CD Spot Welder design, needs comments and suggestions

Discussion in 'The Projects Forum' started by Leorenzo, Sep 21, 2013.

  1. Leorenzo

    Thread Starter New Member

    Sep 21, 2013
    I have already noticed the other posts here about Spot Welder however I would like to get some comments and suggestions on my design. Here is the schematic:


    The discharging is controlled by a microcontroller which can be set from 50ms to 300ms.

    LM317T (1.5V - 17.8V) - this component will dissipate a maximum of 28W. I'm wondering if a heat sink can be used to solve the heating problem. I'm also confused because when I look at its datasheet, maximum power dissipation is internally limited and I don't know what that means. It's Operating Temperature Range is 0-125 °C and its thermal resistance is 5°C/W which I think gives me a 140°C for my operation.

    Resistor for Discharging (R7) - It seems to be that it wastes a lot of energy any better idea? I needed it so that when the capacitor voltage is at higher voltage and I adjusted the input voltage at a lower voltage, it would dissipate to match the input voltage.

    MOSFETs - Since I still don't know what the total resistance of my circuit would be. I'm planning to parallel at least two IRF1104 MOSFET which its pulse drain current is rated at around 400A at the discharging side of the circuit (lower MOSFET).

    2N6547 Transistor - is it okay like that with no resistors around? Its maximum power dissipation is 175W and its total power dissipation in the circuit is 150W so it seems that it would be fine.

    Op-Amp - Could the output of the Microcontroller gained to reach 20V really drive the MOSFET? I still haven't use a MOSFET and just self studied it in the internet.

    6F Capacitor (Maximum Voltage Rating 21VDC) - Does capacitor have maximum current rating? I thought they don't have since it will just fasten or lower its charging time however I found a datasheet that has its rated maximum current. Are they really ok at discharging at around 400A? The capacitor might be this one _trkparms=algo%3DSI%26its%3DI%26itu%3DUCI%252BUA%252BFICS%26otn%3D21%26pmod%3D350866497187%26ps%3D54
    Saves me the effort in displaying its voltage. However, I'm wondering why does it have a feature that once you reach a certain low voltage, it would sound an alarm?I thought capacitor is just alright as long as you just don't exceed its maximum votlage rating?

    7805 - I know it would generate a lot of heat because of the Vi 20VDC and Vo 5VDC so I'm planning to put a series of regulator (eg. 7812,7809,7805) to lessen the heat (or probably another heatsink).

    Flywheel diode - I think at some place here at my design I need to place this diode for protection to eliminate the possibility of reverse current but I don't know where and how.

    Microcontroller - (actually not a concern I just want to state its function) as what the LCD displays, it can control the Pulse Width for the discharging so that it has a fix discharge time instead of timing it in by yourself. At start, it will immediately charge the capacitor. Once the push button is clicked, It will terminate the connection between the capacitor and the Votlage Supply to eliminate sudden current surge. and once the discharging is finished, it will charge the capacitor again. All of this are equipped with delays to ensure safety.

    This really is my first big project and I have used components and concepts that I have never learned before that's why I want to see what are the flaws here in my design.

    Thank you in advance! I really hope you can give me some of your time to look at my design :)
  2. #12


    Nov 30, 2010
    First, I want to apologize. The number of mistakes in this is going to require a small book and it's going to sound like I'm ripping on you. I can see you are working in good faith, and so am I.

    1a) 317 chip. Internally limited means it will shut down if you abuse it.
    1b) 4 C/W, 125C limit. If you are right about the 28 watts, 28 watts times 4c/w = 112C temperature rise. Add 112 to room temperature and you are at 137C with an infinite heat sink. Obviously wrong. You are going to need to go to a TO-3 package or add a TO-3 helper transistor. See page 12 of the datasheet.
    1c) The 317 needs at least 10 ma of load, so R8 is usually 120 ohms unless you can guarantee a load somewhere else.
    1d) the 317 chip needs capacitors on each side to avoid oscillation. Page 1 of the datasheet.
    1e) Page 6 figure 3. Protection diodes.

    2) What is Q4 doing? It's constantly on, shorting out a voltmeter, and has no resistance to limit the base current.

    3) It looks like you have both electrodes grounded.

    4) A 741 chip is a P.O.S. You can't run its inputs within 4 volts of ground and that makes RV2 look like a mistake. If you turn U1 down below 4 volts, your amplifiers are going to go out of spec. Use a different chip. Possibly a gate driver to get a fast turn on time.

    5) That capacitor is not rated for 20 volts. It's rated for 20V to 24V momentary surge voltage. It's really for a car which rarely gets above 15.5 volts.
    5b) It does not have an equivalent series resistance or inductance rating so you can not calculate 400 amps out of it.
    5c) it alarms at low voltage because it is not built for this purpose.

    6) R7 does not seem to be necessary.

    7) You are going to need at least (4) mosfets in parallel to deliver 400 amps, and I'm not sure about the pulse duration limits.

    8) The 7805 chip will not generate heat because it has 15 volts across it. It generates heat because it has voltage across it and current through it. There isn't much current being used if I'm guessing correctly.

    This is only a partial. I'm sure I missed some of the problems. That's why there are dozens of other helpers here.
  3. Leorenzo

    Thread Starter New Member

    Sep 21, 2013
    Thank you for your quick response! It indeed is a long list of correction! But I do like to receive corrections as it gives me new learning about the matter :). I also would like to apologize on some information I forgot to mention and to my carelessness on some parts of the schematic.

    Well, yeah. Since I just have been a part of a team which I am under battery management, We will be welding lots of batteries so I would really like to finish this up. :)

    I read that to compute for the dissipated power on LM317 is Pd = (Vi-Vo) X Io so Vi which is 20V and lowest Vo is 1.25V, Io is at 1.5A which gives me 28W. However I would normally be using it at around 8-15V which at 8V would give me 115°C since thermal resistance is 5C/W. It seems like it still is a high temperature but do you think a TO-3 package attached to a heat sink can do the trick? And also with "infinite heat sink" means that's the best case scenario with a heat sink attached?

    I used it as current amplifier since the output of the 317 is only at 1.5A and I want the capacitor to charge faster. Although from the looks of it, it really looks odd for me.I'll check on this again and add the base resistance that you mentioned. And for the Voltmeter, I was just computing for the power dissipation on the transistor.

    I apologize for this one I kinda hurried this one up. One of the electrode should be at ground and the other at end of R1.So from the picture at my first post, R1 should not be immediately connected to the ground.

    The input to the Op Amp is at 5V constant. I used the pot at the microcontroller for the ADC and to adjust the pulse width.

    Yeah I guess I can't use this one and other car audio cap since most of them have those warning. They seemed perfect for this application as they have high voltage rating, good range of capacitance and low ESR (from others that I have seen) with that bonus voltage display. I guess I should take into consider using low rated voltage supercap and place them in series although it gives a problem of possibility of unbalanced charge in each capacitor and increase in ESR. It sure is hard to find the right capacitor :(.

    I just want to have the ease of changing the capacitor's voltage when I change the input voltage because when the capacitor's voltage is higher than the input voltage, it just stays at its high voltage so I just want a resistor to discharge the extra voltage from the cap to match the input voltage.

    For some other that I did not quoted, I'll be keeping that in mind when revising my schematic :). I don't know why but I feel like I want to apologize for this lengthy post of mines lol. I'm just glad that there are people here who take the time to read my post!
  4. #12


    Nov 30, 2010
    Aha. I just figured it out. There was a guy here last week trying to balance the charge across several lower voltage super capacitors during a 400 amp discharge. It was roundly decided that you can't get there from here. The technology has not arrived yet.

    Rather than spend the rest of my week with you, I am going to say, "Don't be a cheapskate. Buy a real welder." It will be cheaper than buying the parts to make one and it will actually work on a regular basis.

    It might be fun to try to figure this out, but, you are not going to design and build a successful commercial grade welder for your first hobby project. You can learn about the practical side of how to design around components, but I am NOT going to expect you to end up with a proper welder in a week. I suggest you ask about one section at a time. That way it won't be so scary for the people on this end.

    Seriously, it took courage for me to start this. There is so much wrong and there are so many things that need to work together that it is frightening to consider how much typing it would take to fix all of it.
    shortbus likes this.
  5. MaxHeadRoom


    Jul 18, 2013
    I haven't gone into your design in any depth, and I do not know what you intend to weld, but why not take a leaf out of commercial spot welders, use an AC supply and control the primary at a much lower current with a Triac or 2 SCR's?
    The design of which is very simple.
    You need a high current supply anyway.
    This looks more like it was designed along the lines of a capacitive discharge welding tool.
    Just my two cents.
    Oops, I see it is a CD anyway, but why not discharge the energy with a SCR as in the commercial type methods?.
  6. Leorenzo

    Thread Starter New Member

    Sep 21, 2013
    @ #12
    I agree that I made some scary looking wall of texts there and may have been long enough for others to respond to my post. Maybe this project is something I still can't handle (although looking at the prices of CD spot welder which ranges from 1000$ and above drives me to make my own spot welder).

    @ MaxHeadRoom
    The problem with the SCR as what I have seen is that once it starts conducting, you can't control the discharge time and will have to discharge most of the cap's charge. However, I think they work just right with just a lesser capacitance value so that completely discharging is not a problem. which I have seen has work for others. The schematic is very simple also.

    I guess I'll just try again with maybe a less complicated circuit :)
  7. THE_RB

    AAC Fanatic!

    Feb 11, 2008
    Like charging the cap to X voltage (giving a controlled amount of energy) then dumping it with a big old SCR into the electrodes?
  8. Leorenzo

    Thread Starter New Member

    Sep 21, 2013
    Yes! that may have been an easier solution. Maybe I'll just use a laboratory power supply for varying of the voltage then just completely discharge using SCR. :)