High Voltage DC Bus feedback signal. 3

[bobo480]

What methods are employed for measuring a 680 VDC bus to provide a 0-4 VDC feedback signal to a microcontroller ADC. The DC bus consists of 2- 450 VDC capacitors connected in series. I had initially considered a voltage divider measuring from the capacitor center tap to positive, but this would require bonding the center tap to GND. The HV DC bus normally floats, would it be acceptable to ground the center tap of the DC bus to GND and likewise the GND reference of the ADC?

 

[Noway2]

Since you are working with DC you will likely need to use a resistive divider at the DC bus.

There are ADC converters and similar devices, that will sense signals and communicate the result digitally over an isolated link. This way you do not need to have your logic common tied to the power system.

 

[ScottyUK]

If you want something off the shelf, LEM have a few offerings. Here's a couple:

http://www.lem.com/docs/products/cv 3-1000.pdf

http://www.lem.com/docs/products/lv 25-1000 e.pdf

For the benefit of anyone looking in future: the range goes up to about 6400V AC, 9600V DC.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[electricuwe]

A properly designed differential amplifier will allow to measure the signal without creating a low impedance GND connection that would need a lot of additional concerns.

If you fulfill certain design requirements you can even meet strict safety requirements of IEC standards.

 

[OperaHouse]

Ask yourself, what do you really need to know. An opto isolator fed off a resistive voltage divider may give you enough information. Adding some zeners in series will expand up the region you are looking at. The micro controller can convert it to what you want. In this application I can hardly see the need for more than 4 bits of information.

 

[sreid]

Consider Avago (spun off from Hewlett Packard) Miniature Isolation Amplifiers.

http://www.avagotech.com/pages/en/o...mplifier/current_sensing_isolation_amplifier/

Don't forget that resistors have voltage ratings. Use several in series for a HV voltage divider.

 

[bobo480]

The suggestion of using the LEM Voltage transducer was implemented. Now we are experiencing inverter trips on DC bus overvoltage faults. RC filters have not remedied the noise issue due to switching of the IGBTs.

Nominal DC bus voltage is 670 Volts, V-Transducer output 3.5 Vdc @ 670 Vin. Switching transient appears as 7 VDC peak lasting .4 ms

Low ESR DC Bus capacitors and low inductance DC bus design was employed.

Suggestions are welcomed

 

[ScottyUK]

By low ESR bus capacitors do you mean low ESR electrolytics? Have you got any high frequency bypass caps on the bus, e.g. a polypropylene film type of a few hundred nF or maybe a few uF? The relatively high ESL of the electrolytics would account for what you are seeing. A few manufacturers offer low inductance film types designed for pulse applications but which would serve well in this function - Icar in Italy have a decent range for example.


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[bobo480]

Yes, the DC Bus capacitors are of the Low ESR electrolytic type. There are 2.0 uF film caps across C-E at each IGBT half bridge, and also across the input rectifier. The design requires 6800 uF total bus capacitance, switching to film type (pulse) capacitors would require considerable space and added cost.

 

[ScottyUK]

I didn't mean change the whole bank to film type, just more-or-less what you've described. What is the load? Can you slow the switching edge fractionally?


----------------------------------

If we learn from our mistakes I'm getting a great education!

 

[bobo480]

Currently the inverter has not been subjected to loads other than magnetizing & windage for a 10 HP 4 pole motor. The initial inrush has not been measured yet.

I will look into increasing the gate drive resistor value to provide softer switching.

Thank you for the feedback.

Best regards,

Bosko

 

[itsmoked]

He is not talking about inrush. He's talking about the actual inductive kick spikes caused by switching an inductive load.

The faster the switch the larger the kicks. A VFD requires balancing the heating of the IGBTs, the efficiency, and the EMI/kick peaks by controlling the drive to the IGBTs to control the current thru the IGBTs.

Keith Cress
kcress -

http://www.flaminsystems.com

 

[sreid]

Specifically, what LEM Model Number Voltage Isolator are you using? The important spec is the Common Mode Rejection Ratio [CMMR]and spec'd for a fast dv/dt]. Inductive kicks on the B+ can be large but should be of short duration [i.e., easilly filtered].

 

[bobo480]

My apologies, the voltage transducer is a Flexcore VTU-012X5, I did not see a CMMR value listed.

The Positive 7 vdc pulse is lasting approximately .4 mS with 4.47 vdc being the trip setpoint. RC filtering at the controller signal input has not improved the situation.