Some questions about my SMPS TIG Welder build...

[blasphemy000]

I was just looking over the datasheet for these NCP1654 PFC Controllers and the gate output of these devices are meant to directly drive a MOSFET or IGBT. They can source/sink 1.5A and are in a totem-pole configuration. So for my lower power model I will probably just drive the gates directly from the PFC controller(Which is how it's done in the datasheet and application notes examples) and in the larger power controller I've ordered some MC34152-D SOIC-8 devices which are dual 1.5A(Source/Sink) low side gate drivers for use in PFC and other applications.

I'm currently working on the calculations/schematic/board design of the 1000W(400VDC@2.5A) PFC design that I will be using as a test-bed/prototype. As soon as I get everything drawn up I will be posting the 1000W design as well. My chips should be here in a few days so I can get started working on this...I'm also looking for a decent LCR meter so that I can wind my own inductors for these PFCs because the large inductor(s) that I will be needing for the full-scale device just can't be bought.

Also, MicrosiM, I really like your idea of building an interleaved PFC for the full-scale device. I think dual or even quad PFCs interleaved together might be a better choice than trying to build one massive PFC just due to the crazy inductor currents that I have calculated for a single PFC stage. So I will be looking into this option today as well.

 

[blasphemy000]

Well the calculations are done for the 1000W(400VDC@2.5A) PFC using the NCP1654 controller that I will be using as a prototype. Here are the results.

Vin,rms: 240VAC
Vout: 400VDC
Iout: 2.5A
Fsw: 65KHz
EFFmin: 92%
Vp-p,max: 10V (Output Ripple @ Max Load)
Pout,max: 1000W
Vin,ll: 200VAC

(1) - Coil Selection
Lcoil: 560uH
Icoil,rms,max: 5.435A
Icoil,max: 8.839

Cbulk: 1000uF/450VDC


(2) - Feedback Arrangement
Rfbu: 3.7M (1.5M + 2.2M)
Rfbl: 23.2K
Cfb: 100pF - Non-Polar
Cp: 0.22uF - Non-Polar
Cz: 2.2uF - Non-Polar
Rz: 12K


(3) - Input Voltage Sensing / Brown-Out Detection
Vac,on: 199VAC
Rbol: 84.5K
Rbou: 18.2M (9.1M + 9.1M)
Cbo: 0.47uF - Non-Polar


(4) - Current Sense Network
Rsense: 0.05R/15W/1%/TO-126 Case
Rcs: 2.4K
Rm: 110K
Cm: 680pF


These will be the starting values that I will be using to build the 1000W PFC prototype/test-rig. I just started drawing up the schematic and hopefully I can get the board laid out this evening. As soon as I get these things done I will be posting them and in a few days when my controllers and other stuff arrives I will be building this unit and will be proceeding with the testing and will post all of the results. I'm also trying to come up with a way to load-test this PFC and also the full-scale(16-20KW) unit. Load-testing this 1000W unit shouldn't be too hard, except for I want to be able to vary the load, so I'm still trying to work out a good way to do this. If anybody has any ideas on how to build a 1KW-2KW variable load please let me know. The load must be able to function at >400VDC.

 

[MicrosiM]

Hope this will help you also...

This was my first PFC project using UCC28019.

[video=youtube;_cGWwleeb3w]http://www.youtube.com/watch?v=_cGWwleeb3w[/video]

 

[blasphemy000]

Nice project MicrosiM. I'm very happy that you are interested in my project and I'm learning a lot from the things you have posted to help me along. Thank you.

I'm still trying to figure out how I'm going to load test the 1000W PFC prototype that I will be assembling later in the week when my parts arrive. I have a couple ideas that should work, but I really have absolutely _NO_ idea on how to load test a 16-20KW PFC for the full-scale version if all goes well with the 1000W device. Any ideas for this? Also, I'm leaning more and more towards the interleaved PFC idea that you proposed. I think that (3) 6KW or (4) 4.5KW PFC devices will be much easier to build/control than one massive device. Also, if I use 3 or 4 interleaved PFC devices(all powered from the same controller), that will drastically reduce the size of my boost inductors as well as significantly lower the peak and RMS currents flowing through each of the boost inductors. I thank you very much for proposing the interleaved PFC idea.

 

[blasphemy000]

I also have a smaller SMPS, around 1000W@100% Duty-Cycle, 1500W-Peak, that I'm designing as well. I would like to make it a Half-Bridge ZCS Quasi-Resonant design so that I can run a higher switching frequency to increase the efficiency and reduce the switching losses. Do you have any recommended reading as to how to design a ZCS supply. I have some info on the subject but it's very limited as far as being "in depth" discussion. I'm having a hard time finding good information about how to choose the components for such a design. I'm always trying to learn new things and I would like to learn as much as I can about SMPS design. I feel that the best way for me to learn is by building them. Thank you for your help.

 

[blasphemy000]

That UCC28019 controller seems to work similarly to the NCP1654 controller that I'm going to try. Seems like another good candidate to try if the NCP1654 controller won't handle what I need it to.

Also, I have a question. Some of these example PFC circuits I've seen have inductor bypass diodes and some don't. I understand that they are used to charge the bulk capacitors to the max voltage that can be achieved before the PFC is turned on and this helps reduce the inrush current since the PFC only needs to charge the caps from say 340VDC up to 400VDC instead of having to charge them from nothing. Once the voltage across the bulk capacitors rises above the peak voltage coming out of the bridge rectifier connected to the line then this diode becomes reverse biased and no longer has any current flowing through it. For the types of PFC that I'm building, both the 1000W and the massive one I have planned, would you recommend that I use one of these inductor bypass diodes to provide the initial charge to the bulk capacitors? Is there any disadvantage to having this diode in the circuit?

 

[MicrosiM]

Nice project MicrosiM. I'm very happy that you are interested in my project and I'm learning a lot from the things you have posted to help me along. Thank you.

I'm still trying to figure out how I'm going to load test the 1000W PFC prototype that I will be assembling later in the week when my parts arrive. I have a couple ideas that should work, but I really have absolutely _NO_ idea on how to load test a 16-20KW PFC for the full-scale version if all goes well with the 1000W device. Any ideas for this? Also, I'm leaning more and more towards the interleaved PFC idea that you proposed. I think that (3) 6KW or (4) 4.5KW PFC devices will be much easier to build/control than one massive device. Also, if I use 3 or 4 interleaved PFC devices(all powered from the same controller), that will drastically reduce the size of my boost inductors as well as significantly lower the peak and RMS currents flowing through each of the boost inductors. I thank you very much for proposing the interleaved PFC idea.

I am not only interested into your project, But I would like to see it become reality.

You are doing great job by starting from scratch, taking risks, also on the right track, witch is not easy.

You can use heater elements as a load, you can get some from junk parts, I used same method to test 2KW PFC, and for my current PFC, will use same thing.

I advice to start with 500W interleaved PFC then go to KW levels, these test are not easy and one mistake will be deadly and very costly, so take ultra care when you work !

I am waiting your progress, and if you need help just feel free to ask, this is like your home

Good luck again, and this is massive work indeed.

 

[blasphemy000]

Thank you for the kind words, and very warm welcome into the community here, Microsim. I love doing things the hard way and it usually ends up working out, maybe not the first time, but eventually it all works out.

I do have a whole bunch of heating elements here, some small ones I usually use to test my SMPSs, and some very large ones from 240VAC electric water heaters. The large ones I figured I could wire 2 in series to give a higher voltage rating to load down the 400VDC PFC stage. The only problem is I need to be able to vary the load from nothing to full load and back again because while you're not welding the switching stage will be off so there will be no load on the PFC. When you start welding, the current will ramp up in < 1S and when you stop welding the current will drop to 0 almost instantly. This PFC chip is supposed to be able to handle this. The load instantly dropping to zero will cause the output voltage of the PFC to rise which the PFC switching device will immediately shut down preventing a voltage spike. To make sure this works like it is supposed to I have built a really thick plexiglass "blast-shield" to house this PFC for the initial load testing in-case the voltage does spike to much and the caps explode. I'm all about safety when it comes to these things.

Hopefully before the end of the day today I will be able to post up the design for my 1000W single-stage PFC. I got the schematic done but got too tired yesterday to finish the board design. The board is done with SMT components. I'm using the larger 1206 size resistors and caps and some Panasonic "Size D" SMT electrolytics as well. I figured I'd use the slightly larger components since I've never worked with SMT stuff before.

 

[Virus]

Hi Blasphemy000

I take my hat of to you, just getting in and gritting down and getting things done, you are an inspiration to me definitely in getting my little project going.

Will be keeping my eye on this to the end, it is nice to an adventurist person believing in what he is doing.

Still don't know what PFC stands for

Theunis

 

[MicrosiM]

Thank you for the kind words, and very warm welcome into the community here, Microsim. I love doing things the hard way and it usually ends up working out, maybe not the first time, but eventually it all works out.

I do have a whole bunch of heating elements here, some small ones I usually use to test my SMPSs, and some very large ones from 240VAC electric water heaters. The large ones I figured I could wire 2 in series to give a higher voltage rating to load down the 400VDC PFC stage. The only problem is I need to be able to vary the load from nothing to full load and back again because while you're not welding the switching stage will be off so there will be no load on the PFC. When you start welding, the current will ramp up in < 1S and when you stop welding the current will drop to 0 almost instantly. This PFC chip is supposed to be able to handle this. The load instantly dropping to zero will cause the output voltage of the PFC to rise which the PFC switching device will immediately shut down preventing a voltage spike. To make sure this works like it is supposed to I have built a really thick plexiglass "blast-shield" to house this PFC for the initial load testing in-case the voltage does spike to much and the caps explode. I'm all about safety when it comes to these things.

Hopefully before the end of the day today I will be able to post up the design for my 1000W single-stage PFC. I got the schematic done but got too tired yesterday to finish the board design. The board is done with SMT components. I'm using the larger 1206 size resistors and caps and some Panasonic "Size D" SMT electrolytics as well. I figured I'd use the slightly larger components since I've never worked with SMT stuff before.

The minimum load trick may solve things, once you reach that stage we will discuss it

 

[KX36]

This is a mega project and I can't wait to see the results! I admire your guts, I certainly don't know anyone who's made something on this scale.

I'd agree PFC is probably a must here so that you don't draw your current in huge surges. You can scale most of these circuits up as long as you can get the components to handle the currents.

I can't immagine hard switching at these currents, but the alternative requires a high minimum load. Incidentally, what's your minimum load going to be at that power level? Have you considered parallelling up several smaller converters?

Soldering SMD components is the least of your worries compared to dealing with the massive currents in this thing. Here's a TSSOP14 I soldered to a breakout board myself without magnification https://dl.dropboxusercontent.com/u/32751654/P1015078_crop.jpg the pin pitch is half that of SOIC8. It's the first SMD solding I had done and took me about 30 mins mainly because I forgot the flux the first time and I was unnecessarily fussy about alignment. SOIC is barely any more difficult than DIL. The last SOIC8 breakout board I did took longer to get the header pins in than the chip on because they were too close to the chip for me to get the iron in well. Flux and soldermask do all the work.

Best of Luck!
Matt

 

[MicrosiM]

Any good news?

 

[blasphemy000]

My parts should be here Monday according to UPS. I keep making tweaks and changes to the schematic and board which is why I haven't posted anything yet but I think it's getting close to a version I can build for testing...

 

[MicrosiM]

I was thinking all day about the right method you should use for the PFC, since its going to be painful process to use Mosfets for this huge power, this is for many reasons.

1- Cost, Mosfets are very expensive 50A and above

2- It wont be easy to control too many Mosfets in parallel.

3- using high switching frequency in the PFC is also a problem.

I think I came out with the best solution for this huge PFC.

At first, lets see how things will go from your side, since you already ordered parts.

It would be great idea if you let me know witch cores you have chosen for your PFC, size etc...

Wire type for this?

Regards

 

[blasphemy000]

Well I'm still working on the calculations for the massive PFC. Right now I'm concentrating on the 1000W(400VDC@2.5A) prototype test bed. I'll be using IGBT switching devices, they're cheaper in high-current models, the newest generation of IGBTs can do hard-switching relatively fast, and I just personally think IGBTs are better in a welder do to the way a welder works.

For the 1000W PFC prototype, the controller I ordered has a 65KHz fixed switching frequency. For the IGBT I'm going to try a HGTG20N60A4D with a built in anti-parallel hyper-fast diode. This device has very fast on/rise/off/fall times and according to the graph in the datasheet, it will handle 30A@65KHz and 50% duty-cycle. It also says this device has been specifically designed and optimized for high-speed switch mode power supplies so I think it will fit the bill. I used an inductor calculator to size the inductor for this 1000W prototype and it suggested using a PM62/49 core based on my inductance and current requirements, although it didn't say how many turns or the wire size I should use.

The controllers and gate drivers will be here today and as soon as I put money in the bank I will be ordering the other parts I will need. I got the schematic completed but I'm still working on the circuit board design so this could still take a couple weeks before I get this built.

 

[blasphemy000]

Almost done with the board layout for the 1000W PFC. It's proving to be a challenge. I'm still learning the Altium software package and all its features but it's coming along nicely. I only have a couple hours a day to work on it which is why it's taking so long.

As for the full-power model of the PFC that will eventually make its way into my welder design. I've decided to go with 4 interleaved PFC stages, each capable of 400VDC@12.5A which should yield a combined output of 400VDC@50A peak. I've decided to go with the interleaved design for many reasons. It had a lot to do with the huge currents and massive inductors that would be required for a single stage. With 4 smaller stages, the inductor currents are only 23-27% of what they would be in a single stage and the cores are much smaller and more easily attainable for the interleaved version. Depending on how physically small I am able to make the interleaved stages, I might even go with more smaller stages, because with 4 I'm assuming they will still be quite large. I'm sure there will be a point where adding more stages will no longer result in a smaller footprint, but I'm going to design a few different possibilities on my computer until I get the results I want.

Again, as I said before, this project is probably going to take around a year or longer to complete which is why I don't really have much to show right now. All I can say is please be patient. I've been thinking about building this machine for a couple years already and I finally feel knowledgeable enough to be able to move forward and I'm always learning more and changing designs. I just want to apologize that this is going to take so long to make a reality but I want to make sure that I get it right the first time and blow up as few components as possible.

 

[joryds]

hello blasphemy000 , I am aware of your prototype is fine.

If you need help or want a component Altium, can you tell me maybe I can help

Greetings ...

 

[daniel002]

Hello from Germany,

Indeed very interesting project, not only the pfc but whole the SMPS. The ideea of an interleaved PFC for such a high power is very good. Also the main converter can be interleaved. For this you can syncronise the paralleled converter to be out of phase.
I am currious, how can you draw so high current from only one phase of the main grid? If this will be powered from three phase, it may be that you don't need a pfc.

All the best,
Daniel.

 

[blasphemy000]

Hey everyone,
Just wanted to apologize to everyone that has been following this thread, sorry I haven't been around to provide any updates. I recently lost my production job(I worked in an electronics factory soldering together the controller circuits for high current SCR switching devices, no clue what they were used for...I just did the soldering), and had to go back to driving long-haul tractor-trailer(was the only job I could get quickly). With this new driving job, I am out on the road driving for 12-17 days at a time, then I am home for 3-4 days, then right back to driving again. This being away from home so much has really put a damper in this project, I haven't even gotten around to building my smaller prototype unit yet. Being away from home for so long, when I do come home, there is so many other things that I need to take care of, I haven't even set foot in my lab for 2 months. I have a couple job interviews lined up this coming week as I managed to get a whole week off from driving. I'm hoping that I get hired at one of these places that way I can go back to working night shift and being home every day after work.

daniel002: I've recently(a couple months ago) made changes to the design of my switching stage so that it will be interleaved as well. I was(before I had my job change) discussing the design of the system with another member here via email and I decided to split the switching stage into 6 parallel stages because my original design of one massive stage was just completely out of reach and would have been impossible for me to get working due to the massive currents involved. As for the current draw from the 240VAC mains. That's a non-issue here in the US. My old welder(which was transformer based), when cranked up to the max, drew 100Amps from a single phase of the mains. The 240VAC here in the US is actually done with a split-phase system. Our normal wall socket voltage is ~120VAC and is measured between a single phase and a neutral wire. The power comes from the grid in 2 phases(180* opposed) at 120VAC each(between phase and neutral). When using 240VAC here in the states, we use the voltage between the 2 opposed phases and ignore the neutral wire. So as long as I'm using a PFC on my input stage, I can easily draw 50-60Amps from the 240VAC mains with no problems, but the PFC is an ABSOLUTE MUST HAVE. There is just no way possible to draw that much current from the mains without using a PFC. If I were to just rectify and filter the incoming mains power, the harmonics of the current draw from the mains would be so bad I'm sure it would interfere with all kinds of things and that's not what I'm aiming for, it also might upset my electricity supplier.


Again, I'm sorry for my absence and I will return to the forms as soon as I retain local employment and I can get back to working on this project again.
-Brad

 

[KX36]

Here's a video you might find interesting in terms of scale, a 40kVA UPS

http://www.youtube.com/watch?v=-R47zoV2uxI