Non -linear meter movement

[1audio]

Not really Volt-Nut but something some of us will encounter. Hewlett Packard 400L, GR 1394 Microvolter and my current challenge Quan-Tech 420 noise generator all have meter movements that are quasi-logarithmic. I could find nothing on these on the web. Both GR and HP versions have linear dB steps. The Quantech is calibrated in uV/rtHz but goes from 1-10 in a nonlinear scale.  I needed to get inside the Quantech meter (it had suffered some) and was looking for any info on how to set the static zero. I could find no reference. A little fiddling and I got it tracking well enough. In any case HP blocked access to a mechanical zero on the 400L and has no info on setting it, presuming its not needed. The GR does have a zero adjustment. Its manual doesn't mention meter zero.

While the cover was off the Quan-tech meter I took a photo of the magnet (or as much as I could see) and its quite different from the usuall D'Arsonval movement magnet structure. Also a photo of the GR meter showing the nonlinear scale.

Precision analog meters seem to have mosly passed into history today but were once all we had. Some were remarkably accurate (Sensitive Research had 0.1% meters with 12" scales and hand drawn scale markings https://pearl-hifi.com/06_Lit_Archive/15_Mfrs_Publications/Sensitive_Research/Sensitive_Research_Catalog_1980-1.pdf ). They seem quaint but were really pushing the limits of then current technology.

[Roehrenonkel]

Hi 1audio,
 
is it just the scale, or is the meter-movement nonlinear?
Maybe they are "moving-iron"-instruments?
See: https://en.wikipedia.org/wiki/Ammeter

First instrument seems to have a "zeroing-lever",
adjustabe maybe from the back..... (?)

Best

[blue_lateral]

That certainly looks like the throat in the magnet is nonlinear and expanding, in other words a nonlinear d'arsonaval movement. I have a Ballantine 320 True RMS AC voltmeter that uses a movement about like that. I don't recall offhand what the procedure is to zero it, but it definitely does not zero like a linear d'arsonaval movement. Correctly set, the needle is hard against the left stop when unpowered. It may be just be set to correct the lower end of the scale? There has to be a procedure somewhere. I hope you find what you are looking for.

[EC8010]

I see what you mean; that is different. BBC Peak Programme Meters (PPM) had linear scales of 4dB per step, achieved by a series of resistor/diode networks in an op-amp's feedback loop. Early (valve-driven) PPMs had a right hand zero in order to speed their ballistics, but later movements were sufficiently fast to meet the ballistics specifications with conventional left-hand zero. All PPMs were designed to flick quickly to the peak, then decay slowly so that the eye/brain can register it. Wonderful things and necessary to avoid overmodulating/damaging a transmitter. Operationally, I preferred a BBC PPM, but they were expensive and LED bargraphs were slim enough to fit one on each channel of a mixer.

[Poroit]

G'day 1audio,

My apprenticeship was with a company that built Moving Coil and Moving Iron Meters.

You have a Pivot and Jewel Moving Coil Meter and the magnet assembly has been cut away to produce a non linear scale as blue_lateral pointed out.

Check the meter is in balance....very important as I have seen techs chase their tale for hours trying to tweak electronics.

1. Position the meter with the Pivot axis Vertical and manually adjust the meter to zero.
2. Position the meter with the Pivot Axis Horizontal and check that the meter remains on zero with the pointer Horizontal and then Vertical.

Another thing to check is the Hairsprings.

Make sure they are concentric with no shorting turns.

Some manufacturers used to write the FSD of meters on the back of the scale.

[Kleinstein]

The first picture clearly shows a moving coil meter movement with the gap changing with the angle. So this is an intentional nonlinear meter. This can help to have a more even relative resolution.

There are also moving iron movements (for AC RMS without an external rectifier) with a square law. This would however be a different type on nonlinearity - more spread out at the end of the scale.

[1audio]

Thanks for the input. I'm sure it will help others as well.