What I am doing is measuring the resistance of foils, as a way to infer their thickness.
I have other tools to access the thickness of the foils I am measuring, but these are
time consuming, and I want to develop a more rapid assessment of their quality.
I'm going to marry a current source/sink to a 4-point "probe", take millivolt readings, cross
them with my other measuring tools, and develop a data table.

In my initial testing with a 10mA, 100mA, and 500mA current sources, I have discovered
that I need at least 1000mA to produce reliable data across a range of foil thicknesses.

I've been scouring the web for various approaches, and have settle on construction of my
own precision current source/sink, and implementing the construction of a four-point probe
of my required specification. Because a thousand and one possible solutions are stymied
by the CHIP SHORTAGE, I come to my fellow AAC members, hat in hand, with a request
to evaluate the following schematic, and offer a possible equivalent replacement for
the ONE part that I am unable to obtain, the LM385-1.2 voltage reference:

 

You seem to have a reasonable grip on the situation,
but I think what You may be missing is "distance", and Contact-Resistance variability.

The more distance between the probes,
the better resolution and repeatability You will have.

I don't think that having 4-Probes is a particular advantage.

The most important factor is Contact-Resistance-Repeatability,
( increasing Contact force and area against the Foil may radically change the reading )
.
.
.

 

Dimensions: outer probes 65mm, inner probes, 55mm. I agree, contact resistance on the outer/current supplying probes is a concern, but probes are going to be 3mm diameter, gold coated, and mV probes will be perhaps 1mm, gold coated. Probes will have a small spring each, so contact pressure will be uniform from test to test.

Readings will scale between 10mV and 500mv.

 

Let me add, I can substitute a 9v battery, and dispence with the "voltage-converter", and I don't need the battery monitoring led,
but it's a nice to have feature, and it appears both the current injector, and battery monitor relies on the LM385 voltage reference, so, why not include it?

 

I'm going to marry a current source/sink to a 4-point "probe", take millivolt readings,

I don't think that having 4-Probes is a particular advantage.

If by "4-point" probe he means a Kelvin connection, then that is exactly what is needed in this application.

A problem I see is that all of the foil samples being tested must have the exact same dimensions and shape for comparison readings to have meaning. Also, the probes must be in the same locations for each test.

Gotta hand it to Dave - I've never seen a 555 discharge output used as an open-collector driver for a flyback circuit. Still, I agree with deleting the 555 circuit.

ak

 

Samples with be same size and shape. Probes will be held in a caddy that we keep them tightly oriented to each other.

 

For a one amp power source a 555 timer is not part of the discussion. A monolithic IC voltage regulator and a suitable resistor will serve very well as a fixed current regulator. I think that the TO3 5 amp device, ( LM309-5) with about a 5 ohm resistor will provide the refulated one amp quite well. The TS will need to research "three terminal voltage regulator as a current regulator" to find the very simple circuit..
As for the 4-wire resistance measurement, it is mandatory in order to have any hope of repeatability and accuracy.

But the whole concept is rather uncertain because of the variable resistance of alloys that are not perfectly consistent. Besides that, the resistivity of many metals varies with the degree that they have been cold worked, as well as their temperature,

I wonder how accurately the thickness needs to be known.

 

For a one amp power source a 555 timer is not part of the discussion.

The 555 has nothing to do with the one amp power source. If you check the schematic, you will see that the 9 V source is used only to power the comparators and provide a higher max gate drive voltage. The 3 V source does the heavy lifting.

ak

 

How is the 9V from the 555 regulated?
Without regulation, that voltage could get high enough to damage the comparators and/or the MOSFET gate.

Also have you checked the stability of A2?
It's a comparator being used as an op amp, and my simulations show the stability may be marginal (causing oscillations), even with the C5 compensation.
It might be better to replace both comparators with an LM324 or LM358 op amp.

 

Where is the problem with getting an LM385-1.2 ? (provided you're somewhat flexible wrt the chip package)
See https://octopart.com/search?q=Lm385-1.2&currency=USD&specs=0

 

No, I did not thread thru that massive circuit when a current regulator using a 3-terminal IC regulator can do the same with so very much less complexity? It seems that few are able to visualize a two-component circuit, an LM309-K and a five ohm resistor. The resistor will be burning five watts, and so it should be rated at least 7.5 watts, or ten is better. The required supply voltage is 5 volts plus whatever the most the load will drop. And obviously the supply must be able to deliver over an amp.
The complete circuit consists of the voltage supply connected to the regulator input terminal, the 5 ohm resistor connected between the regulator output terminal and the regulator common terminal, and the current output being delivered from the junction of the resistor and the regulator common connection. I understand that many folks are unable to visualize a circuit from a text description.Not understanding text is certainly a serious impairment.

 

Adding a zener and a resistor will turn the TLC555 stage into a regulated (not very well, but at least limited) voltage supply. Using a proper boost regulator would be a better choice.

If possible, consider a ratiometric design. No need for a highly accurate or stable reference as the forcing current needs to only be held in the vicinity of one ampere, maybe using a regulator IC as a constant current source. Measure the 4-point detect voltage (with a good differential amplifier) and divide it by the voltage measured across a grounded sense resistor in the current control loop.

 

Thanks for the input folks. Indeed, I have been able to source the LM385-1.2, as well as a nice selection of other voltage references. I'll begin construction soon. The LM309 or other voltage regulator to current solutions are not optimal because I need a portable battery operated device, and linear regulators just drop too much voltage, burn too much current to be practical. Also, they lack the required accuracy.

 

Adding a zener and a resistor will turn the TLC555 stage into a regulated (not very well, but at least limited) voltage supply. Using a proper boost regulator would be a better choice.

If possible, consider a ratiometric design. No need for a highly accurate or stable reference as the forcing current needs to only be held in the vicinity of one ampere, maybe using a regulator IC as a constant current source. Measure the 4-point detect voltage (with a good differential amplifier) and divide it by the voltage measured across a grounded sense resistor in the current control loop.

I am intrigued, but I am not competent to put that into a complete package. I plan on measuring the voltage drop with a milivolt meter, with seperate coin cell supply, and get a corrolated ohmic value.

 

How is the 9V from the 555 regulated?
Without regulation, that voltage could get high enough to damage the comparators and/or the MOSFET gate.

Also have you checked the stability of A2?
It's a comparator being used as an op amp, and my simulations show the stability may be marginal (causing oscillations), even with the C5 compensation.
It might be better to replace both comparators with an LM324 or LM358 op amp.

With either of those op amps, would I need to change any other component values?

 

With either of those op amps, would I need to change any other component values?

You can eliminate R9 and C5.

 

No, I did not thread thru that massive circuit

Massive? Really? 2 IC's and fewer than 20 misc parts - ?

when a current regulator using a 3-terminal IC regulator can do the same with so very much less complexity? It seems that few are able to visualize a two-component circuit, an LM309-K and a five ohm resistor. The resistor will be burning five watts, and so it should be rated at least 7.5 watts, or ten is better.

Disagree, and I can visualize it just fine. The problem is that a 2-component circuit will not meet any of the requirements in post #1. What I visualize in Post #1 is an *adjustable* constant current sink. If you flip over the application and replace the current sink with a current source, the standard 3-terminal regulator circuit (beginning with the LM109 in 1969) requires that 100% of the output current go through the current-setting resistor. While an LM309 circuit would have fewer parts than the post #1 circuit, a pot rated to handle 2 A (100% margin, just like your 10 W resistor) at any adjustment position is not a common item, especially at a relatively low value.

The required supply voltage is 5 volts plus whatever the most the load will drop.

Since the available high-current supply is only 3 V, that is a problem. Also, that 5 V value does not include the 309's input-output differential voltage.

A similar circuit based on the LM317 would perform much better. The worst case 309 GND pin current is 100x greater than the 317 adjust pin current, and this current degrades the output regulation. Worse, the 309 GND current can vary by +/-5% over temperature, a significant reduction in output regulation.

Still, while a 317 circuit would work much better in this *type* of application, it would not work in this *specific* case because its dropout voltage, while much lower than a 309, still would be too high for a 3 V source.

The circuit in post #1 will outperform a 3-terminal regulator approach. There is only one component with any relevant power dissipation, and any changes in its operating parameters are compensated for by the negative feedback loop. Once adjusted for a desired current, all of the circuit's stability and repeatability comes down to the behavior of R10. I think a 1% tolerance part should have very good thermal performance.

Not understanding text is certainly a serious impairment.

I completely agree. And I still have some LM309's if you would like one.

ak

 

Still, while a 317 circuit would work much better in this *type* of application, it would not work in this *specific* case because its dropout voltage, while much lower than a 309, still would be too high for a 3 V source.

I'm all ears: If I switched to a fixed output of 1A, & increased voltage between 5 to 6V, would an LM317 constant-current solution provide a solid, repeatable, precision 1A source? (1% regulation or better)

 

Maybe. It depends on how much better as individual part is than the performance numbers in the datasheet.

The problem is temperature. At 1 A, a 1.25 ohm sense resistor would drop the required 1.25 V, leaving 3.75 V across the 317 chip. At 1 A, that's 3.75 W. I think the combination of the chip's internal voltage reference and its adjustment pin current changing as a function of temperature (internal heating) probably will be greater than 1%. But if the chip actually is 5x better than the datasheet says (to cover manufacturing margins), then maybe yes.

Separate from that is the 317's initial voltage tolerance. This is the variation in the internal voltage reference over the operating temperature range and from one chip to another. The datasheet says it is +/-4% over temperature. It also says the output regulation error can be over +/-1% on top of that.

My experience is much better than that. Last year I put in some garage door status lights. I used the LM317 constant-current circuit to drive switched groups of LEDs. I was *very* impressed with the output stability as I hit the chips with freeze spray and then a heat gun.

ak

 

If you use a ratiometric measurement as suggested in post #12 by measuring the voltage and current at the same time, then the accuracy/stability of the current source is not critical.