Tek-Tips is the largest IT community on the Internet today!
Members share and learn making Tek-Tips Forums the best source of peer-reviewed technical information on the Internet!
-
Congratulations JStephen on being selected by the Eng-Tips community for having the most helpful posts in the forums last week. Way to Go!
single phase xfmr primary to secondary phase shift 3
- Thread starter ka2vey
- Start date
-
1
- #3
Tinfoil, There is inherent phase shift in virtually any transformer. The magnitude might be very small and negligible for most applications but it is still there.
I would check with the people who make transformer test sets. These are not for power transformers by any means but more for the small xformers. One of these mfg sites should give you the info you seek. I used to know but this was more than 5 years ago and I work in a totally different field. Good luck.
I would check with the people who make transformer test sets. These are not for power transformers by any means but more for the small xformers. One of these mfg sites should give you the info you seek. I used to know but this was more than 5 years ago and I work in a totally different field. Good luck.
ruggedscot
Electrical
One of the biggest bangs that I saw in my career was with a thyristor drive that was configured for the UK - converted from US. They had an auto transformer to drop the control voltage to drive the electronics from the UK supply. 240 to 115 v and there were three of them. Trouble happened and a tranny primary went open - engineer found the problem and changed out the tranny - put in a double wound job and then they went back to power it up. This thing lifted clean off of the ground and blew the thyristors to bits. Seems that the phases shift on a double wound is quite a bit compared to an auto tranny......
Rugged
Rugged
ka2vey.
It's probably in you book, go back and look or as your instructor. I would by "Schaums Outline for Basic Electrical Engineering".
Best $11.50 you'll ever spend if your going to stay in this raquet.
It's probably in you book, go back and look or as your instructor. I would by "Schaums Outline for Basic Electrical Engineering".
Best $11.50 you'll ever spend if your going to stay in this raquet.
-
2
- #6
Draw out the equivalent circuit for the transformer. Include magnetising inductance and leakage inductance. You can make it a continuosly distributed model if you want higher accuracy. Include inter-turn and inter-winding capacitances. You will arrive at a more-or-less complex arrangement of series and parallel elements. If you also know your load, you can solve the equations describing your model for the phase relationship.
You could use something like SPICE or one of the friendlier simulation tools once you develop the equivalent circuit.
Have fun!!
----------------------------------
If we learn from our mistakes,
I'm getting a great education!
You could use something like SPICE or one of the friendlier simulation tools once you develop the equivalent circuit.
Have fun!!
----------------------------------
If we learn from our mistakes,
I'm getting a great education!
- Thread starter
- #7
Folks,
Thanks for the input. The difficulty I am having is this. I have two single phase transformers, and the secondary of the first is feeding the primary of the second. i modeled each xfmr as a T network with great success, but running the two T's together is giving me fits. I want to adjust the total phase shift of the system by adjusting turns in the first xfmr, but I need to keep the total Vin/Vout ratio constant.
For the record, I am not a student, but a physicist who is doing some EE work, which is why I sometimes sound "different"
Regards,
Scott
Thanks for the input. The difficulty I am having is this. I have two single phase transformers, and the secondary of the first is feeding the primary of the second. i modeled each xfmr as a T network with great success, but running the two T's together is giving me fits. I want to adjust the total phase shift of the system by adjusting turns in the first xfmr, but I need to keep the total Vin/Vout ratio constant.
For the record, I am not a student, but a physicist who is doing some EE work, which is why I sometimes sound "different"
Regards,
Scott
CarlPugh
Electrical
- Jul 16, 2002
- 142
You cannot adjust the phase shift (inductance) of a transformer by changing the turns ratio. (If the transformer dimensions are kept constant AND the percent load is kept constant)
You can adjust the inductance by changing the transformer dimensions. The major change in inductance is controlled by how far the primary is from the secondary.
NOTE: At no load the phase difference will be either 0 or 180 degrees. Phase shift will change with load on transformer.
You can adjust the inductance by changing the transformer dimensions. The major change in inductance is controlled by how far the primary is from the secondary.
NOTE: At no load the phase difference will be either 0 or 180 degrees. Phase shift will change with load on transformer.
Scott,
One thought to modify the leakage reactance is to change the core design: a shell-type core will have lower flux leakage than a core-type. Depending on the size of the core, you might use a ferrite or iron dust 'pot core' which has low leakage flux. Both these run at low flux density relative to steel, and are only in small sizes.
Carl,
Your note is only strictly true if the second transformer of the series pair has zero magnetising current and zero parasitic capcitances. Are you over-simplifying the transformer model to only include leakage reactance while neglecting the other effects? If you are referring all the non-ideal parameters of both transformers to a lumped primary equivalent model, this could also lead to the assumption that phase shift only occurs under load if the model places the parallel magnetising branch across the source upstream of the series leakage reactance in an 'L' configuration, when both parameters are in fact distributed.
Use of a uniformly distributed model, or at least a multi-stage arrangement of several series Pi or T sections would give a more accurate approximation and show that a small phase difference will indeed exist due to magnetising current flowing through the leakage reactance.
----------------------------------
If we learn from our mistakes,
I'm getting a great education!
One thought to modify the leakage reactance is to change the core design: a shell-type core will have lower flux leakage than a core-type. Depending on the size of the core, you might use a ferrite or iron dust 'pot core' which has low leakage flux. Both these run at low flux density relative to steel, and are only in small sizes.
Carl,
Your note is only strictly true if the second transformer of the series pair has zero magnetising current and zero parasitic capcitances. Are you over-simplifying the transformer model to only include leakage reactance while neglecting the other effects? If you are referring all the non-ideal parameters of both transformers to a lumped primary equivalent model, this could also lead to the assumption that phase shift only occurs under load if the model places the parallel magnetising branch across the source upstream of the series leakage reactance in an 'L' configuration, when both parameters are in fact distributed.
Use of a uniformly distributed model, or at least a multi-stage arrangement of several series Pi or T sections would give a more accurate approximation and show that a small phase difference will indeed exist due to magnetising current flowing through the leakage reactance.
----------------------------------
If we learn from our mistakes,
I'm getting a great education!
CarlPugh
Electrical
- Jul 16, 2002
- 142
ScottyUK, I think we are saying the same thing.
Ka2vey's post of 11 Mar 05 states that he is trying to change the phase shift by varying the turns ratio. Varying the turns ratio will not change the phase shift if the percent load is kept constant.
percent load includes stray capacitance and magnetizing current.
"The phase shift is 0 or 180 degrees AT NO LOAD"
Ka2vey's post of 11 Mar 05 states that he is trying to change the phase shift by varying the turns ratio. Varying the turns ratio will not change the phase shift if the percent load is kept constant.
percent load includes stray capacitance and magnetizing current.
"The phase shift is 0 or 180 degrees AT NO LOAD"
Hi Carl,
You might well be right!
I'm toying with an idea to cause some phase shift which might work. It might not either! Here goes:
If the mag current of the second transformer is changed by introducing a small air-gap into the core, and the mag current of the second transformer is drawn through the first transformer, it should be possible to effect some small phase shift due to the higher magnetising current of the second transformer flowing through the leakage reactance of the first transformer. I don't think this would be well suited to large power transformers, but could be feasible on smaller designs.
Ka2vey,
How much phase shift are you actually trying to achieve?
----------------------------------
If we learn from our mistakes,
I'm getting a great education!
You might well be right!
I'm toying with an idea to cause some phase shift which might work. It might not either! Here goes:
If the mag current of the second transformer is changed by introducing a small air-gap into the core, and the mag current of the second transformer is drawn through the first transformer, it should be possible to effect some small phase shift due to the higher magnetising current of the second transformer flowing through the leakage reactance of the first transformer. I don't think this would be well suited to large power transformers, but could be feasible on smaller designs.
Ka2vey,
How much phase shift are you actually trying to achieve?
----------------------------------
If we learn from our mistakes,
I'm getting a great education!
- Thread starter
- #12
Gentlemen,
My current product has a total phase shift of -18 degrees, and I am trying to move this to -3 degrees. These are not power transformers, but are used as feedback instrumentation (I am working on resolvers). The primary and secondary of the first transformer are concentric, with the primary outside of the secondary. The core (due to it's geometry) is air, ferrous laminate, steel and plastic. The geometry of the two coils is fixed, as is the excitation frequency. We are hoping to keep the mods to the primary/secondary of the first transformer.
We have modeled the pair as two successive T networks consisting of 3 impedences each. These impedences are calculated from the open and short circuit transformer measurements. I have resorted to empirically mapping out the impedences (open/short) as functions of primary and secondary turns. It's the physicist in me- Earnest Rutherford once said never calculate anything you don't already know the answer to...
Regards,
Scott
My current product has a total phase shift of -18 degrees, and I am trying to move this to -3 degrees. These are not power transformers, but are used as feedback instrumentation (I am working on resolvers). The primary and secondary of the first transformer are concentric, with the primary outside of the secondary. The core (due to it's geometry) is air, ferrous laminate, steel and plastic. The geometry of the two coils is fixed, as is the excitation frequency. We are hoping to keep the mods to the primary/secondary of the first transformer.
We have modeled the pair as two successive T networks consisting of 3 impedences each. These impedences are calculated from the open and short circuit transformer measurements. I have resorted to empirically mapping out the impedences (open/short) as functions of primary and secondary turns. It's the physicist in me- Earnest Rutherford once said never calculate anything you don't already know the answer to...
Regards,
Scott
ka2vey,
Your transformer sounds like it is a real strange breed. The airgap in the flux path, and the effective airgap introduced by the plastic section, are causing your transformer to have a high magnetising current and are almost certainly also increasing the leakage reactance.
To improve matters, do you have any scope to reduce the airgap? This is the single best thing you can do to bring your transformer's characteristics nearer to the ideal condition. You are using a fixed frequency, which leaves some scope to use capacitance to tune out the inductive effects, but I would look at the design of the magnetic circuit first. Airgaps are generally undesirable in transformers unless there is a DC bias in which case they become essential to prevent core saturation. If you don't have a DC bias, try to reduce the airgap to the bare minimum or eliminate it.
----------------------------------
If we learn from our mistakes,
I'm getting a great education!
Your transformer sounds like it is a real strange breed. The airgap in the flux path, and the effective airgap introduced by the plastic section, are causing your transformer to have a high magnetising current and are almost certainly also increasing the leakage reactance.
To improve matters, do you have any scope to reduce the airgap? This is the single best thing you can do to bring your transformer's characteristics nearer to the ideal condition. You are using a fixed frequency, which leaves some scope to use capacitance to tune out the inductive effects, but I would look at the design of the magnetic circuit first. Airgaps are generally undesirable in transformers unless there is a DC bias in which case they become essential to prevent core saturation. If you don't have a DC bias, try to reduce the airgap to the bare minimum or eliminate it.
----------------------------------
If we learn from our mistakes,
I'm getting a great education!
- Thread starter
- #14
It is essentially a rotary position indicator. The primary of the first xfmr is mounted in a stator and the secondary is mounted on a rotor that is physically connected to rotating equipment. The primary of the second xfmr is mounted on the same rotor as the secondary of the first xfmr. The entire "raison d'etre" of the first xfmr is to get power to the primary of the second xfmr without a physical connection (brushes, commutator etc). The secondary of the second xfmr is mounted in the same stator as the primary of the first xfmr, and by measuring the voltage across this secondary, rotary position can be calculated. The engineering associated with the second xfmr is much more involved than that of the first xfmr, which is why we want to change only the first xfmr.
I don't plan on reducing air gap as this is a small lot size and that would involve re-tooling all of our other pieces.
Scott
I don't plan on reducing air gap as this is a small lot size and that would involve re-tooling all of our other pieces.
Scott
Have you heard of a device called an LVDT? LVDT stands for Linear Voltage Displacement Transducer. They are a transformer-based position transducer which can be made to very high accuracy. A relative of the LVDT is the RVDT, where the R stands for rotary. You will find a lot of information on the net about LVDT's, probably less on RVDT's although they work on the same principle. You may find exactly what you are after if you google LVDT or RVDT.
Yokogawa (used to?) make RVDT's - we removed some from a hydraulic servo application because they were a nightmare to set up. Where are you based? I might be able to dig one out if they aren't all in the skip.
----------------------------------
If we learn from our mistakes,
I'm getting a great education!
Yokogawa (used to?) make RVDT's - we removed some from a hydraulic servo application because they were a nightmare to set up. Where are you based? I might be able to dig one out if they aren't all in the skip.
----------------------------------
If we learn from our mistakes,
I'm getting a great education!
As I see it, the only way to have any reasonable control over the output phase is to add some extra external reactive components somewhere in the system. As has already been said, cumulative leakage inductance over the four windings is probably the main cause of the excessively lagging phase. Leakage inductance is not something that can be easily changed without a total redesign of one or both transformers.
But adding some capacitive reactance to it might pull the phase back enough without striking problems of resonance, or excessive voltage/current magnification. Some damping resistors might not help efficiency, but may assist phase stability by lowering circuit Q.
As your synchro load is inductive too, you might be able to get the rotor current phasing right, without worrying too much about voltage phasing. Where you actually fit the capacitors may take a bit of experimentation as well as some circuit simulation and analysis. But adding fifteen degrees of phase lead is probably quite realistic.
But adding some capacitive reactance to it might pull the phase back enough without striking problems of resonance, or excessive voltage/current magnification. Some damping resistors might not help efficiency, but may assist phase stability by lowering circuit Q.
As your synchro load is inductive too, you might be able to get the rotor current phasing right, without worrying too much about voltage phasing. Where you actually fit the capacitors may take a bit of experimentation as well as some circuit simulation and analysis. But adding fifteen degrees of phase lead is probably quite realistic.
- Thread starter
- #17
Greetings all,
An update of sorts. I have successfully modeled the impedances as functions of turns and wire gauge, and have been able to successfully predict (within about 1 degree) output phase and voltage. Although the suggestion of adding a capacitively reactive load was a good one, it was not a step we wanted to take from a reliabitlity/temp range/added manufacture step standpoint. Thank you for your support.
Scotty UK, I am in the US, and I have heard of Yokogawa, lvdts and rvdts. The big difference here is that instead of a ferromag rotating piece that changes coupling between primary and secondary, we are actually producing a field on the rotor.
has a good elementary discussion. Looking at the first schematic, I have tried to alter the phase shift of the rotary transformer.
All the best
Scott
An update of sorts. I have successfully modeled the impedances as functions of turns and wire gauge, and have been able to successfully predict (within about 1 degree) output phase and voltage. Although the suggestion of adding a capacitively reactive load was a good one, it was not a step we wanted to take from a reliabitlity/temp range/added manufacture step standpoint. Thank you for your support.
Scotty UK, I am in the US, and I have heard of Yokogawa, lvdts and rvdts. The big difference here is that instead of a ferromag rotating piece that changes coupling between primary and secondary, we are actually producing a field on the rotor.
has a good elementary discussion. Looking at the first schematic, I have tried to alter the phase shift of the rotary transformer.
All the best
Scott
After reading this,
I would suggest looking into designing,
"pulse and high frequency transformers".
This is not a specific book but a general catagory.
Once you overcome the general losses (noise)
mag concepts (motor generator) should give you a fixed
phase shift, (according to someones concepts that should
be about 90 deg, (if I remember right, same ac field to both coils the generated field powers the second field, answer shift the static field)).
I do know, that mag current and other losses
will shift the apparent input current waveshape and time
of apparent "0" time, but true servo systems
are not dependent on "0" time, only "0" position.
Zero position control is normally done with an
external loop.
Without Knowing the specific fc, energy, purpose,
of this system, the real componets appear odd.
The US navy solved this problem about 1918, for 60hz,
the US air force about 1944 for 400hz.
A 18 deg error in a 40 mile ships gun could
put a round into a different country, much less town.
"some still use the same pointing system
just better targeting systems".
There are tens of sycro and resolvers in the market
that report to do less than 1 deg resolution "some approch .01"
So O K what am I missing.
I would suggest looking into designing,
"pulse and high frequency transformers".
This is not a specific book but a general catagory.
Once you overcome the general losses (noise)
mag concepts (motor generator) should give you a fixed
phase shift, (according to someones concepts that should
be about 90 deg, (if I remember right, same ac field to both coils the generated field powers the second field, answer shift the static field)).
I do know, that mag current and other losses
will shift the apparent input current waveshape and time
of apparent "0" time, but true servo systems
are not dependent on "0" time, only "0" position.
Zero position control is normally done with an
external loop.
Without Knowing the specific fc, energy, purpose,
of this system, the real componets appear odd.
The US navy solved this problem about 1918, for 60hz,
the US air force about 1944 for 400hz.
A 18 deg error in a 40 mile ships gun could
put a round into a different country, much less town.
"some still use the same pointing system
just better targeting systems".
There are tens of sycro and resolvers in the market
that report to do less than 1 deg resolution "some approch .01"
So O K what am I missing.
Please correct me if am wrong, but I believe that ultra precision is usually obtained by using several different synchro systems linked together. It is like the hour, minute, and second sweep hands on a clock.
You don't try to read the hour hand of a clock all by itself down to to one second of time resolution.
If the hour minute and second hands of a clock all had random errors of +/- 2 degrees, You could still read the 24 Hr time to within a few seconds of accuracy quite easily.
You don't try to read the hour hand of a clock all by itself down to to one second of time resolution.
If the hour minute and second hands of a clock all had random errors of +/- 2 degrees, You could still read the 24 Hr time to within a few seconds of accuracy quite easily.
- Thread starter
- #20
wrong100,
What you are missing is that phase shift and accuracy are two different animals. The angular accuracy of these devices is on the order of about 10 arc-minutes. They are excited with ac at about 8 kHz (more or less) and the output voltage lags the input voltage by about 16 degrees. This excessive phase shift is causing problems for the customers specific interface, and they asked us to reduce the lag to about 3 degrees. I have been able to accomplish this by changing #'s of turns and wire gauges (these things use 30 to 40 gauge wire), effectively changing the real and imaginary components of the reactances.
Regards,
Scott
BTW, are you in the Navy? i spent 6 years as an MM submarine nuc.
What you are missing is that phase shift and accuracy are two different animals. The angular accuracy of these devices is on the order of about 10 arc-minutes. They are excited with ac at about 8 kHz (more or less) and the output voltage lags the input voltage by about 16 degrees. This excessive phase shift is causing problems for the customers specific interface, and they asked us to reduce the lag to about 3 degrees. I have been able to accomplish this by changing #'s of turns and wire gauges (these things use 30 to 40 gauge wire), effectively changing the real and imaginary components of the reactances.
Regards,
Scott
BTW, are you in the Navy? i spent 6 years as an MM submarine nuc.
- Status
Similar threads
- Question
- Question
- Locked
- Question
- Locked