On 2/16/2015 4:52 AM, Chris Jones wrote:
On 16/02/2015 16:07, mike wrote:
On 2/15/2015 5:34 PM, Chris Jones wrote:
On 16/02/2015 04:55, mike wrote:
On 2/15/2015 3:46 AM, Phil Allison wrote:
mike wrote:


if you say so.


** Wot a smug prick you are.
Seems to be your opinion of everyone.


As long as you stay out of saturation,


** No chance - you are all wet right now.



the main component of the
primary current is due to the shorted secondary.

** The secondary is not shorted.

Once again, your signature condescending tone declares that
the other guy is always wrong.
Once again, you nitpick instead of attempting to understand.
Communication is difficult enough, even when you try.
When you try NOT to understand, you're just being you.
I'm sure you win a lot of arguments when the other guy
just gives up trying to influence your thinking.

My presence in this thread is to help others get the most
out of their MOT welder. I gave up trying to influence you
long ago. Just trying to give some balance to the view.


I think I'll give my SCR the benefit of switching on when there's
zero
voltage.


** More fool you.
There it is again.


... Phil




Although he is rude, Phil is usually right, and this time is no
exception.

I suggest you get a hall-effect current transducer connected to a DSO
been there done that
AEMC MR461 current probe.
TEK TDS540 scope.
I designed my first production forward converter ~40 years ago.
so
that you can measure inrush current, and try out switching the
transformer on at both the peak of the mains voltage, and also at the
zero-crossing,
with the secondary heavily resistively loaded,
If there is unnecessarily large inrush current that is due to core
saturation late in the first half-cycle, then adding a resistive load on
the secondary won't fix that.

and look at the current waveforms. When you run out of
working triacs you could also google it.

The key to understanding this is to realise that the magnetic flux is
proportional to the time-integral of the applied voltage, and that in
continuous operation the flux is normally close to zero when the voltage
is close to maximum, and the flux is close to maximum when the voltage
is close to zero. Have a look at the first link he

Are we talking about inductors or transformers with load resistors
that cause a steady state primary current 2X their design rating?
It doesn't much matter whether we are talking about transformers or
ungapped inductors, in that the transformer with its secondary
unconnected can still draw a lot of inrush current, and adding a load on
the secondary won't fix that.


http://www.te.com/commerce/DocumentD...=CS&DocLang=EN



http://sound.westhost.com/articles/inrush.htm
http://en.wikipedia.org/wiki/Inrush_...t#Transformers

Chris

Thanks for the links. I like to learn new stuff.
I'm still trying to get my head around why the graphs in the first
link are reversed in time, but if I stand on my head, it looks
like the drive signal is optimized to maximize inrush current.
Yes I don't know why they put the scope plot backwards. It adds
unnecessary confusion, though at least they do mention it in the text. I
can only assume they lacked the ability to flip it easily in their
chosen method of document preparation.

I don't have any argument with that. You can certainly manage
the drive so the core saturates.

My attempts were to arrange the drive signal to MITIGATE inrush
current.
I know, and I was just suggesting that turning on when the mains voltage
goes through zero is not the best way to do that. Turning on at the
zero-crossing of the mains voltage is good if your load is capacitive,
e.g. the input of a SMPS.


The key point is in the wikipedia link:
"Worst case inrush happens when the primary winding is connected at an
instant around the zero-crossing of the primary voltage, (which for a
pure inductance would be the current maximum in the AC cycle) and if the
polarity of the voltage half cycle has the same polarity as the remnance
in the iron core has. (The magnetic remanence was left high from a
preceding half cycle)."
end quote

If you always turn off the current at the current zero crossing with
a positive voltage slope, then always turn on the next weld pulse
at zero voltage on the positive voltage slope, doesn't that leave
you in a remanence position to avoid saturation at the next turn on?
If not, why not?
I think the remanence is clouding the issue. It is a relevant effect but
even without it, there are good and bad times to switch on the
transformer primary, and it would be better to consider remanence only
after the basic situation with a soft-magnetic core is thoroughly
understood. You might be able to use the remaining flux in the
switched-off transformer to choose the least worst of the two zero
crossings to switch it on at, but even then, I think you would do better
to switch on at a different time. Why not seriously try it out with a
current transducer and DSO, (and a vastly over-sized triac or even
better a pair of big SCRs, just in case!). It would be nice to see the
plots, and it is one way to end an argument.

This has been much more of a discussion than the "argument" you typically
see in these newsgroups. I appreciate your use of logic instead of
the usual name-calling. A discussion doesn't need to end with a loser.

I had it in my head that remenence was the primary source of the inrush
problem and all you'd need to do is prevent the first peak from
driving the core into the saturation. If you left the core in the
optimal state given your intent to
turn it on at zero voltage crossing, you'd be done.
I'll certainly examine it more carefully next time I use a transformer.

I dug out the documentation for my welder. Turns out that I couldn't
delay the turn-on if I wanted. All the solid state relays I had
were/are zero voltage turn on units. Best I can do is control the slope
of the AC input at turn on and turn off.

The relay is rated at 600A non-repetitive for the duration of a typical
weld. I was more concerned about drawing 40 amps out of a 15 amp breaker
thru a daisy chain of connections thru outlets along the way from the
breaker to the test bench in the back room. I never popped the breaker,
but I'm sure I wasn't doing it any favors.
I took it out to the garage and hooked it up to the dryer plug
when I wanted to use it.

Along the way, I picked up a 125 joule CD welder off ebay for $20 shipped.
Solved all the battery tab welding problems without stressing the
electrical system.
The MOT is in the pile of superseded projects.



The SSR is gonna turn off near zero current. About all I can control
is the slope of the voltage sinewave when I give the command.
To turn it on it's far easier to sense the zero crossing of the line
voltage than the peak.
The mains frequency (or period) is accurate and stable enough, and
microcontrollers or even 555 timers are cheap enough that as Phil
mentioned, you can figure out the time of the voltage peak from the zero
crossing.
I used a GAL20V8 as the controller. Wouldn't have been hard to delay
the control, but the relay would have waited for zero volts anyway.

Isn't a major portion of the primary current
in phase with the primary voltage due to the resistive secondary load?
Isn't it the leakage inductance that causes the phase shift?
That sounds reasonable, but if the core saturates then that's the least
of your worries. The primary current is not necessarily a good way to
determine the core flux density, as you can make the primary current be
whatever you want by choosing the secondary current.


Under the control conditions described above where we control both
ends of the waveform to manage remanence and have a very low value
resistive load, how much would I gain by waiting for the peak line
voltage at turn on?
Time for an experiment. It depends a lot on the transformer design. I
have read that toroidal transformers produce more problematic saturation
effects than E-I types, and if the core was nominally run at less than
half of its saturation flux density then there will be no problem. Due
to the more uniform geometry I think they can run toroidal transformers
close to saturation in normal operation.


I'd go look, but it's stored behind a bunch of junk in the garage.

As I recall, I didn't make many measurements without load. But
with the secondary (almost) shorted in the weld mode, I don't
remember any horrible input current transients.
I wonder whether they even bother to switch the transformer on at the
right time in a microwave oven. The current that your transformer draws
in steady state might be so high (due to the secondary current) that
saturation doesn't make it all that much worse,

that was my assumption. I'm already running the transformer way past its
rated steady state load capability.
but given the choice (or
given a bigger transformer than your circuit breaker likes), you might
as well make it optimum if that is just a matter of inserting a small
delay.
Yep, I learned something. I'll take a look at that next time I do
something with a transformer.

I have an arc welder that sometimes trips the breaker if I turn it on at
the wrong time (with a mechanical switch), and it would be nice if it
didn't.

I do know that synchronization with the line made a major improvement
in the repeatability of the welds.


I'm up for some education.

My thinking was that, if not for saturation, the SCR would be less
stressed if I turned it on at zero voltage when the primary current was
zero. And from the unpowered state, the voltage and current can't
be anything but zero.
The risk of saturation occurs well after the zero crossing when the
voltage is turned on. If the load is not capacitive (not a big SMPS)
then the SCR won't mind if you turn it on when there is voltage across
it, and later on in the cycle it will be much happier.

And that, if I could arrange the resting place on the B-H curve
from the previous pulse such that the first half-cycle wouldn't
saturate the core, that's the best I could do easily.
Measurements didn't show any horrible first cycle inrush.
Welds got more repeatable.

Let me say the same thing in different words.
If the load is linear resistive, the transformer current and voltage
will be approximately in phase. If the SCR shuts off at zero
current, the voltage will also be near zero volts (plus whatever the
leakage inductance allows).
And the magnetizing current too.

Case 1, you start the next pulse in a nanosecond.
This is not the same thing as I was discussing. If the transformer is in
steady-state operation and if you were able to turn off the primary at a
zero-crossing of the mains voltage, that is a time when there is maximum
flux in the core. If you instantly switch it back on again, sure this
will be pretty much a continuation of steady-state operation.

If instead you turn it off for an integer number of mains cycles that
adds up to a few seconds, the core flux will not be the same when you go
to switch it back on again.
Would be interesting to learn the trajectory of the magnetic characteristics
when you turn off the input with the output (almost) shorted.

Isn't the initial current still pretty near zero?
Perhaps but that doesn't really matter as regards the risk of saturation.

Isn't the point on the B-H curve still about the same?
Yes, if you only switch it off for nanoseconds. No, if you wait a few
seconds until you have repositioned your parts for the next weld.

Case 2, you start the next pulse next week at the zero crossing
of the input voltage headed in the same direction.
What's the initial current?
What's the initial point on the B-H curve?
How is restarting it synchronously significantly different from
just leaving it running?
The flux in the core is different.

Put a big inductor (maybe a car ignition coil) across a 12Vrms AC supply
and make sure you are holding the terminals a nanosecond after you
disconnect the supply, at the instant when the AC supply is at zero
voltage (and the inductor is carrying maximum current).

Then give it to me and I will hold the terminals a week later.

I think you will notice the difference. The state of the flux in the
core matters.
I think that's a red herring. I can't argue the result, but don't think
it's relevant. Short the output and repeat the experiment.

Note that in the case of a transformer, it is possible that some value
of secondary current could result in the primary current being zero (or
any other chosen value) at the zero-crossing of the mains voltage.
If the secondary is loaded heavily with a resistor, why wouldn't the
primary voltage
be pretty much in phase with the primary current? The leakage inductance
should be a small part of the equation?
That
is not relevant to my point, which relates to the flux density in the
core, which won't be affected much by the secondary current if a
low-impedance supply is driving the primary winding.

OK, but the (almost) shorted output should have a major effect
on what happens when you stop supplying primary current and where
the core ends up on it's B-H curve.

I appreciate the principle. But, how much does it matter?
If you have an unloaded transformer and see a 40 amp input surge,
you're likely to be concerned. If your load causes a 40 amp
input current and you do what you can to minimize the surge current,
how much does it really matter? Unfortunately, I have no easy way
to measure it given my SSR characteristics.




I'm not disputing the articles you posted.
I'm not saying anything about the general (worst) case.
I'm suggesting that this is how you engineer a spot welder using a MOT.

Where did my thinking go wrong?

It may just mean that you need a larger rating for your fuse or circuit
breaker and more expensive triac or SCRs than you could otherwise get
away with.

When I tried spot welding, I was never able to get enough current from a
MOT-sized transformer with a few turns on the secondary. I could sort of
weld things if I applied very light pressure so that the workpieces made
poor contact with each other and the resistance was high enough for the
(insufficient) current to heat them, but this wasn't really satisfactory
because getting the force and contact resistance just right was not
reliable.

That was my problem. Slight changes in contact pressure made huge
differences in the energy contributing to the weld.

Here's what I did to improve MOT welds in order of
decreasing benefit.
Control the turn on/off phase of the input voltage and control
the number of integral cycles of 60Hz. When your weld takes
six cycles, being off by one is significant.
Spring load the contacts for pressure repeatability.
Cutting a slot lengthwise down the battery tab forces
more of the current thru the weld points decreasing
sensitivity to contact resistance somewhat.

With all that, I still had trouble constructing a whole
battery pack without blowing at least one weld.
The same fixture with a CD welder gives almost 100% good
welds.

http://i.imgur.com/yd8c0rf.jpg
http://i.imgur.com/er1BqSb.jpg
http://i.imgur.com/OeQZdWH.jpg

If I made a transformer big enough to weld thick workpieces with proper
contact pressure, it might cause excessive drop in my mains supply,
and/or trip breakers.

It's all about time. For battery tabs, you need very localized heating
and a very short weld time.
I have a commercial spot welder. You just push the button and wait several
seconds for the metal to turn red. Works great for angle iron,
but would kill a battery instantly.

I think the best option for me is a series-parallel array of Maxwell
boostcaps. This would eliminate the requirement for a large mains supply
capability. The 3000F ones are rated for 1900 Amps each, so about 5 in
parallel would probably supply about enough current for any normal sheet
metalwork up to a couple of millimetres thick which seems to require
close to 10kA capability. (Aluminium welding requires several times more
current so I won't try that.) Most of the references that I have seen
tend to suggest that the weld itself requires somewhere in the region of
1.5 Volts, but the electrodes etc. will have some resistive drop so I
think at least 2 banks of boostcaps in series will be desirable. Due to
the capacitors holding more energy than the total that you would want
for one weld, it would be necessary to find a way to switch them off,

Just lower the initial voltage on the caps. E = C*V^2/2.
You could also disconnect some of the caps. Or maybe not fire
all the FETs...although that may be easier said than done.
I had great difficulty measuring output current because the huge
magnetic fields generated coupled into
everything.

and it would also be very useful to be able to adjust the current by PWM
during the weld.
See the link below
Therefore a lot of MOSFETs would be required. It seems
that the best current rating per dollar occurs for individual MOSFETs
rated at about 100A, so about 100 of these in parallel would be required
for 10kA. I think a totem-pole style half-bridge topology might work,
using the output cables as an inductor to smooth the output current. A
multi-phase PWM arrangement with multiple output inductors could make
better use of the current rating of the caps. It would be an interesting
project but I don't have time to do it yet. I am somewhat concerned
about what would happen in the event of one failed MOSFET, and I would
like to think of a way to mitigate that. Perhaps the bondwire or package
pin would be an adequate fuse.
Each FET has to be able to handle all the current it can deliver with
the output shorted at the weld. Limit the charge current to the caps
and a short won't cause much problem.
I wonder if you couldn't run a wire from each FET thru a common toroid
before the junction point to help equalize the peak currents??
As for failures, anything that would "blow" under short conditions would
seriously undermine your ability to deliver current to the "short"
at the weld.

Chris





The advantage of CD is that you get a defined energy pulse.
The energy delivered to the weld point is relatively insensitive
to the resistance of the contact. Far less so than with a MOT.

I've watched a lot of youtube videos of successful welds using
racks of low voltage caps and arrays of FETs.

My Unitek uses about 400V discharged into a pulse transformer.
That V^2 factor for energy storage adds up rapidly.
The magic is in the pulse transformer. It's smaller than my
fist and claims to put 7KA into .001 ohms.
Would be interesting to learn how to build one of those.
From this:
http://i.imgur.com/ZeZerGx.png
it appears that they might pass current backwards thru the transformer
to "reset" it.

Tripped over this overview today:

http://www.google.com/url?sa=t&rct=j...,d.cGU&cad=rja

I've given up trying to weld thin aluminum.
With wire feed and Argon, I can lay a bead on THICK aluminum.
I experimented briefly with TIG. Better, but would take practice
that exceeds my attention span and my Argon budget.

Welco 52 rod with a torch makes a very strong bond
on stuff that can take the heat. There's very small margin
between a successful bond and a puddle on the floor.
I learned the hard way that most stuff that looks like aluminum
is really potmetal and has the same melting temperature as the Welco 52.