Royer induction heater and high current power supply

[RocketScienceSmurf]

This thread will be used primarily to share my build of a Royer induction heater and also a power supply capable of driving it with ease. I hope my findings can help others.

A while back I decided to start experimenting with induction heating. I had a very specific application in mind but I won't go into specifics. My goal was to heat up a large stainless pot, but I quickly decided to transfer the energy to a pipe instead. To start experimenting without spending a fortune I ordered a "1kW induction heater" from ebay and went to work.

One thing that was clear from the beginning was to make sure I had a power supply that could deliver quite a lot of current at 12V so I took a large UPS battery I had laying around. It quickly became clear that 20A or even 30A current draw was no problem and I had to find another, more permanent solution than a lead acid battery. Buying a large 2kVA or higher rated transformer cost a lot of money and if you want a three phase transformer to keep the ripple down it's even more expensive, but I managed to find a large welding power supply that I took apart. In it there was a 6kVA rated three phase transformer weighing 80kg that I will use to build my power supply. To build a three phase rectifier bridge I need 6 pcs of high current diodes but there was only 4 in the welding PSU so I bought 6 new ones, 3 "regular" and 3 reverse polarity. The original ones were rated 300A but my new ones are only rated for 150A but I think 150A should be enough.

The diodes in picture are a regular 4007, a fairly large MUR420 and my 45L10

I use the 500V primary which gives me a rectified secondary DC of about 37V

[RocketScienceSmurf]

Most problems people have with this induction circuit is because of the PSU not being powerful enough so DON'T skimp on the power supply and make sure it can deliver at least 12V under HEAVY load. These induction heaters really need to be kick started so make sure your PSU can deliver a lot of current promptly, in other words make sure you have a large capacitor bank.

Here is some really good information about DC bus capacitors: http://kaizerpowerelectronics.dk/tesla-coils/drsstc-design-guide/dc-bus-capacitor/

The reasons I looked for a three phase transformer were lower ripple than a regular one phase and that huge three phase transformer are easier to find than a single phase and also easier to power. Most single phase mains outlets are fused at 10A which gives a theoretical limit of ~2300W and I wasn't sure that was enough for me.

I would recommend anyone looking to build a beefy power supply to look for welding power supplies. They can be found really cheap sometimes the old ones are built like cold war era tanks.

I salvaged all parts from my welding psu, and among the many interesting parts there is a HUGE inductor/choke and 4 high current capacitors that I would like to re-use. Should I connect the rectifier first to the choke and then to the capacitors or vice versa? My thinking was that the choke that I measured to 1,8mH would act like some kind of soft start and make the power supply easier to power up.

For those of you who are interested I have attached the service manual/spare parts list for the welding psu and my diodes. (The spare parts list file was too large so here is the link instead)

http://www.esab.com/gb/en/support/upload/DTA-300-AC-DC.pdf

[Mads Barnkob]

You have really set sails for a large induction heater here!

How many bridges have you planned to use in parallel? I know that RogerInOhio on 4hv made one with 4 or 8 bridges in parallel, that would draw around 4 kW.

To experiment with the power supply design, where to place choke, capacitors etc, if you want a CLC or LC filtering, check out the Duncan amplifiers PSU designer: http://www.duncanamps.com/psud2/

From the part list for the welding machine, I can see that there is a shunt resistors and a digital Ampere meter in the front, you could use those for measuring the DC current drawn from the power supply.

I look forward to see the rest of this monster build and how it performs melting stuff

[RocketScienceSmurf]

First of all I will see how far I can take a single pair of mosfets. I aim for 1500W and if I reach that I don't think I really need to go much higher. Since I am going to use it for heating water and not melting metal it's hard to find relevant data. Everything is focused in crucibles and heating knives and bolts.

I remember psud2 from my vacuum tube days and it's a really nice software.

When it comes to current metering I first thought about using the regular shunt, but since I would like to use an arduino to both measure and control the output (just on/off) I think I will use a ACS712 hall effect current sensor instead. The shunt from the welder has a voltage output of 60mV at 300A and since I will have a maximum continuous output of 150A that translates to 30mV. Even if I use an external voltage reference of 1.023V for my arduino the A/D resolution will be too low, and also I will have to be careful about the ground and positive power supply since the arduino will be tied to to the circuit. The ACS712 will let the arduino float in regards to either the  positive or negative side of the psu and also give a much higher A/D resolution.

I hope to have crude psu ready for testing within a month or so. This entire project will take some time, but I will try and document all useful steps as good as I can.

[RocketScienceSmurf]

Finally got around to wind the chokes. They are INSANE! Took quite a bit of work but the result looks pretty good. I think I managed to get 38 turns of AWG 9 wire around the T-300-cores. Will measure them with my LCR-meter later and post the values.

I put a glove there for size comparison.

Edit: measurements

Here are my measurements:

L1
@100kHz L=183uH Q=112
@10kHz L=184uH Q=142

L2
@100kHz L=180uH Q=114
@10kHz L=182uH Q=154

Pretty similar values. Can't wait to try them out

[Mads Barnkob]

Looks great, I can imagine that it must have been interesting winding those 38 turns of 6 mm^2 wire, that has to done carefully to get a good looking result like you have there.

What do you have in mind for filtering the DC supply? A huge stack of electrolytic capacitors?

Is the bridge/bridges the next part of the project?

[Dylan Burger]

This looks like its going to be an insane machine

What kind of switches will you be using to handle all that power ? I used to push 30A constantly through a set of irfp250's which didnt last very long. After pushing the voltage higher than 40v I had to move up to IRFP460's and as a last resort I switched to IGBT's.

Also , what resonant frequency are you expecting ?

[RocketScienceSmurf]

Yes Mads, rectifier is probably the next step. I asked a friend to help me cut, drill and tap the heat sinks I salvaged from the welding power supply for my new power diodes.

I'm not sure about the psu caps. I would like a bunch of large electrolytic or high power film capacitors, but I can't spend an infinite amount of money on the just the caps alone. Also, since I am using a three phase transformer I don't believe I need as large bank of caps as I would have with a single phase setup.

For turning it on and off I will probaly use a pair of Infineon IRFP4468 driven by a high power mosfet driver, controlled by an Arduino. Initially the IRFP4468 was planned for the induction circuit but the voltage of my upcoming psu forced me to change IH mosfets to IRFP4568 with a 150V voltage rating instead. Expected resonant frequency is around 55kHz.

[Dylan Burger]

For turning it on and off I will probaly use a pair of Infineon IRFP4468 driven by a high power mosfet driver, controlled by an Arduino. Initially the IRFP4468 was planned for the induction circuit but the voltage of my upcoming psu forced me to change IH mosfets to IRFP4568 with a 150V voltage rating instead. Expected resonant frequency is around 55kHz.

Nice  

Your IH mosfets should also handle the resonant rise within the work coil. usually 3x input voltage. So , even if you somehow get inductive spikes from switching you will be covered, considering that 3 x input (36V) is ~108v.  55Khz is also a nice resonant point. I find that most objects heat fairly well at this frequency. I have gone to 200Khz and it doesnt seem to work well on thicker objects.

[RocketScienceSmurf]

I wound the chokes myself by hand. They are for the induction heater circuit. I have a BIG choke that came with the welder power supply that I will use for my power supply. I measured it to 1,8mH and it's most likely rated for 300A continous use since that was the specs for the original welding psu.

I can take a picture of it and post here. I would like to get my hands on a milli or even micro-ohm-meter to check the DC resistance of it but I doubt it's much.

I realized almost immediately when fiddling around with the small ebay-IH that these suckers can draw A LOT of current so a beefy power supply was one of my first priorities. I guess this PSU will weigh in at close to 200kg and will be able to deliver 150A continuously. Should be enough

[edit]

Here are some pics of the choke as promised.

[RocketScienceSmurf]

I received a PM about why I wanted to use a three phase transformer, so I thought I should take some time to explain in as good as I can in case someone else would like to know.

The reason is that with a three phase setup you have three voltages that are out of phase by 120 degrees.



When you rectify this with a full wave bridge rectifier (that requires only 2 more diodes than a single phase full wave rectifier) you end up with a DC voltage with lower ripple voltage and higher frequency which in the end requires less capacitors for smoothing out the ripple.

[Mads Barnkob]

I also tried to summarize the differences between 1 and 3 phase, half and full wave rectification with some pictures and charts here: http://kaizerpowerelectronics.dk/tesla-coils/drsstc-design-guide/rectifiers/

[RocketScienceSmurf]

I have hit a bit of a wall with my power supply

I have found the capacitors I WANT but they are stupidly expensive. Dream scenario is 6000uF DC link film capacitors but that would cost about 1000 USD. Would it be possible to use a large bank of electrolytics bypassed with a few smaller film caps and achieve acceptable performance?

I found these but they are still quite expensive: http://www.ebay.com/itm/Ducati-Energia-DC-Filter-Link-Capacitor-1500-Micro-Farad-700-VDC-165-A-/142169549692

[Mads Barnkob]

I have hit a bit of a wall with my power supply

I have found the capacitors I WANT but they are stupidly expensive. Dream scenario is 6000uF DC link film capacitors but that would cost about 1000 USD. Would it be possible to use a large bank of electrolytics bypassed with a few smaller film caps and achieve acceptable performance?

I found these but they are still quite expensive: http://www.ebay.com/itm/Ducati-Energia-DC-Filter-Link-Capacitor-1500-Micro-Farad-700-VDC-165-A-/142169549692

I have been eye-balling those capacitors for a while too, but that should be used in a large DRSSTC that has to pull enormous amount of energy and dump it into a resonant LC circuit, they are way out of the scope for a high current low voltage power supply as yours.

You should build a high capacity bank of electrolytic capacitors with a voltage rating of no more than 200 VDC, it will be waste to go any higher for your low voltage application, and would only result in a lower capacitance bank. Parallel as many as needed for the capacitance needed for the desired voltage ripple. I guess you can find some inspiration in my DC link electrolytic capacitor part of the DRSSTC design guide to calculate what you need.

[RocketScienceSmurf]

Finally got the rectifier assembled. I received the diodes a while ago but I needed help to drill and tap the salvaged heat sinks. Here is the result

[Mads Barnkob]

Really nice and beefy stud diodes

But are they mounted correctly in regard to be a 3 phase rectifier bridge? They would only be correct right now if you have the DC+ at one heat sink and DC- at the other.

You can see here how its normally done with stud diodes, the 3 inputs at the bottom and DC+/- is the rails at top.

[RocketScienceSmurf]

I think I have mounted them correctly. My intention was to have the DC output at the heat sinks and because of this I was forced to buy new diodes and half of them are reversed polarity. I do however need to double check the color coding of the sleeves before I connect everything.

I still haven't found any suitable capacitors. According to one site I found, a rule of thumb regarding ripple current is that ripple current over the capacitors is double that of the current draw to whatever load is connected. The limiting factor for this power supply is the diodes that are "only" rated at 150A meaning I would like a capacitor bank capable of handling 300A so I keep getting back to film capacitors that cost a small fortune. Just for fun I made some calculations for heat losses on a fairly high rated electrolytic capacitor with ESR of 8mOhm and maximum ripple current of 26A. This translates to roughly 5 watts. I would require 12 pcs of this, quite pricey electrolytic. The cost would be almost the same as using DC link film capacitors.

[Mads Barnkob]

What about if you made something yourself? A MMC of larger film capacitors, paralleling a lot of these should give you the current rating needed, but how much capacitance are you after?

http://www.ebay.co.uk/itm/132219914390

http://www.ebay.co.uk/itm/132219915352

http://www.ebay.co.uk/itm/132221070952

this seller has multiply auctions for more of the same capacitors, its possible to get a good deal with combined shipping.

[RocketScienceSmurf]

I would like to reach somewhere around 4000-6000uF. Together with my DC choke that should give me somewhere near -30dB if I understand things correctly.

I found this one: http://www.ebay.com/itm/371440106874

It is 65 millifarad (65.000 uF). I double checked with the seller. I don't think I have ever wanted anything as badly as that behemoth of a capacitor

[Mads Barnkob]

I would like to reach somewhere around 4000-6000uF. Together with my DC choke that should give me somewhere near -30dB if I understand things correctly.

I found this one: http://www.ebay.com/itm/371440106874

It is 65 millifarad (65.000 uF). I double checked with the seller. I don't think I have ever wanted anything as badly as that behemoth of a capacitor

It will take a LOT of soldering, but it would be possible to reach with around 4-6 boxes of those capacitors

I also want that capacitor now, if only we knew what it looked like on the inside, maybe its just a array of smaller capacitors so that it could easily be taken apart and split between buyers, as no one really needs 65000 uF in hobby use. If you can get a really good offer from the seller, I am in on a group buy!