Author Topic: AC Voltage Calibration Source - Not Super Accurate/Expensive  (Read 2533 times)

0 Members and 1 Guest are viewing this topic.

Offline EPAIIITopic starter

  • Super Contributor
  • ***
  • Posts: 1165
  • Country: us
AC Voltage Calibration Source - Not Super Accurate/Expensive
« on: December 23, 2022, 08:34:30 am »
What would you use as a sine wave, AC Voltage source for a rough check of meters? I am sure there are expensive sources that probably have an accuracy of a very small fraction of a percent, but they probably cost a large fortune. I am asking about something perhaps between 1% and 0.1% and with an affordable price (well under $1,000 USD).

It should at least cover AC power frequencies and Voltage levels. I would assume that precision resistors could be used to divide it's output down to lower Voltages.
Paul A.  -   SE Texas
And if you look REAL close at an analog signal,
You will find that it has discrete steps.
 

Online BrianHG

  • Super Contributor
  • ***
  • Posts: 8275
  • Country: ca
    • LinkedIn
Re: AC Voltage Calibration Source - Not Super Accurate/Expensive
« Reply #1 on: December 23, 2022, 09:17:44 am »
Have you looked into precision function generators.  Though, this will not be a proper high voltage reference, but with the ~15v ac range, you may end up with a multipurpose device.  You should be able to achieve better than 1% precision with a DAC synthesized output unit.

Asking users in the TEA or Test Equipment, or Metrology threads would probably give you more authentic hardware choices to choose from.
« Last Edit: December 23, 2022, 09:21:29 am by BrianHG »
 

Offline AVGresponding

  • Super Contributor
  • ***
  • Posts: 4931
  • Country: england
  • Exploring Rabbit Holes Since The 1970s
Re: AC Voltage Calibration Source - Not Super Accurate/Expensive
« Reply #2 on: December 23, 2022, 11:40:59 am »
If you're patient and have a good spread of search strings on the likes of eBay, and check regularly, you can find something nice and/or useful eventually.
I scored a fully working (with only minor cosmetic issues) Boonton RFL (may also be branded Clarke Hess) 828 AC/DC V/I source for (icr exactly) somewhere around $100. It has 1-1kV, 1mA-10A (5A max) ranges with 4-decade adjustment, and does 50, 60, and 400Hz in AC.
The claimed accuracy is +/-0.05% of setting for voltage plus 0.05% of range for DC and 0.1% of range for AC. For current it claims +/-0.1% of setting plus 0.1% of range for DC and AC.

Since it's a US manufacturer, I would expect it to be easier and cheaper to find on your side of the pond.
nuqDaq yuch Dapol?
Addiction count: Agilent-AVO-BlackStar-Brymen-Chauvin Arnoux-Fluke-GenRad-Hameg-HP-Keithley-IsoTech-Mastech-Megger-Metrix-Micronta-Racal-RFL-Siglent-Solartron-Tektronix-Thurlby-Time Electronics-TTi-UniT
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 17428
  • Country: us
  • DavidH
Re: AC Voltage Calibration Source - Not Super Accurate/Expensive
« Reply #3 on: December 23, 2022, 02:12:24 pm »
For low voltage calibration, I chop a precision DC source into a square wave with exactly 50% duty cycle which can be used for AC calibration with RMS and average responding values which can be calculated.

However a sine wave source is required for calibrating the AC frequency response, but it only needs to be leveled because once the frequency response calibration is done, the precision square wave can be used to calibrate level.
 

Offline mawyatt

  • Super Contributor
  • ***
  • Posts: 4119
  • Country: us
Re: AC Voltage Calibration Source - Not Super Accurate/Expensive
« Reply #4 on: December 23, 2022, 04:47:30 pm »
Awhile back we developed a DMM & Scope Device just to "check" our various instruments.

https://www.eevblog.com/forum/testgear/ac-rms-dmm-tests/

Note the last couple posts where we had to replace the Voltage Reference which died for some unknown reason. Likely we shorted one of the exposed banana plugs to something, as we are not exactly "gentle" and it gets tossed around the lab, but do try and keep it powered 24/7.

This device has been used to "verify"  handhelds Fluke 77 and 87, U1233A & UT210E. Also used for lab bench 6 1/2 digit DMMs HP34401A, AG34401A, SDM3065X, three KS34465A (light and 2 dark) and a Keithley DMM6500 as well as our DSOs.

We were curious as to how the old analog IC RMS converters behaved verses the Digital RMS Computational Methods in the KS34465A and DMM6500. So comparing the 34401A vs. the 34465A and DMM6500 reveled the analog IC RMS techniques held up very well indeed :-+

Of course YMMV.



Edit: Just did quick "Turn on and Read" each instrument, no proper warm up and the lab temp is ~24C and humidity ~58% (normally ~26C & 45%). OK need to warm up the lab it's cold here (~15C), so justifies turning on all these instruments :-+

Handhelds:
35 year old Fluke 87    2.502VAC (never calibrated)
KS U1233A                 2.507VAC
UT210E                      2.45VAC

Bench: Analog RMS Converters
HP34401A    2.50005VAC
AG34401A    2.49978VAC
SDM6500     2.49723VAC

Bench: Computational RMS
DMM6500      2.499945VAC
KS34465A-1  2.5000405VAC
KS34465A-2  2.4998607VAC
KS34465A-3  2.4999653VAC

So think one can surmise that this square-wave technique described in the link above is acceptable to quickly verify typical DMMs used in repair and around the lab with AC RMS, and are working properly, or not!!

Read is NOT a claim to ultimate precision, altho one would have difficulty arguing against this technique and results, but simply an easy, low cost, simple, means to VERIFY a DMMs proper AC operation within the constraints of the squarewave stimulus.

Best & Happy Holidays,
« Last Edit: December 23, 2022, 06:19:12 pm by mawyatt »
Curiosity killed the cat, also depleted my wallet!
~Wyatt Labs by Mike~
 

Online NiHaoMike

  • Super Contributor
  • ***
  • Posts: 9321
  • Country: us
  • "Don't turn it on - Take it apart!"
    • Facebook Page
Re: AC Voltage Calibration Source - Not Super Accurate/Expensive
« Reply #5 on: December 23, 2022, 06:29:36 pm »
Cheap audio amplifier module running into a backwards connected mains transformer, use a known good multimeter for the reference reading.
Cryptocurrency has taught me to love math and at the same time be baffled by it.

Cryptocurrency lesson 0: Altcoins and Bitcoin are not the same thing.
 
The following users thanked this post: Someone

Online vk6zgo

  • Super Contributor
  • ***
  • Posts: 7857
  • Country: au
Re: AC Voltage Calibration Source - Not Super Accurate/Expensive
« Reply #6 on: December 24, 2022, 12:04:42 am »
For low voltage calibration, I chop a precision DC source into a square wave with exactly 50% duty cycle which can be used for AC calibration with RMS and average responding values which can be calculated.

However a sine wave source is required for calibrating the AC frequency response, but it only needs to be leveled because once the frequency response calibration is done, the precision square wave can be used to calibrate level.

Hate to be "that guy", but, by definition, a "square wave" is one with a 50% duty cycle.
Any other duty cycle, & it becomes a "rectangular wave".
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 17428
  • Country: us
  • DavidH
Re: AC Voltage Calibration Source - Not Super Accurate/Expensive
« Reply #7 on: December 24, 2022, 02:45:32 am »
For low voltage calibration, I chop a precision DC source into a square wave with exactly 50% duty cycle which can be used for AC calibration with RMS and average responding values which can be calculated.

However a sine wave source is required for calibrating the AC frequency response, but it only needs to be leveled because once the frequency response calibration is done, the precision square wave can be used to calibrate level.

Hate to be "that guy", but, by definition, a "square wave" is one with a 50% duty cycle.
Any other duty cycle, & it becomes a "rectangular wave".

I phrased it the way I did because many square wave generation methods do not result in exactly a 50% duty cycle, which is required for good accuracy in this case.  Even if a toggle flip-flop is used to divide by two and provide a precision 50% duty cycle to the chopper, the chopper may have asymmetrical switching times, so some care is required to get a precision square wave.
 

Online vk6zgo

  • Super Contributor
  • ***
  • Posts: 7857
  • Country: au
Re: AC Voltage Calibration Source - Not Super Accurate/Expensive
« Reply #8 on: December 24, 2022, 03:14:58 am »
For low voltage calibration, I chop a precision DC source into a square wave with exactly 50% duty cycle which can be used for AC calibration with RMS and average responding values which can be calculated.

However a sine wave source is required for calibrating the AC frequency response, but it only needs to be leveled because once the frequency response calibration is done, the precision square wave can be used to calibrate level.

Hate to be "that guy", but, by definition, a "square wave" is one with a 50% duty cycle.
Any other duty cycle, & it becomes a "rectangular wave".

I phrased it the way I did because many square wave generation methods do not result in exactly a 50% duty cycle, which is required for good accuracy in this case.  Even if a toggle flip-flop is used to divide by two and provide a precision 50% duty cycle to the chopper, the chopper may have asymmetrical switching times, so some care is required to get a precision square wave.

Sorry, David, I was being a bit of a "smartarse", you obviously know the difference, but there is a depressingly large number of people who don't!
 

Offline mawyatt

  • Super Contributor
  • ***
  • Posts: 4119
  • Country: us
Re: AC Voltage Calibration Source - Not Super Accurate/Expensive
« Reply #9 on: December 24, 2022, 03:54:09 am »
One idea we've been bouncing around for some time now is to use both the Q and Qbar from the CMOS FF and a CMOS driver to create an "H Bridge" configuration. This will produce the same RMS Voltage as the Reference VDD voltage, but with a DC average of zero. We would have done this way back when we developed the original PCB but wanted something that could check our DSOs as well as DMMs, and DSOs don't have "floating" inputs, so decided on just single ended as shown.

Believe a fast FF like a 74AC74 driving a fast CMOS inverter like a 74AC04 with the outputs paralleled, or resistively summed to create a given Z out, like 50 ohms differential with 3 inverters with 75 ohms series R on FF "Q" and the other 3 inverters with 75 ohms on "Qbar". Use the extra FF D Input to load the used FF "Q" output to equalize the loading on "Q" and "Qbar". VDD for the FF could be from a 5 volt 78L05 or something, and the VDD for the 74AC04 is from the Precision Voltage Reference.

Anyway, something to think about.

Best and Happy Holidays,
Curiosity killed the cat, also depleted my wallet!
~Wyatt Labs by Mike~
 

Offline EPAIIITopic starter

  • Super Contributor
  • ***
  • Posts: 1165
  • Country: us
Re: AC Voltage Calibration Source - Not Super Accurate/Expensive
« Reply #10 on: December 28, 2022, 10:31:16 am »
Thanks to all.

Yes I am watching E-Bay. I looked at that Boonton RFL 828. It looks like it would do nicely. I set up an Ebay watch for them.

Adjustable source and known good meter. Yes, I thought of that.

Paul A.  -   SE Texas
And if you look REAL close at an analog signal,
You will find that it has discrete steps.
 

Online Kleinstein

  • Super Contributor
  • ***
  • Posts: 15156
  • Country: de
Re: AC Voltage Calibration Source - Not Super Accurate/Expensive
« Reply #11 on: December 28, 2022, 06:27:24 pm »
One idea we've been bouncing around for some time now is to use both the Q and Qbar from the CMOS FF and a CMOS driver to create an "H Bridge" configuration. This will produce the same RMS Voltage as the Reference VDD voltage, but with a DC average of zero. We would have done this way back when we developed the original PCB but wanted something that could check our DSOs as well as DMMs, and DSOs don't have "floating" inputs, so decided on just single ended as shown.

Believe a fast FF like a 74AC74 driving a fast CMOS inverter like a 74AC04 with the outputs paralleled, or resistively summed to create a given Z out, like 50 ohms differential with 3 inverters with 75 ohms series R on FF "Q" and the other 3 inverters with 75 ohms on "Qbar". Use the extra FF D Input to load the used FF "Q" output to equalize the loading on "Q" and "Qbar". VDD for the FF could be from a 5 volt 78L05 or something, and the VDD for the 74AC04 is from the Precision Voltage Reference.

Anyway, something to think about.

Best and Happy Holidays,
The bridge circuit would create a higher output voltage, but also no langer has a well defined ground or lots of common mode signal.  This is inviting problems with EMI and shielding effects.
For most pasts it is not about a near perfect square wave, but more a near sine wave that can be varried over a large range. What voltages are acutally needed depends on the meters.
It would be nice to also have one higher votlage, like 200 V, at least with a low frequency.

The square wave could be used for a crude AC/DC transfer at low voltage (e.g. 5 V DC), but there are still limitations: A meter working well with a square wave may still have a bad RMS circuit (e.g. averaging cap dead) and failing on the square wave could be just an effect of limited slew rate with a perfectly fine, though limited BW meter.
The square wave could still be usefull in combination with a known (e.g. calculated) filter to slow down the steep slopes and limit the BW.

It still makes sense to check the response at a low (e.g. 50 Hz) and higher frequency (e.g. 10 kHz range if the meter supports this), as the AC divider usually have a resistive and capacitive part to get in theory independent gain for the lower and higher frequency part. Cheaper meters may not support 10 KHz with higher voltage - so one may not need this somewhat more tricky point.
 

Offline mawyatt

  • Super Contributor
  • ***
  • Posts: 4119
  • Country: us
Re: AC Voltage Calibration Source - Not Super Accurate/Expensive
« Reply #12 on: December 28, 2022, 07:08:44 pm »

The bridge circuit would create a higher output voltage, but also no langer has a well defined ground or lots of common mode signal.  This is inviting problems with EMI and shielding effects.
For most pasts it is not about a near perfect square wave, but more a near sine wave that can be varried over a large range. What voltages are acutally needed depends on the meters.
It would be nice to also have one higher votlage, like 200 V, at least with a low frequency.

The square wave could be used for a crude AC/DC transfer at low voltage (e.g. 5 V DC), but there are still limitations: A meter working well with a square wave may still have a bad RMS circuit (e.g. averaging cap dead) and failing on the square wave could be just an effect of limited slew rate with a perfectly fine, though limited BW meter.
The square wave could still be usefull in combination with a known (e.g. calculated) filter to slow down the steep slopes and limit the BW.

It still makes sense to check the response at a low (e.g. 50 Hz) and higher frequency (e.g. 10 kHz range if the meter supports this), as the AC divider usually have a resistive and capacitive part to get in theory independent gain for the lower and higher frequency part. Cheaper meters may not support 10 KHz with higher voltage - so one may not need this somewhat more tricky point.

Yeah the bridge would double the RMS value and better "fit" within most DMMs we have which have a 10V range, but is a bipolar waveform with an average of zero.

The fundamental problem with a "filter" is that it's response needs to be known better than the measurement. We prefer to use as close to an ideal squarewave as possible, with a precise known peak voltage and keep the frequency low enough so the DUT meter can have a good chance of representing the True RMS value, as most of the waveform harmonics will fall well within the DUT bandwidth. Of course if we could have a filter of precise known response and/or also a precise known frequency response of the DUT DMM then we could use just about any waveform, of which the sine wave is ideal since it has no harmonics.

Again, the benefit of the squarewave is the ease of creation, ease of establishing a precise amplitude, and of course the simple relationships between peak, rms and average values. If we were still teaching we would get a grad student to formulate the relationship between RMS value and a "squarewave" with unequal exponential RC rise/fall times, the more simple solution is with a linear rise & fall.

Of course YMMV.

Best,
« Last Edit: December 28, 2022, 07:18:30 pm by mawyatt »
Curiosity killed the cat, also depleted my wallet!
~Wyatt Labs by Mike~
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf