Meter overload Q

I wouldn't let a delicate little thing like that ever see over 1 * FS. Maybe 1.05 or 1.1, if you're lucky.

Good Luck! Rich

Reply to
Rich Grise
Loading thread data ...

Take an analog meter - 1mA or 50uA as an example. What is the maximum "allowable" step overload that it would take without damaging the needle? 200 percent (2*FS)? 500 percent (5*fs)? 2,000 percent (20*fs)? Reason: I have a Triplett 420 5uA meter movement that i want to use to build a 200K/V VOM and want to protect the movement with some diodes.

Reply to
Robert Baer

Standard design practice is 2x continuous and 10x non-repetitive pulse for the coil. Needle damage depends on damping, which you can slow down with capacitor if needed, but 5uA doesn't have very much torque and mechanical damage from shock may be more likely. It is somewhat common to arrange for "off" position to short meter so that it is heavily damped for transport. May need low leakage diodes to avoid affecting linearity.

Reply to
Ol' Duffer

On Thu, 04 Aug 2005 18:15:16 GMT, Robert Baer wroth:

A step overload can slam the needle into it's mechanical stops and damage it. A simple overload much larger than the step overload, slowly applied, would not damage the needle. The coil will burn out before the needle is damaged with a simple overload.

Since a D'Arsonval meter is usually fairly slow anyway, I'd suggest using a capacitor actoss the meter movement its self along with a resistor in series between the meter and cap and the voltage source to slow the movement down so that it can't slam the stops on a step overload.

Jim

Reply to
jmeyer

AFAIK there is no known electronic protection scheme that could do that - much less yield reasonably accurate readings up to FS.

Reply to
Robert Baer

That is a good point, and i will incorporate a capacitor into the design, as you suggest. However, i still have the overload / overdrive protection concern.

Reply to
Robert Baer

Thanks for the 2X and 10X info! I did intend to do the short-when-off for protection. And you are very correct about diode leakage! I had thought that use of a schottky diode across the meter would give better overload / overdrive protection than a silicon diode. Well, yes and no. Problems: at 5uA, 4.5uA goes thru the meter, and 0.5uA thru the diode

- making the reading error at least 10% off. And at a large drive, the internal resistance makes for poor protection. Hell, even the reverse leakage screws up the accuracy of the meter reading too much!

So far, the design stands as having a 3.6K resistor in series with the 16.4K meter movement and 1N4007 diodes across the movement (protects both polarities), and 1N4007 diodes across the series 20K. Use of an NPN high beta transistor is intriguing: emitter to meter negative, collector to meter positive, and base to the 20K point:

3.6K 16.4K meter + o--+--/\\/\\/--+---/\\/\\/---+-o - | | | | | >------+ | \\ / | ----- | | +------------+

(done be hand) This is not quite as good as a diode at low drives, but (at the meter) actually decreases the meter drive above a certain input value. Adding a diode across the meter gives the best of both worlds. This is better than nothing, but there should be a better (passive) way...

Reply to
Robert Baer

On Fri, 05 Aug 2005 07:04:36 GMT, Robert Baer wroth:

Assume that the coil is wound with AWG 46 wire. That sets an upper limit for the current that will make it melt like a fuse.

Measure the resistance of the coil and assume that 250 mW is a power dissipation that will cook it. That sets an upper limit for longer current pulses.

Jim

Reply to
jmeyer

The 2N4117 jfet has an Idss mfr'ing spread of 30-90uA. I wonder if you could select a high end device and put it in series with the meter. The Siliconix data sheet sheet suggests that it would look like about a 16k resistor below Idss.

--
Tony Williams.
Reply to
Tony Williams

Use a comparator to turn on a shunt which steers current away from the movement when the current exceeds some limit - or use it to open the path to the movement. I use the shunt scheme on an expanded scale voltmeter I built to measure 10.5 to 12.5 volts. Full scale is 13 volts, but I don't care above 12.5. When voltage exceeds 12.5, the shunt turns on in proportion to the amount over 12.5, and an overvoltage led indicator lights. In your case, a proportional shunt is probably not a good idea. Might be better to shunt the meter with an SCR that stays on until reset by increasing the scale with the range switch.

Ed

Reply to
ehsjr

One can easily *vaporize* a #12 wire, so the reference to melting #46 wire would not seem to be a useable giuideline. The 250mW limit translates to 4mA of current - which is obviously way too much by any account.

Reply to
Robert Baer

One could put a resistor between the gate and source, and use the gate-drain as a 2-terminal current limiter; the resistor could be adjusted for (say) 9uA. An interesting approach that would solve the overdrive problem for the meter itself. Now i will have to sit down and see if that scheme still allows that composite to be useable in a VOM. Thanks for the idea!!

Reply to
Robert Baer

Hmm..The cathode of the SCR would connect to the meter negative, the anode to the meter positive, and the gate to the 20K point. That would give less meter protection (relative to an NPN) below the trigger voltage (a sensitive gate SCR would be best). However, the "reset" attribute is useful.

Reply to
Robert Baer

Sorry for the delayed reply - I was travelling. The idea is to set up the comparator to trigger the SCR so the meter needs *no* protection below the trigger voltage. The gate of the scr goes to the comparator output.

Ed

Reply to
ehsjr

If i add electronics that need a power supply, then i defeat the VOM "passive" concept. If i add electronics, i could go all the way and have zero volts drop for all current ranges, and a 10Kmeg fixed input resistance for most input voltages (from 25KV down to maybe 1V FS). That would give me the solid state version of a VTVM.

Now if you know of a (unpowered) way to "turn on" something in the

100mV to 250mV region, i would be extremely interested (the lower the voltage the better). Be advised that germanium and schottky devices have too much leakage to be practical.
Reply to
Robert Baer

In order to be a VOM, you must have at least one power supply (battery?) for ohms measurement. If this happens to be 9 Volts, it is enough to run a good variety of low-power op-amps, comparators, MOSFETs, etc. And there's some newer stuff that will run at 3 Volts. VOMs, FETVOMs and VTVMs that require battery power for electronics usually have a self-test or confidence range to ensure adequate function. Perhaps not quite what you had in mind, but food for thought.

Meter protection is of course not a new problem, and most VOM makers have settled on 50uA 250mV movements that can be sufficiently protected with garden variety silicon diodes.

Reply to
Ol' Duffer

Point taken concernig ohms ranges. I am not adding resistance ranges, so there will be no battery. I think i may have the best protection possible, but will try a different EPAD MOSFET that has a +200mV threshold (ALD110902PAL). Clearly, any JFET or MOSFET that has a low IDss also has high resistance before pinchoff, so cannot be used in series. And as a shunt unit, the "leakage" current at low voltages must be low enough at 85mV to not add a significant error. That +200mV MOSFET is the only part i have not tested, and the only one that has a remote possibility of being as useful or better than what i have.

Reply to
Robert Baer

Now that you've said it's not a VOM (no ohms range):

Just shunt the damn 5ua movement to make it a 500 ua, or something you are comfortable with, and design to your hearts content. The added shunt is always "turned on".

Ed

Reply to
ehsjr

ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.