USA CAl Club: Round 3 suggestions

[bitseeker]

That sounds like plenty of temperature history, but I'm no expert.

[zekeymonkey]

It's got around 15 KB of FRAM so we could comfortably store a 4 byte time, 2 byte temperature and 2 byte humidity once a minute to FRAM and then once a day copy the FRAM contents to an external flash memory device.  A 32 MB device would store almost 8 years of data.

Looking at the pictures of the packaging, it's hard to believe the thermal time constant of package would be anywhere near as fast as 1 minute. For logging shipping temperatures, once a minute is probably excessive, but having the extra data can't hurt. If the same measurement is used during reference measurement, faster data may be useful.

Of course, when you're talking about tracking the temperature and humidity over such a long period of time and varying conditions, drift of the sensors may become of concern depending on the accuracy desired.

[rhb]

For reference environmental history, 1 minute is more than needed in most cases.  But a package sitting in bright sunlight on a hot day in Texas will get warm very quickly.

For data collection I plan a different solution.  This is *just* for long term monitoring of references.  If someone will do a dedicated board once I have a working prototype, I'd like to mount a dedicated logger on each reference.

For the measurements it's very desirable that all the readings are taken at the same time via a GPIB group trigger event.  So integrating time, temperature and humidity into a GPIB-USB or GPIB-LAN  interface is very desirable.

However, as experience has shown in Round 2, lots of things can go awry.  So a dedicated Pi or Beagle based data collection system might be worth considering.  I think we'll know a lot more by the end of Round 2.

[MaxFrister]

These are supposed to be suggestions, right?

Most of us have not yet seen the round 2 equipment.  It seems vastly premature to be planning (and implementing) round 3 changes.

I think a battery operated shipping logger seems okay in theory, but someone would need to research how that effects shipping.  My understanding is that anything with a lithium battery (including coin cells) must be in the off position for shipping and can not fly on passenger airplanes.

I think integrating environmental monitoring, let alone gpib interfaces, into every reference is a bad idea.  I'll be happy to discuss this further later in the current round.

[rhb]

It takes quite a while for things to actually get done.  Which is the entire reason for starting the discussion now.

CR2032 cells are rated at 235 mAh.  That's not large enough to be an issue.  I don't know the exact limitation, but I received 6 RTCs all equipped with CR2032s via USPS a few days ago from CA.

I'm *not* suggesting building  GPIB-USB or GPIB-LAN interfaces for *each* reference, much less into them.  I'm suggesting that a compact logging system that traveled with the references would improve the quality of the data collection.  Particularly if it logged all data to an SD card which could be mounted as a mass storage device via USB or WiFi without being removed from the data logger.

I fully expect things to change as we gain experience in Round 2.

[FriedLogic]

Different carriers have different rules, but even small cells can have restrictions. Might it simplify things to use small alkaline batteries instead?

[CalMachine]

Most of us have not yet seen the round 2 equipment.  It seems vastly premature to be planning (and implementing) round 3 changes.

Most of us have not yet seen the round 2 equipment.  It seems vastly premature to be planning (and implementing) round 3 changes.

[rhb]

Different carriers have different rules, but even small cells can have restrictions. Might it simplify things to use small alkaline batteries instead?

We are in the USA using USPS flat rate shipping.  Alkaline cells would certainly work.  But the USPS doesn't seem to have an issue with small lithium cells.  Big ones are a different story.  I used the CR2032 as an example for calculating allowable drains for an MSP430 based logger.  I had 6 of them sitting on my bench which had arrived a few days ago via USPS.  235 mAh is a rather small amount of power.  The datasheet says ~1% self discharge per year.  So I thought that an excellent design metric, nothing more.

The point is not the source of power.  The point is to actually have a *complete* temperature history for the references.  It was officially 104 F in Austin yesterday where @cellularmitosis lives.  I'll let you speculate on what the back of a truck carrying mail to Austin is.  It's 4 hours from DFW to Austin driving down I-35.  That mail does not go by air.

Rather than people speculating without even basic information, how about we measure and record actual data?

[CatalinaWOW]

If you want a very small sample set which is virtually guaranteed not to see extreme possibilities build a recorder.  You can at least look back and say something happened and confine your speculation to what effects what happened might have had.  If you want to have a pretty good idea of what the possibilities are the attached document is a pretty good resource.

[rhb]

That's useful information, but it doesn't tell anything about the inside of a truck sitting in the sun.  Or an attic.  Dad and I once found a thermometer in an attic where we were working which had exploded.  It read to 120 F.  The end was blown off.  The cold extreme is well documented by the publication you linked, which I've saved as it's potentially very useful, but it says *nothing* at all about the limits of high temperature.  A flat black box sitting in full sunlight at 120 F ambient will be too hot to pick up with bare hands and probably reach well in excess of 160 F inside.

I grew up dealing with having to enter spaces where the heat was potentially lethal.  And about once a year in the southern US, some fool leaves a child or animal in a car for a few minutes to run in the store.  And comes out to find the child or animal dead.  Animals dying don't make the news, so I'd bet that happens a lot more often.  But the dead baby is far too common.

In any case, what matters is not what *might* happen to a voltage reference, but what *has* happened.  And the only way to know the latter is to record the information.

My interest is the continuum mechanics of the reference device.  In particular the thermoviscoelastoplastic deformation of the metalization layer and the bond wires.  FWIW I did *not* make up the alphabet soup adjective.  That's what the mechanics of materials crowd uses for such things. So I'm just conforming to the standard nomenclature.

On the other hand, TiN, having contributed a very fine piece of gear is justifiably concerned that it not be abused in transit. I think his concerns should be respected.

[Tj138waterboy]

Has any more progress or decisions on round 3 kit or kits been made yet? I would like to know which reference boards
Would be used because im about to bulk order some of Tins designed KX boards possibly as well as ar488 gpib adapters also hamon divider module. Would donate some of each to the cause if needed as well as get feedback on my designed hamond divider.

[KK6IL]

My reference needs are simple.  I have two HP3456's, history unknown, a LM399 reference I built, and had a cal lab measure 15 years ago, and an L&N 10K resistor I've had 50 years. Three circulating PX's would only let me calibrate the 3456's and verify the L&N, but not the LM399.

The Cal Club Kit Instructions need to specify  the output current allowed from the FX. Can it drive a 3.333K Hamon divider to produce 1 Volt?

[ArthurDent]

...The Cal Club Kit Instructions need to specify  the output current allowed from the FX. Can it drive a 3.333K Hamon divider to produce 1 Volt?

I don't know if anyone else has checked to see how the FX behaves under load because measurements are generally done with as close to no load as possible, or poteniametrically with a sensitive null meter like I did with my HP419 null meter set to +/-3uV full scale (diagram below). Using that circuit I checked my LTZ1000ACH reference against the club's FZ with different load resistors and here is what I posted previously.

"I decided to check the 2 supplies to see how they reacted to light loads and I knew mine could survive a direct short for some time because I designed the added transistor buffer within the opamp feedback loop to handle a direct short and also protect the opamp. With a 100K load both supplies dropped 1uV; at 10K, mine dropped 10uV, the club’s 20uv; and at 5K, mine dropped 20uV, the club’s 40uV so the output resistance of mine is less than the club’s but neither would be loaded like that in any testing situation."
 
Any load will cause the output to drop and it looks like a 3K load could cause the output to drop more than 50uV.

[TiN]

FX itself able to sustain high (relatively, 20mA any direction) loads, however it need two additional posts to allow 4-wire connection.
Once USA Club reference make its way to me, I will modify it and add those posts to allow use of FX as current source.

[guenthert]

[..]
Can it drive a 3.333K Hamon divider to produce 1 Volt?

As has been stated before, you'd want to avoid loading the reference.  Instead ('potentiometer method') you'd want to drive the Hamon divider with a beefy (low output resistance), stable, noise-less source (say a lead battery or a precision PSU) which has been reduced (resistance divider or something more fancy) to the voltage of the FX reference (verified e.g. using a null meter or a high impedance long scale DMM).

[rhb]

How many people are interested in RF?  I've acquired quite a bit of T&M kit of late.  Most is uncalibrated, but I'm working on that. 

Is there interest in reference capacitors and inductors, noise sources, etc? There's no point in sending such to people who are just concerned with volts, ohms and amperes, but there may be people with a wider range of interests.  So I have in mind a 2nd set of stuff.

I now have both 4284A and 4285A LCR meters and an 8970B noise figure meter and 346B noise source.

Have Fun!
Reg

[bitseeker]

I'm no RF guru, but have dipped a toe into the waters. Is there a club-friendly way to source a 0, -10, -20, etc. dBm signal to check calibration (and calibration factor) of power meters and thermistor mounts?

My interest here is just to verify that my HP 432A meter and 478A and 8478B mounts are in reasonable nick.

[rhb]

As a first step, circulating a calibrated step  attentuator would provide a start.  A signal source gets a bit more difficult because of weight and cost.

There are plenty of DC to 2.7 GHz JFW step attenuators on ebay.  I've got 4x 0-30 dB by 1 dB and a pair of 0-60dB by 1 dB.  They're cell tower surplus.

I intend to automate a full annual cal via relays and attenuators once I get my lab set up.  So this is an essential step in developing that process.

I'll look into what it would cost to have a commercial lab calibrate an attenuator for cal club use.

Have Fun!
Reg

[bitseeker]

Yeah, a typical sig gen would be too much. I was wondering about something that had a fixed frequency (or set of frequencies) at a given output power. Similar to the cal signal on a spec analyzer. Then, attenuators will get you to other power levels.

[rhb]

I was thinking *exactly* the same thing last night.  While it would be very nice to have a lightweight  DC to 2.2 GHz RF synthesizer based on a pair of ADF4351s with proper LP filters and ALC, that would be a substantial design effort.  Not something I can undertake at this time.  But  an xtal oscillator with a carefully designed and implemented ALC  would be tractable.

I did a bunch of work some years ago on an automated version of the xtal tester Chris Trask designed around a Butler oscillator (differential amp) which incorporated ALC.  I could readily make up one of those Manhattan style neatly sealed up in a nice aluminum case.  All I would really need to do is refine the ALC circuit.  I've got a slew of 40 MHz 3rd overtone xtals with a fundamental frequency around 14.3  MHz.

That would put it right in the middle of the HF spectrum and avoids the complications at higher frequencies.  For this purpose I'd want to put a very good LP filter before the ALC pick off and probably a 2nd ALC on an output buffer.  A single frequency source would be much easier to get a good impedance match on the output.  It would make an interesting project which is within range of my limited skills.  I'm really feeling the limitations imposed by not being proficient with a PCB design tool.

I've started looking for 10 dB and 1 dB step electronically controlled lab grade attenuators for my automated cal system.  With those calibrated commercially I can then use them in conjunction with my 438A/8482A and  8648C to calibrate a JFW step attenuator over the  full frequency range.
I've got several JFW attenuators already.  They are nice solid builds and are rated at 2 watts.  So they should be pretty immune to accidental damage.  They are widely used in cell towers so with patience they can be had cheaply on ebay.

A 0 dBm xtal source, 1 MHz - 1 GHz diode noise source, attenuator,  a bridge for comparing attenuators and a couple of reference capacitors and inductors would be a pretty decent start at RF cal in a small or medium flat rate box.

I have a small ocean  of WW II era RF gear Dad left me so a bit of digging should turn up some parts I can package up to General Radio reference standards.  They probably won't be any specific value, but they will be very stable.  Between my 8753B and 11801 I should be able to characterize any residual parasitic reactances very accurately.  Add an MSP430 to report temperature and humidity over a USB serial port and we ought to be able to get pretty decent uncertainties.

Have fun!
Reg

[bitseeker]

Interesting. Yeah, an RF package for those interested in such things.

[rhb]

Yesterday I got a Fluke 515A portable DMM calibrator while picking up a Tek 577.    Not being at all familiar with it, I didn't realize that the switch knob was missing until this morning.  I drove 700+ miles yesterday and was rather tired when I bought it and *really* tired when I got back home after 12 hours of driving.

Major lesson, do not buy stuff you've never seen without doing more research.  All I knew at the time was it was a Fluke DMM calibrator "as is".

However, it turns out that it works.  It's got 1980 date codes, so it's old, er, "well aged".  But at 10 V my best 34401A reads 10.0033 V.

After I buy or make a replacement knob, I'm thinking about sending it for a factory cal and then circulating it to anyone who wants to calibrate their DMMs. General notion is that the person using it pays shipping both ways plus a small fraction of the cost of the OEM cal.  Unfortunately, it will not fit in a flat rate box and shipping might be too expensive.

I also got an HP 343B-K05 RF reference which provides 0 dBm output at 5 and 50 MHz.  According to my 8560A the levels are off, but it was made in 1983.

Have Fun!
Reg

[bitseeker]

Wow, that's a lot of driving. Neat finds, though.

I know what you mean about buying something on a whim without prior research. It's really hit or miss, but sometimes hard to resist.

[Magnificent Bastard]

My $0.02 on this:

All you really need are some "known" fixed points for voltage, resistance, and temperature.  So, a 10V reference, a 10K resistor, and a temperature probe.  With just these 3 things, everything else can be derived if needed.

Each of these should be designed to be: Relatively low cost (compared to commercial standards), rugged enough to be roughly handled without losing calibration, small enough to be shipped in the lowest cost priority mail package, robust enough to be used by inexperienced people without damage, and stable enough to verify up to an 8-1/2 digit meter (or even calibrate a 3458A if you are lucky enough to own one of those).

This suggests: An LTZ1000-based 10V voltage reference, a quad-VHP100 based 10K resistor (similar to Lymex's design), and a PT100 based temperature probe.

The voltage and resistance standards should have an uncertainty [k=3] of no more than 0.5ppm, and the temperature probe should have an uncertainty [k=3] of no more than 0.01^oC.

If a person needs to have more voltage or resistance cardinal points, they would be responsible to create them on their own, using the supplied references as standards.

Obviously, calibrating these standards would be expensive.  Someone with an SR104 could get that calibrated (about $400 plus shipping and insurance, maybe $800 total), and that should be used to calibrate several in-house 10K standards, as well as the 10K traveling standard(s).  Someone with a Fluke 732 could get that calibrated (about $500 plus shipping and insurance-- maybe $1000 total), which should immediately be used to calibrate several in-house 10V voltage standards, as well as the 10V traveling standard(s).  Someone with a set of fixed-point cells could calibrate the temperature probe(s).  Costs for these calibrations would be shared across the calibration club, and shipping of the traveling standards would be paid (in each direction) by each end user for each use.  Calibration of the traveling standards against lab standards before shipment and after return would be done on a volunteer basis by the central hub (which could be one hub that handles all standards, or as many as three hubs that only handle one type of standard).

There could also be a down-graded kit for people with 5.5-digit (or less) equipment, consisting of an LM399-based 10V reference, a single VHP100-based 10K resistor, and an 100K NTC-based thermistor probe.  This kit would also have a slightly downgraded uncertainty for each standard-- say no more than 3ppm for 10V and 10K, and <= 0.1^oC for the temperature probe.  This would have a lesser participation cost than the higher level standards.

This is how I see it as the only alternative for a serious calibration check among club members-- anything else is just "playing", and provides nearly zero real value other than entertainment.

[rhb]

Perhaps you should read the USA Cal Club Round 2 thread.  That kit is pretty close to what you describe, though neither reference is 10 V.  It's quite a lot of work to design and build a device to achieve 30 ppm uncertainties from 10 microvolts to 100 volts which is the basic DC spec for the Fluke 515A.

What you describe is useless for anything but DC.   Using that to calibrate RF gear would take a lifetime of building instruments.  And without a time reference, not even possible.  The subject of "metrology" covers a lot more than the volt, ohm, ampere and temperature.

I'm not even remotely a volt or time nut.  I'm  concerned with practical calibration of a lab full of LCR meters, VNAs, spectrum analyzers, RF synthesizers, semiconductor testers, etc.  That requires accurate volt and time references, but those alone are quite inadequate. It requires quite a lot more and it is vastly more difficult to get 100 ppm RF measurements than 10 ppm DC measurements. And in many instances national laboratories struggle to get 10,000 ppm uncertainties.  Try to find a 0.1 dB step RF attenuator good from DC to 18 GHz.  After all, at DC that's just a 0.25% accurate  resistor ladder.  I want one, but I'm still looking.  They may not exist above 6 GHz at any price.

I started this thread to discuss aspects of speeding up circulation of the references  among people and encompassing at least basic AC measurements up to RF.  Primarily the latter.

Have Fun!
Reg