Dumb question: why not 2 stage turn on, where 1st is trivial circuit to monitor mains waveform, and then flip main switch at exactly the right moment so no nasty current surge ...?
Edit: probably dumb, because hard to guarantee the timing of the full closure of that main switch?
Edit: probably dumb, because hard to guarantee the timing of the full closure of that main switch?
Last edited:
Sorry I'm late in here, but JN is quite right.
It can be simulated if you are using a FEM sim program.
I am using COMSOL.
The OPTO22 part I mentioned earlier is a solid state switch module that has a built in circuit which prevents the unit from turning on until zero crossing. As it's scr based as well, it naturally turns off at zero crossing as well.
But as Ed pointed out, it may produce unacceptable line noise as it doesn't turn on until a volt or two due to the nature of an scr or triac. I would recommend a relay be used across the device after settling, perhaps a delay unit. Turn off, different story as the relay bypass disables the scr.
I have no idea what was just said..
I STILL have no idea what was just said..
Ok, well now I'm beginning to see some words I recognize..
Ah, nice.
It is important to consider what is happening physically, not just assume that linearity is a consequence.
Go back and examine the current density profile, and the fact that it is not a function of a sine only, but a function of a sine within a sine magnetic field, and that the crowding of current is only in one direction.
jn
That statement is incorrect.
There are two aspects.
jn
ps. Your earlier statement regarding "amateur" was somewhat puzzling coming from you.
Me, I've designed, built, tested magnetics for the last two decades. Fast, slow, DC, half gauss, 10 tesla, room, cold, hot,,..a cm cubed to 50 foot long 45 ton things, called to consult on large things. If "amateur" was meant to classify me, perhaps you need to rethink. If it was not meant for me, my apologies.
Last edited:
If the transformer saturates due to the effect described in the linked reference there is nothing you can put on its secondary that can make any difference.
It is true, however, that once the transformer has been properly energized the inrush current is a function of the load on its secondary, usually a capacitor input peak rectifier.
Regards
G.
I prefer the version of Mr. Curl. It makes for a straightforward interpretation (i.e. everyone else but he)
George
Thanks for the link. I really have to admit, reading the scope pictures from right to left was really throwing me.
It is very important to note that the text I quote here a and it's accompanying diagrams are very misleading in addition to being inaccurate.
""Remanent magnetism in the core can aggravate this surge condition. It
is the nature of core material to retain magnetism to some degree after
magnetizing voltage has been removed. If transformer primary voltage
is reapplied at zero crossover and in such a direction that the increasing
field supports remanent flux, a flux of 2øm +ør results (C of Fig. 2). This
flux, of course, is entirely offset from zero, and the core is in deep
saturation, as shown by the hysteresis curve in F of Figure 2.""
It is important to understand what the actual level of remanent magnetism actually is. That is defined by the coercivity of the core material.
Also note figure 3A and 3B. 3A shows clearly a peak current of 40 amperes, yet the text states 200 amperes peak.
I had thought maybe they just messed up the scales in the paper, but note that the steady state currents to the left in both diagrams are of consistent size despite the scales purportedly being a factor of 5 different.
It is important when reading this stuff, to examine it for internal consistency. This fails that simple check.
edit: I am not saying toss it out entirely. I do point out that what was written exceeds the understandings of the author, so it would be difficult for us "amateurs" to understand which aspects are correct and which are not.
jn
Last edited:
Yes indeed, and you can easily see/measure the difference because the magnetizing current and the current due to the secondary load are 90 degrees apart.
jan
A bit of Googling has enabled me to discover that there is an anomalous skin effect which can cause nonlinearity. However, this is normally only seen in plasmas. To get it in a metal conductor you need high frequency (microwaves and up), fairly high amplitude and low temperatures (so the electron mean free path is longer than the skin depth). It appears to be totally irrelevant to audio, and largely irrelevant to radio. Its effects scale like the ratio of skin depth to free space wavelength. Normal skin effect is linear.
So was jn talking about anomalous skin effect (yet, for some reason, not saying so or noting any of its particular properties) or just misunderstanding normal skin effect?
I'm not trying to repeat the argument as that would be pointless; just trying to clarify what it was about.
So was jn talking about anomalous skin effect (yet, for some reason, not saying so or noting any of its particular properties) or just misunderstanding normal skin effect?
I'm not trying to repeat the argument as that would be pointless; just trying to clarify what it was about.
No, what you speak of is not it.
It is very important to note that you will require a very significant effort to find this via google. Been there, done that. But dont' give up, it's out there.
Since proximity within metal conductors does happen, and at frequencies of interest here, I recommend you keep looking..
You will find lots and lots of papers which come up with simple scaling to cover the magnitude, but I've only found one which derives it properly. How the dissipation actually occurs in time is of less interest to the SMPS designers than the quantity, so most of the papers consider it DC.
jn
Zenfolio | Wizard | Neumann Audiotechnik
100W/4, MOSFET output stage amp.
Has a >300VA toroidal transformer for both channels.
With a hefty resistor soldered between IEC receptacle pin and 2A slow-blow main fuse, and NO 2-step relay.
Guess how many were sold, at the equivalent of >$5.5k now.
Interesting that the electrostatic equivalent is so easy to observe, an unbiased condenser mic capsule emits 2X the input frequency since the force is always attractive.
- Status
- Not open for further replies.