Mark V 125vdc Surge Protection

B

Thread Starter

Barry Thomas

Has anyone put Surge Protectors on the 125VDC Power Supply to the Mark V Processors? If so did it keep the processors from re-booting during voltage surges? We have had this problem a few times during lightning storms. I believe the processors are rebooting because the DC Voltage is exceeding the 90-140 voltage limits of the power supplies.
 
Barry Thomas,

I hope this thread might prompt a bigger discussion on earthing/grounding practices, and on lightning suppression in plants.

Do you have data that indicates the 125 VDC experiences a spike during a lightning strike? Most of the times I've tried to capture this with a digital storage oscilloscope I only see three or four volts DC spike during a strike (when measured positive with respect to negative, not either leg with respect to ground).

I haven't seen anyone try to use surge suppression on the 125 VDC supply to Speedtronic turbine control panels. Most 125 VDC systems used to power Speedtronic turbine control panels consist of an ungrounded battery. and a battery charger, the charger being powered by single-phase 440/480 VAC. That AC supply comes "through" one or two or more step-down transformers from any high voltage bus bar or cable that might have been struck by lightning.

The Speedtronic 125 VDC power is referenced to ground at one point, only for ground detection. The battery charger and the battery actually serve as kind of a filter and surge suppressor for the Speedtronic panel; they are not designed as such but do provide some protection and isolation from many lightning strikes.

The microprocessors and electronics in Speedtronic turbine control panels run on much lower DC voltages, +/-5 VDC, +/-15 VDC, etc. <b>ALL</b> of these voltages <b>ARE</b> grounded (DCOM, PCOM, CCOM, etc.).

It is my personal belief that when a lightning strike occurs it most usually strikes grounded steel (structural steel, pipes, beams, highyard support structures, etc.) and not current-carrying electrical conductors (bus bars/tubes, bare cables, etc.). I don't have any actionable data or maths to support my position, but I do have a lot of personal experience with Speedtronic control systems, and with trying to help plants with lightning-induced problems.

Further it is my personal belief that elevated ground potentials caused by lightning strikes are what causes most problems with Speedtronic turbine control panels, and that it's the lower-voltage DC supplies that are directly grounded that suffer most from elevated ground potential during lightning strikes--not surges from the 125 VDC power supply.

Unfortunately, if I'm right, there's not much that can be easily or inexpensively done to remedy the situation. Poor earthing/grounding grid design or construction is something that requires a lot of review and money to overcome. It's similar to the routing of high- and low-voltage cables and wires too close to each other during construction--which can cause many problems for electronic control systems.

It is, again, my personal belief that it's not problems with the 125 VDC supply as much as it is the elevated ground potential wreaking havoc with the lower-voltage DC power supplies in the Speedtronic turbine control panel. And, I also personally believe that good earthing/grounding grid design and construction practices are also a large contributor to fewer problems with lightning-induced problems for any electronic control system, including Speedtronic turbine control systems.

Hopefully you will provide your experience, and any data you have, and others will also contribute based on their knowledge and experience.
 
I'd recommend hooking up a high speed digital recorder to monitor each DC power supply bus and the voltage to ground on either leg of each DC bus. The sample rate should be at least 10K samples per second. You would really be surprised at what can go on behind your back without knowing it. There are some really nice recorders that can record 16 channels simultaneously at up to 200K per second. You should be able to rent one if you can't afford the ~15 to 25K to buy.

Sudden or intermittent DC grounds suddenly change the voltage of the entire DC system when referenced to ground. Due to capacitance, this can cause very short voltage spikes on the system. The bigger the DC system, the more total capacitance to ground. If the turbine control DC system is also common with lots of other equipment, there is more chance for problems. I've seen short duration DC grounds that occur when a switch contact opens caused by a portion of the arc going to ground.

Cable shields are usually grounded at the turbine control cabinet, and all shields are referenced to one dedicated ground cable to the ground grid. If this one ground reference is lifted, all shields and other ground systems should show open to ground.

This might be the cause of your problem, but it is a common problem. Good luck.
 
daveam,

Very good recommendations.

Shield grounding is another of those construction practice problems that can be quite time-consuming to rectify and can cause various problems for decades.

Do you believe installing some kind of surge protection on the DC power supply to the turbine control panel?

Most Mark Vs are supplied with a <CPF>, Conditioning Power Filter. They are usually well very concealed (hidden) and the middle of the bottom of the back of the control panel enclosure--behind that rats'-nest of cables and wires connected to <QD1>. I have seen lightning strikes cause components of the <CPF> to fail, which probably would cause future strikes to cause even more problems under the right conditions.

These "filters" aren't really there to filter the input power; they are primarily in place to limit electrical noise generated by microprocessors and electronic components IN the Mark V from affecting devices OUTSIDE the Mark V. They do provide MINIMAL protection from incoming electrical noise, but it is minimal at best. Again, they're not there to filter incoming power, but to try to prevent noise from getting "out" of the panel and adversely affecting other equipment that might be powered by the DC supply.

Another particularly <b>BAD</b> source of spikes and problems for Mark V (and Mark VI) control panels are the AC-DC converters supplied with many Mark Vs (and Mark VIs), called <DACA>s. These things are used to convert AC (usually from some kind of UPS or inverter) to 125 VDC, ostensibly to power the panel in the event of loss of the 125 VDC battery. (Some steam turbine Mark V and Mark VI control panels are powered exclusively by <DACA>s since 125 VDC was not a common control system power supply voltage for many steam turbines.)

"Bad" or "noisy" UPS or inverter output can actually be amplified by the <DACA> and make any spikes even worse, as seen by the DC side of the Speedtronic. And, if the UPS or inverters switch between mains and battery sources, sometimes their outputs can be really bad.

While the Speedtronic does a "high-select" of DC inputs (from the battery or the <DACA>), if the <DACA> output has a spike that exceeds the battery voltage then the panel will see that spike and react accordingly. So, even if under normal conditions the <DACA> output is slightly lower than the battery voltage under transient conditions the <DACA> output can exceed the battery voltage for a brief time and play havoc with the panel.
 
> It is my personal belief that when a lightning strike occurs it most usually strikes grounded steel (structural
> steel, pipes, beams, highyard support structures, etc.) and not current-carrying electrical conductors
> (bus bars/tubes, bare cables, etc.). I don't have any actionable data or maths to support my position, but I do have a
> lot of personal experience with Speedtronic control systems, and with trying to help plants with lightning-induced problems.

Concepts were demonstrated over 100 years ago. Further confirmed by Westinghouse and GE research atop what was a new building - the Empire State Building.

Protection is always about where current flows. Voltage is only a symptom of that current path. Surges are from a current source; not from a voltage source.

Protection was always about connecting to earth BEFORE current can enter a building. Once inside, that current will hunt for paths to earth destructively via electronics.

Do not earth equipment. Earth the surge. Equipment must be connected to a safety ground; not earth ground. Those two grounds (even if interconnected) are electrically different. Especially to a surge current.

All electronics already contian significant protection that makes most 'surges' irrelevant. Your concern is the rare transient (maybe once every seven years) that can overwhelm robust protection in electronics.

Either that current is earthed harmlessly outside a building. Or it will find earth destructiuvely via electronics.

Telcos connect a $multi-million computer to wires all over town. Suffer about 100 surges with each storm; without damage. Wire inside every incoming cable connects to earth directly. Or makes a similar same low impedance (ie 'less than 10 foot') connection to earth via a protector. Only purpose of effective protector: connect that current to what does protection - single point earth ground. All four words are electrically significant.

Telcos want separation up to 50 meters between electronics and an earthed protector. That increases protection. Again, protection is about the current path to earth. Separation (increases wire impedance) increases protection.

A minimal 'whole house' protector is 50,000 amps. A typical lightning strike is 20,000 amps. Protection means even a protector remains functional. And every incoming wire connects "short" to earth. If AC has three wires, then one connects directly to earth. Other two make a same 'low impedance' connection via a protector. Protection is about how that current gets to earth.

A protector is simple science. The art is about what actually does protection - single point earth ground. Where hundreds of thoussand of joules harmlessly dissipate. Effective protection is mostly about how each incoming wire is earthed. Protectors are a simple tool for accomplishing what is most important.
 
Dear CSA,

Regarding your question as to whether turbine control DC power supply systems should have surge protection:

I'm definitely not an expert on electrical systems or lightning surges. I can offer the following comments.

* The DC systems I am familiar with are floating relative to ground, so all conductors and equipment should be insulated from any lightning surge current. If the lightning surge current in a building has a low impedance path to ground, the voltage shouldn't build up high enough to bust through any insulation and get into the DC system.

* The DC floating DC systems should have ground detection capable of detecting a path to ground on either leg of at least less than 1 KOhm, or possibly as high as 5 or 10KOhm. The more sensitive -the more likely nuisance alarms will be due to dust, fuzz, etc. Water from a storm could cause a low impedance path into a DC system if equipment becomes wet. The ground detection circuit normally causes the positive bus and negative busses to run equal voltages away from ground. * If you want to prove the electrical integrity of all the cables and field equipment to resist voltage breakdown during a lightning strike you could VERY CAREFULLY isolate the appropriate equipment, install appropriate jumpers, and megger individual cables and components to ground. However, you need to know what you are doing to avoid equipment damage. I have never felt the need to do this on a large scale, but if you suffer equipment damage on floating DC systems during lightning strikes, you could consider this.

* The proper grounding of structural steel, control cabinets, electrical systems, shields etc is extremely important in avoiding nuisance problems. Grounding problems can cause so many other problems that can life miserable.

A high speed recorder can help you assess what kind of unwanted noise or surges you have on all your electrical systems. You will often be shocked at what you see. After you know what is there, you can decide whether to worry about it and what to do. The recorder needs to have fully isolated channels relative to each other and to ground. The impedance between the plus and minus should be very high - 1 MOhm is typical. A good 16 channel high speed recorder can pay for itself many times.

Another thing I like to do when hunting glitches is lightly wiggle on cables, connectors, tap on circuit boards, etc when the unit is in a safe condition and look to look for bad connections. Thermography can be used to search for hot components and bad connections.

It is also not a bad idea to have an electrical check all the terminal for tight connections every 6 years or something like that. You can spend days troubleshooting a problem, and it may take an electrician a few hours to check every electrical connection in the system.
 
Top