SOLENOID QUENCH / Electromagnetic Projectile Accelerators / Forums

Forums

4hv.org :: Forums :: Electromagnetic Projectile Accelerators

« Previous topic | Next topic »

SOLENOID QUENCH

first previous 2 3 4 5 next

Move Thread

LAN_403

Signification

Wed Mar 27 2019, 11:50AM

Registered Member #54278 Joined: Sat Jan 17 2015, 04:42AM
Location: Amite, La.
Posts: 367

In fact, DO foils care about polarity--especially a brief voltage reversal. Another thing came to mind which should apply to electrolytics as well. If an underdamped pulse is applied through a firing SCR, then why, when the current is approaching the minimum "holding current" don't they just shut off before going significantly below zero (or sooner)?? I thought this was one of the normal ways SCR's turn off!!!!

DerAlbi

Wed Mar 27 2019, 03:14PM

Registered Member #2906 Joined: Sun Jun 06 2010, 02:20AM
Location: Dresden, Germany
Posts: 727

Because you are looking at the wrong waveform: SRCs turn off if the current goes towards zero, not if the voltage is small. In an undamped oscillation (triggered by an SCR) the SCR conducts for the first current half wave - leaving the capacitor at maximum negative voltage due to the phase shift.
What you want is to only conduct for the first quater voltage waveform which is a no go, since current is at its peak if voltage is 0, so the SCR is happily turned on.
In a damped circuit the phase shift is less dramatic of course. But damping also means less efficiency - in the end you shape the current waveform by dissipating heat at a vertain rate so that no (or little) oscillation occurs.
Using foil capacitors and accepting the reverse voltage, utilizing the first current half wave in its undamped shape (sinus) is an incredible efficient and easy design. The issue is how to start the second shot from a reverse polarity (and how to recharge it - needs some kind of a bipolar charger).

Signification

Thu Mar 28 2019, 10:30AM

Registered Member #54278 Joined: Sat Jan 17 2015, 04:42AM
Location: Amite, La.
Posts: 367

@Albi:
Um...please excuse me. I was visiting another planet where Planck's constant is a TINY bit different. Of course (assuming R~0) when switched
on, the -LAGGING- SINUSODIAL CURRENT through the SELF-INDUCTOR (L) is going to be zero at the
turn-on instant (t=0+)...by it's very nature. The CAPACITOR (initial VOLTAGE -LEADING-
CO-SINUSOIDAL energy source) has V0=Vmax, Then a Vc(t) form follows.
and will start an exponential envelope cos decrease after t=T0+.

How about an add-on for making high-power SCR's with programmable holding current/inflection polarity cutoff?

DerAlbi

Thu Mar 28 2019, 12:02PM

Registered Member #2906 Joined: Sun Jun 06 2010, 02:20AM
Location: Dresden, Germany
Posts: 727

Aww i lost context.

Sry. Did you just make fun of me because i explained something, that you already know and didnt ask (because i did get the problem wrong) or did you just had a blackout and remembered or... i am lost.
If you wonder why SCRs turn on at all since the start current is 0, then the answer is that they are forcefully conducting as long as you power the gate. If the gate-pulse is too short, they actually dont stay on. (i ran into such issue in my coilgun)

Concerning the SCR turn of: why do you think this is necessary.
First, there are GTO-SCRs already which, with some circuitry, can achieve what you want. However, consider what happens, if you turn off a switch as long as you have inductive current flow. You quickly end up in a topology like a half-bridge or with a freewheeling diode - just with an SCR instead of an IGBT or Mosfet.
You cant just switch off current in an inductor. What you wish for makes only sense for resistive loads.

Maybe you should shift your thinking away from the suck-back issue and the whole current-waveform forming idea.. there are limits to the achievable waveform given a certain drive circuit topology. It might be worth a thought that fighting the inevitable physics is a waste of time and instead engineering a solution with the givens could be more productive.

You still didnt state a goal btw.. do you want to maximize energy transfer or efficiency or weight-power density or volumetric power-density or what? If you just want something moving, there is no need for such discussions, it will work

Signification

Tue Apr 02 2019, 11:56AM

Registered Member #54278 Joined: Sat Jan 17 2015, 04:42AM
Location: Amite, La.
Posts: 367

What if I wasn't looking at the wrong waveform? In particular, at the capacitor VOLTAGE waveform (cosine). It would be ideal to kill the current (sine) through the SCR right here, since it would be large. What about killing the current flow through the SCR by applying a series capacitor, actually a couple of small supercaps (~2.7VDC) in series which are in turn placed in series with an air core coil (~10AWG wire, two or three layers of 20-30 turns/layer)...two series caps since over 3v is safer. I think I heard this referred to as a class D type SCR commutation.
The whole thing will be triggered at the zero-crossing of the capacitor's zero voltage cosine waveform.

Does anyone see why this series LC circuit will not cut off (via briefly shorting) the SCR at peak (or near peak) current?? It will be a temporary short to the SCR current, bringing it below the holding current.

Sulaiman

Tue Apr 02 2019, 12:46PM

Registered Member #162 Joined: Mon Feb 13 2006, 10:25AM
Location: United Kingdom
Posts: 3140

Assuming that a coil does not saturate; current (A) = (integral of applied voltage)/inductance (V.s/H)
(ignoring energy lost to heat and the projectile)
So for a given coil, for the initial and final currents to be zero, the reverse/discharge volt.seconds must equal the forward/charging volt.seconds.

The half-bridge topology reduces the coil current at approximately the same rate as it increases it,
as the voltage across the coil is similar but opposite,
a benefit of the 1/2-bridge is that unused energy in the coil is returned to the power supply.

A coil operated in 'flyback' mode will have the most rapid coil current decay as it has the largest reverse voltage.

One untried (AFAIK) option would be a lower than normal inductance allowing very rapid increase in current over a very short period,
then a diode across the coil to maintain the current when the driving transistor is switched off.
The current waveform would be rapid rise with slow decay, the opposite to a flyback
This would allow a simple topology with high electrical efficiency and low suckback,
but probably only suited for lower velocity accelerators.

DerAlbi

Tue Apr 02 2019, 04:23PM

Registered Member #2906 Joined: Sun Jun 06 2010, 02:20AM
Location: Dresden, Germany
Posts: 727

Signification, HOW do you turn off current in an inductor? You talk about turning off the SCR at peak current... guess what.. at that time, also peak energy is stored inside the magnetic field of the inductor. This energy must dissipate somewhere over time in terms of electrical power, that means, current must flow while a voltage over the coil is measurable.
I mean, congratulations, if you make a circuit that arbitrarily switches off an SCR, but for the moment, dont think about the SCR, but about a general purpose switch, (on system level, not schematic level)
A switch that you just want to open during inductive current flow.. It will arc and destroy it self. This is true for Mosfets that will avalanche, that is true from SCR that will re-trigger due to overvoltage spilling into the gate, that is true to a bipolar transistor's first breakdown - it is true to every current controlling element in series with an inductor: switching off current is impossible. By physics law. You cant violate it. Not even if you start a 1000 more sentences with "What if".

Sulamain: the approach is not untried, but rather the normal operation of some mosfet switched coilguns. If the energy stored in the coil, during turn-off is more than the maximum single pulse avalanche energy stated in the datasheet, the mosfet will die in a very short time and many people have tried to solve this issue by using a free-wheeling diode as you describe. Yet they found that it was not necessarily optimal to choose an extremely low inductance but rather preferred a fast current decay by adding a series resistor to the diode.

Signification

Wed Apr 03 2019, 06:50AM

Registered Member #54278 Joined: Sat Jan 17 2015, 04:42AM
Location: Amite, La.
Posts: 367

So, shorting the SCR A-K terminals briefly with a series pair of supercapacitors (which act as a temporary short as they start to absorb energy-forget the wirewound resistor), won't turn off the SCR by bringing it's current below 'holding current'?
I have tried this on a somewhat smaller scale--A discharged high-capacitance capacitor does indeed turn off the SCR. And in this application it's OK for the SCR's main series cap to retain a high charge after turning off. However, IIRC it worked by far best with low C, large V. I think the C made the difference.

DerAlbi

Wed Apr 03 2019, 01:28PM

Registered Member #2906 Joined: Sun Jun 06 2010, 02:20AM
Location: Dresden, Germany
Posts: 727

Yes Yes Yes, you can of course turn off an SCR somehow. I myself would use a different circuit, but that is not the problem. The issue is, that you then have no path for the current to flow after you turned off the SCR. But current must flow! (Because it cant instantaneous change in an inductor.)
The resulting voltage spike will go so far that it will break through the SCR and turn it on again.

Please understand that i do not really comment on the "turn off the SCR" idea, but i try to tell you that such circuit is useless, since the underlying problem (that you cant turn off current in an inductor) is not solved by the circuit.

Again, you try to fight laws of physics. You will loose. I told you before that the best action is to work with what you can actually achieve.
I am not sure if i will comment further on this topic. Your willingness to deep dive into a topic without understanding the context of the problem and this unwillingness to analyze an issue on a system level will lead to nothing substantial but exotic implementations of some underlying topology.
E.g. any topology that you can implement with a turn-off-able SCR you can implement with an IGBT or Mosfet - just much easier. If the current rating, that an SCR gives you, is actually needed is unclear.

Signification

Thu Apr 11 2019, 08:53AM

Registered Member #54278 Joined: Sat Jan 17 2015, 04:42AM
Location: Amite, La.
Posts: 367

@ Albi: don't quit just yet...
This reference you sent me in another forum concerning 'fast quench of large capacitors'

was most helpful..I had even done (previously, over time) most of this and other related integrals which helped a lot...I think it may also help here:

@ all:
OK, lets forget quenching via the main SCR switch, and use the capacitor "C" in the series RLC circuit. Does anyone have any thoughts on simply shorting the capacitor "C" at the desired time by perhaps...

1) Simply shorting the cap with another SCR.
2) shorting via the SCR circuit with the capacitor in series (or parallel, or both) with the 2nd SCR?

first previous 2 3 4 5 next

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob