Inductance is proportional to turns squared, not square root. Start with a simpler winding with one tap, and look at the directions of current flow. In an autotransformer, the 'secondary' current opposes the 'primary' current.

The following items have been added to the Files area of the LTspice@groups.io group.

• /Temp/Autotransformer.zip

By: Bell, Dave <Dave.Bell@...>

Description:
I need some tutorials regarding autotransformer modelling! I was intrigued by Hassan's AC Stabilizer and tried to model TR1 The model differed from my expectations in several ways. 1: Why do the tap-to-tap voltages differ? 2: Why is the separate 18V winding so different? 2a: Why is it more different when I separate the Kxxx directives? 3: Why do the outputs throw infinities with megohm loads, but not when open??

-- 
OOO - Own Opinions Only
Best Wishes
John Woodgate
Keep trying

Two things I notice instantly:

Primary voltage should be 311V. Voltages in SPICE are peak.

You don't run a 50Hz transient for only 10ms.

The model differed from my expectations in several ways.

1: Why do the tap-to-tap voltages differ, with equal inductances?

Inductance is proportional to N², but voltage is proportional to N.
You must redo your inductance calculation.

Use the cursor instead of MEAS statements, which are so easily wrongly-written.

.TRAN for 1010ms made no difference; same measurements at 1005m as at 5m

 

From: LTspice@groups.io <LTspice@groups.io> On Behalf Of Jerry Lee Marcel via groups.io
Sent: Thursday, December 26, 2024 2:04 PM
To: LTspice@groups.io
Subject: EXTERNAL: Re: [LTspice] Autotransformer deep questions!

 

Two things I notice instantly:

Primary voltage should be 311V. Voltages in SPICE are peak.

You don't run a 50Hz transient for only 10ms.

Le 26/12/2024 à 21:40, Bell, Dave via groups.io a écrit :

Please see my recent upload, Autotransformer.zip

 

I need some tutorials regarding autotransformer modelling!

I was intrigued by Hassan's AC Stabilizer and tried to model TR1

The model differed from my expectations in several ways.

1: Why do the tap-to-tap voltages differ, with equal inductances?

2: Why is the separate 18V winding so different?

2a: Why is it more different when I separate the Kxxx directives?

3: Why do the outputs throw infinities with small or megohm loads, but not when 1G or open??

 

Dave

-- 
OOO - Own Opinions Only
Best Wishes
John Woodgate
Keep trying

-- 

OOO - Own Opinions Only

Best Wishes

John Woodgate

Keep trying

I plotted the currents in L1 and L2: Essentially identical , but + and – 6+amps?!?

Impedance of a 100mH inductor at 50Hz=31 ohms

220V into 31 ohms= about 7 amps

Is that some kind of circulating current in coupled inductors?

Only mA in L6, the separate 18V winding

 

From: LTspice@groups.io <LTspice@groups.io> On Behalf Of John Woodgate
Sent: Thursday, December 26, 2024 2:41 PM
To: LTspice@groups.io
Subject: EXTERNAL: Re: [LTspice] Autotransformer deep questions!

 

Use the cursor instead of MEAS statements, which are so easily wrongly-written.

On 2024-12-26 22:32, Bell, Dave via groups.io wrote:

.TRAN for 1010ms made no difference; same measurements at 1005m as at 5m

 

From: LTspice@groups.io <LTspice@groups.io> On Behalf Of Jerry Lee Marcel via groups.io
Sent: Thursday, December 26, 2024 2:04 PM
To: LTspice@groups.io
Subject: EXTERNAL: Re: [LTspice] Autotransformer deep questions!

 

Two things I notice instantly:

Primary voltage should be 311V. Voltages in SPICE are peak.

You don't run a 50Hz transient for only 10ms.

Le 26/12/2024 à 21:40, Bell, Dave via groups.io a écrit :

Please see my recent upload, Autotransformer.zip

 

I need some tutorials regarding autotransformer modelling!

I was intrigued by Hassan's AC Stabilizer and tried to model TR1

The model differed from my expectations in several ways.

1: Why do the tap-to-tap voltages differ, with equal inductances?

2: Why is the separate 18V winding so different?

2a: Why is it more different when I separate the Kxxx directives?

3: Why do the outputs throw infinities with small or megohm loads, but not when 1G or open??

 

Dave

-- 

OOO - Own Opinions Only

Best Wishes

John Woodgate

Keep trying

 

Virus-free.www.avg.com

You are still struggling with a complicated transformer. Start with a simple winding with just one tap, with the mains supply across the whole winding, and no separate 'secondary' winding. When you get that simple version to work, then expand. Nothing should ever 'blow up' in such circuits.

I later extended the TRAN to over 1 second, then starting data at 1s.

The currents definitely took 6-8 cycles to settle down, but voltages were essentially the same.

 

“But there is zero coupling between L6 and any of L2, L3, and L4.  That is not physically possible!  This might be why it "blows up".”

I tried both K directives, one separate for just L1 & L6, and all (L1..4 & L6) in one line.

There was a small difference in L6 voltage, but similar issues with infinities with small loads.

 

Dave

 

From: LTspice@groups.io <LTspice@groups.io> On Behalf Of Andy I via groups.io
Sent: Thursday, December 26, 2024 4:00 PM
To: LTspice@groups.io
Subject: EXTERNAL: Re: [LTspice] Autotransformer deep questions!

 

Several more things to note:

 

It is never a good idea to simulate for only 1/2 of a cycle, or even 1 cycle - especially when inductors are involved!  A far better thing to do is to simulate for 10s to 1000s of cycles (or more), and then apply your .MEAS commands to only the last cycle.  e.g., you might use ".tran 0 10 9.98" or similar.

 

Because you extracted their amplitudes at only one specific point in time, I strongly recommend disabling waveform compression.  With waveform compression enabled, the saved points miss the peaks entirely, and now your measurements depend on interpolation.  It's far better to use ".MEAS ... RMS" over a full cycle, instead of ".MEAS ... AT".

 

By using "AT", you also introduced errors due to phase shift.  With both resistances and inductances, there is phase shift.

 

You constructed a physically impossible magnetic circuit.  You have L1 and L6 tightly coupled, and L1 is tightly coupled to each of L2, L3, and L4.  But there is zero coupling between L6 and any of L2, L3, and L4.  That is not physically possible!  This might be why it "blows up".

 

Andy

 

 

-- 
OOO - Own Opinions Only
Best Wishes
John Woodgate
Keep trying

Ahh, unloaded, so (little to) no reflected impedance!

In this simplified circuit, L2 is on top of L1, and is the 220V input, so the same current flows in L1.

This is not true. In the bottom inductor you have all the magnetizing current (about 7A), but in the top inductors there is almost no current, because there is a negligible load.

In a typical autoformer, as someone mentioned earlier, output current circulates in reverse in the primary, for a smaller resulting current.

Use Kirchoff at the junction of L1 and L2.

Please see my recent upload, Autotransformer.zip

 

I need some tutorials regarding autotransformer modelling!

I was intrigued by Hassan's AC Stabilizer and tried to model TR1

The model differed from my expectations in several ways.

1: Why do the tap-to-tap voltages differ, with equal inductances?

2: Why is the separate 18V winding so different?

2a: Why is it more different when I separate the Kxxx directives?

3: Why do the outputs throw infinities with small or megohm loads, but not when 1G or open??

It's not quite the same as your schematic, but I uploaded an autotransformer model earlier this year. If you searched the Files section for examples of "Autotransformer" you wouldn't have found it, because I called it a "Variac", which was actually an old trademarked name, but I guess it stuck.

Briefly, I investigated Helmut Sennewald's previous model and found where that was going wrong, and fixed it. (I continue to be amazed how much work Helmut put into this group.)

You might find the workings of the model useful in modelling autotranformers with more taps, or just in general. It was deceptively difficult to accurately model a variac properly, which is presumably why I didn't find any other working models out there.

--