See post 2734
I still think you could and should calibrate any flame sensor to a measured heat. It wouldn't be hard to do. See the note I appended to the bottom of the above post.
Induction brass annealer redux
- Thread starter Gina1
- Start date
I still think you could and should calibrate any flame sensor to a measured heat. It wouldn't be hard to do. See the note I appended to the bottom of the above post.
Those FETs are needed for the board to oscillate. Any "pulsing" would need to engage and then disengage oscillation. It would likely be pretty easy to screw up oscillation if you messed with trying to control via those two paired FETs.
Oliver should really couple an oscilloscope to his board to see how it behaves when he attempts to pulse power to it.
I've pointed this out before:
forums.accuratereloading.com/eve/forums/a/tpc/f/9111080861/m/7681074262
And here's a video from that thread:
vimeo.com/391700531
Note the absence of any large relays. ~1000W is being controlled with a little push button switch on the end of some thin wires.
Indeed because the MOSFET gates don't need current to drive them and if you ground the gates you only need to ground the current through the two 470R resistors. Such grounding could be controlled via a simple, sufficiently fast semiconductor switch and it would not need to handle much if any current (about 2 x 100mA with a 48V supply). Much better than trying to switch the power supply load. But I wonder just how responsive the oscillator is. I suspect these things don't start up and stop as quickly as we'd like them to. Hence I'd like to see an oscilloscope trace of Oliver's unit while he is pulsing it.
(Way back when, and after another discussion with McFred, in this post I posted a schematic and LTspice model of the typical board. I guess I could simulate a pulsing of the gates to ground as well if I had the time.)
(Way back when, and after another discussion with McFred, in this post I posted a schematic and LTspice model of the typical board. I guess I could simulate a pulsing of the gates to ground as well if I had the time.)
When I get a chance I may run a test by grounding out those two gate terminals. I can run an oscilloscope trace of the output. But I would be concerned about is lock up during power on and what rate the circuit can be powered on and off. Mine is oscillating at about 85 khz I presume I can't cycle it any faster than that and it would be ludicrous. So something on the order of 1 Khz or less may work. I could even help the circuit oscillate by powering one side up first delaying and then powering up the other one.
Controlling the on-off cycles by controlling the gates seems like a lot better solution than using some unknown quality SSR.
Controlling the on-off cycles by controlling the gates seems like a lot better solution than using some unknown quality SSR.
"Mine is oscillating at about 85 khz I presume I can't cycle it any faster than that and it would be ludicrous."
?? As has been discussed often before, the frequency of oscillation is dependent on the inductance of the work coil and the capacitance on the board. It was interesting to see that a model of the board using a rough estimate of the inductance of the work coil (based on an air gap inductor calculator) led to a modeled oscillation frequency that was very close to what I observed with the oscilloscope (see graphics in the post I linked to above). 102kHz vs 104kHz.
Switching the PSU is fine for low duty on/off - even if it is less efficient and you have to dissipate/heatsink a good deal of RdsOn. Using a switch to ground the MOSFET gates as an on/off switch makes more sense.
It still has to be shown that pulsing the oscillator is practically possible (be it via the PS to the board or via the board's FET gates). In conjunction to the rise and fall times of the device you use as a switch there's also the rise and fall times of the board's FETs (plus the rise and fall times of the controller running the switch). Oliver seems happy with his results thus far but it would be great to get a scope trace of the output during this short period.
?? As has been discussed often before, the frequency of oscillation is dependent on the inductance of the work coil and the capacitance on the board. It was interesting to see that a model of the board using a rough estimate of the inductance of the work coil (based on an air gap inductor calculator) led to a modeled oscillation frequency that was very close to what I observed with the oscilloscope (see graphics in the post I linked to above). 102kHz vs 104kHz.
Switching the PSU is fine for low duty on/off - even if it is less efficient and you have to dissipate/heatsink a good deal of RdsOn. Using a switch to ground the MOSFET gates as an on/off switch makes more sense.
It still has to be shown that pulsing the oscillator is practically possible (be it via the PS to the board or via the board's FET gates). In conjunction to the rise and fall times of the device you use as a switch there's also the rise and fall times of the board's FETs (plus the rise and fall times of the controller running the switch). Oliver seems happy with his results thus far but it would be great to get a scope trace of the output during this short period.
The results I got from the one test that I ran on a 44 magnum case getting about 85 KHz. As far as you can cycle the device on and off the answer is obviously yes. The question is at what rate can you do it effectively and efficiently. I'm sure if you cycle it on and off every second it would work perfectly. Given it oscillates around 100Khz I would think that you can cyclic probably at least as fast as 1Khz using gate control. But it will have to be tested.
Once stable and oscillating it is at circa 100kHz. I'm not so sure it starts and stops so cleanly. Dunno.
You will definitely need to beef up those resistors. If you ground them, they're each passing, assuming you're using the 48V PSU most are, 48/470=102mA and dropping 48V in the process. So 4.9W and hence requiring a resistor rated for circa 10W. I think your calculation is incorrect as you have assumed their right side is still held at 12V which it won't be. Maybe I'm wrong as I haven't had my morning coffee yet.
The resistance could be increased substantially to lower the current (and even the zeners could be changed) and so that would help.
You will definitely need to beef up those resistors. If you ground them, they're each passing, assuming you're using the 48V PSU most are, 48/470=102mA and dropping 48V in the process. So 4.9W and hence requiring a resistor rated for circa 10W. I think your calculation is incorrect as you have assumed their right side is still held at 12V which it won't be. Maybe I'm wrong as I haven't had my morning coffee yet.
The resistance could be increased substantially to lower the current (and even the zeners could be changed) and so that would help.Right about my calculations being wrong. The zener diode is still in operation however my calculation is only for the current going through the transistor controlling the gate. ZThe zener drops the other part the current. It won't be a problem for me anyway because I'm using the 30 volt 20 amp power supply. I ordered some transistors to test this because I didn't have any that I knew would work. I had were some transistors left over from a Radio Shack grab bag that is over 40 years old. No specifications I only know they are a to 92 case and have a hfe of 1000.
Perhaps McFred's hint is to drive the gates not from the 48W or other high PSU but from the 12V supply.** You could possibly go even more directly if the "controller" unit can hold the gates high at greater than the 4V Vgs(th) needed to turn them on. You can rip out a few parts. The gates draw no current and so the controller only has to supply that going through the resistors R1/R3 and the diodes D3/D4 limited by new suitably sized R2/R4. Controller "on" output logic level high 5V and the oscillator starts. Controller "low" and it stops. Drive the gates directly (even from a Sestos timer).
** This mod would be a good improvement for even the most basic GinaEric design.
** This mod would be a good improvement for even the most basic GinaEric design.
Last edited:
McFred's hint has a lot of merit. Much more energy efficient. Looks to me like 4 levels to control the on/off state of the ZVS.
1. Do as most everyone has done so far to externally control it with a relay or SSR. Large relay or expensive SSR should be used. A SSR is needed if you want to turn off and on very quickly. Good SSR are expensive, waste power and slow to cycle compared to direct control of the gates on the PCB. Simple with no mods to the PCB.
2. Control it by tapping into both gates, ground them to turn off, NC to turn on. Simply to tack solder some control wires to the PCB, cheap and faster switching time, but wastes power on the off state and with a 48V supply is right at the limit of the 5W resistors, wasting 10W at idle.
3. Cut the wire shown by McFred and connect them to turn on, disconnect them to turn off. No power wasted, fast switching, but the mods to the PCB are much more drastic. On my version 2 traces that are hard to reach would have to be cut and then 2 wires solder on the backside. If parts are removed access to the traces wouldn't be a problem.
4. Similar to 3, but change or rip out parts not needed for on/off control with a lower 12V or 5V control signals. More work but simplifies the control logic to the board.
In my case, using only a 30V power supply, the power wasted in #2 is 3.8 watts total and only 1.9w per 5 watts resistor. I can live with that loss for ease of the mod. Also it gives the flexibility to control each half of the oscillator with the MCU. This may turn out to be useful if you plan use a flame sensor and keep a constant temp by vary duty the cycle and during startup the oscillator takes too long to stabilize. Maybe delaying the connection of one side may help the oscillator reach stability sooner. A scope trace would tell the story.
1. Do as most everyone has done so far to externally control it with a relay or SSR. Large relay or expensive SSR should be used. A SSR is needed if you want to turn off and on very quickly. Good SSR are expensive, waste power and slow to cycle compared to direct control of the gates on the PCB. Simple with no mods to the PCB.
2. Control it by tapping into both gates, ground them to turn off, NC to turn on. Simply to tack solder some control wires to the PCB, cheap and faster switching time, but wastes power on the off state and with a 48V supply is right at the limit of the 5W resistors, wasting 10W at idle.
3. Cut the wire shown by McFred and connect them to turn on, disconnect them to turn off. No power wasted, fast switching, but the mods to the PCB are much more drastic. On my version 2 traces that are hard to reach would have to be cut and then 2 wires solder on the backside. If parts are removed access to the traces wouldn't be a problem.
4. Similar to 3, but change or rip out parts not needed for on/off control with a lower 12V or 5V control signals. More work but simplifies the control logic to the board.
In my case, using only a 30V power supply, the power wasted in #2 is 3.8 watts total and only 1.9w per 5 watts resistor. I can live with that loss for ease of the mod. Also it gives the flexibility to control each half of the oscillator with the MCU. This may turn out to be useful if you plan use a flame sensor and keep a constant temp by vary duty the cycle and during startup the oscillator takes too long to stabilize. Maybe delaying the connection of one side may help the oscillator reach stability sooner. A scope trace would tell the story.
Last edited:
Good Morning, oliverpsmile,
Thank you for posting your code. It is a big help in studying Arduino and C.
I am interested in using Arduino and a flame sensor to pulse the coil for 1.88 seconds. This must be the line in your code where the flame sensor input is identified: "int PinIR = A3; // IR Sensor". However, I cannot see where that input is read and used to turn the coil on and off.
Can you point me in the right direction? Thanks again,
I hope you became proficient in Arduino programming.
int Glow=0; int AnnTime = 1880;
int F=XX; // Your sensor calibration for 1000 degree
// Use the annealing code from my code for AnnTime
Glow = Glow + analogRead(PinIR); // n times for a period
// For each period
if (Glow/n > 1.02*F){ Switch SSR OFF}
if (Glow/n < 0.98*F){ Switch SSR ON}
Good luck
int Glow=0; int AnnTime = 1880;
int F=XX; // Your sensor calibration for 1000 degree
// Use the annealing code from my code for AnnTime
Glow = Glow + analogRead(PinIR); // n times for a period
// For each period
if (Glow/n > 1.02*F){ Switch SSR OFF}
if (Glow/n < 0.98*F){ Switch SSR ON}
Good luck
Thanks for the quick response. I appreciate your help.
Closed-loop temperature control by current limiting
I'm planning an annealer build using a power supply that supports control of current limiting via CanBus, so I'm thinking of using this in conjunction with an IR/Flame sensor to implement closed-loop temperature control. What do the induction-heating experts think of this idea? Is current control a valid way to do this?
I'm also very interested in the idea of switching off the MOSFETs directly rather than using an external relay/SCR.
Main components are a 2500W ZVS board, 48V 2kW supply. Eventually I plan to mount a Dillon case feeder on top for continuous operation.
I'm planning an annealer build using a power supply that supports control of current limiting via CanBus, so I'm thinking of using this in conjunction with an IR/Flame sensor to implement closed-loop temperature control. What do the induction-heating experts think of this idea? Is current control a valid way to do this?
I'm also very interested in the idea of switching off the MOSFETs directly rather than using an external relay/SCR.
Main components are a 2500W ZVS board, 48V 2kW supply. Eventually I plan to mount a Dillon case feeder on top for continuous operation.
It may work, but I far from ideal. You would have to test how fast the power supply can control the current. Same problem with using an SSR to control current. Also you should insure that the ON state reguardless of current is more than 12V. These device have a time delay to respond. Usually the delay isn't symmetrical either. These delays limit the frequency that you can adjust heating coil. The typical Arduino can do at least 10K ADC conversions per second and with better coding you can get over 70K. For annealing 1K is probably more than enough. The controlling switch or power supply should have a response time at least 4x better, or you risk oscillation and or wasting a lot of power. Many SSR have off time of >=300 usec. The ZVS board also will have a delay from power on to reaching a stable oscillation.
Oscillation implies a very rapid process. The temperature fluctuations of the brass are very slow, thus maintaining annealing temperature for a period of time does not require rapid switching of the power.
I used a SSR to switch ON and OFF the annealing power in about 500 cycles/sec. As quoted, the SSR requires more time to react to the control signals. Thus, my cheap SSR naturally went busted. A good mosfet SSR requires at least 1 millisecond to do the switching.
I modified my program and replaced the busted SSR with another one. At the moment, the programed time for switching is 25 milliseconds.
It works. My average neck temperature is 1,000 F for a period of 1.88 seconds. True, I have some fluctuations of the neck temperature (950-1090 F), which I think is OK. However, I'm sure with the new SSR (1 millisecond switching time) on order, I could smooth the temperature curve.
A note - not all brass are created equal. I monitor the time when the neck reaches 1000 F. The time is not constant - +/- 0.3 seconds. The time also changes with the position of the brass. Therefore, in order to achieve consistency, I think it would be better to cut off the annealing power based on temperature level, not on length of time. Actually, VenatusDonatus is doing exactly that.
Great auto feeder...Where can I purchase?
Upgrades & Donations
This Forum's expenses are primarily paid by member contributions. You can upgrade your Forum membership in seconds. Gold and Silver members get unlimited FREE classifieds for one year. Gold members can upload custom avatars.
Click Upgrade Membership Button ABOVE to get Gold or Silver Status.
You can also donate any amount, large or small, with the button below. Include your Forum Name in the PayPal Notes field.
To DONATE by CHECK, or make a recurring donation, CLICK HERE to learn how.
Click Upgrade Membership Button ABOVE to get Gold or Silver Status.
You can also donate any amount, large or small, with the button below. Include your Forum Name in the PayPal Notes field.
To DONATE by CHECK, or make a recurring donation, CLICK HERE to learn how.