Input reactors upstream to VFD rectifiers

Status
Not open for further replies.

LMAO

Senior Member
Location
Texas
I know line reactors are used to isolate harmonic voltage, reduce harmonic current and ensure proper load sharing between rectifiers (if more than one in parallel used).
But I am also vaguely hearing that they are used to "match transformer impedance" and "suppress voltage surge"... :? ...have you heard these too? Can you elaborate?
 

petersonra

Senior Member
Location
Northern illinois
Occupation
engineer
I know line reactors are used to isolate harmonic voltage, reduce harmonic current and ensure proper load sharing between rectifiers (if more than one in parallel used).
I buy into 1 of them. The other 2, not so much.

But I am also vaguely hearing that they are used to "match transformer impedance" and "suppress voltage surge"... :? ...have you heard these too? Can you elaborate?

I hear all kinds of things. There are as many old wives tales among electricians as among any other group of people.
 

Jraef

Moderator, OTD
Staff member
Location
San Francisco Bay Area, CA, USA
Occupation
Electrical Engineer
The issue of adding impedance is a long story. Hard to explain without my PowerPoint presentation graphics, but I'll give it a shot.

When you have a transient event from the line, such as a grid switch, there is often a ringing effect in which the source voltage drips to or close to zero for a cycle, often a series of them. For many consuming devices, this is irrelevant. But for anything with a rectifier bridge AND a fast switching device behind it, it can present a dangerous situation. We think of diodes as "one-way valve" style conductors, but in reality they do NOT conduct full time even in the one direction. There is what is called a "Forward Conduction Voltage"(FVC) threshold below which they do not conduct. That's why, when you have a diode bridge rectifier, the current drawn through the diodes happens in "gulps" only at the tops of each sine wave coming in, what is collectively called a "non-linear" current draw. This, as a side note, is what causes the harmonic distortion as well. But the issue here is with regard to what happens during a transient event.

If, during that ringing transient, the line voltage sine wave dips below the FCV of the diodes, they cease conducting on that cycle. So effectively, that diode is NOT contributing current to the DC bus during that event. At the same time, if the motor is still running, the transistors are sucking energy OFF of the DC bus and can do so for a few seconds because of what is stored in the capacitors, so the motor never sees the transient. But capacitors discharge and recharge almost instantly, so when one diode in the bridge stops conducting, the capacitors will pull the current to replace themselves from the next one that does conduct, and they will do so at the Available Fault Current level of the circuit, right ahead of the drive.

If your transformer feeding the drive is more than 10X the kVA rating of the VFD, the current that is drawn though the diodes during an event can exceed the current rating of the devices by a factor of 3-10X. As an example, we did a test on a 25HP drive connected behind a 1000kVA transformer at our factory. The drive is rated for 40A, the diodes are rated for 60A, but during a grid switching transient, the diodes conducted 805A for short bursts. So although they survive because it's brief, it's like chipping away at a wall with a pick; eventually you cause enough damage that it causes a cascading failure of the silicon layers inside of the diodes and they conduct full time, then blow the drive. In our experiment, we added a simple 3% line reactor ahead of that drive, and the transient events never exceeded 55A, well within the capacity of the diodes. The reason it works this way is because you are adding inductance to the circuit, thus adding what is called the "inductive time constant" which states that, in an inductive circuit, there is a fixed rate at which a change in current can take place. So by adding that delay in the change in current, both on dropping and rising, the short duration of the transient event has less of an effect on the depletion of the DC bus voltage, so the next diode in the sequence does not need to make up for the loss.

So here's the thing with needing or not needing reactors with regard to impedance: If your transformer is less than 10X the kVA of the drive in question, you probably don't need it. If you have multiple drives, one line reactor sized for all of them provides the exact same benefit as an individual reactor in front of each one. Will your drive survive without the reactor? Yes, but for a shorter time. I call them cheap insurance.
 
Last edited:

LMAO

Senior Member
Location
Texas
The issue of adding impedance is a long story. Hard to explain without my PowerPoint presentation graphics, but I'll give it a shot.

When you have a transient event from the line, such as a grid switch, there is often a ringing effect in which the source voltage drips to or close to zero for a cycle, often a series of them. For many consuming devices, this is irrelevant. But for anything with a rectifier bridge AND a fast switching device behind it, it can present a dangerous situation. We think of diodes as "one-way valve" style conductors, but in reality they do NOT conduct full time even in the one direction. There is what is called a "Forward Conduction Voltage"(FVC) threshold below which they do not conduct. That's why, when you have a diode bridge rectifier, the current drawn through the diodes happens in "gulps" only at the tops of each sine wave coming in, what is collectively called a "non-linear" current draw. This, as a side note, is what causes the harmonic distortion as well. But the issue here is with regard to what happens during a transient event.

If, during that ringing transient, the line voltage sine wave dips below the FCV of the diodes, they cease conducting on that cycle. So effectively, that diode is NOT contributing current to the DC bus during that event. At the same time, if the motor is still running, the transistors are sucking energy OFF of the DC bus and can do so for a few seconds because of what is stored in the capacitors, so the motor never sees the transient. But capacitors discharge and recharge almost instantly, so when one diode in the bridge stops conducting, the capacitors will pull the current to replace themselves from the next one that does conduct, and they will do so at the Available Fault Current level of the circuit, right ahead of the drive.

If your transformer feeding the drive is more than 10X the kVA rating of the VFD, the current that is drawn though the diodes during an event can exceed the current rating of the devices by a factor of 3-10X. As an example, we did a test on a 25HP drive connected behind a 1000kVA transformer at our factory. The drive is rated for 40A, the diodes are rated for 60A, but during a grid switching transient, the diodes conducted 805A for short bursts. So although they survive because it's brief, it's like chipping away at a wall with a pick; eventually you cause enough damage that it causes a cascading failure of the silicon layers inside of the diodes and they conduct full time, then blow the drive. In our experiment, we added a simple 3% line reactor ahead of that drive, and the transient events never exceeded 55A, well within the capacity of the diodes. The reason it works this way is because you are adding inductance to the circuit, thus adding what is called the "inductive time constant" which states that, in an inductive circuit, there is a fixed rate at which a change in current can take place. So by adding that delay in the change in current, both on dropping and rising, the short duration of the transient event has less of an effect on the depletion of the DC bus voltage, so the next diode in the sequence does not need to make up for the loss.

So here's the thing with needing or not needing reactors with regard to impedance: If your transformer is less than 10X the kVA of the drive in question, you probably don't need it. If you have multiple drives, one line reactor sized for all of them provides the exact same benefit as an individual reactor in front of each one. Will your drive survive without the reactor? Yes, but for a shorter time. I call them cheap insurance.

wow, your explanation was so clear. You should consider teaching. You are more understandable than most of my professors back in college.
 

Jraef

Moderator, OTD
Staff member
Location
San Francisco Bay Area, CA, USA
Occupation
Electrical Engineer
wow, your explanation was so clear. You should consider teaching. You are more understandable than most of my professors back in college.
Thanks. Around here you can't teach without at least a Master's Degree, which I don't have. A colleague used to teach at a nearby Wyotech where they have an Electrician program, and he said they are desperately seeking instructors, no Master's necessary. But the reason they are so desperate is because they are now owned by Corinthian Colleges who is on the brink of insolvency. Still I've been considering it, not as a full time job (because I have one) but just to teach maybe one or two classes for the experience and fulfillment I might get out of it.

I already make more money than most teachers I know anyway... :roll:
 

AZElectrical

Member
Location
Arizona
The impedance of line reactors increases linearly with frequency (X = 2*pi*f*L), which is beneficial to preserving the power electronics of the drive. As an example, when utilities switch capacitors into lines, a high frequency transient over-voltage can occur on the nearby system. To this high frequency transient, the line reactor presents a very high impedance and thus has a large voltage drop across it, which prevents the majority of the transient overvoltage from reaching the sensitive drive components.
 

LMAO

Senior Member
Location
Texas
Thanks. Around here you can't teach without at least a Master's Degree, which I don't have. A colleague used to teach at a nearby Wyotech where they have an Electrician program, and he said they are desperately seeking instructors, no Master's necessary. But the reason they are so desperate is because they are now owned by Corinthian Colleges who is on the brink of insolvency. Still I've been considering it, not as a full time job (because I have one) but just to teach maybe one or two classes for the experience and fulfillment I might get out of it.

I already make more money than most teachers I know anyway... :roll:

I have a fill time job too but I always look for something part time on the side as well, maybe occasional consulting, like doing analysis studies (harmonics, load flow, short circuit, coordination, fault, etc...). I am in Houston but someone like you needs all the money you can get, you live in SF!
 

Jraef

Moderator, OTD
Staff member
Location
San Francisco Bay Area, CA, USA
Occupation
Electrical Engineer
I have a fill time job too but I always look for something part time on the side as well, maybe occasional consulting, like doing analysis studies (harmonics, load flow, short circuit, coordination, fault, etc...). I am in Houston but someone like you needs all the money you can get, you live in SF!
I moved out of SF to the Burbs years ago, but it's still expen$ive to live here for sure. A co-worker just transferred to our office and she's single, mid 30s, so she wanted to live in SF proper. She pays $3,800/mo for a 1 bedroom appt., 750 sq. ft. She's renting in the same general neighborhood of SF I used to live in, I paid $500/mo for a place that size. It's nuts.

Lucky for me I bought a house before it went crazy. I can no longer afford my own house, even to rent it!
 
Status
Not open for further replies.
Top