V3 Coil Carbine / Electromagnetic Projectile Accelerators / Forums

Forums

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

« Previous topic | Next topic »

V3 Coil Carbine

first previous 6 7 8 9 next last

Move Thread

LAN_403

ben5017

Thu Mar 07 2013, 02:57PM

Registered Member #3315 Joined: Thu Oct 14 2010, 04:23PM
Location:
Posts: 156

Yan:

About my firing topology. I actually have the sensor triggering a 555 which in turn, puts the output low then to the inverting igbt gate driver.

As for my bug. Turned out I had a bad 555 timer, (a good portion of the 555's I just got of mouser were defective so I kept replacing a bad one with a bad one.... as well as a shaky at best breadboard).

That being said could the sensor still cause the same issues on the 555 as on the igbt driver?

What is a good way to protect the sensors?
-Would making the sensor housing out of powdered ferrite help isolate the phototransistor from the emi?

MAXWELL:
I just bought a few microprocessors (mega 8 I think couple others) to mess around with, but I have never used one and donâ€™t know where to start... Maybe taking the time to learn and replacing all my firing circuits with a microchip might be the way to go, I am a bit intimidated to try as I have no idea where to begin.

NEW ISSUES:
1) Injector sometimes (maybe 1 out of 10) fires when I release the trigger? ( Topology, adjustable pulse length 555 edge trigger mono stable, connected to 422 driver)
Any thoughts?

2) I hooked up my injector and fired a few shots, ever since then the Resistance on my igbts when open is 14K ohms, (same on all igbts (3 stages +injector))?? As you can imagine this drastically reduced velocity. I have tried everything to figure out whatâ€™s going on (changed 555, drivers, IGBT's) I still cannot figure it out.

Maxwell

Thu Mar 07 2013, 06:26PM

Registered Member #8497 Joined: Tue Dec 04 2012, 06:24PM
Location:
Posts: 74

ben5017 wrote ...

MAXWELL:
I just bought a few microprocessors (mega 8 I think couple others) to mess around with, but I have never used one and donâ€™t know where to start... Maybe taking the time to learn and replacing all my firing circuits with a microchip might be the way to go, I am a bit intimidated to try as I have no idea where to begin.

If you have $25 and a USB cable laying around, buy any arduino you can easily get your hands on. I'll write the timing code for you. Real simple coding. FYI - the DUO utilizes the microUSB connector - which breaks easily.

Atmega8's will require an external programmer - something I don't venture off into.

ben5017

Thu Mar 07 2013, 06:35PM

Registered Member #3315 Joined: Thu Oct 14 2010, 04:23PM
Location:
Posts: 156

That would be awesome Maxwell. I will buy one today after I get out of work. Will any arduno work? Which one would you recommend? Itâ€™s long over due for me to start playing with the microchips. You think I could "reverse engineer" the code to learn how to program them?

Do you think changing to digital would have any significant effect on performance?

ben5017

Thu Mar 07 2013, 08:18PM

Registered Member #3315 Joined: Thu Oct 14 2010, 04:23PM
Location:
Posts: 156

just picked up an arduino mega 2560, now just waiting for that code maxwell haha

Yandersen

Thu Mar 07 2013, 09:25PM

Registered Member #6944 Joined: Fri Sept 28 2012, 04:54PM
Location: Canada
Posts: 340

I gave up with optical sensing already as even full metal shielding doesn't solve the problem completely. So the sensor must be as far from coil as possible. Either use laser diodes to stretch the beam, or optical cable to drive it away from coil to somewhere where EMI is low enough to not interfere with sensor. I would recommend optical cables as you can stretch all optowires in one place where all sensors and microcontroller are located - no low-current electronics works near the high pulsed magnetic fields.

Issue 1 is not surprising - common problem caused by mechanical switch. Hardly solvable - switch always glitches a little when contacts are connected or disconnected giving some false signals due to microvibrations. You can involve microcontroller to filter those false pulses: after triggering occurs make it insensitive for any other state changes for some period of time. For example...
My solution is two push buttons and a 555 timer - one button switches power supply on, another - switches it off and initiates firing sequence.
If you think smg may be wrong with yours IGBT, measure it outside the circuit, having gate and emitter connected. 14K may be given by other elements or some small floating positive voltage on gate.

Maxwell

Fri Mar 08 2013, 02:40AM

Registered Member #8497 Joined: Tue Dec 04 2012, 06:24PM
Location:
Posts: 74

Quick Reply: Modifying my code for you to be able to understand. I'll have it posted (edited here) before midnight.

Code is posted below.
I haven't used the Mega in a year or two - but it should like this code.

Serial communications are established thru USB (make sure the baud rate on the arduino side is 9600)

variable delta x is the distance between photogates in mm.

Digital pin 3 is declared as an input. It will accept 0-5v (use a high impedance voltage divider just for safety for your trigger). When a HIGH signal is read, a de-bounce code pauses for 225ms (what I found to be fool-proof for my simple mechanical switches. IF you are doing a direct contact you may drop this down to 1 or 2ms).

Digital pins 10,11 are the photogate pins. They have a timeout in the speed_trap function that will declare an error if nothing passes between within 500ms. When the signal goes high to a pin from the photogate, the timers are triggered.

Variable 'engage' is the modular part - you can set the Arduino to jump to different functions by the number assigned to engage. I assumed engage 0 to mean waiting for trigger, engage 1 is trigger actuated.

ben5017

Fri Mar 08 2013, 03:14AM

Registered Member #3315 Joined: Thu Oct 14 2010, 04:23PM
Location:
Posts: 156

You are the man. I love this place...

Maxwell

Fri Mar 08 2013, 03:58AM

Registered Member #8497 Joined: Tue Dec 04 2012, 06:24PM
Location:
Posts: 74

//Okay, so I down-compiled my code into this small format.
//Ask if you have any questions.
//It's not running 100% efficient but it is modular.

volatile byte engage; //Declare 'interrupts' as volatile
byte tabA;
byte darkA;
byte darkB;
unsigned int deltax; //Timer spacing in mm
unsigned int thrutime;
unsigned long dlaytimeA;
unsigned long dlaytimeB;

void setup()
{
Serial.begin(9600); //Reserve Digital Pins 0 and 1 for Serial
//on Duemillanove but not on Leonardo.

darkA=0; //Set initial Speedtrap ADC values
darkB=0;
engage=0; //Trigger status flag

pinMode(3,INPUT); //Pseudo-Interrupt pin (trigger)
pinMode(10,INPUT); //PhotoGate Entry
pinMode(11,INPUT); //PhotoGate Exit

deltax = 80; //Speedtrap light spacing in mm

}

void loop()
{

if(digitalRead(3)==HIGH && tabA==0)
{dlaytimeB=millis()+225; tabA=1;} //digital de-bounce script (225ms)
if(millis()>(dlaytimeB))
{tabA=0;}
if(tabA==1 && millis()>dlaytimeB && digitalRead(3)==HIGH)
{Serial.println("Trigger Action Detected."); engage=1; tabA=0;}

if(engage==1)
{
speed_trap(10,11); //digitalPin# entry, digitalPin# exit
}

if(engage==2)
{ if(thrutime==0)
{ Serial.println("Timing error. No timer activity detected."); }
else
{ calculations(); }
engage=0; //reset fire sequence
}

}

void speed_trap(int vinpinA, int vinpinB)
{ unsigned long enter; unsigned long exit;
unsigned long error; int errorcount = 0;
unsigned long errortime;

errortime=millis()+500; //500ms timeout for non-speedtrap triggering.

while(darkA==0 && errorcount!=1)
{ darkA=digitalRead(vinpinA); error=millis();
if(error>errortime){errorcount=1;}
}
enter=micros();

while(darkB==0 && errorcount!=1)
{ darkB=digitalRead(vinpinB); error=millis();
if(error>errortime){errorcount=1;}
}
exit=micros();

if(errorcount==0)
{ thrutime=(exit-enter); } //time in microseconds (1000us in 1ms)
else
{ thrutime=0; }

engage=2;
}

void calculations()
{
float velocity;
velocity = (deltax/1000.0) / (thrutime/1000000.0);
Serial.print("Thrutime (us): "); Serial.println(thrutime);
Serial.print("Velocity (m/s): ");Serial.println(velocity);
Serial.println("calculations function completed.");
}

ben5017

Fri Mar 08 2013, 02:56PM

Registered Member #3315 Joined: Thu Oct 14 2010, 04:23PM
Location:
Posts: 156

Thanks alot. What do you think about me changing all my firing circuit into arduino? worth it?

Maxwell

Fri Mar 08 2013, 04:29PM

Registered Member #8497 Joined: Tue Dec 04 2012, 06:24PM
Location:
Posts: 74

My full code controls the charging, triggering, speedgates and efficiency calculations.

I have a high-impedance voltage divider measuring charge voltage.

I use the Atmega's interal SRAM to take ~100 instantaneous voltage measurements during the discharge (make-shift oscilloscope) before my code jumps to measure velocity.

False triggering and spontaneous charging is somewhat an issue. In my next build, I will implement optocouplers to the HV and charging circuits. For now, i have some transistors that ground out all critical gates after the firing process completes.

The mega is very very capable. Lots of I/O and a large amount of memory.

Try not to fry pins - it makes programming a little more complicated. ;) I found out that they don't like direct exposure to 400v.

first previous 6 7 8 9 next last

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