EEVblog #544 - Fluke 5450A Resistance Calibrator Teardown

[PA0PBZ]

Mica is one of the few materials that can provide a physical barrier, yet allow alpha radiation through. It's pretty much the exclusive material used in high sensitivity geiger probes.

Also used in microwave ovens to prevent the food from entering the wave guide, and supporting the heating wire in toasters (at least the older ones).

[EEVblog]

Why aren't the resistors mounted inside a thermal oven to reduce the effect of ambient temperature and sealed in a hermetic chamber with desiccant to prevent humidity from changing the values?

Because they are already low enough tempco to meet the desired spec.

[rs20]

Bit of constructive criticism -- I think an opportunity was missed in the video to point out that, when the 4-wire method is used, low contact resistance relays are pointless. Doesn't matter on the current supply side, and there's practically zero current on the sense side, so all around it doesn't matter. And if you're using this bit of kit without using four wires*, you lose out on performance in a big way: referring to the manual, using the 5450A in 2-wire mode at 1 ohm introduces a 25 milliohm error (90 days, after the 2 wire compensation, using test leads that were included in the calibration procedure, etc, etc), which works out to 25000 ppm, a thousand times more than the 4-wire mode specs you pointed out.

Throughout the video, whenever you were talking about contact resistances, or surprise and the non-soldered relay sockets, or the mention of amazing gold plated relay sockets, you should have heard me shouting "only in 2 wire mode!?!?". Those are not 25ppm in 2 wire mode relays or sockets. And 25ppm in 4 wire mode is achievable with a rusty nail.

Along the same lines, the relays on the top and bottom of that schematic aren't the rows and columns of a matrix (without a link to that earlier video you mentioned, I can't verify what you meant by "matrix", so I may be putting words in your mouth here, sorry), they are the separate output and sense lines, specifically separately relayed to make relay contact resistance completely, completely irrelevant**. You could connect a 1 ohm resistor using 5 ohm contact resistance relays all around, and it'd work perfectly fine**.

That minor nitpick aside, outstanding video as always; amazing components on show in there.

* Sure, you can calibrate a multimeter that only supports 2 wires, but in that case, all the plug contact resistances and lead resistances are probably going to dwarf even a half-decent relays' contacts? So again, what's the point in using million dollar relays?
** In 4 wire mode.

[BravoV]

Really surprising to see that new "branded"  benchtop DMM just can't do auto ranging decently, isn't that strange ? 

No. What's the problem? Any decent autoranging system will be designed with some hysteresis to prevent the meter from nervously switching between two ranges. In general a stable reading with one digit less is preferable to the meter switching range every few seconds due to noise. If you want to force the meter to choose one particular range, use manual ranging. Calibration procedures will always call for manual ranging.

So we just can not use this expensive benchtop DMM say like in logging mode at a DUT that it's resistance may wildly jumps across ranges ? 

[rs20]

So we just can not use this expensive benchtop DMM say like in logging mode at a DUT that it's resistance may wildly jumps across ranges ? 

He said hysteresis, not a complete refusal to ever change.

What that means is that is that the measurement ranges have slight "overlaps"; if the measured value is within that overlap region, it won't change range and just stay put. On the other hand, if the measured value changes 10x, 100x, or probably even 2x, it will most certainly change and follow.

[EEVblog]

Along the same lines, the relays on the top and bottom of that schematic aren't the rows and columns of a matrix (without a link to that earlier video you mentioned, I can't verify what you meant by "matrix"

I'm used to using the word "matrix" because in a former day job I spent years designing countless "crosspoint" type relay matrices to do all sorts of stuff for production test systems, it's just a term that generically pops out of my mouth whenever I see any form of relay switching system with a lot of relays.

[cengland0]

Dave said something that interested me.  There were four 40 ohm resistors in parallel to make 10 ohms.  Each resistor was at 0.05% tolerance and he said having four of them makes that better than 0.05%.  Why would that be?  Obviously later in the video he discovers they are all specially matched which could make this true but he said this before knowing that detail.

So each 40 ohm resistor could have a resistance between 39.98 and 40.02 ohms and still be within tolerance.  If each resistor just happened to be 39.98, you would have a total resistance of 9.995.  That is still 10 ohms at 0.05%.

[Vgkid]

^^^ I believe that these would be binned to reduce the errors.

[free_electron]

the electrols are used in the theithley electrometers as well.

[SeanB]

I got a Google itself ad. and I am running this site as the only one not adblocked ( gives the clicks to Dave). Looks like they are monetising almost every video these days.

[SeanB]

The boards most likely were made, tested and then stored in a CA store until needed, and then they were pulled out of storage, cleaned and assembled.

The Mica used to make those low value resistors was likely mined in South Africa, funnily enough in a mine situated in the Lowveldt right next to a small village called appropriately enough Mica. Nice town, with some beautiful scenery and some lovely road cuttings through the actual mica bands.

[cengland0]

The boards most likely were made, tested and then stored in a CA store until needed, and then they were pulled out of storage, cleaned and assembled.

The Mica used to make those low value resistors was likely mined in South Africa, funnily enough in a mine situated in the Lowveldt right next to a small village called appropriately enough Mica. Nice town, with some beautiful scenery and some lovely road cuttings through the actual mica bands.

Doing some cave exploring in Italy, I found some naturally occurring mica.  It was one of my best geological finds and it remains in my rock collection today.  It's only about a square inch, oddly shaped, about 2mm thick, and very fragile.

[wavebits]

I'm thinking the shielded box around the low ohm wirewound resistors is not so much to prevent electrical noise but to decrease thermal differentials caused by drafts within the enclosure. The noise suseptability of this low impedance part of the circuit would be minimal. Temperature gradients however, would be an important consideration.

[Stonent]

I'm thinking the shielded box around the low ohm wirewound resistors is not so much to prevent electrical noise but to decrease thermal differentials caused by drafts within the enclosure. The noise suseptability of this low impedance part of the circuit would be minimal. Temperature gradients however, would be an important consideration.

I figured it was to just keep something from falling on them and shorting or affecting the measurement.

[RadoK]

Here is what I mean

Hi Dave,
I had the same Ad problem in Firefox a week ago. If your problem is firefox related then check firefox Add-ons menu. In my case there was one add-on that was showing ads below videos. After removal everything was ok.

Rado

[nivek]

Here is what I mean

Are you per-chance using the smooth gestures plugin in Chrome?  They recently updated it to include ad's that look identical to that... you can disable them in the settings for the extension (Hidden at the bottom of course).

[Razor512]

How do companies come up with a reference to begin with, for example how did they come up with the first resistance reference and what makes it so special and accurate?

For a bunch of calibration equipment, they seem to be calibrated off of a bunch of other calibration equipment is a weird cycle, but there is never ant decent info on how the reference came to be in the first place and how was that first one made to be accurate when there is nothing else to compare it to?

[alm]

This gets into experimental physics and metrology. You could define 1 Ohm as the resistance of a certain length and volume of mercury (artifact calibration), which is what they did in the past. You can also define the Ohm in terms of current and voltage, assuming you have standards for 1 A and 1 V. The current definition of resistance is based on quantum physics (search for integer quantum Hall effect). Since this is not a very practical definition, they will make a (supposedly) very stable standard resistor and characterize its value and stability using the quantum Hall effect. They will then use this resistor to calibrate other resistance standards.

[John Coloccia]

Hey, did anyone notice what looks like silver Mica caps around the Z80?  I wonder what function they serve that they would need to use mica caps?

[codeboy2k]

This gets into experimental physics and metrology. ...they will make a (supposedly) very stable standard resistor and characterize its value and stability using the quantum Hall effect. They will then use this resistor to calibrate other resistance standards.

This stuff is pretty cool, actually.  I just read up what I could on it.  You start with a MOSFET with high electron mobility (HEMT), sometimes also called an HFET. For making a quantum hall resistor (QHR), they will make a device similar to a typical MOSFET, using GaAs and AlGaAs, but with specially constructed terminals for making a 4-wire resistance measurement.  At ~1K temperature and in the presence of a magnetic flux density of 7-9 Teslas (the actual value depends on the silicon structure), then the 2D electron gas (2DEG) layer (which is tightly confined beneath the gate) will completely quantize, and a resistance between the source and drain is formed that is exactly 25,812.807 ohms !!

In other words, applying a fixed, known accurate constant current source from drain to source, it will develop fixed voltage plateaus (called the Hall Voltage V_H) across the drain to source at each quantization level as the magnetic flux density is increased from 0T up to 8 or 9T and higher.  Each voltage plateau V_H(i) actually represents a different resistance called the Hall resistance R_H(i) (because the current is constant). Later experiments related this to the universal constant h, Planks constant and e, the elementary charge on an electron, as  R_H(i) = h/ie². The physicist who did this research and related this quantum hall effect to the universal constants was Klaus von Klitzing. He received a Nobel prize for this discovery in 1985.  In 1990 they (SI standards body) chose i = 1 and thus 25,812.807 ohms as the resistor standard and called it R_K-90 (K in honor of von Klitzing).

And then, as alm said, this resistance is not practical to use, so they transfer it to a multitude of standard resistors like 1 ohm, 100 ohm, 1000 ohm and upwards using various methods, usually involving a very accurate null bridge and cryogenics.

[alm]

And the fun thing is that current is currently defined in terms of Ohm's law and the definition of resistance (in terms of the QHE) and voltage (defined based on the Josephson effect). As far as I know they haven't come up with a QM definition of current, yet. So just put a known voltage from a Josephson junction array across a QHR (via some transfer standards) and you get your current standard .

[AG6QR]

Dave said something that interested me.  There were four 40 ohm resistors in parallel to make 10 ohms.  Each resistor was at 0.05% tolerance and he said having four of them makes that better than 0.05%.  Why would that be?  Obviously later in the video he discovers they are all specially matched which could make this true but he said this before knowing that detail.

So each 40 ohm resistor could have a resistance between 39.98 and 40.02 ohms and still be within tolerance.  If each resistor just happened to be 39.98, you would have a total resistance of 9.995.  That is still 10 ohms at 0.05%.

The resistors will have something like a "bell curve" normal distribution.  If they've got a standard normal distribution, the chance of getting all four of them at the very bottom or very top of the range is very slim.  The very worst case would be that the percentage error in the combined resistance would be as bad as the percentage error on one of them.  But the average case is much better than that, even if you're not specially selecting the resistors in matched sets.

If you're randomly selecting the resistors, you can only put a statistical confidence level on the likelihood of the worst case not happening.  But if you individually select the resistors, combining high ones with low ones and medium ones with each other, you can guarantee that the worst possible case doesn't happen.

[rarthur22]

I noticed that you didn't mention that the relay contacts were bifuricated.  We use those when running very low current inputs on control systems.  Standard relays have too high a minimum load to keep the contacts clean.  When switching low mA loads these type are necessary.   The minimum current for single contact relays would be too high for this type of application to get reliable operation.

[Bored@Work]

And the fun thing is that current is currently defined in terms of Ohm's law and the definition of resistance (in terms of the QHE) and voltage (defined based on the Josephson effect).

The SI ampere is still defined via the force between two conductors. But that is impractical to realize, so standard labs derive it via Ohm's law. There is some work going on to redefine the SI ampere based on the charge of a proton. But the BIPM members are discussing that for years now, and they don't seem to be in a hurry to do the change.

[alm]

The resistors will have something like a "bell curve" normal distribution.  If they've got a standard normal distribution, the chance of getting all four of them at the very bottom or very top of the range is very slim.  The very worst case would be that the percentage error in the combined resistance would be as bad as the percentage error on one of them.  But the average case is much better than that, even if you're not specially selecting the resistors in matched sets.

You can't assume a normal distribution (although the Fluke engineers might have had additional information), especially not one centered around the nominal value, so worst case is that you should assume a uniform distribution within the stated tolerance. Still, the standard deviation (and any confidence interval you care to calculate) of the parallel (or series) combination will be smaller than the single resistor. I believe the freely available book 'Analog SEEKrets' by Leslie Green contains a decent discussion of this topic.