Author Topic: What would be the most efficient way to step down and deliver 1A from a battery?  (Read 1090 times)

0 Members and 1 Guest are viewing this topic.

Offline WarFreak131

  • Contributor
  • Posts: 20
  • Country: us
I am building a project to regulate the temperature of my Hakko FX-901 portable soldering iron.  In it's stock configuration, it puts 6 V across a heater of 4.8 Ohms.  Assuming you're using a fresh set of 4x AA batteries (so the voltage is close to 6.4 V), the temperature immediately jumps up to about 950 F and then decreases linearly to about 575 F over a period of 70 minutes, at which point the batteries drain to a point of being unusable.

I don't like this setup.  First, 950 F will wear out the tips (which are expensive) quickly, and it spends about half of it's life at higher temperatures (further reducing tip life).  Second, 70 minutes isn't a lot.  Third, I don't want to have to buy AA batteries that much.

To resolve #3, I bought two 1100 mAh lithium ion AA-size batteries.  My approach to this is to step down the 7.4-8.4 V from the Li-ions to whatever voltage will cause the iron to heat up to 350 C only.  I used my power supply and found that the tip has a negative resistance coefficient, and will settle at 350 C @ 1.01 A with a resistance of about 4.3 Ohms.  Keep in mind, these tips have no feedback mechanism.  I cannot simply read a thermistor/thermocouple value and adjust accordingly.  All this testing was done empirically with a knock off Hakko tip thermometer.

My initial idea was to use an NPN transistor as a current source, but that would dissipate a lot of heat in the transistor.  My second idea was to use a MOSFET, but due to the negative resistance, the voltage that I set the gate to won't be sufficient once the heater resistance changes.  I could have a digital potentiometer that dynamically changes the voltage at the gate, but that would require a microprocessor and a current sense resistor.

My third idea is to use a buck LED driver.  As opposed to a typical voltage-mode buck converter, these are current-mode buck converters, which will step down the voltage as much as is necessary to deliver 1.01 A, usually at around 90% efficiency.  This should automatically account for the negative resistance changes in the tip.  My only hesitation is that since the tips have a rather low thermal mass, they respond to current changes quickly, so would I need a large output capacitor to smooth the current out?

Does anyone else have any suggestions on simple and effective ways to source current accurately and efficiently?
« Last Edit: September 16, 2020, 02:31:35 pm by WarFreak131 »
 

Online ledtester

  • Frequent Contributor
  • **
  • Posts: 958
  • Country: us
I think you're asking a lot out of such a simple soldering iron and AA batteries.

Any kind of of constant voltage or constant current regulation is going to result in a lot of inefficiency. What would really help is constant temperature regulation, but that needs a temperature sensor in the iron - preferably the tip.

If you want a better battery-powered iron I would look at the TS100 which lots of people power from lithium-ion battery packs or the TS80 which can be powered from a QC 3.0 power bank. Just search for "TS100 battery" or "TS80 battery".
 

Offline WarFreak131

  • Contributor
  • Posts: 20
  • Country: us
I considered that, but this is purely for my own edification and enjoyment, and as a challenge.  I already have a decent soldering station.
 

Offline MosherIV

  • Super Contributor
  • ***
  • Posts: 1489
  • Country: gb
Hi

If you just want pure fixed voltage or current  then option3 (some kind of buck converter) is your best bet.

The other option is to build your own temperature control system but use pulse width modulation to vary the power. In effect you would be designing your own TS1000
 

Online ledtester

  • Frequent Contributor
  • **
  • Posts: 958
  • Country: us
Maybe replacing the slide switch with a push button would be a good mod to give you more control of how hot the iron gets.
 

Offline WarFreak131

  • Contributor
  • Posts: 20
  • Country: us
Hi

If you just want pure fixed voltage or current  then option3 (some kind of buck converter) is your best bet.

The other option is to build your own temperature control system but use pulse width modulation to vary the power. In effect you would be designing your own TS1000

Using PWM was one of my original ideas, but someone brought up the point that having doing so may cause inductive or capacitive coupling at the tip which would put the tip at a (possibly) high enough potential to damage sensitive components.

Maybe replacing the slide switch with a push button would be a good mod to give you more control of how hot the iron gets.

That's an interesting idea, I may just roll with that.
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 11896
  • Country: us
  • DavidH
Low frequency pulse width modulation will be the most efficient method, and it is simpler than a buck regulator since it requires no inductor or second switch.
 

Offline magic

  • Super Contributor
  • ***
  • Posts: 2618
  • Country: pl
Adding an inductor is series with the heater will smoothen the switching edges without much energy loss. Be sure to add a flyback diode. And by the way, you now have almost a buck converter.

There is a possibility of using the heater's thermal coefficient for sensing. I have seen a station which used PWM and sensed heater resistance during off time by pulling it up with a resistor paralleled with the power switch. If you choose variable DC drive, I suppose you could add a small resistor in series with the heater and compare a fraction of heater+resistor voltage against resistor voltage.

I might be the one who scared you about voltage transients at the tip ;)
Recently somebody provided compelling evidence that they may not be much of a problem if there is no long cable and the controller is in the iron itself. The only way to know for sure is to measure it on your unit. Easily done if you have a scope.
 

Offline WarFreak131

  • Contributor
  • Posts: 20
  • Country: us
Adding an inductor is series with the heater will smoothen the switching edges without much energy loss. Be sure to add a flyback diode. And by the way, you now have almost a buck converter.

There is a possibility of using the heater's thermal coefficient for sensing. I have seen a station which used PWM and sensed heater resistance during off time by pulling it up with a resistor paralleled with the power switch. If you choose variable DC drive, I suppose you could add a small resistor in series with the heater and compare a fraction of heater+resistor voltage against resistor voltage.

I might be the one who scared you about voltage transients at the tip ;)
Recently somebody provided compelling evidence that they may not be much of a problem if there is no long cable and the controller is in the iron itself. The only way to know for sure is to measure it on your unit. Easily done if you have a scope.

Yes! It was you who told me about inductive coupling. XD

You'll have to explain a little more, perhaps with a schematic, of your explanation about testing the thermal coefficient.  But for now, I think I found something to solve my design issue.

https://www.ebay.com/itm/1-8A-SXD-Super-X-Drive-Laser-Driver-M140-PLTB450B-NDB7875-NDG7475/183726019793?ssPageName=STRK%3AMEBIDX%3AIT&_trksid=p2060353.m1438.l2649

It's a programmable current source that is small enough to fit into the battery compartment of the soldering iron.  I think I may just buy this.  But sometime in the future, I may make something that measures the voltage/current to work out a resistance, which can be mapped to a temperature.
 

Offline magic

  • Super Contributor
  • ***
  • Posts: 2618
  • Country: pl
Wheatstone bridge, basically.
The two voltages are equal when heater resistance is exactly 4.3Ω.
The lower resistor will dissipate a bit of heat, tough luck.
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 11896
  • Country: us
  • DavidH
You'll have to explain a little more, perhaps with a schematic, of your explanation about testing the thermal coefficient.

The heating element has a rather high temperature coefficient of resistance so if powered with pulse width modulation, the heating element resistance can be measured and sampled during the off period.
 

Offline WarFreak131

  • Contributor
  • Posts: 20
  • Country: us
Wheatstone bridge, basically.
The two voltages are equal when heater resistance is exactly 4.3Ω.
The lower resistor will dissipate a bit of heat, tough luck.

Wouldn't this mean that once the soldering iron reaches temperature, both the heating element and the resistor are both dissipating ~4.3W?

You'll have to explain a little more, perhaps with a schematic, of your explanation about testing the thermal coefficient.

The heating element has a rather high temperature coefficient of resistance so if powered with pulse width modulation, the heating element resistance can be measured and sampled during the off period.


What method do you suggest to measure the resistance using a uC like an Atmega chip?  Off the top of my head I can imagine passing a known current through it via a voltage regulator in current mode and measuring the voltage at the output of the regulator.
 

Offline David Hess

  • Super Contributor
  • ***
  • Posts: 11896
  • Country: us
  • DavidH
What method do you suggest to measure the resistance using a uC like an Atmega chip?  Off the top of my head I can imagine passing a known current through it via a voltage regulator in current mode and measuring the voltage at the output of the regulator.

That is what I would do, with the provision that the low sensing current can be in parallel with the PWM drive so there is no need to turn it on and off.
 

Online ledtester

  • Frequent Contributor
  • **
  • Posts: 958
  • Country: us
If you're going to employ a uC, maybe use a thermocouple and a MAX6675:

https://youtu.be/RxZJujS1znU
 

Offline magic

  • Super Contributor
  • ***
  • Posts: 2618
  • Country: pl
Wouldn't this mean that once the soldering iron reaches temperature, both the heating element and the resistor are both dissipating ~4.3W?
What makes you think so? Power dissipated in a resistor is current times voltage. Current through the two series resistors is equal but voltages very much are not.

What method do you suggest to measure the resistance using a uC like an Atmega chip?  Off the top of my head I can imagine passing a known current through it via a voltage regulator in current mode and measuring the voltage at the output of the regulator.
No need for a regulated current source, just a resistor. Measure what fraction of VCC is the voltage at the heater then, do the math.
 

Offline WarFreak131

  • Contributor
  • Posts: 20
  • Country: us
If you're going to employ a uC, maybe use a thermocouple and a MAX6675:

https://youtu.be/RxZJujS1znU

Thanks, I'll give this a watch.

Wouldn't this mean that once the soldering iron reaches temperature, both the heating element and the resistor are both dissipating ~4.3W?
What makes you think so? Power dissipated in a resistor is current times voltage. Current through the two series resistors is equal but voltages very much are not.

In the diagram you sent (ignoring the small resistors, and I'm assuming they're supposed to both be 100m?), there's 4R3 resistor and the heater.  Once the voltage across the bridge equals zero, that means that the heater has reached the proper temperature, due to the current passed through it at some voltage V.  But if the heating element reaches 4R3 when its passing 1 amp, that must mean that the other leg of that bridge is also passing 1 amp and I2R = 12*4R3 = 4.3W.

Scratch that, I understand.  I'll give it a shot.
« Last Edit: September 22, 2020, 02:10:10 am by WarFreak131 »
 

Offline magic

  • Super Contributor
  • ***
  • Posts: 2618
  • Country: pl
Note that this was for driving the heater with variable DC when power is always applied to it.
With PWM, just connect 1kΩ or something like that from VCC to the heater and measure output of the resulting divider when power to the heater is off.
 

Offline WarFreak131

  • Contributor
  • Posts: 20
  • Country: us
Note that this was for driving the heater with variable DC when power is always applied to it.
With PWM, just connect 1kΩ or something like that from VCC to the heater and measure output of the resulting divider when power to the heater is off.

Can you provide a diagram?  I'm not quite following what you're saying.  VCC>1k>heater?  Wouldn't that cause the total current to drop, so then I'd never make it to 1A?
 

Online ledtester

  • Frequent Contributor
  • **
  • Posts: 958
  • Country: us
Here's an idea:

Code: [Select]
                   ADC
                    |
  Vbatt --- heater -+- MOSFET --- GND
                    |
                    \
                    /  R
                    \
                    |
                   GPIO

Normally the GPIO is an input. When the MOSFET is off, turn the GPIO into an output pin with a value of LOW. Then take an ADC reading. Finally turn the GPIO back into an input.
 

Offline WarFreak131

  • Contributor
  • Posts: 20
  • Country: us
Here's an idea:

Code: [Select]
                   ADC
                    |
  Vbatt --- heater -+- MOSFET --- GND
                    |
                    \
                    /  R
                    \
                    |
                   GPIO

Normally the GPIO is an input. When the MOSFET is off, turn the GPIO into an output pin with a value of LOW. Then take an ADC reading. Finally turn the GPIO back into an input.

Got it, thank you.  What is a safe way to determine whether the PWM value is low, and will remain low for long enough to read the voltage correctly?  The Atmega328P PWM runs at 490 Hz, and I believe it's independent of the rest of the code.  The PWM period is 2ms, and the low period will be half of that, 1 ms.  The analog read function takes 100 us, so I can take 10 reads in the period of the low phase of the PWM signal.  There's a reasonable chance that the voltage will eventually read during the high phase if I were to just chance it.
 

Online ledtester

  • Frequent Contributor
  • **
  • Posts: 958
  • Country: us

IIRC, on the atmega328 automatic PWM is done by setting an output compare register to a certain value. When the timer hits that value a GPIO pin is toggled.

You can also generate an interrupt when the timer hits that output compare value. So you should be able to tell exactly when the GPIO pin is set to zero by the timer unit. All of the details are in the 300 page datasheet - a very worthwhile read.

In any case it probably would be a good idea to disable the MOSFET gate pin while you are doing the ADC reading. The reading itself might only take 20 microseconds so even if you attempt a read when the MOSFET is supposed to be on it won't affect the heater that much.

 

Offline magic

  • Super Contributor
  • ***
  • Posts: 2618
  • Country: pl
Don't bother with the GPIO, wire the resistor permanently to ground.

That means there will always be a few mA flowing through the heater so it will never really, fully turn off, but who cares :-DD
 

Offline Haenk

  • Regular Contributor
  • *
  • Posts: 245
  • Country: de
If you have a defined targetvoltage and targetcurrent, maybe this dirt cheap LED supply board will do:

qs-2405ccbd-3a

(version with 3 potentiometers)

It's not large and will well go with some batteries.
Might be worth a try and see how it behaves. Advantage: It works on a large range of input voltage.
 

Offline WarFreak131

  • Contributor
  • Posts: 20
  • Country: us
Thank you everyone for the input.  If I'm driving this by PWM through a MOSFET, what is the heat dissipated through the MOSFET?  Is it just I2RDsON?
 

Online ledtester

  • Frequent Contributor
  • **
  • Posts: 958
  • Country: us
My understanding is that RdsON is the resistance when the MOSFET is ON and thereby represents the minimal value. The actual power dissipation will depend on how fast the MOSFET is driven ON and OFF by the gate signal.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf