Hyundai dhy12500se Generator Issues with 1000w MH Lights

[Dan121]

Hi, I was wondering if anyone could pass on some much needed help.

The local rugby club have just built a floodlit area to use for training on. It consists of 3 12m columns with 1 Philips 1000w metal halide floodlight on each column with running gear etc with in the columns.

The power comes from the above diesel 12.5kva 230v 50hz generator. When testing today all appeared fine but after about 2 minutes the generator shut down with a high voltage alarm which is set at 250v.

We fired it all back up again to monitor what was going on. starting current 18 amps after 1 minute this reduced to around 12 amps I also noticed the voltage dropped to 225v over the next minute I noticed the volts steadily increase all the way to 250v and tripping out the generator with the high voltage alarm setting.

I initially suspected the generator regulator was at fault but I tested it against 3 500w led lights and it works perfectly fine. I also tried each metal halide floodlight individually on the generator but the same outcome happened only difference was the increase in voltage happened a lot faster. Also the generator is brand new.

Any help or guidance would be hugely appreciated.

Thanks
Dan

 

[ImpededLoop]

I'd suspect its power factor thats the issue.

 

[Dan121]

I'd suspect its power factor thats the issue.

Sorry if this sounds stupid but why would that affect the fault we are seeing? I know the factor is 1 on this generator

 

[DPG]

Sorry if this sounds stupid but why would that affect the fault we are seeing? I know the factor is 1 on this generator

I think he means the power factor of the load!

 

[Deleted member 105166]

Hi @Dan121 & welcome to the forum.

The voltage readings you have taken - I guess they are from the onboard meter on the Hyundai set. Be that the case, first off I would verify the readings (and the progressive voltage increase) with a trusted MFT or multimeter (not sure if you're an electrician).

My only experience with Hyundai was not a good one, with one of their DHY6000se sets about seven years ago, which suffered a completely fractured engine block with less than a couple of hundred hours on the set. I also recall the onboard voltage meter was never particularly accurate.

 

[Dan121]

@ImpededLoop & @DPG Ok so im no expert but the power factor being a issue I presume would be too low and the AVR might not be RMS rated giving a not true reading on the display? Adding more load would help that theory then? say adding in a heater or something?

@TonyMitchell , thanks I have been back and fore on the forum for previous info and its always been a handy place to come to. Struggled to find much similar issue of mine this time around though. So I took voltage readings before switching everything on with a fluke and it matched up very near perfect. However I did not actually take live readings while switched on to confirm the display volts was accurate. That's a test I will do tomorrow morning. I'm no electrician unfortunately, actually from a Hydraulic mechanical background but in my job I work alongside a electrician as a ''multi skilled team'' so I am exposed to it every day and pick up the basics and try to learn as much as possible

 

[PEG]

Hi,did you take any frequency readings,at these increased voltages?

 

[Dan121]

Hi,did you take any frequency readings,at these increased voltages?

Hi, PEG..no I actually didn’t and I was thinking about that earlier on. I will measure that tomorrow also. It was sitting at 50hz but never checked it while the voltage was increasing. The engine was solid as a rock no major fluctuation in revs etc

 

[Megawatt]

Hi, PEG..no I actually didn’t and I was thinking about that earlier on. I will measure that tomorrow also. It was sitting at 50hz but never checked it while the voltage was increasing. The engine was solid as a rock no major fluctuation in revs etc

@Dan is the generator bonded to the frame and do you think earthing it would help

 

[Lucien Nunes]

It does sound like the AVR cannot correct for the abnormal power factor of the load. There are two aspects to that; displacement pf due to the inductance of the ballast and capacitance of any integral correction, and the distortion pf due to the non-sinusoidal lamp current. Typically, overvoltage occurs with a strongly leading pf as this causes the output winding to partially self-excite, over which the AVR cannot exert any control. If the fittings contain parallel correction capacitors, disconnecting them may solve the problem. If the frequency is high, the capacitors will have too great an effect and the net pf may be leading despite the inductive ballast, so check that carefully.

Adding resistive heating load will probably help or even solve the problem (although it's a rather clumsy method to be called a true solution). It will tend to stabilise the feedback loop of the AVR, partially correct the pf of the load and absorb some of the output that results from self-excitation of the output winding, . Only a few weeks ago I had a 50kVA set that was unstable with about 10kVA of SMPSU load, that calmed down reassuringly when loaded up with a few fanheaters heating the great outdoors.

 

[PEG]

It does sound like the AVR cannot correct for the abnormal power factor of the load. There are two aspects to that; displacement pf due to the inductance of the ballast and capacitance of any integral correction, and the distortion pf due to the non-sinusoidal lamp current. Typically, overvoltage occurs with a strongly leading pf as this causes the output winding to partially self-excite, over which the AVR cannot exert any control. If the fittings contain parallel correction capacitors, disconnecting them may solve the problem. If the frequency is high, the capacitors will have too great an effect and the net pf may be leading despite the inductive ballast, so check that carefully.

Adding resistive heating load will probably help or even solve the problem (although it's a rather clumsy method to be called a true solution). It will tend to stabilise the feedback loop of the AVR, partially correct the pf of the load and absorb some of the output that results from self-excitation of the output winding, . Only a few weeks ago I had a 50kVA set that was unstable with about 10kVA of SMPSU load, that calmed down reassuringly when loaded up with a few fanheaters heating the great outdoors.

View attachment 52941

Don't let that Swedish school girl,see that.....her face will turn inside-out...

 

[Dan121]

@Megawatt its all bonded to the frame and earthed fine, I checked all the earths and insulation at the start and all perfect.

@Lucien Nunes ok that all makes sense I have attached a picture of the gear tray with the capacitor. Would the floodlight not require the capacitor to work or find the current goes through the roof with out it?

I tested one of floodlights and gears on the bench with a old kubuta generator that has no AVR. Set at 240v 50hz it all fired up ok ran for 10 minutes will no voltage increase had a 10v drop over 10 minutes but the generator was struggling so that’s what I put that down too.

One thing I noticed on the Hyundai is it is split load.
Live - earth = 115v
neutral - earth = 115v
live - neutral = 230v
that’s something I wasn’t expecting from this generator. Would this be having any impact? I have attached the wire diagram for the Hyundai also.

The kubuta is more what I expected
live - earth = 240v
neural - earth = 0v
live - neutral = 240v

 

[Lucien Nunes]

The fitting will work without the capacitor. Its supply current will increase to the same as the lamp current, 9-10A or so, harmonics may circulate with the other fittings and it may radiate increased amounts of electrical interference, so there are cons to this approach too. However you might be able to prove that with the cap disconnected in maybe just one fitting the output becomes stable, which at least pinpoints the cause. FWIW 100μF produces 1.6kVAR at 230V 50Hz, 7A flows through the capacitor.

The split-phase output should have no effect on the circuit behaviour however it does require double-pole fusing. If any connected circuits have only SP MCBs (&/or switching) or it is used to supply appliances with BS1363 plugs, they are unprotected in L2 (presently masquerading as a neutral). The N-E link in the generator might be reconnectable to one end of the winding. Consult the MIs.

 

[Dan121]

@PEG the frequency stays solid at 53hz through out the whole process.

@TonyMitchell the RMS fluke multimeter replicated the exact voltage increase as the volt meter on the generator

@Lucien Nunes our next test is to try it with a capacitor removed from the gear tray. We realised that issue last night and currently installing double pole mcbs and twin fuse cut outs at the base of the poles as I type this. We did contact Hyundai to see if we can link N + E but got no response

 

[Dan121]

So testing today..removing one of the capacitors from the gear trays has settled out the voltage issue

Many many thanks for the information and help. Hopefully this is the fault sorted out. We have a few modifications still to do but all is looking ok just