Forums

Harmonics in pole power transformers?

Started by Joerg April 17, 2016
On Monday, 18 April 2016 16:42:56 UTC+1, Joerg  wrote:

> There is no room. Just that last pole, the transformer, a meter and > breaker box and the pumps. Other than that maybe some cattle.
If the meter is an old electro-mechanical one there will probably be a useful B field leaking from the voltage winding, even when there is no load. Also, the input and output wires are well separated at the meter terminals which could make it easiler to attach metal foil for voltage sensing. John
Joerg <news@analogconsultants.com> wrote:
> On 2016-04-17 15:08, whit3rd wrote: >> On Sunday, April 17, 2016 at 1:54:36 PM UTC-7, Joerg wrote: >>> Folks, >>> >>> A system needs to be able to detect whether a pole transformer has >>> failed and stopped feeding power to a pump. The transformer will be >>> 100ft or more away. The detection has to work even without any power >>> flow meaning we can only detect the presence of a 60Hz electric field >>> that leaks from the various boxs such as a breaker box. >> >> If you could depend on a little electric 'hum', maybe a microphone would >> do. Acoustic output from any inductor usually has a bit of 120 Hz, >> For that matter, maybe you could bypass the switch on the pump, >> with a safety-rated capacitor ad/or resistor, >> so there's always a small (1 mA) current, and do your acoustic sensing near >> the pump motor? >> > > No wiring desired and acoustic is too prone to false alarms. Has to be > very robust.
It's funny everything has to be real good, but hiring an electrician is just out of the question. funny project.
On Sunday, April 17, 2016 at 5:14:08 PM UTC-4, Joerg wrote:

What about using a phase monitor relay?
Here's one at random from Grainger

http://www.grainger.com/product/DAYTON-Phase-Monitor-Relay-31EE08?searchBar=true&searchQuery=31EE08

Check around for the relays that are used with automatic generator transfer switches.  These monitor/detect the utility power (and phase, which is where I got the idea to mention this), so that a running generator won't transfer over to a dead utility supply.

If I recall, there are some models of these relays that require no separate power source.

If I haven't understood your requirement, well then,.. nevermind.
On Mon, 18 Apr 2016 10:43:22 -0700 (PDT), jrwalliker@gmail.com wrote:

>On Monday, 18 April 2016 16:42:56 UTC+1, Joerg wrote: > >> There is no room. Just that last pole, the transformer, a meter and >> breaker box and the pumps. Other than that maybe some cattle. > >If the meter is an old electro-mechanical one there will probably be >a useful B field leaking from the voltage winding, even when >there is no load. > >Also, the input and output wires are well separated at the meter >terminals which could make it easiler to attach metal foil >for voltage sensing. > >John
Old-fashioned (and many new) electric meters leak mag field like crazy. Stick some little pickup thingie to the glass. Try this maybe: http://www.amazon.com/Telephone-Microphone-Suction-Cup-Pickup/dp/B0034I75IK -- John Larkin Highland Technology, Inc picosecond timing precision measurement jlarkin att highlandtechnology dott com http://www.highlandtechnology.com
On 2016-04-18 08:51, Phil Hobbs wrote:
> On 04/18/2016 10:58 AM, Joerg wrote: >> On 2016-04-18 07:57, legg wrote: >>> On Sun, 17 Apr 2016 17:16:12 -0700 (PDT), George Herold >>> <gherold@teachspin.com> wrote: >>> >>>> On Sunday, April 17, 2016 at 4:54:36 PM UTC-4, Joerg wrote: >>>>> Folks, >>>>> >>>>> A system needs to be able to detect whether a pole transformer has >>>>> failed and stopped feeding power to a pump. The transformer will be >>>>> 100ft or more away. The detection has to work even without any power >>>>> flow meaning we can only detect the presence of a 60Hz electric field >>>>> that leaks from the various boxs such as a breaker box. >>>>> >>>>> In order not to get fooled by 60Hz from other overhead lines I was >>>>> wondering if classic pole transformers generate enough 3rd and 5th >>>>> harmonics to distinguish the field emanating from secondary wires from >>>>> that coming from overhead HV lines. Does anyone know where to find info >>>>> about that? >>>>> >>>>> -- >>>>> Regards, Joerg >>>>> >>>>> http://www.analogconsultants.com/ >>>> >>>> Magneto-striction causes the transformer to vibrate at 120 Hz, >>>> I don't know if you could sense that. >>>> >> >> Audio buzz isn't reliable here because the transformer will be far away. >> All we have is any E-field leaking out from conduit and the breaker box. >> The good thing is, we can be right at the breaker box, just not in it. >> >> >>>> George H. >>> >>> Transformers can fail and still buzz. >>> >>> There are many reasons why a pump might fail. Expending effort at >>> sensing just a single one of these is misdirected. >>> >> >> No, it is not. The fact that transformer or HV line failure is by far >> the highest concern in this application is based on clear statistical >> evidence. >> >> >>> Concentrate on the pump, not the power source. >>> >> >> Not possible because that turns on and off automatically. Any attempt to >> log that relationship is expensive and requires an electrician to come >> out for installation and that is not desired. Many other reasons, too. >> > > As John says, you don't have a B field and the E field is shielded out > by the box. >
Our hope is that enough E-field leaks out to be detectable. But ideally we'd like to distinguish it from other 60Hz fields from far away and that may be via harmonics that are hopefully stronger out of a transformer than on the other fields (usually from HV power lines).
> How about gluing the sensor to the inside of the box, the receiver to > the outside, and using two piezo buzzers to communicate? >
That's our plan B, sans the buzzer since low bandwidth RF would go through at that short distance. But it's expensive, might nto be palatable to have two units. -- Regards, Joerg http://www.analogconsultants.com/
On 2016-04-18 10:11, Jeff Liebermann wrote:
> On Mon, 18 Apr 2016 07:52:30 -0700, Joerg <news@analogconsultants.com> > wrote: > >>> POWER SYSTEM HARMONICS. A Reference Guide to Causes, Effects and >>> Corrective Measures. >>> <http://literature.rockwellautomation.com/idc/groups/literature/documents/wp/mvb-wp011_-en-p.pdf> > >> Their cores all saturate a little. I only need to know how much that is >> with pole transformers. > > Offhand, I can see some problems with using harmonics: > 1. Most of the distortion (harmonics) are cause by non-linear loads, > over which you have no control. Worse, they vary. If you can see > transformer induced harmonics under all that junk, I would be > impressed by your DSP programming. > 2. Core saturation is a function of load current, which can vary. I > suspect that the distortion is much less with a light load, than with > a heavy load. If the transformer fails while it is lightly loaded, > you probably won't see a change.
My experience is different. There is always a compromise between idle losses and cost in a transformer and most of the idle losses are due to saturation. Lots of detectable harmonics. But I don't know if it's the same with pole transformers, hence my quesion here in the NG.
> 3. Increased distortion from core saturation is only a function of > perhaps one failure mode, over current.
No, it's always there. This is sometimes why you see US equipment failing in Europe, core saturation (before SMPS came about).
> ... There are plenty of other > failure modes, such as coolant loss, various leakage faults, phase > balancing, space weather, etc that do not necessarily generate a > corresponding increase in distortion. > 4. Grid tied (solar) generators are really noisy and tend to produce > harmonic spikes on the power lines. The better inverters have line > filters to reduce the problem, but most do little or nothing. > 5. HomePlug and various carrier current communications schemes apply > RF onto the line. While this is unlikely to affect low order > harmonics of 60Hz, it still should be considered and tested. > 6. If you put your finger on the probe tip of your oscilloscope, > you'll see a fair approximation of the type of junk found on power > lines. Seeing small changes in harmonic content is going to be > difficult. >
Different out in the boonies. There's only AC motors out there. Or sometimes no load and we must not get an alert upon no-load.
>>> Suggestion: A failing transformer usually like to overheat before >>> boiling the cooling oil and exploding. I would think an IR >>> thermometer and telescope would work to detect overload type failures. >>> However, it won't do anything if the line is simply cut and the power >>> goes away. > >> That's too iffy in terms of reliability. Also, we need to catch HV line >> failures and primary fuse failure as well. > > It's more reliable than trying to detect xformer distortion. If your > purpose is to anticipate a transformer failure, there's nothing better > than temperature. Think about attaching a stick on label that changes > color at some critical temperature: > <https://en.wikipedia.org/wiki/Thermochromism> > Or just slop some thermochromatic paint on the xformer case. >
Not allowed.
> HV line failures and fuse failures will produce an instantaneous loss > in xformer output voltage. I would think you could by just monitoring > the local line voltages and waveforms. > > Back when I was building server farms, we were learning all about > reliable power. PG&E would pay for damaged equipment, but it was my > responsibility to prove that they screwed up. So, we invested in > power line impairment equipment. As I recall, it was made by Dranetz, > but I don't recall the model. It gave us a paper tape printout of > just about anything that might go wrong on the power line. It paid > for itself in about 4 years and inspired PG&E to replace some failing > equipment. I still do some of that at various mountain top radio > sites. > > This is probably overkill for what you need, but comes close to what > you're trying to detect. It might be useful to look at some of this > equipment, see what they can detect, determine how they do it, and see > if it can be used. If you invent a new method of doing all that by > induction or no wire contact, I want to invest in the company. > <http://www.dranetz.com/power-quality-analyzers/> > <http://en-us.fluke.com/products/power-quality-analyzers/> > <https://www.google.com/search?q=power+quality+analyzer&tbm=isch> >
However, we cannot plug in anywhere and there may not be any current. -- Regards, Joerg http://www.analogconsultants.com/
On 2016-04-18 10:43, jrwalliker@gmail.com wrote:
> On Monday, 18 April 2016 16:42:56 UTC+1, Joerg wrote: > >> There is no room. Just that last pole, the transformer, a meter and >> breaker box and the pumps. Other than that maybe some cattle. > > If the meter is an old electro-mechanical one there will probably be > a useful B field leaking from the voltage winding, even when > there is no load. >
Now that is a good idea! As long as there is a meter.
> Also, the input and output wires are well separated at the meter > terminals which could make it easiler to attach metal foil > for voltage sensing. >
All in conduit and inaccessible. -- Regards, Joerg http://www.analogconsultants.com/
On 2016-04-18 08:59, mrdarrett@gmail.com wrote:
> On Sunday, April 17, 2016 at 2:14:08 PM UTC-7, Joerg wrote: >> On 2016-04-17 14:00, John Larkin wrote: >>> On Sun, 17 Apr 2016 13:55:05 -0700, Joerg >>> <news@analogconsultants.com> wrote: >>> >>>> Folks, >>>> >>>> A system needs to be able to detect whether a pole transformer >>>> has failed and stopped feeding power to a pump. The transformer >>>> will be 100ft or more away. The detection has to work even >>>> without any power flow meaning we can only detect the presence >>>> of a 60Hz electric field that leaks from the various boxs such >>>> as a breaker box. >>> >>> If the transformer fails, won't that stop power to the pump? >>> >> >> Yup. The system is a mote network that then needs to send out an >> alarm. The pumps aren't always running and the alert should not go >> out when no power is drawn but mains power is available. > > > If the pumps aren't always running, what triggers the pumps to turn > on and off? >
All kinds of triggers: Float switches, people, timers, et cetera. It's beyond out influence and we can't rely on any of that.
> Would it be acceptable to determine that power has failed when the > command to turn the pumps on fails? >
Nope :-( -- Regards, Joerg http://www.analogconsultants.com/
On 2016-04-18 11:42, mpm wrote:
> On Sunday, April 17, 2016 at 5:14:08 PM UTC-4, Joerg wrote: > > What about using a phase monitor relay? Here's one at random from > Grainger > > http://www.grainger.com/product/DAYTON-Phase-Monitor-Relay-31EE08?searchBar=true&searchQuery=31EE08 > > Check around for the relays that are used with automatic generator > transfer switches. These monitor/detect the utility power (and > phase, which is where I got the idea to mention this), so that a > running generator won't transfer over to a dead utility supply. > > If I recall, there are some models of these relays that require no > separate power source. > > If I haven't understood your requirement, well then,.. nevermind. >
We can't have anything installed beyond sticking a plastic box somewhere in the outside because that would require truck roll (electrician). -- Regards, Joerg http://www.analogconsultants.com/
On 2016-04-18 11:17, Cydrome Leader wrote:
> Joerg <news@analogconsultants.com> wrote: >> On 2016-04-17 15:08, whit3rd wrote: >>> On Sunday, April 17, 2016 at 1:54:36 PM UTC-7, Joerg wrote: >>>> Folks, >>>> >>>> A system needs to be able to detect whether a pole transformer has >>>> failed and stopped feeding power to a pump. The transformer will be >>>> 100ft or more away. The detection has to work even without any power >>>> flow meaning we can only detect the presence of a 60Hz electric field >>>> that leaks from the various boxs such as a breaker box. >>> >>> If you could depend on a little electric 'hum', maybe a microphone would >>> do. Acoustic output from any inductor usually has a bit of 120 Hz, >>> For that matter, maybe you could bypass the switch on the pump, >>> with a safety-rated capacitor ad/or resistor, >>> so there's always a small (1 mA) current, and do your acoustic sensing near >>> the pump motor? >>> >> >> No wiring desired and acoustic is too prone to false alarms. Has to be >> very robust. > > It's funny everything has to be real good, but hiring an electrician is > just out of the question. >
Very normal. Calling out an electrician can mean a 2-3h truck drive.
> funny project. >
They always are but in the end generally work as desired :-) -- Regards, Joerg http://www.analogconsultants.com/