Induction brass annealer redux

How much current is circulating in the coil? The short answer is: a lot. This is why the thick copper conductor gets hot. The high current and frequency also means it is not trivial to measure. Tesla coil builders have ways to measure the current on their primaries, but it is tricky. Trying to calculate it with the V=IR approach is also tricky, because R is difficult to determine. Skin effect and proximity effect are significant in this situation. If you want to calculate it, I think using the capacitor reactance is a better approach. For this parallel resonant circuit the capacitor current is roughly:

Ic = 14 * Vin * f * C

So for example at 48v input, 100kHz, and 2uF, the RMS capacitor current is ~135A. The coil current would just be this plus losses/load. If the "no load" current draw of the circuit is say 10A, then you could estimate the coil current at 145A. This has some assumptions built into it of course, but it should get you close. Why do you want to know what the current is?

How much current do you suppose is flowing in a 308 case neck?

 

Might want to check that formula, no?

 

Be my guest. You can start here:
https://en.wikipedia.org/wiki/LC_circuit
Further discussion unrelated to the thread topic should probably be taken to PMs.

 

This is the reason I posted the question:

Starting a new batch of annealing, the cartridges did not get hot enough. The DC current from the Power Supply was at the usual level (at my set up - using ferrite core and coil, the current practically does not change).

The problem was resolved by tightening the coil connections. Fortunately the induction board was OK. It was obvious that the R component in the impedance (Z=R+jX) of the coil changed , respectively the current.

So, i was looking for a simplified method to detect/monitor/alarm any annealing power reduction without monitoring the glow of each cartridge.

Yes, a current clamp and oscilloscope will do the job, but to me, it is not practical. Maybe monitoring the voltage at the coil connection would be an alternative (the voltage (pretty low, I suppose)) should be indicative.

 

Take that thought further and you likely have the essence of Aztec. An ammeter with shunt would work assuming it can measure alternating current. Currents are very large though.

 

Very Very, Nice

 

Fred is right.

The discussion with Fred prompted me to do some work I'd wanted to do some time ago but never got around to doing and also to get out my oscilloscope and probe my induction board again. Previously I had scoped my induction board rapidly to check the frequency of oscillation. I saved that scope pic and it has mislead me since. It had a much lower Vpp. I honestly can't remember the conditions under which I made the measurement. The scope pics below are from my finished build. The voltages shown are indicative of even higher currents flowing in the work coil.

I also did a rough LTspice model of the circuit. This circuit draws roughly 10A from the 48V supply without a load (RLoad) and 14.4A with Rload attached (roughly the current drawn by my build midway through annealing a 308 case). I've attached a screen shot showing the modeled current across the work coil. To this one needs to add the 2A flowing through Rload. (Using a resistor to model the case in the work coil is obviously a rough approximation.)

I've included the LTspice model in a zip file for those who like to play with these things.

 

A couple more screen grabs from the Spice model

 

Thanks for the LC topic. Ironically, it's been a very long time since my separation from calculus.

I'm planing to build an auto feeder and run the annealing without supervision. Given the fact, that the induction modules are not so famous for their reliability, I would like to avoid using improperly annealed brass.

The exact level of the coil current, for my case, is irrelevant. I need to pinpoint the moment, when the high frequency annealing power is reduced or disappears at all (please note - when it happens the DC current from the power supply does not change much). Eventually, I could use temperature detection IR etc.

For now, I envision a current clamp probe (toroid) and monitoring the secondary voltage/current (David101 suggestion). Apparently, I would need calibration and ways to generating a DC signal to be fed in the Arduino micro controller. In my particular set up (the magnetic field is concentrated in a ferrit core), I can add easily a secondary coil onto the ferrit core.

 

I still think you could monitor the main PSU current. If a case isn't in the coil or the LC tank oscillation fails for any reason there will be a very significant drop in current from the PSU. With nothing in the work coil, and a working tank circuit, I get a c10A draw from the main PSU versus something several amps higher (a rapid rise to 15A over 5 secs) if there is. You could monitor the main PSU draw and if the current didn't hit a selected threshold, 10% less than 15A for example, you could shut the system down. (I guess you have to leave that case in the coil else you wouldn't find it in the bucket of many.)

 

You could ask how @LR88 monitors current.

See #114

And design your own version of Aztec while doing so.

 

SHEEEEESH..... As one of the founding parents of the "Basic GinaErick" you guys have come a long way. My hat is off to you.

Regina

 

Once the Curie Temp is reached, current drops.

 

When using a Ferrite the current drawn from the DC power supply is much less especially when there is no case in the field. The method you are suggesting is probably the "best way" to do it but not sure if it really matters that the best way is needed. When or if the induction board fails it is pretty obvious most of the time with either no current flowing on the DC side or an overload current happening depending on exactly what has failed and how. I would take the easy way. I think the most important thing when going automated is to put in a little bit of a time delay and to build in ways to cool the components. I think the components that get the hottest and induce heat in other components are the filter Inductor coils. The PCB esp the connections around the standoffs and then the capacitors. A fair bit has been discussed about cooling these components. With the reduce currents due to the Ferrite less heat is generated.

 

Good morning
By compiling the program with the Arduino IDE version 1.8.10, I get the following message:

C:\Program Files (x86)\Arduino\libraries\SoftwareSerial\src\SoftwareSerial.cpp:45:30: fatal error: util/delay_basic.h: No such file or directory

#include <util/delay_basic.h>

^

compilation terminated.
exit status 1
Erreur de compilation pour la carte Arduino Due (Programming Port)

Can you tell me what a mistake I made.
Cordially

 

Yes. Of course the question remains as to just when to stop before then.

 

Monitoring temperatures is important with any autofeeder. That's why I built in monitoring of temperatures at 4 points, the scaling of fan speeds with temperature and auto pause if temps rise above a selected temperature. With a ferrite concentrating the flow of current across the air gap the work coil is more efficient, but you lose the ability to flow coolant through it. (My chokes don't get hot. The capacitors definitely heat but the placing the heat diffusers on the bottom of the board really helps a lot. My coolant really heats up quickly considering its volume. You can use the LTspice model to gain insight into just how much relative heat is emitted in watts by various components.)

BTW you can also do a few things to improve the cross coupling of the driver. Change the diodes to Schottky and place a 0.01uF capacitor across each of them. A resistor across each (200 ohm) may help as well. One guy who did a lot of work on these boards found that twisting the board's supply wires tightly to reduce parasitic inductance also helped improve the performance of the MOSFETs. (Given we don't need the board to run across a broad range of input voltages you could also look at rejigging the voltage divider to reduce the power burnt by the resistors and zener diodes as the MOSFET gates don't need current; you just need enough flow for the zeners to work properly.)

 

Little modification of "Anneal Shelf for Bill" 3D parts, inclued direct mount for solenoides and a trap door.

 

the trap door

 

Have you tested the motion of that set of pivots? I found that my coupling needed fine-tuning, and that while originally I felt I needed a loose coupling between the solenoid rod and my coupling to the trap door component it turned out that a rigid one was better. So having the flexibility to place the solenoid for best effect was useful. Some strong double-sided foam tape was all that was necessary to attach the solenoid to the bottom of the shelf. Unless you implement a return spring the solenoid needs to be able to push the door closed as well as pull it open. (And you need it all to be light because these solenoids have weak return springs on their rods.) (Fusion can help you test freedom of motion.)