On 30.07.16 12:43, Gene Heskett wrote: > On Saturday 30 July 2016 08:10:20 Erik Christiansen wrote: > > > On 23.07.16 18:03, Erik Christiansen wrote: > > Armed with that info, one unit might suffice, especially if the > > junkbox includes reasonably equivalent spare MOSFETs. > > Such as those in the common atx power supply?
They'll have pretty good voltage specs, e.g. 400 to 500v if for 220v (>300Vp), but have an RDSon in the general vicinity of 100 - 200 mOhm or so, rather than the <50 mOhm of a lower voltage device of current manufacture, I figure. So they'll run hotter. Good results were achieved in one of the youtube clips, using the Banggood unit, drawing 15 A. At 0.2 Ohm that's 15^2*.2 = 45W, shared between two half-cycle-conducting MOSFETs. Add some for switching losses, and bigger heatsinks or a gutsy little fan would definitely be called for, unless the cooling water for the tube coil were first routed through some an Al heatsink slab. Easier might be to parallel a couple of salvaged MOSFETs on modestly larger heatsinks, for mounting space. I'm sure you're ahead of me here, ready to point out that at half the current per device, the dissipation will be reduced to a quarter. > Anyway, it seems they need a hard switched power supply, so turn it on > first, then hard switch it to the unit. Yeah, a "Count to three" label on the switch could substitute for timer and relay, but the auto-dropout of a latching relay would be rather handy at power-off, so we're safe for the next turn-on. Might as well add the timer, then. > My problem is sourceing a suitable supply. The RSP-1000-48 SMPS this guy uses: https://www.youtube.com/watch?v=1wwlbN-9jsU answers to a quick google, with multiple suppliers. It just costs around 9 times as much as the 1 kW coil driver. What some people do is retrieve an old microwave oven from the recycling area at their local dump, extract the ~ 1 kW power transformer, hoik out the high voltage secondary, and replace it with a low voltage one. The rest you're likely to find close to hand. We have a hard rubbish day twice per year, and there's always a few old microwave ovens amongst the goodies left on the kerbside for a few days, until the truck makes its way round. A cruise around the neighbourhood late on the weekend before pickup day can be rewarding for a dedicated repurposer. > I'd assume if I use A2, which is air hardening steel, that if I wanted > full length hardening, I could simply pass a long piece thru it, paced > to obtain the temp needed, that I would wind up with a full length > hardened piece. But that generates a couple questions. I'm not sufficiently informed on the properties of A2 steel to answer the questions which you've now answered anyway, but the ursine part of my brain looks for bear traps, and I wondered if the zone-refining effect of passing an incandescent zone down the bar would alter its metallurgy, but I guess not. As you'd not be melting the zone, alloying components couldn't migrate. (I.e. not in solution.) On 30.07.16 14:43, Gene Heskett wrote: > On Saturday 30 July 2016 12:43:55 Gene Heskett wrote: > > 1. Is the austenitic state hot enough for A2? > > > Couple answers from the practical machinist blog. > This on is no, the heat range for A2 is 1725 to 1750F. Thats a fairly > tight spec that would need to be verified with a pyrometer, something > thats not on my tool shelves. That's 940 to 954 °C, around bright cherry red or a bit more. Now, is there a cheapo pyrometer with analogue output? > > 2. Does A2 need a faster cooling profile than I'd get by doing the > > slow transport like this: > > Again no, A2 needs to "soak" at that temp for 20 to 30 minutes in order > for the chemical changes to do their job. So it must be protected from > the oxygen in the air, either by a controlled atmosphere, or by wrapping > in SS foil, or possibly by dipping it in a flux powder once its hot > enough to melt and make a coating of flux to block the air. OK, the pyrometer could measure the temperature of the flux layer, as it'd be pretty close, I guess. It'd be possible to regulate the temperature by gently oscillating the bar along its axis, increasing the length of the heating zone as temperature rises. That'd increase productivity rather than throttle the power supply back - something that might be awkward with a fixed voltage SMPS. > That also demands a water cooling setup for the coil if its going to > be excited for the several hours it might take to do a 6" long piece > given the needed soaking time. For 6" I'd possibly make a smaller diameter coil with smaller Cu pipe, and wind it 6.5" long. Spacing the turns a little bit more would stop the inductance from increasing too much, and for a few dollars, might do that job at one hit. See the length of this higher power job: https://www.youtube.com/watch?v=0vL-sArhmkI Now you need to be able to drop the input voltage though, for temperature regulation. Simpler might be to actively drive the MOSFETs, and either shut them both down at the top temperature, or try duty cycle regulation, as it does alter the energy pumped into the tuned circuit. Finicky isn't as easy as all the other firehose applications. Erik ------------------------------------------------------------------------------ _______________________________________________ Emc-users mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/emc-users
