it is bridged.

 

[nwsoundman]

Thanks again Dave and PRR.

I have been looking at my buffer amps and am so confused because Vcc+ and Vcc- are tied to what seem to be rectifying diodes. The diodes are two different types, a pair made up of one of each type, going to Vcc+ and Vcc-. Pins 4 and 8 on a TL072CP. I am unsure if there is a rectifier in the meter itself? If I ohm out the meter I only see a minor change in resistance when switching my test leads around. The difference is only like 200 ohms and the meter ohms out at about 39K. When my DMM is set to diode test the VU meter pegs out when the DMM is applied in either polarity. The DMM shows like 1.2 volts.

I think I will build PRR's "line level" circuit and add a rectifier to it. Would you install the recifier between R7 and the meter? Can I make R7 a trim pot to be able to calibrate the meter?

With these circuits the meter will behave more like a PPM meter instead of VU. Is this tue?

Thanks for helping me out with this.

NWSM


UPDATE - I took apart my VU meter. It is a Dixson model 220. It has a rectifier and 1/2 watt resister inside. It does however have it's polarity marked where you connect the meter to the buffer amp. There is a + sign for one of the two terminals. How does that effect things?

 

[nwsoundman]

Ok. I have it all figured out except how I can trim the meter to calibrate it.

Thanks for the help.

NWSM

 

[PRR]

> the meter will behave more like a PPM meter instead of VU. Is this tue?

No, it is still a VU meter. But it isn't a medium-Z non-linear load on the line being metered.

> It does however have it's polarity marked where you connect the meter to the buffer amp. There is a + sign for one of the two terminals. How does that effect things?

It is probably a universal case for all meters, DC or AC. If you know it is an AC meter, ignore the mark. If you don't know, put some audio in it. A DC meter will vibrate with bass but never swing much away from the rest-stop. An AC meter will respond proportional to the audio.

In my buffer with gain, the variable resistor is about no-connection for unity gain, 20K for 6dB gain, 7K for gain of 12dB, 2.2K for gain of 20dB.

 

[nwsoundman]

Thank you PRR. The metering was holding me up a bit on this box.

NWSM

 

[electronaut]

Hello,

I built PRR's simple buffer amp and it worked well enough for me. Now I'm trying to find a way to avoid the +/- 15v power supply requirement for the op-amps, and instead run it on something like +30v only.

Here's what I came up with. I haven't built it yet, so I don't know how well it works, but it seems logical.

Any comments, suggestions, etc. are greatly appreciated.

-E.

 

[NewYorkDave]

That will work, but I would cap-couple the output. You only need one coupling cap; you can put it on either side of the VU meter. The main thing is that you don't want any direct-current offset to flow through the meter.

The (nominal) impedance of the VU meter with its series resistor is about 7500 ohms. With a 4.7uF coupling cap, the response will be down 3dB at 4.5Hz.

You could increase the value of the V/2 biasing resistors on the noninverting inputs in order to provide a higher-impedance input. The TL072 has FET inputs so there's no base current to worry about. You could use 1M resistors if you have 'em handy. After all, the whole purpose of a buffer is to be as "invisible" to the circuits being measured as possible.

The standard series resistor for a VU meter is 3.6K, not 3.3k. But if you're getting proper indication with a 3.3K, that's fine. Most of the better VU meter setups provide some kind of trim, anyway. You could add a 1K trimpot in series for fine-tuning.

 

[Mbira]

Hey Dave,
I just read thru this thread. I'm confused why this is the case?

(Rf/Ri)+1 for noninverting stages

I thought that if a feedback loop didn't have a resistor, then the negative side would equal the positive side on the input and therefore create no gain? What am I missing?

Joel

 

[NewYorkDave]

What you say is correct, there is no gain.

ASSUMING that the input impedance at the negative input terminal is very high (and we'll make this assumption if it's a FET-input opamp like a TL072), then it's true that if there is no Ri, there is effectively 100% negative feedback and no gain. Let's look at our noninverting gain formula again:

G= (Rf/Ri)+1

So in the case of no Ri...

G= (Rf/infinity)+1

= 1

 

[PRR]

> you don't want any direct-current offset to flow through the meter.

You are thinking that the 5%-10% tolerance resistors in the voltage divider could give 3V different DC voltages, and slam the meter with DC.

So since the TL072 has small offset voltage and negligible bias current, use one divider to make around 15V, and two 1Meg input resistors? Same number of parts, and both sides get the same DC voltage (within a couple mV). Input caps could be 0.02uFd, and no output cap.

With a very noisy supply and an unbalanced source, you might want a bigish cap on the voltage divider.

 

[Mbira]

G= (Rf/infinity)+1

= 1

So what is this +1?

Are you just saying it is 1 in and 1 out?