BILL OF MATERIAL

DUALPULSEbom4

Please note that the following components were used for version 1 and are not necessary to install for version 2. 
Q1, Q2, Q3,
D6, D7
U4
R4, R8, R9, R10, R12, R13, R27
C9

Finished Dual Pulse Welder

The oscilloscope photo above shows a 12.6 Volt, 5Ms and 15.1Ms weld. Both pulses and the charge times is measured between the two purple lines and is 189Ms which is more than 5 welds per second at this setting.

Photo's

 Left to right : Pulse1, Pulse2 and Voltage adjustment

 

This took me way longer than expected and I was hoping to have it done by the end of this weekend. The aluminum brackets that the pcb's are on took me most of Saturday to make. I broke two brackets before getting it right and then had a hard time mounting the pcb without making a short. I finally cut all my hard work off and put a piece of wood between the the two cut offs. Bending, drilling and fitting the copper bars took up all of today.
I did however test the dual pulse before tackling this job and I am very pleased with the results. It works better than expected and you can use any one of the two pulses by turning anyone off or use both. At this moment you get two bleeps when you use the dual pulse and one bleep when you use one pulse, it is very fast though and sounds more like a long and a short bleep. It's things like this that needs a little attention.
I will try and have it finished by next weekend and hopefully get it on youtube.

Just a quick update

I'm done writing the code for the dual pulse and ran it on a simulator but haven't tested it on the welder yet. I took my old welder apart and are in a process of building the new design into that box. I had to cut the Mosfet board and the heat sinks smaller to fit and will post the pictures when I'm done. I was looking into using Joules and Watt Second and not time to control the welder and I will definitely use that to control this welder with a version 3 chip after this one. 
What is the difference? Joule and Watt Second will look at the resistance and adjust the time automatically to make better controlled welds and preventing burning holes when the pressure on the electrodes are not same. It will also give you the exact energy release when you are welding different resistances. 

ALMOST DONE

Well, I'm busy with the dual pulse code for the processor and also a couple of bugs that is p***ing me off :)) I guess everyone is asking "why so many Mosfets?". Many Mosfets in parallel lowers the resistance that lowers the heat and increase the amps. The difference between six of these Mosfets and eighteen is a couple of hundred amps more and a welder that is not going to overheat and burn out. You can also see that I get excellent controlled welds at much lower energy rates than commercial welders out there. I'm also turning this Mosfets on and off at  precise calculated speeds according to the data sheets to prevent them from going up in smoke.
I'm in the process of ordering more boards and will hopefully be done with the programming by the time they come in to make it available to you guys.

CHARGING SPEED

The voltage on the left is the actual capacitor voltage and the voltage on the right is the set point. This was just a test with a 19Ohm resistor connected to the discharge terminal causing the discharge to be very slow when the set point is lowered. This will be much faster with a resistor bank with a lower resistance.

PCB TEST

I've spent this weekend putting the new PCB together( I actually had it built before Friday) and it was very easy to build. You will need magnifying glasses, a good light and a hot air soldering iron for the surface mount components. The rest of the components are soldered at the bottom side and the solder gets pulled into the holes so no soldering is necessary at the top.  There are two problems however, I screwed up on the solder mask on the Mosfets and you have to scratch the mask off around the solder points to get a bigger solder area and Diode 10 needs to be turned around.
The charging speed is so fast with a 12 Amp supply that when the capacitor reached it's set point and the mosfet kicks in when the voltage drops 0.1V that it actually overshoots the set point by 0.4V before turning the Mosfet off. I've been writing some really complicated code to try and fix this issue.

THE PCB'S ARRIVED

Well, the PCB's  finally came in. I have a very expensive camera and as you can see I'm not very good with taking photo's. This board is very solid and absolutely of high quality. I don't know why, but my camera makes the middle of the board look like its bent and some of the silver on the board look like it's black.

I will build this board this weekend and post more photos.

Just a peek at the Mosfet part of this welder. I am still waiting for parts

CAPACITOR DISCHARGING EXPLAINED

I was browsing the internet and found many sites where the guys think that they will get the same or better welders by making the voltage higher and the capacitance lower. Yes, it will be cheaper and easier to increase the voltage to get more Joules or Watt-second (1/2x capacitance x voltage x voltage) BUT here is the catch: The capacitance of your cap and the resistance that you weld, determine the length of your weld and not the WATT-SECOND. In other words if you have a 1 farad cap, charged to 35Volt (612Ws) then it will completely discharge in 15ms if you weld metal with a resistance of 0.003Ohm (0.005 nickel).
If you increase the voltage to 50V then you will have 1250Ws and your capacitor will still discharge in 15ms at the same resistance.
With a bigger capacitor, lets say 3farad at 20Volt (600Ws), will discharge over a 45ms period.
You can already see what is happening here: A higher voltage will give you more current discharged into your work piece in a shorter time. It will produce more heat, is very hard to control and will make welding thinner metals very difficult, meaning that 0.1ms can give you a nice weld and 0.2 might burn a hole in your work piece.
 With a lower voltage and bigger capacitance your weld will be more spread out and will have less current than above but it will give you a much better controlled weld and fuse your metals better and stronger together.
 The higher Ws makes a big difference but the capacitance should also be bigger to give longer and better controlled welds.The Volts and Capacitance needs to be balanced to make a good welder.
For battery packs, I will suggest the higher capacitance and that is what I'm building here.

ANY VOLTAGE, 60V AND OVER, AC OR DC CAN KILL YOU! VOLTAGES BETWEEN 50 AND 60V   MIGHT GIVE YOU A TICKLE BUT IT WON'T HURT YOU. THIS WELDERS WITH VOLTAGES OF 20V IS TOTALLY SAFE AND YOU CAN HOLD THE PROBES IN YOUR HAND WHILE 10000AMP IS FLOWING THROUGH IT WITHOUT FEELING A THING.

NOVEMBER 6Th

I finally have 3 boards on order. I ordered only 3 because of the overall high price (tooling amount) and will test them first before ordering more at a cheaper price and also to see how many of you guys will be interested.
I am waiting for parts to build I guess the part of this project that everyone is waiting for. I am trying to make this welder as cheap as possible and this weekend my calculator kept on spitting out data that I didn’t like. I have a circuit that turns my SCR’s off but to build it to work with all big SCR’s out there will be more difficult and expensive than just using Mosfets for the actual weld. I never wanted to use Mosfets but looking at the numbers made me change my mind.
Big SCR’s have an ON resistance of about 1.5mOhm (It differs). If we do a quick calculation you will see that if we make the resistance of the metal that we weld 0(dead short because we are trying to figure out the max current to protect our components) then we only need the Capacitor's ESR and the SCR’s ON resistance to work out the actual max current at the moment the switch is closed (we are not going to use the wire and weld probe resistance because they are variables)
Let’s say our Cap have an ESR  of 1.6mOhm. If we use the above SCR and charge our cap to 20V (max) then 1.6 + 1.5 = 0.0031Ohm. I = V/R 20/0.0031 = 6452Amps.
The Mosfets in parallel that I’m going to use will have a much lower ON resistance of 0.0002Ohm which will give us about 11111Amps for a split second(much higher than the SCR). The current will be less if we add the wire and probe resistances and also the metal that we are going to weld. The number of Mosfets in parallel will also give us a much bigger surface area to work with.
I should receive the PCB’s by the 17Th. I am busy designing this new Mosfet circuit and will post pictures and more information Tomorrow when I'm done.