Stamford generator in double delta 3

[GrahameB]

thread237-405800. I'm in a similar situation with a small Stamford unit bought 2nd hand by a client with no manuals or a maker's name plate showing model number, etc; supposedly about 12KVA, 240VAC, 1Ph, 50Hz and because I could see the single red, black and green cables coming from the terminal box (they'd been cut about 500mm outside the box) I never thought anything of it until I came to wire it up and took the terminal box cover off - it's a 12 wire, 3ph unit wired in double delta to provide a "single phase" supply. It does indeed produce 240VAC and I ran it into a 1.5Kw resistive load to prove that it does work but I have to wire it for an MEN system (neutral and earth connected). During the test run I tested between "neutral" (being the original black wire) and "active" (being the original red wire) to "earth" (being the generator frame) and got about 70VAC N - E and about 170VAC A - E. Seemingly a winding end is connected to the generator frame somewhere but it isn't either of the ones providing the 240VAC as originally wired; does anyone know if this type of generator can even be configured to produce 240VAC for an MEN distribution system?

 

[waross]

The generator will not be grounded unless and until someone grounds it.
If the set is not grounded then voltage tests to ground may be meaningless.
The voltages that you are seeing are the voltages developed across the leakage resistance to ground, in parallel with the winding capacitance to ground and in parallel with the impedance of your voltmeter.
Those readings are mostly meaningless.
There may also be some interaction with the AVR.
Alternatively there may be a ground fault in the windings.
1. Disconnect the AVR.
2. Do a continuity check from N to the generator frame. This test should show an open circuit.
3. Do a megger check of the windings. This test should show greater than 1 megOhm from any terminal to ground.
If the set does not pass these tests take it to a rewind shop.
If the set passes these test what grounding scheme do you want.
The North American standard would ground the center point so that the voltages would be:
L1 - N 120 V
L2 - N 120 V
L1 - L2 240 V
N - G 0 V
For this configuration, install a jumper from N/G to the generator frame.
I have seen sets factory wired for use in the UK where the standard voltage is/was 220 - 380.
In this case the single phase voltages would be:
In this case the N terminal is not used.
The jumper would be installed from L1 to the generator frame.

Grounding the windings.
1. Reconnect the AVR
2. Determine which terminal that you wish to ground.
3. Connect an incandescent lamp from the selected terminal to ground (eg: the generator frame).
4. Start the set and bring up the voltage.
5. The lamp should not light.
6. As a last check, measure the voltage across the lamp. You should read zero volts.
7. Remove the lamp and install a grounding jumper in place of the lamp.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[GrahameB]

Thanks for all that Waross and yeah, I was a bit liberal with terms like earthed, etc - that's why they were in parenthesis - and I had pretty much already done the tests you suggest in the following way. With the AVR still connected I did a continuity test from the generator frame to all 6 terminals in the connection box (9VDC multimeter on ohms) and as expected the readings were all infinity. I then disconnected the (2 wire) VAR and did the same continuity test from winding end to winding end (U1 to U2 etc); these were all low (as expected) and although the meter I was using wasn't going to give highly accurate readings at such a low resistance, they were all consistent enough to convince me that there were no issues with shorted or open circuit windings. I then did an insulation resistance test on the windings from each terminal to the generator frame (test voltage 500VDC) and each reading was above 100 megohm. The only conclusions I could draw from that are (1)- winding continuity wasn't abnormal, (2)- no winding end had been connected to the generator frame on purpose, (3)- there was no winding damage that had allowed a winding to short circuit to the frame, (4)- that because the windings have no reference to the frame, there cannot be any potential difference between any winding and the frame. But there is a potential difference once the VAR is reconnected and the generator is running, so it must be reasonable to assume that the windings are being somehow being connected to the frame via the AVR. What I can't determine is whether that type of AVR in that particular generator just does this, or if it's faulty. If it is faulty, it's still doing a remarkably good job of holding the voltage stable even with fluctuating loads. Frustration is setting in !

 

[waross]

What your high impedance meter is reading is the voltages developed by the leakage currents and the capacitive currents to ground.
Pick the line you want to call neutral and connect it to ground through a 100 Watt incandescent lamp. (low impedance.)
The light should not light with the set running normally.
Now connect a grounding jumper from the generator frame to the line of your choice.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[GrahameB]

Thanks once again, waross; when I did the lamp test earlier (generator still in double delta) I only used a 10w LED one, figuring that any "stray" currents wouldn't be of much magnitude but maybe they are. The 10w lamp did light up and the terminal to frame voltage didn't plunge when the lamp was connected; it will be interesting to see if the 100w load does cause the voltage to drop off.

In a moment of desperation more than anything I reconfigured the windings into series high zigzag with the intention of using T3 as the neutral (which would be connected to the generator frame and therefore to earth as well, but for this test that connection to the generator frame wasn't made) and the junction of T10 and T11 as the phase (what would you call that? Active? Hot?) I connected one lead of the AVR (8) to T3 and the other (7) to T10/11. Once again I got a very stable 239VAC out of the generator from T3 to T10/11 and once again, a voltage from each of those to the frame. The voltages this time were slightly different from the double delta configuration though; T3 to frame was still about 160VAC but T10/11 to frame went down to 30VAC. When I put the 10w lamp between T3 and the frame it glowed and the voltage didn't change; when the lamp was between T10/11 and the frame it didn't glow and the voltage did change from 30 to 29.

I know an LED lamp isn't ideal for this sort of test but incandescent lamps were withdrawn from sale here many years ago in the interests of energy conservation; initially, anyone with a blown incandescent lamp had to replace it with a compact fluorescent but now the replacement lamp is LED, as is all forms of lighting. It is still possible to buy small "special purpose" incandescent lamps (like small 5w ones that got used in range hoods, for example) but there's no hope at all of finding a 100w one; I guess I'm going to have to make up some form of resistive load bank to do this sort of testing.

One other puzzling aspect of this adventure is generator frequency; I tested my meter on the Hz scale by plugging the leads into a standard 240VAC 10A 50Hz power outlet and in less than a second it had settled down and showed a stable 50Hz. When I then put the meter across the generator terminals it MIGHT show 50Hz for maybe half a second then just went crazy by giving fluctuating readings up to 550Hz - or just gave crazy readings as soon as the meter was connected. Thinking that the meter might be faulty I did the same with another one and got the same results. I only did this test when it was connected in double delta, and while the generator was on load - one of those loads was a portable fan and it ran at the speed I'd expect it to on a 50Hz supply so I have no idea why my meter was busy fluctuating between 250 and 550Hz at the same time. I wouldn't have thought that the sine wave from the generator would somehow be distorted enough to confuse a standard frequency meter but apparently it is.

 

[waross]

An LED lamp is a High Impedance load. You need a low impedance load.
In place of the lamp you can use any load that draws an Amp or two.
The leakage impedance is so low that your meter is probably forming part of the leakage circuit.
If you replaced the high impedance digital meter with an old moving coil Simpson multi-Meter the reading would be much lower.
Don't worry about the frequency meter. I have often found that frequency meters intended for radio frequencies are somewhat erratic at grid frequency. Try twisting the frequency meter leads. This may help but no guarantee.
BTW the correct frequency is 51.5 Hz at no load dropping to 50 Hz under load.
For this reason some models of UPS do not like to run on a generator. Either find a frequency agile UPS or live with it.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[GrahameB]

Yeah mate; for a number of years after completing my apprenticeship moving coil meters were all we could get and they had their place; the ohms/volt on many wasn't especially high. All our test gear now is digital - a sign of the times I guess.

Anyway, I bit the bullet on this job because all I could rattle up on the boat in terms of a resistive load was a 1.8Kw electric kettle so with the unit still configured as high zigzag I put the kettle between T3 and the frame (being the points where I got the lowest voltage reading). I put a voltmeter across it, and a clamp on ammeter, then started the unit on no load. T3 to frame gave me 30VAV with the true RMS ammeter fluctuating 0.00 and 0.01 (as they do). I then turned on the kettle and the voltage went down to zero with the ammeter still reading the same. Flushed with that success I made the connection a hard wired one, terminated the generator leads into the boat's switchboard, restarted the unit and ran it at a 20A load (mixture of resistive and inductive loads) for 20 minutes and observed a 2 degrees C temperature rise in the discharge air from the generator cooling fan. That convinced me it wasn't overloaded, and I've always taken it as a comforting sign when a generator under load doesn't start producing smoke after 10 minutes running.

Many thanks for your assistance on this waross; I hope I'm able to contribute to the forum in such a way in the future.

 

[FreddyNurk]

A halogen bulb is still available as a downlight fitting, and is just as effective as an older incandescent bulb for these sorts of things.

EDMS Australia