How to hack this? Ideas?

[Moby]

I have a moving coil ammeter with FS sensitivity of 20ma. Max current going through is around 10ma and I need some idea how to "push" scale to go max at this value. Is it possible to do some active circuit to "boost" current through meter and make it 20 instead of 10ma. Of course, "real" current have to remain as it is :-\

 

[PRR]

It would help thinking to know what your "real" circuit is. A general solution for a floating current multiplier is non-trivial; a to-ground multiplier is as simple as this:

 

[Moby]

It would help thinking to know what your "real" circuit is

Variable Mu triode. I know that easiest way is to put more sensitive meter but 

 

[Svart]

hmm.

If you are indeed working with 150vdc, then it's going to be hard.  PRR is right that the only way to do it would be to actively buffer it with something very high impedance but your voltage seems to rule out most active, solid state devices.  Anything passive will increase your current draw from your primary node and you'll skew your true measurement as will using a divider network to divide down your voltages for an active component.

maybe use an inductive type sensor on your voltage rail?  maybe you could hack a common mode choke(or a transformer) by adjusting the windings, using one side as an inductor with only a couple turns and the other side's turns are adjusted to give you the reading you desire?  Maybe even just winding some magnet wire around the voltage source's output wiring could give you a reading?

At least with the inductive sensor you will have isolation for safety and isolation from affecting your voltage source adversely.

 

[Speedskater]

We have a series resistance of 15k to 300k.  The series resistance could be a fixed resistor selected so that 10ma. could be full scale on a voltage range of your meter.

 

[JohnRoberts]

You might consider taking the meter apart and looking inside. 20ma seems high for a simple meter movement, so here is probably a shunt inside.  If yes, remove and rescale the shunt for your desired range.

JR

 

[PRR]

> right that the only way to do it would be to actively buffer it with something very high impedance

NO, there are many ways to do it. A very low-impedance input is the abstract goal. First-order trivial to implement a grounded zero-Z input: we call it "summing amp". The requirement for 2X current multiplication forces a few more parts. And meter-resistance is uncertain. And we now learn that the "10mA" goal is uncertain (tubes may sit at 8mA or 13mA), and that resistance only needs to be "small", not Zero.

And with that 5 ohm resistor, no simple fault from a "150V" line is going to smoke the 19-cent chip.

> We have a series resistance of 15k to 300k.

We now know that is a Test Rig. The actual application is "grounded cathode" with a varying grid voltage and cathode current. The ab-max voltage at Zero current is about B+/Mu, maybe 30V (though some brutal low-Mu stages can run higher). The actual happy-working cathode voltage MUST be dang near zero, or GR won't happen and THD is not canceled. It might be a couple volts, though under 1V might be better. It can't ever go negative. The idle current is the maximum possible current. For any reasonable number of suitable tubes, it is unlikely to flow 20mA; for most reasonable tubes, 2mA to 10mA.

The plan above, drawn before my second cup, still works, with one caveat. If indeed the meter is 20mA (no internal shunt as JR guesses), then the chip needs +5V to +35V power (not shown) at up to 20mA, which is generally awkward in tube gear. If your outputs are P-P hot 6V6 or bigger, you can tap the cathodes for power, but the 0-20mA draw will skew the output bias. That's probably a general problem for this 20mA meter in simple tube gear.

Frankly it might be as simple to double-up or triple-up your vary-Mu tubes to bring VCA draw up to 20mA: same total demand from B+ compared to a drop-and-waste meter-buffer supply, but some improvement in deep-compression performance.

To trim sensitivity, adjust the 5 ohm resistor. Or since few-ohm pots are rare, and the LM324 input is infinite, and vari-Mu tubes won't care about a dozen dead ohms, raise the 5r to 10r and shunt it with 1K or 100K pot, trim to suit.

 

[Svart]

We have a series resistance of 15k to 300k.  The series resistance could be a fixed resistor selected so that 10ma. could be full scale on a voltage range of your meter.

In his schematic, he already states that the load resistor is a variable resistor for setting the current on the meter.  The problem lays in the fact that he wants to monitor 10ma currents on a 20ma scale but changing the load will also change the current you are trying to measure!  Since we don't want to change the 10ma current at all, the only solution is to buffer the monitoring load or otherwise decouple it using some other technology.

I know this is not as easy as it could be but I work in test and measurement and one of the underlying rules is DO NOT CHANGE the original signal!  

IF you can live with affecting your signal, then simply parallel your ammeter with the load and use a load resistor of 1/2 of your true load.  This means that for 15k, load your ammeter tap with 7.5k.

Of course at 150v, 15k of load is roughly 1.5W of dissipation and that means your 7.5k load will be 3W of dissipation,  yet another reason I would use a decoupled solution if you can figure one out.


EDIT:

NO, there are many ways to do it. A very low-impedance input is the abstract goal.

Two sides of the same coin I think.  I'd rather have the tap be high impedance towards the power rail that way you don't load it down, then the buffer would drive the meter and do the multiplication for you.  Anyway, I KNOW that there are many ways to do this, I just don't like affecting a signal or rail if I don't have to, especially when measurements are involved.  Measurement is like being a doctor: Do no harm.


If quick and dirty is ok, then by all means brute force it.  I just think the dissipation would be high that way.

 

[sahib]

maybe use an inductive type sensor on your voltage rail?  maybe you could hack a common mode choke(or a transformer) by adjusting the windings, using one side as an inductor with only a couple turns and the other side's turns are adjusted to give you the reading you desire?  Maybe even just winding some magnet wire around the voltage source's output wiring could give you a reading?

At least with the inductive sensor you will have isolation for safety and isolation from affecting your voltage source adversely.

Known as current transformer.

There are also  Toroidal Current Sensors which are in the shape of a doughnut with two wires coming out. You pass  wire through the hole and get a proportional voltage from the sensor. Scale the voltage to suit. I have few high current versions on my drawer that I used for current sensing on an R&D project. Actually for sensing weight on an end effector of a robot arm. The heavier the payload became the higher the holding current went up on the joint drive motor. That gave me a proportional voltage output which I could process.

Farnell and RS stocks them and I am sure you can get them from US suppliers too. Just google.

 

[Svart]

Ah yes, thanks, I knew I had heard of something like this before.

 

[Moby]

Thanks guys.

You might consider taking the meter apart and looking inside. 20ma seems high for a simple meter movement, so here is probably a shunt inside.  If yes, remove and rescale the shunt for your desired range.

No, the impedance of moving coil is 2ohm (sorry for previous 2.5 ohm value I checked the resistance again). Shunt is already removed.

The plan above, drawn before my second cup, still works, with one caveat. If indeed the meter is 20mA (no internal shunt as JR guesses), then the chip needs +5V to +35V power (not shown) at up to 20mA, which is generally awkward in tube gear. If your outputs are P-P hot 6V6 or bigger, you can tap the cathodes for power, but the 0-20mA draw will skew the output bias. That's probably a general problem for this 20mA meter in simple tube gear.

I have a +17v separated from 150V, so is it worth of trying the above circuit? But I still don't understand the part valued with 5v. Is that series resistor of 500ohm? With 10ma flow I will have 5 V across but still cant figure where is the 0 point?

 

[Moby]

There are also  Toroidal Current Sensors which are in the shape of a doughnut with two wires coming out

Yes I know about but isn't that supposed to be for AC measurements? I have a 150V DC  :-\

 

[JohnRoberts]

While not very elegant, perhaps simpler to implement for one time measurements would be to make a floating, battery powered, current mirror that drives the meter with some multiple of direct path current. At tens of mA battery life wouldn't be great, but if you could do it in 1.5V a C or D cell could handle it.

JR

 

[sahib]

Doh. Sorry, I should have read yours and other posts properly.

 

[Moby]

While not very elegant, perhaps simpler to implement for one time measurements would be to make a floating, battery powered, current mirror that drives the meter with some multiple of direct path current. At tens of mA battery life wouldn't be great, but if you could do it in 1.5V a C or D cell could handle it.

JR

Yes floating PSU will be the solution but battery is not the option . It has to work relatively safe

 

[JohnRoberts]

Ok, if the meter is sitting up at 150V just add a current mirror between the meter and load that senses the current drawn by your variable load, then dumps an equal copy of that current through a high voltage pass transistor to ground. The meter now sees 2x the load current.

.01A x 150v means a power transistor with heatsink will be needed for the pass element to ground at 1.5W dissipation and more than 150v breakdown.  This assumes your 150v power supply has the extra current.

JR

Note: a floating current mirror referenced to the positive supply is similar to referenced to ground just flipped upside down.

 

[sahib]

Check this link out. PDF data sheet is located right-bottom. It might be an expensive solution though.

http://www.datasheetcatalog.com/datasheets_pdf/D/F/-/C/DF-C10.shtml

 

[Moby]

.01A x 150v means a power transistor with heatsink will be needed for the pass element to ground at 1.5W dissipation and more than 150v breakdown.  This assumes your 150v power supply has the extra current

Unfortunately no. There is not enough juice . I was thinking about using separate winding with 17v but it seems impossible .

Check this link out. PDF data sheet is located right-bottom. It might be an expensive solution though.

Looks interesting but I can't find a seller. BTW, if it's too expensive (and I'm afraid it is) then changing meter is way to go 

 

[sahib]

I am at the office now and the one I used is made by U_RD. They also have other versions that operate from DC. these are small things. I don't think they would be very expensive if you want to keep the valuable meter. There is the link.
http://www.u-rd.com/english/products/dc/dc_1.html

 

[PRR]

> I still don't understand the part valued with 5v

Sorry for my bad hand. That is an "r", shorthand for "ohms".

Now that you say the meter is 2 ohms, that part would be 4 ohms. We measure the voltage across 4 ohms, force it across your 2 ohm meter, get double the current. By adjusting the resistor we can force any ratio of current. Simple math.

> I have a +17v separated from 150V, so is it worth of trying the above circuit?

Sure. If you want to read cathode current in a conventional Variable Mu stage, and have 7V to 35V at 22mA, all these other guys are leading you off-track, just do the above plan.

Post your REAL plan, so these guys know what you are doing.

> one of the underlying rules is DO NOT CHANGE the original signal!

Are you reading Moby's posts? That is a test jig, NOT the actual application.

The actual app is a "Variable Mu". This is almost certainly metering the cathode current in a "grounded cathode".

The voltage is low, nearly ground. There is no need for "inductive type sensor". There is no safety issue.

Anyway, the GR meter (which is what this is) may sit at 10mA all day long. The app is DC!! While there are ways to "pass DC through a transformer" (modulators, Hall cells), there is NO reason to go there. And for someone who has to ask how to hack 10mA into 20mA, the design of an inductive coupling would be overwhelming.

The Variable Mu cathode is "grounded", but does not have to be NO ohms as discussed in another thread here. 10 or more ohms is negligible. The 2X multiplication and 2.5 or 2 ohm meter suggests 5 or 4 ohms on the 2X side, which will not upset the Variable Mu.

But if you believe a Variable Mu cathode must be GROUNDed..... I've had 15 cups coffee since this started, and finally flashed on "the answer". It is not as attractive as the voltage buffer, needing one more resistor plus a negative supply (rare in tube gear). There are of course other answers: current-mirrors, Norton amps, etc. I don't see any being cheaper/simpler than 4 ohms plus 1/4 LM324, and the "fault" (4 ohm series resistance) is beneath notice.