Induction heater: hybrid drivers strange behaviour / Projects / Forums

Forums

4hv.org :: Forums :: Projects

« Previous topic | Next topic »

Induction heater: hybrid drivers strange behaviour

previous 1 2

Move Thread

LAN_403

parasole

Sun Dec 20 2015, 09:38AM

Registered Member #54647 Joined: Wed Mar 18 2015, 02:52PM
Location:
Posts: 13

well, every thing is possible, it just require some more engineering...
Did you check Wolfram thread? He use as the basis same PLL however different phase detector which is found to be much more stable for resonant circuits...

With regard to Campbell project, in one of his intermediate project he used an micro controller inserted in the PLL analog control loop, that variant in my opinion is quite interesting for its simplicity and more possibilities added by MC use. You may try to combine new phase detector with Campbel MC variant, the code he provided is very easy to understand and apply some tuning if required.

Wolfram

Sun Dec 20 2015, 10:41PM

Registered Member #33 Joined: Sat Feb 04 2006, 01:31PM
Location: Norway
Posts: 971

The new phase detector makes it possible to make a universally stable PLL, but also some other points are important to get a working system. This turned out to be really tricky, and I understand why so many people have trouble with PLLs. Working with a PLL induction heater has been my main hobby project for one year now, and I'm finally feeling like I have everything under control. The main important points are:

The VCO range needs to be limited to a range where the bridge and gate drivers work correcly and without being damaged. The VCO does not need to be completely linear, but there should be no abrupt frequency steps when the VCO control input is changed (this is a problem with some types of 74HC4046 if the VCO control voltage goes outside the acceptable range, which is typically 1 to 4 volts when the supply is 5 volts)

The current (or voltage) feedback signal needs to be stable for the whole range of expected VCO frequencies and bridge voltages, and the phase shift between the actual current and the measured current needs to be minimal and independent of primary current. It is easy to get a good phase feedback signal when the tank is at resonance, but it is more difficult to get a feedback signal with minimal additional phase shift when the VCO is running far from the resonant frequency. Limiting the VCO range makes this simpler.

As already discussed, the phase detector needs to work correctly for the whole range of phase shifts in normal operation. It is not possible to make a totally stable solution with any of the normal phase detectors, and I have made a separate thread about this:

It is a bit simpler with a microcontroller, because you can use clever algorithms to limit the VCO range to around the resonant frequency. I prefer to do it with pure logic to get a complete understanding of the system, then I can simplify it later.

Ramonn

Mon Dec 28 2015, 05:00PM

Registered Member #54436 Joined: Sun Feb 08 2015, 10:33PM
Location: Spain
Posts: 9

Hi friends; Thanks for posts; but as I said before, regretably my knowledge of electronics is so small to get your level :( I've read nearly everything about it in the net, and also tried to build some of them, but as I need a high power unit for my work, finally found that Campbell one and decieded to follow it, seeing it has few components, and awaiting the day he released his total project, so I could buy him the "difficult-to-build-parts". No day so far. So I'm trying to be back to previous tries, as that board in my first picture, belonging to mindchallenger.com project, used in PLL and manual mode because I have not the code for the uP, and waiting till I make it work in that way, and then asking for someone to write a code for it. So complicated, yeah, depending on other's work... But sure you can understand how induction heating rises fascination in diyers!, and when the beast bites you, difficult getting rid of...
About my first problem, I will use a gate transformer to duplicate and invert the signals, and connect it to M57962L drivers as if they were usual optocouplers; not feeding the input pins with constant 5V. Hope it works, will keep informed.
Happy new year!!

parasole

Thu Dec 31 2015, 09:44PM

Registered Member #54647 Joined: Wed Mar 18 2015, 02:52PM
Location:
Posts: 13

Ramonn, I don't think you need transformers to drive M57962L, they are intended to exclude transformers and to provide easy way for bridge control. I would suggest to use ether or transformer or M57962L, latest being a better choice. However, in this case you need to do some more preparation for driving signal, and for better understanding I would suggest once again to investigate Wolfram schematic...

Happy New Year!

parasole

Mon Jan 04 2016, 03:09PM

Registered Member #54647 Joined: Wed Mar 18 2015, 02:52PM
Location:
Posts: 13

Hi again Ramon,
I am working now on the hybrid IH solution, and thought the very same chip for bridge control may help you to solve your issues, just switch to use TI UC3875/3879, it is a ready full bridge phase shift controller which easily could be included in PLL loop. In my case I want to use ADPLL, however I don't see any problems in using classical phase detector with filtered output to tune up internal oscillator of uc3875, in this case you will have full solution for bridge control with PLL frequency tracking. In my opinion this combo is much better than 4046 with all additional logic for bridge control, with lots of additional functions which would be complex to implement around 4046...
Another way is to use bridge current as control input for frequency adjustment to the resonant point (as on picture), I have an academic paper on this, let me know if you want it.

1451921184 54647 FT174126 Uc3875 Frequencycontrol

Ramonn

Tue Jan 05 2016, 03:40PM

Registered Member #54436 Joined: Sun Feb 08 2015, 10:33PM
Location: Spain
Posts: 9

Thanks Parasole! -Later days I've solved my first problem about the overlapped signals; faling using transformer, I discarded it and tried an IC 4013 flip flop to duplicate and oposite the signals; firstly same problem appeared, but I discovered that putting a resistor in the outputs, and another one in the 4013 voltage supply, the signals decreased in period time, or duty cycle, even getting some "deadtime" between signals... so playing a little with resistor values I found the appropiate ones. Now I have to test it all connected to the tank and workcoil.
-Yes I read the very interesting Wolfram threads and posts, but as I said, that level is still difficult to reach to me; first I want to explore the limits in the project I have built and followed, and if failing (most probably!! ;) considere migrating to Wolfram schema, or to anyone's proved ok...
-Also the UC3875 you point seems nice, I'll let it in consideration for those future tries.
-About that paper you say, ok if you share it to me, but I'm affraid its complexity will be again out of my knowledge reach...

Josh Campbell

Sat Dec 16 2017, 07:50PM

Registered Member #5258 Joined: Sun Jun 10 2012, 10:15PM
Location: Missouri - USA
Posts: 119

Hello Ramonn. First off my apologies for not getting back to you. Over the past few years whenever I would receive an email about Induction Heaters (which was very often) I would flag it so I could respond in a timely manner. However, I missed yours. I just searched your name and boom there it was, it never got flagged! :( Thank goodness for the awesome minds on this forum!

Are you still doing work on your IH?

I should mention that the adjustments possible when using the power stage controller (AT90PWMXXX) are very important when used with a driver of this type. Coming from the world of gate transformers and delay logic no longer applies, it must all be in code now. Reading your OP the first thought that came to mind was, I wonder if he was just had his logic reversed, active high vs low. Which would mean the duty cycle adjustments were on the wrong side of the signal. I doubt that was the case because that seems too obvious especially if you are scoping it.

Anyway hope all is well and I am continuing work on the ReactorForge!

previous 1 2

Moderator(s): Chris Russell, Noelle, Alex, Tesladownunder, Dave Marshall, Dave Billington, Bjørn, Steve Conner, Wolfram, Kizmo, Mads Barnkob