Announcement

Collapse
No announcement yet.

testing machine autotransformer

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • testing machine autotransformer

    I have a 3 phase autotransfomer here I'd like to check out before plugging in, that steps my 380v down to 220v 3 phase to run a machine (the gloat subject edm has arrived) but my multimeter giving screwy readings, how do I test it? just hook it up and measure voltage output on the 3 phases to the neutral pole with a bulb connected as a load to make it transform?
    I've tried measuring the resistance between phases or end to end between input and output, and everything reads high, drops to zero and then rises to 0.1 ohm, so I think my meter is trying to saturate the core and failing and this is messing with how it works. I tried another dvm and a analog too to be sure. The only thing that checks out satisfactory is anywere->earth point has a suitable Mohm number.

    I have a current limiter on my supply which I can set low, but I'd like to know the proper way because just hooking something up and seeing if it goes bang or not doesnt seem very scientific.

  • #2
    The plate says its star wired input and out, and it has neutral points input and output too.
    The neutrals are what worry me the most, going between a neutral and a pole on either side still shows 0.1 ohm, and I would expect to see a few ohms for the resistance/impedance of the transformer coils.

    Comment


    • #3
      A three phase autotransformer is just three single autotransformers on a common shaft, assuming it is a variable one. But yours sounds like a fixed one. You say it claims to be star connected, so that means that the low Voltage ends of each of the three transformers are connected together and that common point will also be common to the input and output circuits.

      First, you must disconnect ALL input and output connections before taking any readings. If you leave any connections, you may be reading external circuits in addition to the transformer and that will completely confuse the situation. Disconnect everything!

      So, you should read the same, low resistance from the common to each of the input terminals. If any of these three readings is either very high or a direct short, then you have a bad winding. When you read across any two input terminals you should read about two times that low resistance from common to any one in terminal.

      Your 380 to 220 is about a 58% stepdown so the output taps are between the common and the input terminals. This means the readings for the outputs to the common should be about 58% of the input readings. And it also should be the same to each output terminal.

      If you have a variable autotransformer, complications occur if the brushes are bad or the coils are pitted from arcing. I have seen transformers like this where it was almost impossible to find a place where they worked due to years of arcing. That was on autotransformers that were motorized to allow adjustment by the circuit. Too many small adjustments over a period of years just plain wore them out.

      As for what those "low" resistance readings are, that depends on the individual transformer. The higher the current they are designed for, the larger the wire and the lower the resistance. I would expect that on some high current transformers it would be difficult to see a reading above 0 or the least count on many Ohm meters. You may need a meter that measures milliOhms or even lower.

      Your changing readings are probably not due to any saturation issues. Ohm meters use very low currents and would almost never saturate any transformer cores. It would have to be a microscopic core. If you have been trying to take measurements while it is still connected, those readings are totally worthless: disconnect it and try again. If it was disconnected, you are probably seeing bad contact between the meter lead probes and the transformer terminals or at the meter's jacks. Compare your readings to just shorting the meter leads. And clean off any corrosion on the transformer terminals or meter probes.
      Paul A.
      SE Texas

      And if you look REAL close at an analog signal,
      You will find that it has discrete steps.

      Comment


      • #4
        Place a resistor across the output terminals and read through from the input terminals.

        It will read proportional to the winding ratio

        Comment


        • #5
          As Paul A. said, if this is a transformer of any substantial size, and I'd guess it is if it is for your EDM machine, then the resistances you are likely to find in the windings will be very low and can be difficult to resolve even with a DMM.

          If you have a variac, you could power the thing up at some low AC voltage and confirm that you get the proper ratios at the output side. If all you have is a single phase variac, you can do each winding separately. As stated above, this is really just three transformers that share parts of the core.

          you could also hook it up to your 380 line power with the secondary disconnected and take verify that the no-load output voltage looks reasonable. You do not need a load to make it transform. Measure the line-to-line voltage, it should be something near the 220V you are looking for, likely a bit higher without a load.

          Comment


          • #6
            Its rated at 15KVA, stating 40amp at 200v (would have to check, I keep getting confused with 220) output per phase, so yes its quite a substantial beast. There's no variac on it, but there does appear to be a on/off switch at the front, but it has only thin gauge wiring to it from the cores, so I don't really know how that's wired. And three phase equipment that needs a neutral connection is new to me also as if it references the input voltage against the neutral its going to see 220v input rather than the 380v between phases so I will have to experiment and adjust the jumper if needed.

            Its currently completely disconnected out of circuit. The dvm shows 0.1 ohm difference between probes shorted and connected the terminals, which if its enormous copper leads inside could be plausible, just not used to putting probes on terminals of a transformer and at least seeing something recordable, but a bit thrown by the way all the meters have to settle down first too, thats just odd and normally something I'd associate with a capacitor somewhere having to charge first but I cant see anything like that just the 3 seperate cores with the wires disappearing into the windings.

            I read somewhere that autotransformers need a load to transform & if they are open circuited the outputs run in purely conductive mode and just pass input voltage to output, which is why I was thinking of the light bulb as a load so it had a load to start the transformer action off. I have a variable resistor (dimmer switch) I can put in the input and do it phase by phase to test.

            Treading carefully, its rather larger than what I have dealt with in the past. And the seller of the edm told me they had had a tech in who diagnosed the transformer as the reason it was a non runner.
            Last edited by MrFluffy; 08-16-2014, 12:42 PM.

            Comment


            • #7
              That will not work. Transformer action only occurs with AC. Meters usually use DC so you will just see some combination of the DC resistances. The addition of an extra resistor will just make the readings smaller and harder to read: you are probably well below one Ohm to begin with.


              Originally posted by JoeFin View Post
              Place a resistor across the output terminals and read through from the input terminals.

              It will read proportional to the winding ratio
              Paul A.
              SE Texas

              And if you look REAL close at an analog signal,
              You will find that it has discrete steps.

              Comment


              • #8
                Note also that if you try to power this at reduced voltage using a dimmer, it is possible that you will get erroneous meter readings unless it is a true-rms reading DMM, as the dimmer output is not a complete sine wave.

                Comment


                • #9
                  I drew a schematic diagram of what your transformer should look like. I omitted any additional Neutrals as I have no idea how they would be included. More information would be needed there.



                  I would take all readings from the common point (center of the Y) to the various terminals. The three readings to the three input phases should all be the same. Those three readings alone should be enough to determine if and where a defect exists. In other words, if those three readings do not match within 5 - 10 %, then you have a bad transformer.

                  The three readings to the output or Load terminals should also match each other and they should be about 58% of the input readings.

                  You say this transformer is rated for 40 Amps. That is fairly large. A quick calculation shows that at 40 Amps and neglecting any reactance, the DC resistance alone is a significant factor for losses this device. For example, if one coil had a DC resistance of only one Ohm, the power lost in that coil alone would be I^2 R = 40 * 40 * 1 = 1600 Watts. That IS a room heater. All three coils would have a total of 4800 Watts. You could weld with that amount of power. Seriously.

                  Since the current rating is 40 Amps, I must conclude that the DC resistance MUST be a lot lower than one Ohm in order to prevent this device from melting. 1/100 Ohm = 0.01 Ohms = 10 milli Ohms would still produce 48 Watts of heat, more than my soldering iron. I do not know the exact size of the wire used, but I would have to expect that the DC resistance would be somewhere between 1 and 10 milli Ohms. And to take any reliable, usable readings, you would need a meter that can read at least down to that range. Lower would be better. Of course, an open circuit from a burnt out winding would show up on almost any meter, but that does not seem to be the situation.

                  As for the meter showing a changing reading, I can only speculate that it may be an artifact of how it operates or there may be other elements inside the autotransformer that produce this performance, perhaps some filter capacitors of some type. Are you using an analog meter (much more reliable for this kind of testing) or a digital one? Digital meters are prone to doing almost anything before they have time to settle. I would not trust anything a digital meter shows until it has had that settling time. On the other hand, a lot can be inferred from the way an analog meter acts in the instants just after it is connected.
                  Paul A.
                  SE Texas

                  And if you look REAL close at an analog signal,
                  You will find that it has discrete steps.

                  Comment


                  • #10
                    Mr Fluffy,
                    If you are using a digital meter to measure the winding resistance you need to be aware that some meters will give erroneous readings simply because of how they work. If you have access to a good old fashioned AVO, preferably a model 7. That will work properly.
                    Digital meters use analog to digital converters and the conversion pulse rate causes weird readings due to the large winding inductance.
                    Paul twigged that in the previous post.
                    Best Regards:
                    Baron J

                    Comment


                    • #11
                      Thanks Paul, diagram makes sense to me. Yes, I'm still not sure if they mean 40amp total, or 40amp per phase. I'm hoping its the former as I only have a 30amp/phase supply (20KVA) and my workshop plugs/sockets for 3 phase are only rated at 25amp/phase but logically its rated at 15KVA so a 20KVA source is going to be enough. I'll have to hardwire this in to get it a neutral, because the normal 3 phase plugs + sockets I have don't have a neutral pin, just 3p + earth though I have a neutral in the box as its needed to split the 220v single phase socket rail off it. Once the machine is in location in the workshop it will be hardwired anyway, theres little point putting a plug/socket on a 4,400lbs machine.

                      I have one real analog multimeter with a mirror behind the needle to eliminate parallax error and that did the same dip and rise back afterwards routine, because I went and grabbed it after I put it down to the dvm taking time to average readings initially too. I don't have a 4 wire meter, or any other way to measure milliohm readings, so I can forget checking it offline I guess with the tools I have to hand.

                      Time to stop prancing round it, and wire it up and see if 220 comes out.

                      Comment


                      • #12
                        Originally posted by MrFluffy View Post
                        Time to stop prancing round it, and wire it up and see if 220 comes out.
                        Yep, Just note that autotransformer = No isolation. Expect your 220 outta the transformer to be just as deadly as any other 220V you might encounter.
                        Play Brutal Nature, Black Moons free to play highly realistic voxel sandbox game.

                        Comment


                        • #13
                          Hi MrFluffy
                          I don't have a 4 wire meter, or any other way to measure milliohm readings, ...
                          A 1 amp DC current source and a meter capable of measuring millivolts would work. At 1 amp 1 millivolt equal 1 milliohm. A 5 ohm 5 watt resistor in series
                          with a 5 volt power supply could serve as a current source.
                          Location: Long Island, N.Y.

                          Comment


                          • #14
                            If you apply single phase power thru a large light bulb or other current limiter to each transformer in turn you can check for proportional outputs. If there is an ON-OFF switch with small wiring there must be a contactor relay somewhere in the system.
                            Don Young

                            Comment


                            • #15
                              Originally posted by MrFluffy View Post
                              Hi MrFluffy,

                              I have one real analog multimeter with a mirror behind the needle to eliminate parallax error and that did the same dip and rise back afterwards routine, because I went and grabbed it after I put it down to the dvm taking time to average readings initially too. I don't have a 4 wire meter, or any other way to measure milliohm readings, so I can forget checking it offline I guess with the tools I have to hand.
                              The reason that you see that dip and rise behavior is due to the inductance of the windings. The small current flowing causes a magnetic field which opposes the current causing it. Its similar to the effect that you see when you connect an ohmeter across a capacitor. When the needle comes to rest that will indicate the resistance of the winding.

                              If its any help, I use a small 6.3 volt heater transformer (used for radio tube filaments) to identify transformer windings along with an AVO.
                              Best Regards:
                              Baron J

                              Comment

                              Working...
                              X