SCR Question / Electromagnetic Projectile Accelerators / Forums

Forums

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

« Previous topic | Next topic »

SCR Question

previous 1 2 3 4 next

Move Thread

LAN_403

klugesmith

Wed Apr 14 2010, 06:28AM

Registered Member #2099 Joined: Wed Apr 29 2009, 12:22AM
Location: Los Altos, California
Posts: 1716

Well I just learned something here.
Was about to caution that clamp diodes in parallel with a coil will switch on at maximum current. Then, since they prevent a large "braking" voltage, the coil current falls off slowly. Can be a problem when you want to quickly open a relay or avoid suckback in a coilgun.

That's a fair description when the RLC circuit is lightly damped, but is misleading in circuits (like jaysun's) that are almost critically damped.
The coil's own resistance does a fine job of slowing down the current, even without the help of reverse voltage on the terminals. And -- the voltage/current phase angle is now much less than 90 degrees. When voltage crosses zero and the diodes turn on, current is already well below its peak value.

Here is a simulation intended to illustrate these points. WE see current and voltage for four circuits: Original, original with clamp diode, reduced resistance (from 700 to 70 milliohms), and reduced resistance with clamp diode.

[edit]oops - in chart, the D labels correspond to the non-Diode circuits.

Saz43

Wed Apr 14 2010, 08:46PM

Registered Member #1525 Joined: Mon Jun 09 2008, 12:16AM
Location: America
Posts: 294

Turkey9 wrote ...

I think you'd be better off putting it in anti-parallel with the coil. That's where the voltage reversal builds up. This will protect your SCR as well as your caps.

What is the danger for the SCR? I ask this because my coilgun has the diode placed across the cap terminals. It works just fine but am I somehow damaging my SCR?

jaysun92

Sat Apr 17 2010, 03:52AM

Registered Member #2114 Joined: Fri May 08 2009, 11:58AM
Location: London, Ontario
Posts: 36

I made my own circuit diagram combining Uzzor's boost converter and the firing diagram and i was wondering if this is the best way to do the switches. I also wanted to make sure that there are no mistakes in the diagram.
Thanks to anyone who can help!
Diagram

Sorry about the bad quality, I had to take a screen cap of the program because my scanner isn't working right.

[Edit: Image size]

jaysun92

Sun Apr 18 2010, 12:20AM

Registered Member #2114 Joined: Fri May 08 2009, 11:58AM
Location: London, Ontario
Posts: 36

Ok, now that I actually made my coil I got more accurate numbers for the Inductance and RCL simulators and these are the values i get now. I wanted to know if my SCR that I posted before will be able to handle the increased amperage.
Thanks to anyone who can help!
Graphs

[Edit: Image size]

GhostNull

Sun Apr 18 2010, 05:59AM

Registered Member #2648 Joined: Sun Jan 24 2010, 12:45PM
Location: Australia
Posts: 291

In your RCL sim you can clearly see that the peak current 739 Amps, your SCR has an AV on current of 25 amps. As a rule of thumb you use an SCR that has an AV on current of 1/10 of your peak current. if you use this you will blow it up. It's very much worth it if you buy a more expensive screw terminal SCR that you can reuse later.

jaysun92

Sun Apr 18 2010, 03:51PM

Registered Member #2114 Joined: Fri May 08 2009, 11:58AM
Location: London, Ontario
Posts: 36

Ok, I found this SCR and it should be able to handle the current.http://ca.mouser.com/Search/ProductDetail.aspx?R=TYN840RGvirtualkey51120000virtualkey511-TYN840

I am still wondering if the schematic I made is ok.
Thanks for all the help

klugesmith

Sun Apr 18 2010, 10:59PM

Registered Member #2099 Joined: Wed Apr 29 2009, 12:22AM
Location: Los Altos, California
Posts: 1716

In your RCL sim you can clearly see that the peak current 739 Amps, your SCR has an AV on current of 25 amps. As a rule of thumb you use an SCR that has an AV on current of 1/10 of your peak current. if you use this you will blow it up. It's very much worth it if you buy a more expensive screw terminal SCR that you can reuse later.

I claim that the proposed current pulse is still within datasheet limits for S8040R (a nominal 800 volt 40 amp device), though the margins are less than we'd want for a commercial product design. I suggest you get a couple of those parts (they are easy to connect), but agree that you should get a bigger SCR as a backup. At this point you have no measurements of anything, just calculations and simulations, not even using worst-case component values, and your design might change again.

Here's how we could figure this.
1. Continuous duty maximum I_T is 25A (DC) or 40A (RMS on-state, with device on for half of each cycle).
2. Non-repetitive single cycle rating (actually single half-cycle) for 60 Hz is I_TSM = 520 A.
Your expected current pulse width is just over 3 ms, corresponding to 162 Hz.
Device should be able to handle I_peak of 520 A * sqrt(162 Hz / 60 Hz) = 854 A.
3. The current pulse in SCR will be truncated if your voltage clamp diode goes across the coil (as you drew) instead of across the capacitor. Let's figure the I^2T integral and compare it with datasheet limit of 1122 A^2s.

Here the solid curves are voltage & current without clamp diode. SCR turns off when current reaches zero, at the moment of maximum voltage reversal. Cumulative I2T is 788 A2s, a bit less than a half-sine pulse of same height & width (whose exact formula I2T = Ipeak^2/2*pulse_width gives 827 A2s).
The dashed lines show voltage and coil current with clamp diode to prevent voltage reversal.
I2T eventually reaches 921 A2s.
But -- when the diode switches on at 1.86 ms, it diverts I_coil away from the SCR. I2T (at the SCR) is limited to 630 A2s! The reduced stress is a benefit of keeping SCR outside of the coil-diode loop. On the other hand, the sudden interruption of current through SCR and capacitor will create a voltage spike due to parasitic inductance in -that- loop. An R-C snubber in parallel with SCR might be advised.

It will be easier on the SCR and caps if you charge the bank to less than its rated voltage. I bet the capacitor datasheet has nothing supporting discharge at well over 700 amps.

As for review of your schematic and simulation pane -- they are hard to read after resizing.
The bleed resistor on HV side should be hardwired to the capacitors, and divided into quarters (one per capacitor) to balance the voltage during charging -- otherwise the voltage will distribute unevenly due to mismatched leakage and C values. Also, the clamp diodes across coil will not prevent voltage reversal on the individual capacitor with the lowest C value. The hard switch (to disengage charger and connect load ckt) must handle full impulse current. You could skip it and protect the charger with a series resistor (be sure to have a small C directly on the boost converter output).

GhostNull

Mon Apr 19 2010, 08:07AM

Registered Member #2648 Joined: Sun Jan 24 2010, 12:45PM
Location: Australia
Posts: 291

Klugesmith wrote ...

The hard switch (to disengage charger and connect load ckt) must handle full impulse current. You could skip it and protect the charger with a series resistor (be sure to have a small C directly on the boost converter output).

If you don't want the switch to handle the impulse current you can configure it like this:

And I think it even if your SCR can handle the pulse current, its still a good idea to over-rate and get something that you could reuse in later projects.

jaysun92

Mon Apr 19 2010, 11:19PM

Registered Member #2114 Joined: Fri May 08 2009, 11:58AM
Location: London, Ontario
Posts: 36

Thanks to everyone who has helped me, I think I am going to go with this SCRhttp://ca.mouser.com/Search/ProductDetail.aspx?R=S6070W_virtualkey57610000virtualkey519-S6070W
You have all been very helpful.

GhostNull

Tue Apr 20 2010, 12:17AM

Registered Member #2648 Joined: Sun Jan 24 2010, 12:45PM
Location: Australia
Posts: 291

Well, I wish you luck on your project

previous 1 2 3 4 next

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