Gyraf 1176 Meter

[spoontex]

Hi,

I'm building the Gyraf 1176 compressor:

https://www.gyraf.dk/gy_pd/1176/1176.htm

And I use a DC meter instead AC meter. So, for use a DC I need a bridge rectifier ( I use 4 germanium transistors as a diodes ).

Gyraf says:

The 1176 meter MUST be an AC-reading type - the one with an integral diode bridge. The SSL-Meter on the other hand, is a normal DC-reading type.

To use a DC-meter with the 1176, you'll need to mount it with a diode-bridge, four (preferably germanium) diodes rectifieng the signal to the meter. And then add a resistor in series, so that the sensitivity is around showing 0dB VU at +4dBM signal level.

But, how to know this resistor value? And how can I inject a +4dBM signal level? I only have a daw and a multimeter.

Thanks.

[Rob Strand]

To use a DC-meter with the 1176, you'll need to mount it with a diode-bridge, four (preferably germanium) diodes rectifieng the signal to the meter. And then add a resistor in series, so that the sensitivity is around showing 0dB VU at +4dBM signal level.

But, how to know this resistor value? And how can I inject a +4dBM signal level? I only have a daw and a multimeter.

You would be best calibrating the resistor value yourself as the diode drops are unknown.

Based on the AC scale where the meter connects to the X,Y terminals and reads 0dB on the VU meter with +4dB and +8dB levels, my best estimates are:
- no diode drop  Rseries = 665 ohm
- 92mV diode drop Rseries = 0 ohm

I also estimate a DC current of 200uA full-scale for the analog meter.  A 200uA full-scale analog meter is going to be about 600 ohm series resistance on it's own.  So on a real meter no external resistor is like Rseries=600 ohm; could be a lower on some models.  Which means you need no diode drops.   Clearly no diode drop seems unrealistic, in fact germanium diodes are going to have a voltage drop of about 200mV at the 0dBVU point.  So we can conclude the least error would be with no external resistor in series with the meter.

I know for a fact a lot of the old VU meters with the internal rectifiers used tiny selenium copper oxide rectifiers.   I can't remember the voltage drops on the those things.  If it's lower than germanium then it may support the no external resistor theory.   I seem to remember there were no resistors in series with the meter on those things.

If you want to calibrate using your DMM as a reference reading then I recommend no higher than 400Hz,  maybe 200Hz.   Fairly clean sine wave as the reference signal.
I back engineered the above from the circuit.  You can see it matches up with wikipedia,
https://en.wikipedia.org/wiki/VU_meter
However, not all VU meters are the same especially these days.
Another point is zero deflection is -20dB essentially accounting for the diode drop.  So if the scale and diodes don't match 0dB will be correct but -20dB will be off a bit.  (I think the GR meter driver has a way to fix that but then it won't track the AC mode!)

[PRR]

not all VU meters are the same especially these days.

The true-VU rectifier is Copper Oxide. Selenium was just starting to scratch at the time.

Both, and Germanium, are essentially vanished.

[Rob Strand]

The true-VU rectifier is Copper Oxide. Selenium was just starting to scratch at the time.

I should have fixed that.  I realized it later.

The rectifier is definitely full-wave.   (That's what I assumed in my estimates.)

I don't have a lot of accuracy expectations from VU meters.   If you print your own scale you can make it accurate but then it's only going to match your diodes!

There are whole lot of dynamic performance specs for VU meters as well.  If you don't emulate those then in practice one VU meter can show something different on real program material.   Of course at least one point should be calibrated and agree with a sine input.  I'm pretty sure the VU specs say what that frequency is - your extract mentions the 1kHz which is the most likely.   However, if you use  anything but a wide-bandwidth DMM as your reference then the accuracy can start to fall-off at 400Hz - that's why I suggested 200Hz to 400Hz.

[Chillums]

https://michaelfidler.com/articles/practical-vu-meter-circuits/

[ElectricDruid]

There are whole lot of dynamic performance specs for VU meters as well.

This is true. The attack/decay (rise/fall) times are all defined. The other problem I found is that there never was one definitive standard, so different organisations or companies did it different ways. I looked into a bit when I was working on my Audio Meter chip, and in the end I decided that since everyone else had already made up their own standard, I could do the same, and went for something that suited my application (which is what the muliple standards out there do).

<edit>However, the OP is trying to copy a specific device, so they presumably have a single specific set of parameters they need to match. This should make life much easier!

[PRR]

> never was one definitive standard

The true VU meter was standardized. There are part-percent numbers in the spec.

However it is really specified "by example": the movement that Simpson sold to the telephone and radio networks. It is high-performance (speed and damping) and was expensive. Most copy-cats (and even Simpson) despaired of making a right-on copy and resorted to cheaper movements. And radio or tape tech in the day knew a true-VU from a copy-cat from the first few bounces.

(Today: while still expensive, the meters on early TASCAM consoles are IMO a fair imitation of true VU. They were a hair slow on the uptake and a little fast to fall, but you rarely had un-indicated overload.)

HOWEVER! In the 1176 the meter's main function is to show Compression Ratio. Unless you are prepared for heavy redesign you can't use those clever "VU drivers", you need an honest floating naked low-current meter with reasonable speed (not full VU speed cuz R-C time-constants in the sidechain). The 1176 also let you monitor audio level as VU, but VU is 1939-era monitoring, today barely tolerable as "loudness" (Dorr is 1000% better) and about useless to avoid digital "over"s.

[Rob Strand]

https://michaelfidler.com/articles/practical-vu-meter-circuits/

"The meter should read 0VU when a current of 200μA flows through the coil."
Perhaps should read at full scale not at 0VU.

Which seems to be confirmed by wiki,
https://en.wikipedia.org/wiki/VU_meter
"The original VU meter is a passive electromechanical device, namely a 200 µA DC d'Arsonval movement ammeter fed from a full-wave copper-oxide rectifier mounted within the meter case."

There are whole lot of dynamic performance specs for VU meters as well.

This is true. The attack/decay (rise/fall) times are all defined. The other problem I found is that there never was one definitive standard, so different organisations or companies did it different ways. I looked into a bit when I was working on my Audio Meter chip, and in the end I decided that since everyone else had already made up their own standard, I could do the same, and went for something that suited my application (which is what the muliple standards out there do).

If you just buy a 200uA (or whatever meter) you can't guarantee it has the same dynamics as a VU meter.  So what you often had to do is add a fairly large cap across the meter (for some ckts a bit more - actually see Chillums's link for an example).  Many moons ago I did some spice models for physics of the meter movement so you can tweak the caps.  I'm fairly sure it was a damped second-order system.  The LED bar meters don't inherently have any dynamics so they add RC filters but many only used a first order RC filter.

Like a lot of this stuff I don't play with it often anymore.  I used to remember all the numbers  but now have to resort wiki and references .

It's always good to stick the to standards since it means different equipment will tend to produce numbers that agree but there's plenty of examples where it's only a guide reading and you can get away with something calibrated at one point with a steady-state sine wave.

HOWEVER! In the 1176 the meter's main function is to show Compression Ratio. Unless you are prepared for heavy redesign you can't use those clever "VU drivers", you need an honest floating naked low-current meter with reasonable speed (not full VU speed cuz R-C time-constants in the sidechain). The 1176 also let you monitor audio level as VU, but VU is 1939-era monitoring, today barely tolerable as "loudness" (Dorr is 1000% better) and about useless to avoid digital "over"s.

It's a good point.  What you *should* see is a completely different thing!   I don't think the short term vs long term perceived loudness issues are resolvable by any metering  they are different things and need different meters.    I've spent *many* hours on algorithms for leveling MP3 files.  Even the new ITU standards (BS1770 etc.) are not perfect - I think I did a bit better.    Even under the constraint of listening to music the statistics still show quite a wide variability.

[spoontex]

I think finally I will buy and original Sifam AL29...