Getting Stuff Calibrated?

[creepyoldenj]

Hey all,

I have several pieces of used test equipment for a bench I'm trying to put together.
They all seem to be in good working order, but as they are
decades old,  I'd feel a lot warm and fuzzier, before I do some serious
bench work, to verify their accuracy.

I'm sure there are a lot of people out there in a similar fix:

I can't afford to go to a calibration house, where the cal.
on one piece of test equipment would cost me more than most
of the test equipment I own.  Also I don't need traceability.

I live on a fixed income, I can't afford to buy, even used cal. equipment.

I live in a small town.

I don't live close to a school or university that would take on cal-ing
my stuff as a project for students.

I can certainly build circuits to cal. my stuff, which would be sort of fun,
if it didn't cost me too much in expensive or hard to obtain precision components.

I could rent or borrow for a week or two cal./ precision test equipment if it's affordable.

I have a Tek 2235 that has drifted out of spec.
and requires a proper cal. I have the documents
need to perform a cal. on it, no big deal if you have a Fluke or Tek
scope calibrator.

Same goes for my frequency counter, signal gen., and a fluke 8050A.

I can do the cals myself, I just need precision signals to make it happen.

Any ideas would be appreciated.

BTW, I know I shouldn't, and just suck it up, but I do get a little jealous watching videos of people
who have piles of new, high end test equipment, and go on about about their specs.,
while I have knowledge to use the stuff and want to do what they do with it,
but I simply don't have the resources to make it happen.
Frustration & sour grapes, I'll admit it. But then this is getting off topic, sorry.

[robrenz]

I take your avatar is an indication of these feelings?

[jpb]

I think you can go quite a long way with a cheap GPS and maybe voltage references from http://www.voltagestandard.com/Home_Page.php

It depends on your counter, but mine (a TTi 930) will measure sub Hz so I can check it on the 1pps from a cheap (approx $30) GPS module:

http://proto-pic.co.uk/ultimate-gps-breakout-66-channel-w-10-hz-updates-mtk3339-chipset/

I use the above because you can use an external antenna with it. I find that the 1pps jitter is less than 10nsecs so for a 100sec gate it is good for 10 figures on the counter.

It all depends on how precise you want to be - there are lots of threads on this forum on people building serious calibrators.

[SArepairman]

Sometimes I think its alot easier to accumulate precision equipment then make use of its 45 digits of accuracy. When you get into precision it often becomes a cumbersome burden to take good measurements.

But On the other hand its a great "sanity check", and its great to have crazy equipment so you feel confident about approaching difficult problems. I know before I had certain pieces of equipment that were described as necessary I would shy away from those solutions. Once you spent the 500$ or 30 hours design time on some calibrator etc you suddenly have alot more initiative to work on more complicated things because you feel guilty about how you spent your time/money.

check dave jones videos about precision resistor standards. As for an oscilloscope keep in mind the accuracy is not great, you are reading it off a picture, its not like there is 6 digits of numbers to look at. You can calibrate it off any half decent function generator for frequency and probobly any old power supply for voltage (I mean, a cro has the amplitude accuracy of a handheld DMM at best).

[qno]

When I started electronics we only had analog meters with 3 to 5% accuracy full scale 20kohm/V .
To read the needle you could make another 2% mistake.

In the company we had one very precise 1% electronic VA OHM meter.

Still we could produce nice amplifiers radio's and TV sets.

The first 3.5 digit multimeter was a quantum leap forward in accuracy but the radio's TV's and amps where not getting any better.

You have to ask yourself if you really need the accuracy.

This reminds me of an old saying: A man with one thermometer knows the temperature. A man with 2 thermometers does not.

[KedasProbe]

What about buying a new Rigol DM3058E Digital Multimeter?

You get a new toy and you can use it to check your other devices.

[pickle9000]

For voltages I use one of these.

http://www.ebay.ca/itm/AD584LH-4-Channel-2-5v-7-5v-5v-10v-High-Precision-Voltage-Reference-Module-/300937695293?pt=LH_DefaultDomain_0&hash=item461148d03d This one has a sheet and that is what I looked for when I purchased mine.

One of Daves videos has a homemade resistance box with a single, very accurate resistor in it. You can also get one of those basic DMM checkers. It's just a pcb with a few resistors and such.

Frequency is more of an issue. Good frequency standards are hard to deal with.

Things like this are always handy, if only to verify your gears stability.

 

[HKJ]

http://www.ebay.ca/itm/AD584LH-4-Channel-2-5v-7-5v-5v-10v-High-Precision-Voltage-Reference-Module-/300937695293?pt=LH_DefaultDomain_0&hash=item461148d03d This one has a sheet and that is what I looked for when I purchased mine.

Some of them works well, but not all.
I have a couple of them, some of them matches my Fluke 8846A, but not all.

[DJ]

GPS and trimmed references are Godsends compared to the old days. Do a quick error budget before applying dividers and such.

Two point calibration should suffice,  unless the equipment is horribly non-linear.

[iDevice]

Well, virtues of calibration are highly overrated.
Every common sense statements cited in this thread are true, precisely calibrated or even high precision test instruments are not really needed in the context of a hobby or even in most professional applications.

Coincidently, one of my hobbies as EE is to restore vintage instruments and one of them being a calibrator, I could come to the conclusion that most of these old guys are still in spec. even after sometimes 30 years since last cal.
I'm not sure it will still be true with today's stuff though...

So unless something went wrong, like a dead calibration storage battery or something, I guess your instruments are still accurate enough for any non mission critical purposes.

But I understand the compulsive need of perfect calibration. I do it each time I restore or repair something, it's just fun and I've spent numerous hours trying to achieve the perfect 1.000,000,000 mark on all ranges.
Completely useless, but entertaining 

[jlmoon]

@Cree...  I know you probably know this, but.... don't forget to keep in mind that the instruments you have are only as good as (for time) your timebase reference (accuracy/drift over time), and when measuring volts, current or ohms only as good as your voltage reference within your device.  A good calibration is based on uncertainties of a known standard.  It is so easy to become consumed with accuracy and precision while forgetting about the actual resolution and accuracy (in ppm measure) of the given equipment.. hope this helps.

JLM

[jlmoon]

But I understand the compulsive need of perfect calibration. I do it each time I restore or repair something, it's just fun and I've spent numerous hours trying to achieve the perfect 1.000,000,000 mark on all ranges.
Completely useless, but entertaining   

  hence the inception of the 'volt-nut' and 'time-nut' classes.. there's a little bit of that in all of us! 

JLM

[pickle9000]

http://www.ebay.ca/itm/AD584LH-4-Channel-2-5v-7-5v-5v-10v-High-Precision-Voltage-Reference-Module-/300937695293?pt=LH_DefaultDomain_0&hash=item461148d03d This one has a sheet and that is what I looked for when I purchased mine.

Some of them works well, but not all.
I have a couple of them, some of them matches my Fluke 8846A, but not all.

From a stability standpoint the one I have has been pretty much spot on after about 15 months. I'm retired now so have no need to have things calibrated as I did when contracting but still like to check. That said stability is critical in calibration. As for accuracy that is another argument.

[PaulAm]

You can easily make a time mark generator to check the sweep accuracy on your scopes.  Here's one:
http://www.davidbridgen.com/calbox.htm

Unless you have had to repair something, good quality equipment generally doesn't drift much.  I've had 50 year old scopes still spot on.  Still, it's nice to be sure.

[xwarp]

Not trying to be facetious or anything, but if you don't have anything to validate the cal of your scope, how do you know it's, (or any of your other gear), has drifted?

Like this, but then begs the question of how can you certain of it's accuracy if you have calibrated equipment to verify it's signals?

*****as stated above, I believe in a hobbyist garage, calibration is something that is nice, but are working on things that require that level of accuracy?

You can easily make a time mark generator to check the sweep accuracy on your scopes.  Here's one:
http://www.davidbridgen.com/calbox.htm

Unless you have had to repair something, good quality equipment generally doesn't drift much.  I've had 50 year old scopes still spot on.  Still, it's nice to be sure.

[G0HZU]

As you are in the USA have a look at W2AEW's youtube video for calibrating your counter against WWV.



In terms of checking your DMM I would suggest the most you should spend is about $12 on buying a few $3 voltage reference chips and a few precision resistors.

As for your scope, the best I can suggest is that if you 'know' it is drifted out of spec then why not just tweak it until you 'don't know' if it is accurate or not. Basically, for the vast majority of hobby users I don't think an old analogue scope needs to be formally calibrated.

Sig gens (especially decent sig gens) are a different matter if you want to calibrate them to be close to the original specs. But for most hobbyists the sig gen level accuracy will be good enough as long as the sig gen ALC stays in closed loop and the attenuators haven't been damaged or abused (in which case calibration is pointless anyway)

You can calibrate the sig gen for frequency accuracy  with your counter that you calibrated to WWV. You can then at least check your scope timebase with the counter and sig gen.

When I started electronics we only had analog meters with 3 to 5% accuracy full scale 20kohm/V .
To read the needle you could make another 2% mistake.

In the company we had one very precise 1% electronic VA OHM meter.

Still we could produce nice amplifiers radio's and TV sets.

The first 3.5 digit multimeter was a quantum leap forward in accuracy but the radio's TV's and amps where not getting any better.

You have to ask yourself if you really need the accuracy.

Totally agree!

My advice to people who's hobby is actually centred on calibration (to obsessive levels similar to a calibration house) is to go and find a very, very dark room and sit in there for a while and think about the relevance of their pursuits. Maybe if they stare long enough into the gloom they might see 'common sense' in the darkness and then go and do something useful/creative with their time and test gear instead

[KC0PPH]

My advice to people who's hobby is actually centred on calibration (to obsessive levels similar to a calibration house) is to go and find a very, very dark room and sit in there for a while and think about the relevance of their pursuits. Maybe if they stare long enough into the gloom they might see 'common sense' in the darkness and then go and do something useful/creative with their time and test gear instead

Tried that and drove myself crazy in there too... Now im out doing the same things as before...

TBH i would Love a decent place to send gear off to and get it cal'ed.

[grumpydoc]

Careful - this can lead to be an obsession!

The 'scope is easy because it isn't a precision instrument - you're looking at 2% or so horizontal and vertical for a 2235 to get it within the Tektronix specification.

For the timebase get a 10MHz TCXO from ebay and couple it to a couple of ÷2 and ÷5 counters (74xx90 or equiv) to get you 10,5,2,1MHz -etc  the David Bridgen link takes this a bit further and adds a pulse vs 50% duty output as well.

A decent TCXO should be accurate to 1ppm to 0.1ppm so that's way more precision than you need for the timebase and should do for checking the sig gen master oscillator and the frequency counter.

If you want better than that then a rubidium unit or GPS disciplined oscillator is the way to go for a frequency standard - that will get you into the 0.1ppb (or a little better) territory. Unfortunately prices have gone up a lot recently.

For voltage standards get a DMMCheck - or one of Doug's other reference sources, that will give you some idea of how the 8050A is doing (but might not be enough to adjust it) on at least one voltage and current range and three resistance ranges.

The 2235 has one principle gain pot per channel so you can check the vertical deflection with a DC source such as the DMMCheck.

Or look at the "DMMCheck plus" which has some square wave outputs.

The two pots are towards the front left of the 'scope under the CRT. Don't fiddle with the front end attenuator board pots unless you are feeding in known clean square waves and it looks like the transient response is off (and then have a read of the manual - you need a square wave source with fast edges that can do a couple of frequencies - 1kHz, 10kHz, 50kHz and 1MHz from memory).

You can now check the amplitude accuracy of the sig gen - up to a point.

You don't say what sig gen you have or whether it's a "function generator" or "signal generator". Assuming "function generator" means low frequency (typically max of 50MHz or lower) sine+square+triangle outputs with relatively limited amplitude range then the spec is likely to be 1-5% on amplitude. Also, assuming "signal generator" means a beast intended for RF testing with higher frequency (> 100MHz often >1Ghz eg Marconi 2019, 2022, 2024 etc) but sine only output then the accuracy is more likely to be 1dB (12%) or so.

So, if you check your generator output and it looks to be within 5 or 10% over a range of frequencies chances are your 'scope and gen are close enough. If it's off you won't know whether it's the 'scope or the gen but chances are you'll be OK.

I have no affiliation with voltagestandard.com except that I bought a DMMCheck from them and was perfectly satisfied - my only regret is that living in the UK I can't really take advantage of the free re-checks that they offer.

[David Hess]

The oscilloscope sweep can be calibrated with any frequency counter and signal source.  The fastest sweeps need higher frequencies to do right but usually only the overall horizontal gain needs to be adjusted.

The oscilloscope vertical calibration can be done with a function generator, AC voltmeter, and maybe switched attenuator.  Again, usually just the overall gain has drifted so setting it on one range is sufficient.

Oscilloscope calibrated accuracy is only in the 1% range so the accuracy of the calibration signals is not great.

[KedasProbe]

http://www.ebay.ca/itm/AD584LH-4-Channel-2-5v-7-5v-5v-10v-High-Precision-Voltage-Reference-Module-/300937695293?pt=LH_DefaultDomain_0&hash=item461148d03d This one has a sheet and that is what I looked for when I purchased mine.

Some of them works well, but not all.
I have a couple of them, some of them matches my Fluke 8846A, but not all.

FYI  (everything within specifications, assuming DM3068 or 8846A or Reference box is correct)


edit: a 'funny' thing to notice: the DM3068 has a small offset error while the 8846A has a small scaling error compared to this reference box.

[mamalala]

I'm wondering that the meaning of "calibration" is here? Over here, a calibration of some equipment is "simply" using a known good source, meassure it with the device in question, and record the difference. That is, a calibration does exactly _nothing_ to the device that is to be calibrated.

However, there is also the adjustment. This one does indeed adjust the device so that it's readings agree with the known good reference.

In short, a calibration just tells you how far off the device is compared to a known good source, that difference will be recorded (for all ranges in question), and thus you know when you meassure an exact 8.000000 source, your device reads "8.000034V". When you meassure the same source, and want the device to show that, then you need to adjust it.

Is that different elsewhere? Or have i been misinformed for the past decades when listening to people explaining these differencies to others? Or is that a case of words directly translated having a different meaning?

Greetings,

Chris

[electr_peter]

Useful definitions of calibration steps:

• Calibration (measuring against a standard)
• Verification (checking declared tolerance range)
• Adjustment (adjusting to fit tolerance ranges - manual, automatic or none)

"Calibration" term is usually used in verification (option 2) sense.

[retrolefty]

Useful definitions of calibration steps:

• Calibration (measuring against a standard)
• Verification (checking declared tolerance range)
• Adjustment (adjusting to fit tolerance ranges - manual, automatic or none)

"Calibration" term is usually used in verification (option 2) sense.

I agree but it does seem a little non-intuitive.

In my world the definition would be:
1. Verification = Compare with known standards and note and publish differences.
2. Calibration =  Perform adjustments needed to meet specified accuracy

[David Hess]

Don't fiddle with the front end attenuator board pots unless you are feeding in known clean square waves and it looks like the transient response is off (and then have a read of the manual - you need a square wave source with fast edges that can do a couple of frequencies - 1kHz, 10kHz, 50kHz and 1MHz from memory).

These adjustments are for input compensation and input capacitance and not for transient response.  They can use the same signal source that would be used for compensating a x10 probe.

Three of the transient response adjustments (the service manual refers to them as high-frequency compensation) are in the vertical CRT amplifier and one is in the channel 2 vertical preamplifier for balancing the channels.  There are also a pair of 2mV/div peaking adjustments in the last stages of the input amplifiers.  These all require a fast transition time flat level pulse generator for calibration.

[poida_pie]

get a DMMcheck, cal the dmm and then ebay the DDMcheck. net cost maybe $20?