SCR problem
Part Scavenger, Tue Jul 25 2006, 12:26AMI designed a ring launcher yesterday completely out of my head and spare parts (interchangeable
). I have a bar-voltmeter and everything, but this is the part that needs work. It works great up until 110VAC input, but after that, it starts to draw a bunch of power and C2 craters. Best guess is that the SCR isn't turning off and it's nuking the doubler? After C2 croaks, it works just fine (only the doubler puts out 1.5 Vin).What can I do to fix this?
PS, I guess it'd be worthy to note that I'm using 10uF's in the doubler instead of 5's like the schematic says.
Second question and possible answer:
Why is my doubler cratering? It should only put out something like 10mA and stop I should think. Nobody ever talks about overloading the doubler and my simulations when shorted don't draw alot of power, what gives?
Answer-- I'm using electrolytics in the doubler, so when it's loaded down too much I should see a sine wave -/+ 170V across the capacitor. Because it's an electrolytic, it's only rated one way and it fries. If it wasn't an electrolytic, I wouldn't have this problem and could short the output if I wanted. The output would only be a predetermined amount. Is this correct?
Re: SCR problem
dan, Tue Jul 25 2006, 01:38AM
Are you inputting AC or DC? You have the inputs marked +/-. The doubler needs AC input to work. Also why is there an additional diode on the output of the doubler? The doubler should provide DC anyway. Also I don't see any current limiting on the input. Capacitors act like a dead short when they are discharged so the sudden inrush current would probably smoke the diodes.
I have used electrolytics in a doubler (quadrupler actually) before but using the Cockcroft-Walton configuration.
btw.. You should disconnect the charging circuit before you trigger the SCR.
P.S. w00t my 69th post.. Sorry I just really like that number..
dan, Tue Jul 25 2006, 01:38AM
Are you inputting AC or DC? You have the inputs marked +/-. The doubler needs AC input to work. Also why is there an additional diode on the output of the doubler? The doubler should provide DC anyway. Also I don't see any current limiting on the input. Capacitors act like a dead short when they are discharged so the sudden inrush current would probably smoke the diodes.
I have used electrolytics in a doubler (quadrupler actually) before but using the Cockcroft-Walton configuration.
btw.. You should disconnect the charging circuit before you trigger the SCR.
P.S. w00t my 69th post.. Sorry I just really like that number..
Re: SCR problem
Hazmatt_(The Underdog), Tue Jul 25 2006, 08:54AM
Where the heck is your current limiting resistor?
If you're going to double the mains you really need a limiting resistor before your big cap. Remember when it has No charge its a SHORT!!!
I want to see a 6K resistor in there pronto! and make sure it has some serious wattage behind it!
Hazmatt_(The Underdog), Tue Jul 25 2006, 08:54AM
Where the heck is your current limiting resistor?
If you're going to double the mains you really need a limiting resistor before your big cap. Remember when it has No charge its a SHORT!!!
I want to see a 6K resistor in there pronto! and make sure it has some serious wattage behind it!
Re: SCR problem
Part Scavenger, Tue Jul 25 2006, 10:18AM
Thanks guys, for some reason I was under the impression that the doubler limited itself.
Part Scavenger, Tue Jul 25 2006, 10:18AM
Thanks guys, for some reason I was under the impression that the doubler limited itself.
Re: SCR problem
Steve Conner, Tue Jul 25 2006, 10:48AM
Connect a resistor of about 100 ohms-1k between the SCR gate and cathode. It's bad practice to let it float, as leakage from the anode can trigger it. I'm guessing that's what's happening to you.
If you use small capacitors, the doubler will limit itself. I tested this means of current limiting (though at high frequency) in my ZVS based coilgun charger a while back, and it worked fine. But if you rely on the caps for limiting, for goodness' sake don't use electrolytics: they'll see reverse voltage during the early stages of charging, and ripple current way too high for their size. Try some motor run caps or something.
Steve Conner, Tue Jul 25 2006, 10:48AM
Connect a resistor of about 100 ohms-1k between the SCR gate and cathode. It's bad practice to let it float, as leakage from the anode can trigger it. I'm guessing that's what's happening to you.
If you use small capacitors, the doubler will limit itself. I tested this means of current limiting (though at high frequency) in my ZVS based coilgun charger a while back, and it worked fine. But if you rely on the caps for limiting, for goodness' sake don't use electrolytics: they'll see reverse voltage during the early stages of charging, and ripple current way too high for their size. Try some motor run caps or something.
Re: SCR problem
Part Scavenger, Wed Jul 26 2006, 02:02AM
Thanks! I added a 100 ohm resistor to the G/C junction, and it works great! After a few minutes of use, the capacitors get warm, so I added a resistor. Works wonderful.
When I get it all into my oak box and finished, I'll post it.
Part Scavenger, Wed Jul 26 2006, 02:02AM
Thanks! I added a 100 ohm resistor to the G/C junction, and it works great! After a few minutes of use, the capacitors get warm, so I added a resistor. Works wonderful.
When I get it all into my oak box and finished, I'll post it.
Re: SCR problem
benbradley, Wed Jul 26 2006, 03:12AM
Firstly a complaint, you say "C2" in your original post, but you don't have the capacitors numbered in your schematic!
But anyway, I can see what might happen to give one of the doubler capacitors reverse voltage. The current-limiting resistor others suggested isn't a total fix, it only reduces the problem enough so the cap doesn't (immediately) explode with 120V input.
Presuming you just plug this thing into the wall (or an isolation transformer and/or autotransformer) and the top input goes through the positive half-cycle up to +150V peak, this will charge the top 5uF (okay, you said they were 10) cap to +150V (with perhaps a lot of instantaneous current if plugged in at the middle of the half-cycle, but that's not the worst problem I see). It will also go through that other diode to the 470uF cap, but the return path for that is through the BOTTOM 10uF cap. You've got 150 volts across two caps in series, the 470uF, and the 10uF. About 98 percent of the voltage will appear across the smaller cap, but with reverse polarity! It won't actually have that much voltage across it - with reverse voltage, it will start to pull extra current, heat up, and get damaged.
Adding a couple more of those 1A 1000V diodes, reversed-biased across each of those 10uF caps will fix it. About a 100 to 150 ohm appropriate-power resistor in series with the fuse will limit initial current to within the diodes' current ratings. It should still take only a few seconds to charge up. You said you use a bar-graph voltmeter, that should show you the charging speed.
And I agree about disconnecting the source voltage before firing the SCR. That saves some complications.
Also (while I'm analyzing your circuit), also put a reverse-biased rectifier across the 470uF cap. When the SCR fires, the coil and cap make a tuned circuit. The coil current will increase as the capacitor voltage decreases, and it will not only discharge the cap but continue to pull the current "out of" the cap, charging it negatively. The diode might instead go across the coil.
Do you know the inductance and resistance of the coil? These would be neccesary to calculate the max current and stuff - this last diode might need to be rated at more than 1 amp.
This circuit looks a lot like a coil gun. I dare say it IS a coil gun with the projectile outside the coil instead of inside. I've read a lot on the Electromagnetic Projectile Accelerators forum and seen some really good stuff on designing, simulating and optimizing just this sort of thing.
benbradley, Wed Jul 26 2006, 03:12AM
Firstly a complaint, you say "C2" in your original post, but you don't have the capacitors numbered in your schematic!
But anyway, I can see what might happen to give one of the doubler capacitors reverse voltage. The current-limiting resistor others suggested isn't a total fix, it only reduces the problem enough so the cap doesn't (immediately) explode with 120V input.
Presuming you just plug this thing into the wall (or an isolation transformer and/or autotransformer) and the top input goes through the positive half-cycle up to +150V peak, this will charge the top 5uF (okay, you said they were 10) cap to +150V (with perhaps a lot of instantaneous current if plugged in at the middle of the half-cycle, but that's not the worst problem I see). It will also go through that other diode to the 470uF cap, but the return path for that is through the BOTTOM 10uF cap. You've got 150 volts across two caps in series, the 470uF, and the 10uF. About 98 percent of the voltage will appear across the smaller cap, but with reverse polarity! It won't actually have that much voltage across it - with reverse voltage, it will start to pull extra current, heat up, and get damaged.
Adding a couple more of those 1A 1000V diodes, reversed-biased across each of those 10uF caps will fix it. About a 100 to 150 ohm appropriate-power resistor in series with the fuse will limit initial current to within the diodes' current ratings. It should still take only a few seconds to charge up. You said you use a bar-graph voltmeter, that should show you the charging speed.
And I agree about disconnecting the source voltage before firing the SCR. That saves some complications.
Also (while I'm analyzing your circuit), also put a reverse-biased rectifier across the 470uF cap. When the SCR fires, the coil and cap make a tuned circuit. The coil current will increase as the capacitor voltage decreases, and it will not only discharge the cap but continue to pull the current "out of" the cap, charging it negatively. The diode might instead go across the coil.
Do you know the inductance and resistance of the coil? These would be neccesary to calculate the max current and stuff - this last diode might need to be rated at more than 1 amp.
This circuit looks a lot like a coil gun. I dare say it IS a coil gun with the projectile outside the coil instead of inside. I've read a lot on the Electromagnetic Projectile Accelerators forum and seen some really good stuff on designing, simulating and optimizing just this sort of thing.
Re: SCR problem
Quantum Singularity, Thu Jul 27 2006, 04:57AM
Using a design similar to FastMHz I built a doubler that didnt require extra caps by dividing my cap bank for the gun in half and using them in the doubler circuit. If you have the ability to do this I would consider it since you elimate extra components and charge time by directly driving you cap bank. Definatly look into putting a clamping diode across your coil to protect your scr though. It was mentioned here that a 10kA scr was blown with a smaller cap than yours by not using a clamping diode.
Here is a schematic of my launcher design if you havent already seen it:
C1 and C2 comprise my cap bank and are actually made up of 12 caps in series/parallel. R1&2 are current limiting resistors (I forget their exact rating but over 1K and over 50W atleast), R3&4 are bleeders. The variac and step up xfrmr are just there so I can control the output going into the doubler from 0-280ish VAC rms. With the doubler I am getting just under 750V DC on the caps right now. Works well and I have left the power supply connected and on during firing without any problems.
Good luck
Quantum Singularity, Thu Jul 27 2006, 04:57AM
Using a design similar to FastMHz I built a doubler that didnt require extra caps by dividing my cap bank for the gun in half and using them in the doubler circuit. If you have the ability to do this I would consider it since you elimate extra components and charge time by directly driving you cap bank. Definatly look into putting a clamping diode across your coil to protect your scr though. It was mentioned here that a 10kA scr was blown with a smaller cap than yours by not using a clamping diode.
Here is a schematic of my launcher design if you havent already seen it:
C1 and C2 comprise my cap bank and are actually made up of 12 caps in series/parallel. R1&2 are current limiting resistors (I forget their exact rating but over 1K and over 50W atleast), R3&4 are bleeders. The variac and step up xfrmr are just there so I can control the output going into the doubler from 0-280ish VAC rms. With the doubler I am getting just under 750V DC on the caps right now. Works well and I have left the power supply connected and on during firing without any problems.Good luck
Re: SCR problem
IamSmooth, Thu Jul 27 2006, 09:38PM
Steve, how and why does putting the resistor between the gate and cathode solve the problem?
IamSmooth, Thu Jul 27 2006, 09:38PM
Steve, how and why does putting the resistor between the gate and cathode solve the problem?
Re: SCR problem
Steve Conner, Fri Jul 28 2006, 11:02AM
Like I said in my previous post: SCRs tend to leak some current between anode and gate. Also, they have stray capacitance there, that can couple crud on the anode back into the gate. The resistor gives a low impedance path to drain these stray currents away before they can trigger the gate.
Steve Conner, Fri Jul 28 2006, 11:02AM
Like I said in my previous post: SCRs tend to leak some current between anode and gate. Also, they have stray capacitance there, that can couple crud on the anode back into the gate. The resistor gives a low impedance path to drain these stray currents away before they can trigger the gate.
Re: SCR problem
IamSmooth, Fri Jul 28 2006, 12:50PM
I use a SCR for my electric fence. A capacitor is charged to 400-600v and discharged into a coil using the SCR as a switch. I don't have this resistor across the gate/cathode and I don't have any problems. Is it that the stray current manifests itself only with certain SCRs; or, is it a problem if the SCR isn't robust enough? Mine can handle about 50 amps continuous and 500 peak.
IamSmooth, Fri Jul 28 2006, 12:50PM
I use a SCR for my electric fence. A capacitor is charged to 400-600v and discharged into a coil using the SCR as a switch. I don't have this resistor across the gate/cathode and I don't have any problems. Is it that the stray current manifests itself only with certain SCRs; or, is it a problem if the SCR isn't robust enough? Mine can handle about 50 amps continuous and 500 peak.
Re: SCR problem
Steve Conner, Fri Jul 28 2006, 01:36PM
"Probably yes to both". Some SCRs will leak more than others. Some circuits will put more dV/dt across the SCR and hence make false triggering more likely.
It doesn't matter if there's no actual resistor, as long as your gate drive circuit holds the gate down somehow.
Also, the SCR will still work without its gate held down: it just becomes more susceptible to false triggering. If your circuit doesn't do anything to provoke the false triggering, then you'd get away with it.
Steve Conner, Fri Jul 28 2006, 01:36PM
"Probably yes to both". Some SCRs will leak more than others. Some circuits will put more dV/dt across the SCR and hence make false triggering more likely.
It doesn't matter if there's no actual resistor, as long as your gate drive circuit holds the gate down somehow.
Also, the SCR will still work without its gate held down: it just becomes more susceptible to false triggering. If your circuit doesn't do anything to provoke the false triggering, then you'd get away with it.
Re: SCR problem
IamSmooth, Fri Jul 28 2006, 06:52PM
I just looked at my schematics and I can see that I am using a transitor to drop V+ through a resistor/transistor circuit. When the transistor is off, V+ triggers the gate. This is why I don't have the problem since I am keeping it down as you suggested. Thanks.
IamSmooth, Fri Jul 28 2006, 06:52PM
I just looked at my schematics and I can see that I am using a transitor to drop V+ through a resistor/transistor circuit. When the transistor is off, V+ triggers the gate. This is why I don't have the problem since I am keeping it down as you suggested. Thanks.
Print this page