How it Works - ReactorForge Induction Heater High-Level Overview
ฝัง
- เผยแพร่เมื่อ 29 ธ.ค. 2017
- This video is a high-level overview of how the ReactorForge induction heater works. This video is not a guide on how to use the induction heater but an inside look at how the machine operates.
ReactorForge Updates: reactorforge.com
Updates on this project are being posted on the Patreon page: / reactorforge
You can support the future of this project there.
For details on how to build your own visit:
github.com/ThingEngineer/Reac... - วิทยาศาสตร์และเทคโนโลยี
![ความรักไม่ได้น่ากลัวขนาดนั้น (TRUST ME) - LYKN [ OFFICIAL MV ]](http://i.ytimg.com/vi/9np-lAtCRpQ/mqdefault.jpg)
![CAMPปลิ้น | EP.77[2/2] "รับทราบ" แก๊งเพื่อนรัก 10 ปี ที่ไปไหนก็มีตำนาน!](http://i.ytimg.com/vi/ifc0ofD51B0/mqdefault.jpg)
Hello! what models of transformers used to measure the current?
great project josh, you should try to use litz wire for the primary on the matching transformer, its better for the high frequency high current transfer because it doesn't heat up, therefore you don't need to water cool the primary
Thank you! Litz wire is an excellent idea, it is used in induction cooktop stoves and other high frequency high current application. But there is a problem, Litz wire is costly. Particularly the type that would carry the required current. I did use homemade Litz wire (about 50 strands of 22 awg magnet wire) to feed an earlier version of the matching transformer, but it still got hot at high power levels. I also tried flat copper foil insulated with various materials and smaller copper tubing. Here is a photo of a few of them: reactorforge.com/wp-content/uploads/2017/12/high-current-windings.jpg
So far water cooled 1/4 inch copper tubing is the minimum conductor type/size that can operate at 100% duty cycle and stay cool.
I fed that small diameter tube coil with Litz as well. Forgot about that one.
reactorforge.com/wp-content/uploads/2017/12/small_coil_litz_fed.jpg
I also experimented with directly winding the transformer with Litz wire.
reactorforge.com/wp-content/uploads/2017/12/litz_transformer_windings.jpg
Hi Josh. Thanks for your great videos.
I have a question.
Have you ever measured the output power or power supply efficiency by measuring the generated heat in your piece of metal?
like using the formula Q=mC(T2-T1) for iron or steel or ...
I have a 10KW power supply. I measured the generated power by measuring voltage and current at the output, however, the piece of metal that I want to melt, just show 1.5KW.
Hey Hamidreza, I remember reading this and though I replied to it but obviously I didn't. I have not done that but since seeing how efficiently and quickly it can heat a work piece this has been something I would really like to do. I have some crude plans jotted down for a calorie meter based on an industrial type that I'd like to build once the project is complete.
It sounds like maybe your power supply is not not fully coupled to your work piece. There are dozens of reasons that could cause miss-coupling if that is the issue so I won't even attempt to guess without knowing more about the setup.
Josh Campbell Thanks man for your reply!
Now I have another problem which I am working on!
great project sir, is it still active?
hi josh
great job
just a few questions
what is the rating of your IGBT s?
what is the maximum frequency of your hybrid gate driver circuit(i searched the part number and it says 20 khz is the maximum frequency!)
and how insulated the cooling system between high voltage nods it makes some issues for some cases
Thank you very much!
The IGBT's currently in the inverter are 2-pack 300A 600V modules with a maximum power dissipation of 1200W.
On the driver, 20kHz is the maximum "Recommended Range" and is the frequency used during testing to create the characteristic profile charts and current ratings. What matters are the high-to-low and low-to-high propagation, rise, and fall times. Add up the maximum times to see the worst case absolute maximum on/off time and divide one by that value (1 / .0000043) to get the maximum switching frequency. Now we still have to factor in dead time to prevent cross conduction, but as you can see the actual maximum switching frequency is much higher than 20kHz, well over 200kHz in fact. Results may vary, but I set up a quick test rig when I was looking at these drivers. With a nominal dead time, I was able to reach switching speeds over 150kHz, which is much more than I need in the ReactoForge tank circuit. If you needed something faster than about 200kHz you need to use MOSFETs anyway.
The cooling system isolation and resistance depends mainly on two factors. The level of coolant conductivity along with the size and length of the conductor or cross section x length of tubing in this case. The best coolant solution is always one that is premixed for your application. But deionized or distilled water with an anti-corrosion and growth inhibitor is a good alternative if one is not available. The conductivity of any coolant is going to increase over time as fluids naturally grab ions and dissolve metal and other particulates. I added extra length to the tubing on the primary winding connections of the matching transformer. This method is standard practice in high voltage water cooled applications. I also made the water intake and output manifolds out of copper so they could be grounded. Doing so allowed the DS18B20 digital thermometers to monitor the input and output coolant temperature as well. Other parts of the coolant path also have a high surface area to ground path such as the inverter cooling plate. The short answer is, it is very isolated from mains current. I will post specs and show tests down the road, but no matter how isolated it is by design, the most significant problem will be and usually is whether or not it is appropriately used. You could just hook a garden hose up to it, that would keep the unit perfectly cool. But I'm going to say that is not a good practice, is not supported and would void any warranty relating to any component in the cooling loop.
ok tnq for your clear answer
but it seems that the IGBT's that you have choosed are too much strong for this application
is that beacause to have good relability ?or else?
hope to see your new videos with testing this product
The IGBTs "currently in the inverter" are indeed overrated, by about 500%. This is in part because during early tests I did not know exactly how far I was going to push current draw, this allowed for plenty of headroom. The final part that I choose for the kit will still be overrated by design. Maybe not quite as much, but I prefer that switches in high power circuits operate at about 50% of their rated current and 25-50% of their rated power dissipation. This also leaves room for "hacking" the inverter to run at much higher current levels and or different matching transformer impedance levels. I have run this on 240 volts pulling right at 100 amps on a large workpiece. The energy in the room when pulling nearly 25,000 Watts from your breaker panel is electric, to say the least!
I ordered a medium run batch on the hybrid driver boards (version 1.3) since it is very likely not going to change beyond this point. I'm also ordering bulk parts and passives for the drivers. At this point, I may go ahead and lower the IGBTs to 1 or 200 amp modules as well.
Josh, whats the secondary current ? Just judging by the video if U have 4 strands of 1/2" copper tube then it makes for about 90sqmm or 3/0AWG therefore at 5A/sqmm You could go up to 450A with very minimum heating, and that got me thinking that the heating in the secondary wasn't caused by high current but rather by 'skin effect' of the high frequency as the tube wall thickness is 0.6mm.
What are the values of the secondary cap and the primary DC bias cap ?
A bit more about PFC ?
gr8 job !
Hey Mark, I don't have accurate up to date calculations or readings of the oscillating tank current. Once I get this up and running again with the new drivers, I can do that in software. The last time I measured it I used a parasitic capacitor across the tank capacitor that was a multiple of 100 or 1000 smaller and adjusted the current measurement by that amount. Those readings were in the 200-800A range, although that was a parallel LC tank and this is a series resonant tank.
The tubes are 1/4" copper and don't forget to factor in the constant cooling (no heat build-up) into your current calculations.
The tank capacitor is a 10uF 100V 1000A (500kVAr) Celem.
The DC bias cap is a 3.1uF 600V 150A TPC.
Here are a couple photos of the current measurement process I talked about:
reactorforge.com/wp-content/uploads/2017/12/current_mesure_1.jpg
reactorforge.com/wp-content/uploads/2017/12/current_mesure_2.jpg
And BTW, thank you so much for becoming a Patron! I saw the message pop up while I was replying to this. :D
I feel like a virgin again (just started with Patreon) :)
Been following Your tube channel for Years now, hoping that one day this project will make its way to the public market as an open source... and ol the sudden You started uploading more vids... awsome !
My pleasure.
Ha, that's good to hear I really appreciate the support both as a subscriber following the project and now the monetary. I've begun bulk ordering some of the primary components and am excited to get the first few kits out in the wild!
Oh and about the PFC. That is a broader topic that I will go into. For now basically, by not filtering the DC, the current envelope is entirely encapsulated within the 120Hz DC pulse train from the rectifier output. As the voltage raises and lowers with each cycle (60Hz, 50Hz, or whatever), the current also increases and decreases meaning we are using power in a manner very similar to a resistive load. This results in a near-perfect power factor since the voltage and the current are in phase. So there is no Power Factor Correction, we just aren't degrading it in the first place. As is the case with most switch mode power supplies and all filtered, rectified AC power supplies.
So if the inverter were running at 53Khz, and you looked at the output of it feeding the tank. You would see a 60Hz carrier frequency modulated at 53Khz. The tank doesn't care that there is a 60Hz carrier as this is far from the resonant frequency. We do however need the line filter to ensure no high-frequency component of the inverter feeds back into the mains.
Wiring diagram is missing
Sorry but why my comment was posted 4x times?
Hello! what models of transformers used to measure the current?
Sorry but why my comment was posted 4x times?