Novice-Built Induction Heater (In 14 days?) / General Science and Electronics / Forums

Forums

4hv.org :: Forums :: General Science and Electronics

« Previous topic | Next topic »

Novice-Built Induction Heater (In 14 days?)

previous 1 2 3

Move Thread

LAN_403

Steve Conner

Thu Aug 28 2014, 06:40AM

Registered Member #30 Joined: Fri Feb 03 2006, 10:52AM
Location: Glasgow, Scotland
Posts: 6706

At this sort of level you should be using the same part that the designer of the circuit called out. If the schematic says to use a LM393 then that is what you use. Once you understand the theory behind what is going on, you can make substitutions. (For example, the LM393 in Uzzors' circuit is a comparator, not an opamp, and if you had substituted an opamp the circuit would not work.)

Popular parts are often "second-sourced" by several manufacturers. The first two letters denote the manufacturer. So the 555 timer is available as a NE555, LM555, uA555, KA555 and so on. Again these are supposed to be identical. There are some minor differences, but a good circuit design will work the same with any of them. (But watch out for CMOS ones like the ICM7555.)

Opamps are often available in a premium grade selected for low offset voltage. This is usually indicated by an A in the part number. Again if the circuit needs the A grade to perform, it should be called out.

Outside of these considerations, just pick the package style you want (I use DIP for prototyping too when possible) then pick the cheapest one or whatever is in stock.

AwesomeMatt

Thu Aug 28 2014, 01:26PM

Registered Member #4454 Joined: Sun Feb 26 2012, 12:47AM
Location: Western Canada
Posts: 74

wrote ...
Again if the circuit needs the A grade to perform, it should be called out.

Yes, this much I get, if the circuits I received were from paid professionals with expectations for documentation. When my sources for schematics are volunteer hobbyists who're sometimes amateurs... it's not reasonable to assume the same thoroughness was taken (or expected, they're taking time out of their day to share for free).

wrote ...
Outside of these considerations, just pick the package style you want (I use DIP for prototyping too when possible) then pick the cheapest one or whatever is in stock.

Good, this is what I had been doing and was feeling rather uneasy about it.

...

I sorted about 20 pounds of salvaged components tonight, found all kinds of things I didn't know I had. I've got a half-dozen Schottky diodes in TO-247, and a dozens more in TO-220. Lots of MOSFETS, some that appear to actually be powerful enough to avoid needing to purchase... but they're salvaged so I'll spend the big bucks while I'm already ordering ($22 each). Some of almost every small component. A stack of 4046s. Suitable pots, suitable power resistors. Feeling confident.

I'm also not sure whether to go with Uzzor's gate driver (UC3710T) which is both low and high side... or IAmSmooth's TC4422 & TC 4421.

Smooth seems to have more paranoia and robustness built into his design. His fets are overbuilt, he has zeners across them, his caps on the drivers are much larger too, but his caps on the MOSFET side are 1/10th as big. Uzzors's also has a 4-pack of 1N5819 schottkys on his drivers, Smooth has none. Smooth actually runs his at 10+kW, Uzzors designed for 1600W and IIRC that non-obvious (not just bigger mosfets/support components) design issues start to creep up on a design past 3kW.

I'm not sure which is smarter to emulate.

AwesomeMatt

Tue Sept 09 2014, 06:31AM

Registered Member #4454 Joined: Sun Feb 26 2012, 12:47AM
Location: Western Canada
Posts: 74

UPDATE - FAIL.

Naysayers, commence your toldjasos.

I think I could have completed it on time, had my (borrowed and rush shipped) capacitors not shown up the afternoon before. That, and had I not knocked over my partsbins the same day, spilling tens of thousands of components all over the carpet. That, and had I not chosen to spend my time getting more important faire-related tasks completed.

IT'S OKAY.

I had a blast at our Makerfaire anyway and showed off some other projects and spent my time volunteering for many other things instead. I almost would have been too busy to be stuck supervising the induction heater.

I never cashed the cheque for the funding grant, and I've offered it be returned if they would prefer. I don't think anyone was too upset, no one else got their funded projects done on time either.

Either way, still continuing the build. This time with a less frantic pace and ghetto approach. In fact, I will probably build 2 units, a 120V and a 240V version.

But first, I promised a documented build, and I've been slacking. Here are more build pics:

2008

Power components. Schottky's I found in my partsbins, I have dozens. I'll probably twin them up just to be safe, since I'll be pushing near 30A.

2009

Heatsinks. This and several cubic feet more. I'll probably go with the copper bottomed ones left-mid. Not sure if I'll use fan or water cooling for the fets.

2010

For water cooling, a "heater core". I mentioned what I was making, and a friend said we should make a trip back to his place to pull a heater core out of some car scraps under his deck. Looks good to me. Fins are so thin and dense it probably needs a heafty fan. Not an issue.

2011

To smooth input power, some bigass rectifiers (top right), and some old caps. 208V smoothed is 300VDC, I'll be planning for 400V just to be safe. These are all actually kinda crappy, large for how puny their values are, so...

2012

71 of these, new and unused. 50v, 4700uF. Means I'll use strings of 8 in series for ~600uF, and then start chaining them up in parallel. I've been told to put in as many caps as I can, somewhere around 5000uF @ 400V should suffice. This box could do, nice because they're all matched.

2013

... Or, my higher voltage electrolytics. Lots of 400+V in here, lots of 200V as well. Bit more of a pain, since they're mismatched sizes.

2014

Failing that I still have a another cubic foot of electrolytics left over, I'll make it work.

2015

The electrolytics provide the bulk power to the inverter and are fine for the 60hz they're trying to smooth, but the inverter still needs a supply of power that can be drained quickly and rapidly in the tens of kilohertz... electrolytics can't handle that. Up steps the MKPX2 caps.

On the left are all 250v 1uF+. On the right are my bin full of smaller ones.

I'll probably stack the 250v in series pairs for 500V. It'll still be good for 10uF.

2016

Big (huge, 1/4") terminal blocks. Good for my main wiring. I've got lots of these for when I need them.

2017

Tank caps. 600Vrms, 158 Aeff, 3.1uF. I borrowed these and rush shipped them. I will probably keep them and just pay to replace his when he finds a good deal, since they weren't too much what he needed.

2018

12V 2000mA wall wart for the PLL/driver circuit, and a 3A (chinese amps, so, needs a heatsink for more than 1A or the solder melts) buck regulator to hold it there.

Also, someone suggested to me that on Uzzor's circuit (

), the 12V rail and the 15V rail would be just fine at both being the same 12V rail so that's how I wired my PLL.

More on the PLL later.

2019

PLL/driver board, and all the stuff that's getting mounted on the case: Drive Enabled LED, PLL LOcked LED, Reset button, pot for Peak Tank Cap voltage, pot for Phase Adjust, pots for R1 and R2 (for testing, may hardwire).

2020

Closeup of the PLL/Driver baord on stripboard.

I don't know how to turn a schematic into a physical circuit, so I just tried to think ahead a little bit and otherwise went one component and one connection at a time. My partsbins landed on it and bent some of it.

.......

End photos.

So far, when I plug it in, without any GDT connected and pots set roughly to the range I want them... nothing happens to the LEDs. Also though, the chinese buck regulator flatlines at 2.47 amps (probably it's max) and the output voltage on the buck is only 0.25V (input is ~12V).

So, my guess is that the power is sagging down to almost zero because:

1 - There's a short somewhere.
2 - The PLL/driver board takes a hell of a lot more than just 2.5A to power.
3 - Maybe this is normal behavior when no load is connected?

That's where I am today.

Wolfram

Tue Sept 09 2014, 10:41AM

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

You don't want too much smoothing capacitance across the DC bus, it will only serve to ruin your power factor. Unsmoothed rectified mains will heat just as well as smooth DC. Some film caps across the DC bus to take care of the ripple at the driving frequency is all that is needed.

previous 1 2 3

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