New Induction Heater circuit with no center tap.

[tcmtech]

The idea of stepping up to a real H bridge with proper driver H/L side driver IC's just scares the crap out of you huh?

I thought it was possibly due to the circuit complexity but looking at that schematic clearly that not the case so what gives?

Psychological block or stubborn pride? What are you so afraid of in using a H-bridge circuit is what I am asking?

 

[gary350]

I bought 2 IGBTs 18 months ago thinking this will do the job. Never used the IGBTs been in too much pain with bone degeneration, pain pills and other medication every day it is very hard to get motivated anymore. It is hard to stay on task these days too. It is less painful to type with 1 finger than to type like I should with all 10 fingers. It less painful to think about building a project than actually doing it. Some days i am better than others but i rarely get more than 1 hour work done per day if that much. Some days i double up on medication then I can get more things done but doctor says that is bad for my liver. So what I am going to die anyway some day. LOL. I will look on ebay see if I can find more IGBTs with the same part number. I notice today data sheet says my IGBTs are only good up to 25KHz. My experiments with mosfets show 85KHz is very good much higher Hz there is a skin effect problem it gets hot on the metal surface but not down deep inside the metal. Now I need an H circuit for 4 IGBTs. I my be slow and getting slower but I am not giving up yet. It took about 18 months to find the correct combination of medication not to need the wheel chair every day.

 

[spec]

Very sorry to learn about your medical position Garry, and I gotta admire you dogged perseverance.

But can I say that TCM is right: you really need to get the system design sorted before you start building/testing. That way you will have more success with your projects. Hope you didn't mind me saying this.

spec

 

[tcmtech]

Yes. Sorry to hear as well.

 

[gary350]

I ordered 2 more IGBTs on ebay same brand name PRX. All i need now is a circuit drawing that works.

I put 2 of the IGBTs in parallel and the other 2 in parallel also. Hope this will work with 1 diode for 2 IGBTs? I am not sure the 1N5819 will work on IGBTs?

The 170VDC is the 120VAC wall outlet after the bridge rectifier and filter. 15a to test later i can use the workshop circuit box and have any thing up to 100 amp. I can also have 240VAC from the circuit box that will = 340VDC after the rectifier and filter. A voltage doubler circuit can make 680VDC. I need to get it working on low power with cooling first.


**broken link removed**

 

[tcmtech]

A common H-Bridge circuit design in many high powered inductive loads puts the main capacitor in series with the inductor so that it can work as a capacitive impedance current limiter of sorts.

It's a common design in inverter welders and some plasma cutters so that simple drive frequency changes will allow for stable and simple current limiting or regulation of the output.

There is also a variant of that can be made where the second half of the H-Bridge uses two capacitors in place of the second two switching devices.

Also from what I see in your circuit it wont work as you have it drawn. You cant run a high side switching device off the same driver as a low side switching device is connected to. You have to use dedicated highside drivers to run the upper devices.

 

[spec]

Another variation is a half bridge driving the inductor via a single serial capacitor, with the other end of the inductor connected to 0V.

And yet another version is where the a half bridge drives the inductor direct but the bridge has plus and minus supplies.

By the way, a non center-tapped inductor makes maximum use of the primary winding copper, while the center tapped version only uses 50% of the copper.

spec

 

[spec]

**broken link removed**

Gary, you need two low side/high side driver chips. It is not a good idea to drive two gates with the same driver line.

You would also be wise to provide a dead zone in the gate driving waveform. Dead zone is where all bridge transistors are off. This increases efficiency, gives the bridge transistors an easier time and reduces current spikes in the power line, like the top inductor would in the circuit above.

spec

 

[tcmtech]

This is how Miller designs their high powered H-bridge systems in their equipment.

From here.
https://www.millerwelds.com/files/owners-manuals/O146741B_MIL.pdf

As you can see the output transformer's primary is in series with a capacitor (a pair of them actually being this unit runs on a ~600 volt DC rail voltage).

On their 230 VAC only input units all the way up to the XMT 450 series, that can do up to 600 amps @ 38 volts output (~23 KW output power) , they use the same .34 uF 1000 VDC nonpolar poly capacitors for C1 and C2 but in parallel for the lower input voltage.

 

[spec]

This is how Miller designs their high powered H-bridge systems in their equipment.
View attachment 102090

From here.
https://www.millerwelds.com/files/owners-manuals/O146741B_MIL.pdf

As you can see the output transformer's primary is in series with a capacitor (a pair of them actually being this usint runs on a ~600 volt DC rail voltage).

You can see what a fully production design looks like from that schematic- lots of extra compensating, shaping, protecting, and optimizing components. This is the difference between a theoretical circuit and a practical realisation.

spec

 

[spec]

Gary,

The TI UCC21520 and UCC21521 are dual output chips designed for driving the gates of MOSFETs and IGBTs. For an H bridge two chips are needed so that there is an individual drive signal for each transistor. The drivers are isolated from the input, and the chips provide dead-time control. These chips are state-of-the-art and would help you to get a high-performance, reliable induction heater working. They cost around, $6.00US each.

My advice is to use a separate isolated power supply module for the bootstrapped high side power rails, rather than use a bootstrap voltage generator as is normal.

As far as I have been able to determine, the chips are identical, apart from the sense of the inputs on pin #5: 'Disable' (UCC21520), 'Enable' (UCC21521).

DATASHEETS
https://www.ti.com/lit/ds/symlink/ucc21520.pdf
https://www.ti.com/lit/ds/symlink/ucc21521.pdf
APPLICATION REPORTS
https://www.ti.com/lit/an/slua778a/slua778a.pdf
spec

 

[tcmtech]

Now for the real head scratcher.

On Millers larger capacity and dual voltage welding units they use two H-bridges but only two pairs of High/Low dual switching device units rather than four as shown here for a typical XMT 300 suitcase welder unit but also applies on their higher capacity units as well.

From here. https://www.millerwelds.com/files/owners-manuals/O2205_MIL.pdf

Each H-bridge use two switching devices in one package for the active half of the H bridge and a pair of large HV poly capacitors to make up the second passive half of each H-bridge circuit so depending on which input voltage, 230 VAC or 460 VAC they are either running in parallel or series.

Now from a engineering standpoint that's a pretty clever way to get dual voltage operation in a HV high power H-bridge based SMPS system!

However from a experienced service techs standpoint I absolutely hated working on them being they were extremely prone to catastrophic burndown of the whole power and control systems. When one died everything related to power handling and control inside the unit got replaced.
There were no partial fixes. Ever. Input rectifier, both switching devices, main contorl board, output rectifiers and usually every power system related capacitor had to be changed out as well putting the off warranty repair bill as high as just buying a brand new machine.

I have two of their big uits sitting on a shelf in my shop that I could probably make one good uinti out of but I have never gotten around to it simply because to do so would require a total machine teardown to bare frame for both plus more than likely a few hundred dollars in aftermarket parts anyways just to have one functional machine that could still burn down again at any time.

 

[spec]

... and a pair of large HV poly capacitors to make up the second passive half of each H-bridge ...

I have had some trouble using capacitors for high voltage and high current, mainly because of cost constraints. But in the end we had to bite the bullet and get really expensive capacitors. What are the capacitors like on welders that you mention above? Are they physically large? Do they fail often? Are they ridiculously expensive. When running do they get very hot?

Apologies for all the questions.

spec

 

[tcmtech]

They are HV low ESR poly film units that are roughly the size of a pop can and cost $150 - $200+ each new.

Pretty beefy and spendy considering you're only dealing with a by the numbers .5 uF 1000 volt or less nonpolar poly film capacitor.

 

[spec]

They are HV low ESR poly film units that are roughly the size of a pop can and cost $150 - $200+ each new.

Pretty beefy and spendy considering you're only dealing with a by the numbers .5 uF 1000 volt or less nonpolar poly film capacitor.

Thanks- Wow components to be avoided by the sound of it- back to a full bridge or half bridge with plus and minus supply lines.

spec

 

[gary350]

Can I use my existing circuit at a driver for the IGBTs?

 

[spec]

Can I use my existing circuit at a driver for the IGBTs?

No

spec

 

[tcmtech]

In a commercial application they need to be used since that what is specified in the design but in reality making an equivalent capacitor out of many smaller value ones would do the same job for far less cost. Or just find who makes them and buy ones that are not Miller branded or packaged.

When I worked on those machines every day and thusly ordered overhaul parts buy the case work to keep up with them for warranty alone I figured out who made their parts and components like their 300 amp 1200 volt IGBT blocks that they retailed for ~$600 a piece were just rebadged common brand name units that sold for $100 - $150 tops without the Miller stamp on their side.

 

[spec]

In a commercial application they need to be used since that what is specified in the design but in reality making an equivalent capacitor out of many smaller value ones would do the same job for far less cost. Or just find who makes them and buy ones that are not Miller branded or packaged.

When I worked on those machines every day and thusly ordered overhaul parts buy the case work to keep up with them for warranty alone I figured out who made their parts and components like their 300 amp 1200 volt IGBT blocks that they retailed for ~$600 a piece were just rebadged common brand name units that sold for $100 - $150 tops without the Miller stamp on their side.

Yes, that is true for many parts. A non-related area but, when I had a Ford Granada, the track rod ends were, say, £10UK each, but an identical track rod end for a Ford Escort (much smaller automobile) was £2.50UK.

A BC109 transistor, with manufacturer's part number, for some audio amps can cost, say, £5.00UK, but, as we all know, a BC549 is the same, and costs £0.05UK.

And in the military field- don't ask!

spec

 

[gary350]

I use to run into that all the time at work name brand parts were unreasonable expensive but common identical parts were cheap, some companies were more expensive than others. Honeywell was the worse a $125 micro switch was only $1 from the Micro switch manufacturer, if you know what the replacement part is and where to buy it, it saves a lot of money. Now days companies call those, proprietary parts, I call it greedy deceptions.

What IGBT driver do I need? Where do I buy it? How much does it cost?