180Hz Induction Furnace?

Status
Not open for further replies.
B

big john

Senior Member
Location
Portland, ME
Ran into a very interesting trouble call today: 40 year old induction furnace system.

It appears to be converting 3 phase 60Hz into single phase 180Hz using nothing more than a transformer. I did not know this was possible, and was hoping someone could shed some light:

The transformer primary is a normal 480V delta. But the secondary has all three windings in series, additive polarity.

The beginning and the end of the series secondary are the single-phase feeders directly to the furnace.

My assumption is that each cycle on the 3 phase side is being used to generate a cycle on the common secondary, so multiply 60Hz x 3 = 180Hz. Is it really that simple?

Also, on the primary side there is what appears to be s big tuned LC circuit of inductors and capacitors in series, wired in parallel with the primary. Any idea the function?

Thanks.
 
Possibly the transformer is designed with a "weak" magnetic circuit that saturates each phase core, leading to a voltage waveform on the output that consists only of widely spaced peaks of the 60Hz waveform. Those peaks from the three phases in series would combine to make a crude 180Hz waveform.

That would have an enormous harmonic component in the primary side which could play havoc with other installed equipment, so the LC circuit would be a tuned series circuit that would have near zero impedance at the harmonic frequency and so would be able to supply the harmonic current to the primary without feeding back into the supply.

Sheer SWAG, of course, since I have never seen such a thing.
 
If this were a 1:1 ratio wouldn't you expect a significant voltage reduction on your secondary, because you would never be seeing the peak of the sine?

There are also series reactors on the primary and it seems like that lends credence to the saturation idea, because doesn't curcuit impedance drop like a rock during saturation, so you'd need something to prevent overcurrent?
 
big john said:
If this were a 1:1 ratio wouldn't you expect a significant voltage reduction on your secondary, because you would never be seeing the peak of the sine?

There are also series reactors on the primary and it seems like that lends credence to the saturation idea, because doesn't curcuit impedance drop like a rock during saturation, so you'd need something to prevent overcurrent?
Click to expand...

Yes, the turn ratio would not be 1:1 if you expect to get the same nominal voltage out. But you may not need that. It depends on the design of the induction heating coils.

The series inductors you describe would limit the line current at 60 Hz while the resonant circuit would provide the harmonic current to make the system work.
 
big john said:
......180Hz
...... secondary has all three windings in series, additive polarity.....
The beginning and the end of the series secondary are the single-phase feeders directly to the furnace.

Thanks.
Click to expand...

Where and how are you measuring 180Hz? Can you post a scope waveform and a schematic where you are taking the measurement?
Connecting the 3 phase outputs (if 120 degrees) in series (without inverting one phase) gives ZERO volts.

The induction furnace rectified output will have a large 3rd harmonic.

The reason for a delta input is that the 180Hz content on the primary is a short circuit, that is why delta wye xfmrs are used for harmonic reduction, and why the transformer needs to be oversized relative to the load to be able to dissipate the short circuited primary triplett harmonic currents.

You initial description appears to be incomplete or misleading.
 
The furnace is offline for a fault, which is why I was called, so I can't verify anything about how it's reported to function.

The 180Hz output comes from what documentation I have.

The wiring of the secondaries in series is what I saw on the schematics and it appeared to be supported by the schematics.
 
Linear addition of 60Hz AC voltages will only give another 60Hz voltage.

However 'non-linear' elements allow frequency multiplication. For example, if you rectify single phase 60Hz the unfiltered output is pulsing DC at 120Hz. If instead of filtering this to get smooth DC you block the DC (say with a suitable transformer) you will get 120 Hz AC.

If you can post the schematic, showing the transformer and any other elements (diodes, capacitors, etc) someone here can probably help figure out the circuit.

You might also search the web for 'frequency multiplying transformer' to see if anything that pops up looks similar. I saw a possibly relevant patent when I did the search: https://www.google.com/patents/US2849674

Good luck
Jon
 
I will try to post what photos I have on Monday. I can say with certainty there are no semiconductors in this thing. Whatever it's doing, it is accomplishing with nothing more than a transformer, reactors, and capacitors.

Thanks for the suggestions thus far.
 
winnie said:
However 'non-linear' elements allow frequency multiplication.
Click to expand...
Just remember that a saturating magnetic core can produce a non-linear response from both an inductor and a transformer.
In the case of an inductor, an increase in the voltage across the coil will cause a higher change in the current once the flux passes the saturation point.
In the case of a transformer the coupling to the secondary winding drops dramatically as the current passes the saturation point.
 
winnie said:
Linear addition of 60Hz AC voltages will only give another 60Hz voltage.

However 'non-linear' elements allow frequency multiplication. For example, if you rectify single phase 60Hz the unfiltered output is pulsing DC at 120Hz. If instead of filtering this to get smooth DC you block the DC (say with a suitable transformer) you will get 120 Hz AC.

If you can post the schematic, showing the transformer and any other elements (diodes, capacitors, etc) someone here can probably help figure out the circuit.

You might also search the web for 'frequency multiplying transformer' to see if anything that pops up looks similar. I saw a possibly relevant patent when I did the search: https://www.google.com/patents/US2849674

Good luck
Jon
Click to expand...
As has been posted, they are over-exciting the core to produce triplen harmonics to get 180 Hz for the furnace, not series connecting to get 3*60 (i.e. the series connection is not a frequency multiplier).
 
mivey said:
As has been posted, they are over-exciting the core to produce triplen harmonics to get 180 Hz for the furnace, not series connecting to get 3*60 (i.e. the series connection is not a frequency multiplier).
Click to expand...

And just as we are used to seeing triplen harmonic current add instead of cancelling in the neutral of a wye connection, the triplen harmonic voltages of the three secondary windings add up when they are put in series while the fundamental voltages cancel out.
This is, I think, part of the explanation of why even if 1:1 turn ratios are used the output "single phase" voltage can be as high as or higher than the input L-N voltage.
 
Ajax Magnethermic

Ajax Magnethermic

big john said:
Ran into a very interesting trouble call today: 40 year old induction furnace system.

It appears to be converting 3 phase 60Hz into single phase 180Hz using nothing more than a transformer. I did not know this was possible, and was hoping someone could shed some light:

The transformer primary is a normal 480V delta. But the secondary has all three windings in series, additive polarity.

The beginning and the end of the series secondary are the single-phase feeders directly to the furnace.

My assumption is that each cycle on the 3 phase side is being used to generate a cycle on the common secondary, so multiply 60Hz x 3 = 180Hz. Is it really that simple?

Also, on the primary side there is what appears to be s big tuned LC circuit of inductors and capacitors in series, wired in parallel with the primary. Any idea the function?

Thanks.
Click to expand...

Up until last year I worked at what originally was Washington Steel Corporation's hot strip mill in Houston, PA. This facility had exactly what you describe. The furnaces (three) were built by Ajax Magnethermic in the 1970's. The transformer (aka, tripler or more commonly, multiductor) for each was made by a Canadian company, Maloney (spelling?). The primary voltage was around 25,700 volts. Ahead of the primary was a capacitor bank and the LC circuit you mentioned. I know some of it had to do with preventing sending harmonics back to the power company and for power factor control.

The furnaces were designed to heat stainless steel slabs, approximately 270" long, 6" thick and anywhere from 28" to 52" wide. The 7 Megawatt furnaces acted like a giant, inverted toaster. It was comprised of four vertically stacked induction coils. The coils were part of an LC circuit that could be tuned somewhat to control output. The slabs were charged edgewise from the bottom.

This is bringing back lots of good memories. I worked there for nearly 38 years. If you are interested, I'll try to find more info for you. I have a digital copy of the mill databook that my very capable supervisor put together years ago. I think I also have a video of a titanium slab exiting a furnace before the plant was permanently closed last year.

Those furnaces and that process made a lot of money. In addition to stainless steel, they heated titanium and zirconium slabs with very good control and no surface scale. The Washington Steel Hot Strip Mill was well known in the industry.
 
big john said:
... transformer primary is a normal 480V delta. But the secondary has all three windings in series, additive polarity....
?

Thanks.
Click to expand...

Aha, the Ajax reference reveals the misunderstanding:

The primary is likely WYE, the secondary is delta to get triplett harmonics. Patent link has some schematics that make the system understandable.

Now, it is possible the system John is looking at has an Additional transformer (or winding) at the utility interface with delta input to attenuate the triplett harmonics put back out on the power line.

https://www.google.com/patents/US3913005
Presentation1.jpg
 
junkhound said:
...The primary is likely WYE, the secondary is delta to get triplett harmonics....
Click to expand...
My mistake, I double checked what info I have and you're correct, the primary is a wye.

If the secondary is a delta, then it's an open delta with power being taken from the ends. All I can tell is that they are in series.

At a glance it does look very similar to the patent schematic you posted, except the center secondary only connects to the secondaries on either side of it: There is no load connection there.
 
Last edited:
big john said:
My mistake, I double checked what info I have and you're correct, the primary is a wye.

If the secondary is a delta, then it's an open delta with power being taken from the ends. All I can tell is that they are in series.

At a glance it does look very similar to the patent schematic you posted, except the center secondary only connects to the secondaries on either side of it: There is no load connection there.
Click to expand...

Was the transformer made by Maloney? Our originals used to vibrate themselves apart and fail. They were modified and VPI'd at one point by Westinghouse (?). That made them very reliable.
 
I don't know the cause of the trip. They went out on simple overload, no GFP on the breaker, instantaneous set to 9X. They also blew all the capacitor fuses that contribute to one of the transformer primaries.

My theory is an internal fault on one of the primary windings that blew itself open. Especially since they told us that they reset the main breaker and it held. Further, the suspect transformer has some very suspicious looking discoloration, and they know they recently had a fault on the same unit for reasons unknown.

That said, this thing looks like it will be a mightmare to work on because literally every component has built in water cooling that would have to be disassembled to even do any troubleshooting.
 
You may want to suggest to the plant owners/opertors to invest in a modern solid state 100+ kHz power supply.

Might be only a few years payback due to higher efficiency energy savings - more of the power goes into the furnace, esp if an 'unobtainium' transformer has failed.
 
Status
Not open for further replies.
Top