Free Wheeling Diode(s) / Electromagnetic Projectile Accelerators / Forums

Forums

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

« Previous topic | Next topic »

Free Wheeling Diode(s)

Move Thread

LAN_403

ZakWolf

Fri Dec 04 2015, 01:56AM

Registered Member #3114 Joined: Sat Aug 14 2010, 08:33AM
Location:
Posts: 608

Got the idea here

I want to use these diodes

to make a series string to reduce the ringing and EMF kickback. I choose these especially since there fwd V drop is 5V's!! Since the drop is so high, when the diode conducts it will require that 5v and thus lowering the voltage much faster then just having one single diode.
Vf - Forward Voltage: 5 V
Vr - Reverse Voltage: 1000 V
If - Forward Current: 6 A
Max Surge Current: 400 A

1449193762 3114 FT0 10 Segment Diode Array

1449193762 3114 FT0 Coil A Commutation Current Solid Projectile

I dont know how many I would need to faster then normal completely bring the current to 0v. They're only $.49 so I was going to put at least 8 in series.

Coil Specs. (Goals)

18 awg wire (@400v 1800uf )/per stage... 3 Stages
Stage 1: ~250 turns .267 ohms/ .222mH = 700A in 2.14ms (all values are calculated and are not considering loaded coil)
Stage 2/stage 3: slightly less turns and shorter coils, higher amps and faster decay time.

"If each diode demands an induced voltage of e in order to conduct then as long as energy is available in the field the current will decay at a rate of:

where n is the number of series diodes. The actual decay isn't purely linear due to the exponential behaviour of the diode forward conduction. A more thorough analysis will be added to the theory section as time permits."

How do I use this formula ?

DerAlbi

Fri Dec 04 2015, 11:04AM

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

I suggest you dont use this forumla at all. Or at least not directly - you just learn from it and the theory behind it.
This formula assumes that there is a constant voltage drop across the diode (which is an ok approximation).
The formula simply says:
1) "the more diodes (n) you use, the faster the current decay"
2) "the higher the forward-drop (e), the faster the current decay"
3) "the lower the Inductance, the faster the current decay"
The whole formula is derived from an expession that basically states:
4) "the hgiher the voltage across a coil, the faster is the current change"

The thing is: every diode will have a resistive component to it. Meaning the textbook exponential function can not applied and this effect becomes more noticable at high currents.
I measured those diodes a long time ago and found the Silicon resistance to be ~4mOhm. Meaning at your 700A there is some drop across the actual junction plus resistive 3V drop. (it can be much more.. who knows)

All this must be seen in the context of the whole circuit:
If you coil has 267mOhm then the voltage across the internal ESR is allready 190V. Now you have to question yourself: does it matter much putting some diodes in series to have an overal maybe ~220V drop?
In percent, the change issnt that big considdering the effort.

Of course this argument counts only at high currents. At lower Currents like 30A then the ESR-Drop is 8V while the diodes still provide 25V more. This is a noticable improvement!
However: since the force is quadratic to the current... think about it again: if your cirucit feature is optimizing your current decay mostly at low currents, is the actual suckback created from this optimized protion of the current waveform actually worth mentioning?

What i imply is that your concept is fundamentally flawed

What the diodes do is basically disipate the energy stored in the coil. So what component comes to mind when you want to disipate energy? Resistors!
Of course a resistor only leads to an exponential decay which has a looooong current tail.. and it seems ugly in the diagramm.
Think again: square current graph (to get an idea about the qualitive force-behavior), and notice how insignificant the actual tail becomes.
What will strike you is that the force decays extremely fast at the relevant time if you accept the exponential decay...

So maybe a resistor is much better in disipating energy than diodes. And its much better in decaying the current when it still causes significant suckback.

ZakWolf

Sun Dec 06 2015, 05:34AM

Registered Member #3114 Joined: Sat Aug 14 2010, 08:33AM
Location:
Posts: 608

DerAlbi wrote ...

I suggest you dont use this forumla at all. Or at least not directly - you just learn from it and the theory behind it.
This formula assumes that there is a constant voltage drop across the diode (which is an ok approximation).
The formula simply says:
1) "the more diodes (n) you use, the faster the current decay"
2) "the higher the forward-drop (e), the faster the current decay"
3) "the lower the Inductance, the faster the current decay"
The whole formula is derived from an expession that basically states:
4) "the hgiher the voltage across a coil, the faster is the current change"

The thing is: every diode will have a resistive component to it. Meaning the textbook exponential function can not applied and this effect becomes more noticable at high currents.
I measured those diodes a long time ago and found the Silicon resistance to be ~4mOhm. Meaning at your 700A there is some drop across the actual junction plus resistive 3V drop. (it can be much more.. who knows)

All this must be seen in the context of the whole circuit:
If you coil has 267mOhm then the voltage across the internal ESR is allready 190V. Now you have to question yourself: does it matter much putting some diodes in series to have an overal maybe ~220V drop?
In percent, the change issnt that big considdering the effort.

Of course this argument counts only at high currents. At lower Currents like 30A then the ESR-Drop is 8V while the diodes still provide 25V more. This is a noticable improvement!
However: since the force is quadratic to the current... think about it again: if your cirucit feature is optimizing your current decay mostly at low currents, is the actual suckback created from this optimized protion of the current waveform actually worth mentioning?

What i imply is that your concept is fundamentally flawed

Thanks for the lengthy reply but everything you said about the diode & resistor pair is actually shown to lessen(improve compared to nothing) performance over a diode string.

"Fig 2 shows that the series diode method is faster at reducing the commutating current and the subsequent retarding impulse. The decay time for the diode-resistor method suffers more from the projectile induced current boost. Table 1 lists the muzzle speeds for the different projectile types resulting from each method. It is interesting to note that the current decay rate is initially similar for both methods but a divergence occurs as the projectile begins to interact with the decaying current."

1449379607 3114 FT174220 Coil D Commutation Current Comparison

Although these measurements were taken at a lower voltage, the author stated that more research should be done to see iff higher voltage effects any of this. I can get BEEFY diodes and BEEFY sub ohm resistors so that the diode conducts and the resistor dissipates everything. His tests used a 70 ohm resistor with the diodes.

DerAlbi

Sun Dec 06 2015, 04:52PM

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

Well.. if the maximum current is 140A and the difference begins at 40A, then the pull force where this optimization works is only (40/140)^2 * 100% = 8% of the maxmimum pullforce.
The 8% are further less noticable because both methods do not eliminate the suckback comletely. Additionally again the suckback happens in a region of the coil where the force issnt that big no matter what current.

If this has any noticable improvement at all then the assumption F ~ IÂ² does not hold - meaning deep saturation. But if you work in such operating area, one clearely does not care about efficiency anyway... so why the diodes then, buy a bigger cap - makes a better youtube title

Electrically the current decay can in no way be equal in any given time when using a resistor versus a diode. A resistor causes a exponential decay while a diode introduces a linear decay. Even a exponential governed linear decays looks different. The negative slope should have been changed, even it were only 5 diodes.
I mean.. its looks a little bit like it works but i dont know... its not much, if this is no measurement error. The minimal different slope can also be explained by a variing projectile speed.. its backed by by the fact that the red line allready falls fasster before 21ms..

In my optinion the 2 shots had different boundary conditions.
Premagnetized projectile, wrong oriented projectile, different starting position... those factors could have change the shot so much that the timing missmatch would cause the different tail currents. At the end: its 24ms versus 25ms.. this is is waaaay inside the pure differences that mentioned effects generate.
Single measurements cant show the difference here, you need tens of shots to give that graph any meaning.

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