I was asking myself to solve a little problem what appends in a small project.
In this project, I want to drive a heating element (resistor) by PWM with a Arduino. The resistor (6Ohms) is powered with 12V (security question) by a Darlington transistor (TIP120). The frequency of the PWM is very low (1Hz, but it have to be like that), and resistor' temperature is controlled by the PWM ratio/cycle.
All's working very fine, except the resistor makes strong voltage variations in all the 12V line when it's powered on/off, and there's other electronics stuff power by the same line. I recorded variations around 1V and I need <0.1V.
I learned it's usual in switch-mode, but what's the best way (simple/sort place to make something) to isolate the rest of the line and other electronics parts?
I found some texts speaking about inductance, inductance + capacitor (low pass), but it's not very clear for me, and I'm not an engineer.. :-[
I've made a test circuit like this one:
(Pi filter bridge?)
With a 1000µF polarized capacitor and a 0.1µF capacitor (right side) but for the inductance it's write:
The inductance value is not too critical, but generally as big as you can manage, a few tens of mill Henrys is normal.
It makes me... Thinking :-X
Is a spare part from an ATX power supply (I'v an old one to use as parts) from the line +12V could be fine for my case?
I don't know the Henry value, but.. It's just before wires, associated with a 2200µF polarized capacitor (it looks like a small vertical cylinder).
Or a regulated one rated for less than the 2A you need to use a 6 Ohm resistor as a heater? Maybe it's going into current limit.
The battery is a dryfit 65A and power wires are in 2mm², so I think there's no problem with power.
I think it's append something like that:
I found this graph from a similar issue, not from my project.
PWM command is on the bottom of the graph and power line is on top.
A lot determines where you take your scope reading from, as long wiring runs can cause voltage drops. I would suggest first to place the scope probes across the battery terminals. If you still see a 1volt drop when the load is switched in I think it just means your battery has too high an internal impedenace either due to age or being undersized for the application.
Large filter caps, located at correct points, could compensate for the situation, and may be your only option?
A lot determines where you take your scope reading from, as long wiring runs can cause voltage drops. I would suggest first to place the scope probes across the battery terminals. If you still see a 1volt drop when the load is switched in I think it just means your battery has too high an internal impedenace either due to age or being undersized for the application.
In this case I think "disturbance" is propagated in all the line, up to the battery terminal.. Maybe we could see a small attenuation at the end of a long wiring, but the Arduino is located next to the switch source (<2") in my project.
Large filter caps, located at correct points, could compensate for the situation, and may be your only option?
In some "pro" applications (rail network, home air-conditioning automation), when load is very huge, I could see cap' only solution is not used. Because size (and cost maybe). Inductance + small cap are used.
And in my little case, I think same; to compensate 2A load during 1s I'll have to use very very huge caps (>100,000F?). Am I wrong? :-/
So it's not a solution (?).
For gross power-surge type decoupling, it is also common to use isolation diodes and "reservoir" capacitors to isolate the more delicate electronic load from the heating element load. Connect a diode between the battery positive side and the electronic loads (Arduino, etc.) The anode end goes to the battery and the cathode end points toward the electronic loads. Then at the cathode end of the diode, add the largest capacitor you can fit into the available space.
Hwoo, very interesting!
I already saw something like that, in a SMPS... Mhm.. Like in this schematic:
Is it what you speak about?
Diodes are Schottky, isn't it?
And I could use inductances like this one:
Something on the order of 5000 or 10000 uF at minimum. Note that the faster the PWM, the smaller you can make this capacitor and still be effective.
Yes, I know.. :-/
But 1Hz is very high for my project. I'll start with this frequency but I thought set the PWM ~1/8Hz to be fine... When a solution will be found, in future. Question of thermodynamics.
And side of this, the problem is not to keep the same voltage in the heating element during all the PWM cycle; it's just to solve voltage variations in the power line when switching.
The "switching line" (by which I assume you mean the current path from the battery to the heating element) IS the "main power line" by my terminology. The delicate loads (like Arduino) have a "clean branch" circuit off of the main path.
Yes, exactly, it's what I mean.
And I think I found a solution, it seems to work fine with an inductance + caps; like we can find in automotive decoupling devices. It's closer to what we can find in ATX power supply too, so I'll test it tomorrow with spare parts I have.
Thanks for helps, I'll post 'solution' when it'll confirmed.
I'm not sure I will be able to see it. I am currently in Johannesburg South Africa. (...)
Haaa South Africa is a nice country; push to Cap Town if you can, very cool and sexy place with the Table Mountain, incredibly blue sea, friendly people, fresh sea food restaurants, etc. It's one of my best trip' souvenir.
