Meters suitable for permanent connection to VFD DC bus

[winnie]

I need some help selecting a meter for permanent connection to a VFD DC bus, for improving our lab instrumentation.

I am looking for the sort of accuracy expected from a 5.5 digit multimeter, and will be measuring up to 700VDC. I also need computer output at a rate of 100 readings per second or greater.

I have been able to find a number of meters which meet all of my specifications...except that the 'CAT' rating is CAT I-1000V.

The VFD bus is _not_ power limited but that big capacitor will act to substantially eliminate transient over-voltages.

Can anyone point me to an instrument with a CAT-III rating for 700V DC, or can someone point me to a document which identifies a CAT-I instrument as suitable for the DC bus of a VFD, or to a standard protection technique that would make a CAT-I instrument suitable without degrading meter accuracy.

Thanks
Jon

 

[Jraef]

Are you wanting this external meter in order to cross-check what the VFD says the DC bus level is at? Any meter you get is not likely to be as accurate. And by the way, 700VDC is not going to be good enough, many 480V drives can get to 850VDC before tripping on Over Voltage, so your meter would need to tolerate at least that much.

By the way, typically on a 1000VDC meter, you use an external transducer anyway, so at that point, any meter will do.

Or are you using a VFD that doesn't have that function? Might be easier to just change VFDs...

 

[mpoulton]

If this is really permanent then it's not so much a test instrument as it is a circuit component. Category ratings are probably not applicable. You're looking more for a UL-listed or recognized device to go into your panel assembly. Since you need a digital data stream output, I'd probably consider PLCs with A/D conversion inputs or other process control devices which you can connect to a separate display and configure to send data in whatever format you need. No matter the solution, you may need to scale the voltage with an external divider to match your device. If you didn't need the digital output then there are UL-recognized panel meters available that will do this easily. Some of them even have 4-20mA outputs which you might be able to use instead of digital. Here's one that I think meets the specs except it's a 4-digit display (which, realistically, should be plenty - you don't need to know the DC link voltage to 10mV precision). https://www.predig.com/pd765?field_...402&field_transmitter_power_tid_selective=355

 

[winnie]

You are correct, we don't need 10mV...more like 200mV. What is more important is having something that can be calibrated by a service on a yearly basis, so something that I kit together is less desirable than a stand-alone box that can simply be sent out.

The entire VFD will be part of the test instrument; we need the input DC power known to about 1/2%, tighter is preferable if affordable.

Our existing instrumentation uses an external resistive divider and an isolation amplifier. The accuracy ratings of the isolation amp are much too loose; so we could go down the path of fixing that part of the system, but we would then have separate components kitted together.

Thanks for the suggestion of Precision Digital; I don't think they have the reading rate but it looks like they have the isolation built in, so a simple resistor network would work for scaling.

-Jon

 

[mpoulton]

You are correct, we don't need 10mV...more like 200mV. What is more important is having something that can be calibrated by a service on a yearly basis, so something that I kit together is less desirable than a stand-alone box that can simply be sent out.

The entire VFD will be part of the test instrument; we need the input DC power known to about 1/2%, tighter is preferable if affordable.

Our existing instrumentation uses an external resistive divider and an isolation amplifier. The accuracy ratings of the isolation amp are much too loose; so we could go down the path of fixing that part of the system, but we would then have separate components kitted together.

Thanks for the suggestion of Precision Digital; I don't think they have the reading rate but it looks like they have the isolation built in, so a simple resistor network would work for scaling.

-Jon

Unless the DC bus is extremely well filtered and smoothed, measuring the power through it to 1/2% accuracy will require more than just a voltage reading on the bus and an independent current value from the VFD (I assume that's where you get the current reading from). The bus voltage and current waveforms are complicated and nonlinear. To accurately calculate power, you need voltage and current readings in phase with each other. Since voltage and current are multiplied to give power, the fractional uncertainties of the measurements add, so you need voltage and current measurements to 0.25% accuracy to achieve 0.5% accuracy for power. This will require a current sensor and a voltage sensor both being sampled by the same device, at a sampling rate significantly higher than the frequency characteristics of the waveform. With a 3-phase rectifier, you have 360 pulses per second on the DC bus. The VFD output waveform may be desynchronized from that, causing a DC bus current/voltage waveform with its fundamental frequency around 720Hz. That suggests a sample rate of several thousand per second to capture the waveform and calculate the RMS power accurately. The 0.25% accuracy requirement is not difficult for an A/D converter. Dataq and other instrumentation suppliers have off-the-shelf solutions for this sort of thing. A medium-speed multichannel A/D converter with isolated input amplifier modules will do this for a few hundred bucks. You still need a calibrated temperature-stable voltage divider to scale the high voltage down to something reasonable, plus a current shunt placed between the rectifier and the VFD's bus.