> Is a PPM meter essentially like a VU but doesn't have a buffer (wrong word?) so it reacts quicker to transients, etc?
No. Aren't they teaching you anything in that EE lab?
Put one of these old meters on an audio signal generator. Start with a very low frequency, like 10Hz, and sneak-up the level until the needle moves. This should happen at a very low voltage, less than 1 volt.
Because the input is alternating current, the needle will swing both ways. Do NOT try to get a large deflection, just enough to see. You can read "level" by eye-balling the edge of the blur of the wavering needle. That is awkward. Also as you increase the frequency, the blur-width for constant input level soon decreases, and will be very small above 100Hz. Over-powering the heavy needle is not the answer: the forces needed to cause a large swing will actually throw the moving parts out of their pivot, ruining the meter.
So it is good for DC, poor for 10Hz AC, and useless for audio.
What if we turn the AC into a varying DC, so the needle does not try to swing both ways? Get a bridge rectifier. Now both + and - swings of signal flow as + through the meter, so ut just swings "up". Now you probably have a pretty useful AC meter from around 20Hz up to 5KHz.
One problem: the input is very low voltage and fairly high current. Probably less than 1 volt (maybe 0.1V) DC for full meter deflection, at 0.030 Amps, or around 33Ω. Maybe useful for headphone jacks, or low-level speaker amps. Way too heavy a load for an line level.
Also the 1V or 0.1V voltage sensitivity isn't accurate: the meter was calibrated for current, not voltage. The falling voltage response at the top of the audio band is due to coil inductance causing rising impedance, or less current for a given voltage. Also the meter voltage is low compared to the 0.6V dead-zone of Silicon rectifiers (1.2V for a Bridge).
Put a resistor in series with the movement. Ideally we want resistor voltage drop to be much-much greater than meter movement or rectifier voltage drop, so the resistor dominates the voltage sensitivity. If the diode/meter drop is around 1V and we want the total to be 100 times higher, we want about 100V sensitivity. This may be awkward, and we don't need 1% accuracy. Try 10X, or 10V sensitivity. 10V/30mA= 333Ω in series with the meter and rectifier. For signals from around 1V to 10V the meter input impedance is around 330-340Ω. Low, but we could dedicate an amp to it. The 10V full scale sensitivity is awkwardly high, though we could ask that amp to boost it.
You see why, in days when amps were expensive, they did not use 30mA movements on audio meters. The VU meter is only a 0.2mA movement and still causes significant loading on 600Ω lines. You might use a 30mA on speakers, but you won't see much below 1V signal because common Silicon rectifiers drop a large part of a Volt. You could find Germanium or Copper-Oxide rectifiers with lower drop, and indeed that's what VU meters use. But they are typically rated 100mA and melt easy. They would be OK with 30mA of steady tone, but audio peaks are over 10X higher, so there is a large risk of burn-out from 300mA peaks in 100mA diodes.
About those audio peaks: you already know from AC tests that this needle won't move much faster than 100 milliSeconds (10Hz). In audio metering we are often interested in large transients that are 50mS to 1mS long. SO this meter will read the longer-term average, averaging-in the transients, which will hardly change the reading on the heavy needle.
The VU meter compromises. It uses a higher force/mass movement than the average utility meter, so it will respond to sub-50mS peaks with an extra high twitch. It smooshes 1mS transients, so when using a VU meter to avoid peak clipping you must keep the meter peaks far (10dB) below the system overload. The VU meter is a passable "loudness" meter, and gives a guide for setting average level so most peaks won't clip. It never actually shows peaks.
No mechanical meter can show audio peaks. Just too heavy.
But tubes and transistors can respond faster than audio peaks. And capacitors can hold the levels of those peaks long enough for a heavy needle to indicate.
So most good audio meters are a full-wave peak detector, a leaky sample-hold, and buffering so the meter movement does not load the hold-cap.
It is possible to pick the peak detector and leaky-hold rates to simulate a VU meter on a lesser movement. I think that is dumb: the VU is a terrible meter, it was just the best they could do in 1937. Today I consider VUs more decorative than useful. More modern thinking meter specs use a shorter peak-catch time, often a longer hold. Also audio swings over 1,000:1 while it is hard to read more than 20:1 on a needle-meter. Most recent good-meters use a log-amp to compress a wide range of signal into the usable part of the scale.