Current control for SMPS Arc Welding Machine

[tAhm1D]

Hi all,

I am working on making a switched mode arc welding machine with maximum 150A current and 30-40V. I have done most of the circuitry but the problem I face is the output current control. I do not want to use the bulky old linear type inductor, so what can be the solution?

I have seen in some SMPS welding machine (in schematics), current is controlled with a small ferrite core inductor. I tried to find out the calculation and procedure, but could not find out from any book or internet the required procedure for controlling output current through ferrite core inductor.

I require to control the output current to control the inrush current and also to use different type of welding rod with controlled current from 75A to 150A.

In some discussion on some forums, I found the idea of controlling output current through hall effect sensor. But nowhere it is discussed in detail. So, no concrete idea could have been formulated.

Help in this regard will be appreciated.

[Jehan]

Linear Current control in SMPS is difficult, you can use a hall effect sensor from LEM. in lower current application upto 50 Amp i had used a Current sense resistor and an OP amp OP37 as the amplifier stage , with a low offset op amp the shunt dissipation can be kept low with the sense voltage in the range of 10 millivolt, this will be a lower cost approach than a Hall based design.

[sohel]

u can read this topics for ur requerment

http://www.sonsivri.com/forum/index.php?topic=13375.0

[tAhm1D]

Linear Current control in SMPS is difficult, you can use a hall effect sensor from LEM. in lower current application upto 50 Amp i had used a Current sense resistor and an OP amp OP37 as the amplifier stage , with a low offset op amp the shunt dissipation can be kept low with the sense voltage in the range of 10 millivolt, this will be a lower cost approach than a Hall based design.

Hi,
This is current sensing, but I require to control the current flow with my requirement to use different sizes of welding rod and current control should be from 50A-150A. I think with your proposed system sensing can be done but current control can not be done. However I appreciate your reply and please let me know if any other ideas are there for switching type control. I tried with Pic Micro controller and Triac by Line Phase Angle control but in that case Voltage also reduces, which is not acceptable as Voltage should be constant for good Arc Welding.

For Sohel,

I appreciate your reply but the Link you have provided  will not serve my purpose and I have consulted that earlier.

[bbarney]

You can check Allegro MicroSystems, Inc. for hall effect products and apps
http://www.allegromicro.com/en/Products/Design/Tech_Pubs.asp#B
http://www.allegromicro.com/en/Products/Categories/Sensors/index.asp

[tAhm1D]

Hi,
Yes, I am working regarding hall effect products and think the links will be useful.

Does anybody having any idea regarding saturable reactor/core current control system?  I am working on it also and find it interesting but for designing an workable prototype requires more information/knowledge. Concrete ideas will be appreciated.

[DTiziano]

Hi

From your state " .. have done most of the circuitry" and previous post, I suppose you design the SMPS.
I saw some IC (linear technology) that use a double feedback for regulation, one from voltage or-ed with one from current. Why not use a similar approach ?
Use two signals and or then to the feedback. The voltage one will limit the maximum output the current will work during welding. More or less like a power supply when limits the output current.
The setting/adjust may be made on the reference or better on the feedback signal.

P.S.
You did not specify if your design output is AC, DC or both.

[tAhm1D]

Hi DTziano,
I appreciate your Idea. But it is not fully clear to me. If you elaborate in detail, it will be useful for me to grasp.
My design output is DC and it is SMPS type.

[beque]

Look at these patents...

http://rapidshare.com/files/180410915/Patents_-_Arc_Welding.rar

Size:10Mb

[tAhm1D]

Hi beque,

I have downloaded it and found it very resourceful. I am carefully studying it and hope will get something appropriate to fulfill my requirement. I appreciate your cooperation. If you have any other resources or knowledge regarding smps based ARC/Butt/SEAM Welding Machines or Inverter, please don't hesitate to inform. I will be highly obliged.

[okk]

Here are a few more welding power source patents (there may be the odd duplicate).

I have found them very useful over the years.

Download Link: http://rapidshare.com/files/180500549/Welding_Patents.rar

[tAhm1D]

Hi okk,
Appreciatable effort on your part to help me. This is the beauty of Sonsivri. I will study and confident that something useful will be there. If you have anything regarding saturable reactor current control system, please inform/upload.

[DTiziano]

Hi DTziano,
I appreciate your Idea. But it is not fully clear to me. If you elaborate in detail, it will be useful for me to grasp.
My design output is DC and it is SMPS type.

I did not find yet the Linear Technology reference.
Anyway, any power supply has set point and feedback, or should be.
Suppose a configuration with “voltage” set point.
Add a current measurement circuit  and a comparator with a threshold set point (current set point) that when the current exceed (current set point) will reduce the voltage set point. That is the basic idea.
A linear simple conceptual implementation can be made by OP AMP (as a comparator) driving an open collector transistor that shunt the voltage set point.
On SMPS the power stage (pulse generation) cut off for one or more cycles may be easy to do and simple to balance (loop stability).

[tAhm1D]

Hi DTiziano,

Your Idea is to Power Control. But I want output current control, keeping the output voltage constant for proper Arc Welding. I have formulated an Idea to do that with the help of Saturable Core Reactor made with Ferrite Torroid and it will be placed after the Main Transformer and before making DC, therefore before the Rectifiers. Hope to post the schematic soon. More Ideas are welcome.

[Taner]

Hi tAhm1D,
I can't understand what is the exactly electrical schematics of your SMPS welding machine.
What type is your switching power supply circuits: Push-Pull (with two power MOSFETs and one center-tapped transformer), Half bridge (with two transistors, two center tapped capacitors and one transformer) or Full bridge (with four transistors and one transformer). Depending on circuits you can use a ready PWM IC's (TL494, SG3524 or SG3525) or some digital switchmode regulator made it with some MCU's (PIC or AVR).
For output current feedback you can use one of the known solutions: current sense resistor, current transformer or Hall-effect sensor. Choice is your.
And you must maintain a constant output current not a voltage. And this can be made with PWM control of the output voltage (by the switchmode power supply circuit).
There is a one paper related with current measurement and current sensors in SMPS:
http://www.4shared.com/file/79637554/be589d7c/Current_Sensing_Solutions.html

Best regards.

[DTiziano]

Hi DTiziano,

Your Idea is to Power Control. But I want output current control, keeping the output voltage constant for proper Arc Welding. I have formulated an Idea to do that with the help of Saturable Core Reactor made with Ferrite Torroid and it will be placed after the Main Transformer and before making DC, therefore before the Rectifiers. Hope to post the schematic soon. More Ideas are welcome.

 

Hi tAhm1D
If the arc characteristic is stable you do not need any current control, constant voltage is ok and produce constant current or vice versa.
I am’t a welding expert, but quite sure they change, even only because the user change the torch distance, the equivalent result is:
- at constant voltage, the current change
- at constant current, the voltage change

I do not know any other condition.
The saturable solution you mention, should use a variable impedance device in series to the load (but for AC line only), and by changing the impedance  change the voltage on the load and so the current.
Tell me if anything wrong in my arguments.

[tAhm1D]

Hi Taner and DTiziano,
I appreciate both of you for your kind effort to help me. I am also not an expert in welding. But I am interested in the project as it will help ultimately in numerous fields in Electronics in controlling current. Let me clarify you all regarding the Modalities of my Project.

It is smps type. First, I will rectify the Line Voltage to DC and then convert it in required voltage by Full Bridge Topology and before again converting it to DC, I will channel this 100khz Ac through saturable core transformer/inductor to ensure required amount of current and then it will be converted to Dc and will be supplied.
For proper Arching, min voltage is required, it should be 40-45v for using any size of welding Rod,but with the reduction of the size of welding rod for welding less sized material, less amount of current is required. That is why, I want to keep the voltage constant to around 40v and current should be changed from 50-150A according to requirement.
Yes, saturable reactor is variable impedence device, a Ferrite Torroid Core which will have the winding of the main wire and at the same time different wiring will be there with Dc controlled supply,which will saturate the core according to requirement to control the flow of current.

Hope, I could make you understand my modality. I am at the final stage of preparation, and it will take some time to finalize.
Your all positive suggestions will be accepted with gratitude.

[okk]

I think I can be of assistance.

I assume that you wish to build a welding power source suitable for MMA (stick electrode) welding. If this is correct the power source will have to provide a roughly constant current output characteristic (I will expand on this later).

For ease of arc striking a no load voltage voltage of >50 volts would be preferable, the arc running voltage will be dependent on a number of factors (I will also expand on this).

My main concern is with the operating frequency of 100Khz.

Unless you intend to make a complex resonant ZVS or ZVC system you will have to limit the frequency to 20-30Khz for an IGBT design. Any higher than this then you will have to use power MOSFET's.

I will post some more information over the weekend

[tAhm1D]

Hi okk,
Yes, you are correct. I intend to use Power Mosfet in ISOTOP Package from ST Microelectronics if I intend to use 100khz Freq or IGBT, if I use 20-30khz. You have understood my concept clearly. I will be waiting for your reply.

[tAhm1D]

Hi All,
I am almost at the final stage of my Project. Here, I have used the Saturable Core Reactor,to control the Output Current Flow. The principle of which I tried to show with the following schematic.

         

Controlling the AC current across the lamp L with the DC current,
regulated by variable resistor R. B- battery, G - AC source.

Still I am facing some problems to finally control the output current. Further suggestions are invited to overcome the problem.

[okk]

Sorry for the delay in replying until now tAhm1D.

With an output current of 150 amps you will easily be able to weld with 3.2mm electrodes and you will just have enough current to weld with 4.0mm electrodes.

The power source will have to provide enough voltage to maintain an arc at the maximum output current. According to specification EN60 974-1 the required voltage at 150 amps will be 26 volts.

This is given by: -

V=0.04 X I_max +20 (for MMA only)

It is possible to to weld with lower output voltages but it becomes more difficult to maintain the welding arc. 

Maximum output power will =  3900 watts.

With an overall conversion efficiency of 80-85% the primary inverter will have to be able to provide about 4500 watts of power (a lot!).

Have you thought of using a half bridge design? I have been involved in 3 phase half bridge designs producing more than 25 kilowatts out. A half bridge design will certainly be easier to implement (at least that's my experience).

One more question, I assume the input voltage will be 230 volts single phase?

[tAhm1D]

Hi okk,
I have studied your reply and found it useful. I am thinking about 230v or 380v Ac input or both. I have planned to use Full Bridge Topology but after your suggestion, will think regarding half bridge also. I have decided to put no load voltage= 45v and with load according to requirement. If you kindly provide me voltage and current requirement for Arc Welding (MMA) according to specification EN60 974-1 or source,from where I can collect the data, it will be of great help for me. I am just calculating basing on datas provided by an Expert welder. Hope for your further help.

[tAhm1D]

Hi Okk,
I have considered using half-bridge topology, but there are some disadvantages to using half-bridge.
Opposed to full-bridge where peak current is 1.4 times nominal current, in half-bridge topology peak current is 2.4 times nominal current. So here the MOSFETs will have to handle greater current.
But the main disadvantage I find is that, half-bridge topology relies greatly on the input bulk capacitors as opposed to full-bridge where  the circuit relies on all the MOSFETs. Thus I find it is riskier to use Big Capacitors at such high powers and MOSFETs are safer and cheaper as well.
So my decision stands at using full-bridge topology. I will be looking forward to your suggestion.

[tAhm1D]

Hi,
Though I am highly inclined to Saturable Core Reactor method to use in Current controlling, I am also getting some knowledge regarding the latest mode, that is Hall Effect Sensors. From bbarney's Link in this Post regarding Allegro Microsystem and the beautiful Book on Hall Effect Sensor posted by good friend ravenfeather in the link:http://www.sonsivri.com/forum/index.php?topic=20093.0  helped me to formulate a concrete Idea regarding current control through Hall Effect Sensor. I am working on it and hope to get more ideas/suggestions in this regard, so that I can utilize those in my Arc welding Project.

[abcsantosh]

Hi tAhm1D, I am also working on inverter power supply for tig/mma welding. Currently I have built the prototype smps of 1500W. Today, I am going to test it and verify all design parameters. I have not yet implemented the linear current control for simplicity. But, sooner or later I am going to work on it. For prototype I am using fix output current control using current sense transformer. The circuit is comprising of UC3879 main control, Ir2110 driver and center tap output configuration. On this center tap winding, I have placed the current transformer. In future, I am planning to change the CT's fix gain to variable gain using Opamp, to achieve variable current control. Yet I have not implemented it, but I thought it is the best method to control output current. As for current sense purpose you can use any method appropriate for your application. But I think, CT is best for large current output.

[MAXPAYNE]

If u r sensing current by placing CT in the primary side, then u can't accurately control output current.u can place a CT for over current protection but for accurete current controlling u have to sense the output current directly. r u using full bridge topology ?

[tAhm1D]

Hi abcsantosh,
What is the frequency you used? Is your test completed? What is the efficiency of your inverter? What Topology you have used? If you don't have any objection, you can post your schematic. I will also post my schematic when I finish my final tie up. Please keep up your good effort.

[abcsantosh]

Hi tAhm1D, I am using main fullbridge stage at 200Khz. The estimated efficiency is 80+ because of phase shifted ZVT . The topology above 1KW is of course Full Bridge. Yes, although I have tested the prototype successfully but there is still some problems. Too much noise, voltage spikes are generated. So that I am testing it at 25 % load. After correction of switching waveforms I will go to full load testing. The schematic is still on papers only. I am going to verify some basic building block. If every thing goes well sure I can. (Actually, I am developing it for commercial purpose) :-) May be I can show you photographs of test assembly.

[tAhm1D]

Hi abcsantosh,
Please don't disclose anything which is related to commercial interest. My Topology is also Full Bridge. I am actually interested regarding your output current control process so that for different requirement, different Amperage can be supplied to the output. Keep apprised regarding your problems/success with design parameters so that we can consult for upgrading those. Mutual help will ease our burden.

[abcsantosh]

Hey tAhm1D, your understanding is appreciable. We can share our knowledge. To help you, I found this Image on net. http://www.china-power.net/psta/dzkw/504/tu/yj2.bmp. Take a look at it. I think this is the best solution to fit your needs. Soon I will draw a block diagram to suggest you secondary side current sensing to connect to above mentioned Control scheme. First I will have to study it. :-)

[newboy]

Too much noise, voltage spikes are generated. So that I am testing it at 25 % load. After correction of switching waveforms I will go to full load testing. The schematic is still on papers only.

How do you know there are voltage spikes generated but you said 'The schematic is still on papers only' ? Do you simulate it in PC ? What software you use simulation ... hope I can learn from both of you.

Just newbie in this area

[abcsantosh]

newboy, voltage spikes are generated. I know because I tested and viewed it on O'scope. 'The schematic is still on papers only' - By this statement I mean schematic is not final, although I build it, test it. but, it is at conceptual level. may be I can change the part of circuit or completely change the design. I do not simulate it on PC because I don't have enough of spice simulation (Spice models of Components). You can get the SPICE Simulation Fundamentals Series by following this link. I hope it is useful for you.

http://zone.ni.com/devzone/cda/tut/p/id/5413
 

[abcsantosh]

During the tests, I blew up the Full bridge board. Actually it is my mistake. I am trying to improve the waveforms. In this attempt, I remove the fast turn off diodes of mosfet gates. I thought it might slow down the turn on and turn off. Attached are the Photographs of my Test Prototype.

[tAhm1D]

Hi,
Proper simulation of smps circuit as close to practical one is very difficult to formulate and I have not found one that useful. I have studied the schematics. Good effort. You have blown up the circuit- that means you are in the right track. It is a proverb in China that the person who has not destroyed minimum 1doz Fets and 1doz Drivers, can not earn proficiency in Full Bridge circuit in smps. I am now very confident with this type of circuit but I had  destroyed many. I find that,if properly use, this type of circuit is very reliable. You are probably testing with less power as I could see the Transformer you use is small. What type of transformer you used? EE or ETD? What is the number? One advice to you if you don't mind. When using frequency > 150 KHz, be very careful to use Flux density much less than higher flux density possible for that core and calculate turn ratio basing on that or your core will attain saturation easily and the Fets and circuit will be destroy. Another thing- when you use> 100KHz Frequency, check whether the Freewheel diodes inside the Mosfets can withstand that speed. Better to use 4 more high frequency diodes with 4 Mosfets as freewheel diodes, discarding internal ones. Please let us informed regarding your progress. Hope for the best.

[abcsantosh]

 Hi, you suggested very good points. I have to use external freewheel diode. I am using IRFP460 and it's internal body diode is bit of slow. The full bridge transformer is EE42/21/20 N87 material from Epcos. Also the pfc inductor is built using same core with air gap. I am using the cores considerably inside the boundaries.   The core underneath the big capacitor and in front of mosfets is the full bridge Tx. Output power is 1500W, but when I tested, I put the load for 25% power, as you can see from photo. The green one is my load resistor. I will make the changes and tell you about further results.

[Walkura]

Good evening Tahmid .

I have a few schema's for you ,but unfortunatly the text is in Polish and Russian .
On the other hand i know your understanding of smps's by now so i'm confident you will understand by the schematics itself.
The first one is from a Polish magazine ,its a article where they build a welding inverter .
http://www.ep.com.pl/pobierz.php?id=12163
http://www.ep.com.pl/pobierz.php?id=530
The schema is complete with pcb ,for translating segments of the text you can use the poltran website .
(Sorry but since Polish isn't my native language, i can read some of it but my Polish is far from perfect)
The second schema is from a russian website and its also a welding inverter .
http://imajr.com/Original.aspx?Id=spawark1100_393678
Somewhere in the coming week i will post another schema (although that would be photo's from the magazine)
Last month they published a welding inverter in another Polish magazine but the second part of the article comes out this month .
When i have the second part i will post the schema .
Last but not least i have some other schema's of welders (mostly Polish or Russian) but they are on another computer .
Those i will also post when i get to my other computer monday .

Good luck and be carefull .

[tAhm1D]

Hi Walkura,
Good material. The Russian one seems to be very interesting. I am studying those. If you have more schematics, please post. If you have any other material regarding Polish and Russian smps and welding inverter, please don't forget those also. I will consult you regarding the schematics you posted after detail study.

[Walkura]

Good evening Tahmid .

Sorry i reply so late but i didn't have a chance earlyer .
I uploaded you another welding inverter ,again its in Polish but its a well documented schema with pcb,s and pictures .
http://rapidshare.com/files/204508135/dokumentacja_spawarki_inwertorowej_140a.rar

Some time ago you asked about calculating current transformers .
In the book bij Ron Lenk they explain how to calculate and design them (chapter 5 from what i recall) .
The book was already uploaded here so i didn't put it again ,if the links are dead just let me know and i'll reupload it .

Also i would consider to use those allegro sensors i use them a lot and am very pleased by them .
(i would use them primary side followed by opamp to fit it to the smps controller)
Saves you the hassle of calculating a current transformer .

When i found the rest of the schema's i'll upload them here .

Good luck and be carefull .

[9thwonder]

why not to use a DCCT which were used in the sine wave inverters for load sencing.
i think this is better solution to sence and control dc current.

[MAXPAYNE]

HI 9thwonder,
       please give details info abt DCCT.....

[tAhm1D]

Hi all,
If I use AC phase control to control current in welding machine, is there any problem? In one German schematic, I found that in a DC smps welding machine, they use high current but very low voltage(As low as 10v) but in AC type of welding machine,voltage less than 30v create problem in Arching. I am not sure whether in DC type welding machine, that type of low voltage is enough or not. Can anybody provide any suggestion/idea on this?

[tAhm1D]

I did not find yet the Linear Technology reference.
Anyway, any power supply has set point and feedback, or should be.
Suppose a configuration with “voltage� set point.
Add a current measurement circuit  and a comparator with a threshold set point (current set point) that when the current exceed (current set point) will reduce the voltage set point. That is the basic idea.
A linear simple conceptual implementation can be made by OP AMP (as a comparator) driving an open collector transistor that shunt the voltage set point.
On SMPS the power stage (pulse generation) cut off for one or more cycles may be easy to do and simple to balance (loop stability).

Hi,
Your Idea is correct. I am trying this using shunt resistor and external comparator driving the compensation of SG3525. I have tested a simple circuit like this and found to be working properly according to the Oscilloscope reading. If actually it works properly,it will be an easy and simple method to control current.

[tAhm1D]

Hi all,
I have successfully applied the Phase Angle Control and Pulse Skipping Method in Input side for effectively control the current in SMPS Arc Welding Machine. Both process works satisfactorily. I applied these methods successfully in Butt Welding Machine and Spot Welding Machine also. These processes are not for precision control. can anybody suggest any other process for precision control of Current in SMPS Welding Machines?

[lillbear]

But I think, CT is best for large current output.

or

http://en.wikipedia.org/wiki/Rogowski_coil and from IEEE there is lots of papers about Rogowski Coil.
And commercial one ->http://www.powertekuk.com/cwtmini.htm
if we are talking about large currents.. 

yours

[tAhm1D]

Hi,
Current control is a process which have 2 types of mechanism. One is sensing current and other is controlling the amount of current according to requirement. We can control the upper limit of current by using shunt resistor and external comparator driving the compensation of SG3525. Or we can sense the current by Rogowski Coil as stated by lillbear and many other process. But the second part is important- that is controlling current precisely according to requirement. Precise control of current in the circuit is the name of the game. That is exactly require for application precisely in many smps circuits like micro butt welding machine etc.

[alien]

Hi Tahmid,
Please provide some more details on your phaseangel controlled/cycle skipping based spot welding implementation....I just completed a project on same....commercially available units have many features....many programmable schedules is among one of them...

[arash_tah]

Hi Tahm1d
Did u reach final idea? Current control in SMPS is a very chalengable issue and it takes u alot of effort to achive the gole please inform us what stage are u in? I also want to say u to use current mode control I mean average current mode controlin this control mode you will have a system that inherently limit the current and protect your switches have look at ti.com and search average curent mode control I am sure u can fine mane helpful articles

[tAhm1D]

Hi Tahmid,
Please provide some more details on your phaseangel controlled/cycle skipping based spot welding implementation....I just completed a project on same....commercially available units have many features....many programmable schedules is among one of them...

Please elaborate regarding your project. I hope to provide some details regarding my application of phase angle controlled/cycle skipping based spot welding implementation soon.

arash_tah,
I have not decided finally but testing different processes in practical application. You are correct that by using current mode technique, current control can be done. I will test with those soon and will test with current mode I.C. and with Pic microcontroller.