temperature swinging

Here are my extruder-files.
It took some time, because in my first variant I had troubles with the serial-communication. (packet timeouts). the 0,1uF is also a bad choice. because the duration for the measurement for e.g. 153

The extruder description describes how to calculate beta from measurement and then calculate Rz. Their is a calculator that takes these values and tell you the value you need for your capacitor. Here is the site [reprap.org]

I was always kind of interested in how Simon managed to the the 16F628A to read an analog signal off of that thermistor when the 16F628A doesn't have an A/D circuit. I know he had some neat, sneaky way of doing it, but I can't remember what it was. I wonder if the swinging effect is somehow happening in the code that does that A/D read.

I know from the A/D work I've done on the 16F877A and the 18F6510 that you have to wait a while after you read the A/D circuit for the reading to stabilize. I also seem to remember that if you have open pins to either side of the A/D pin you are using that that can cause you problems, too.

I'm not saying that any of that is happening, but those are some problems that I've run across doing A/D work on PIC's.

He does the A/D conversion by charging a capacitor through the thermistor and timing how long it takes. The erroneous values are caused by the timing measurement being interrupted.

---

[www.hydraraptor.blogspot.com]

Oh Lordy! You're right! I remember now. Struck me as a lot of bother to go through at the time so that you wouldn't have to change PIC chips. There are, after all, a lot of inexpensive 16F's that have A/D.

BTW, I think I have worked out why the RepRap extruder is thermally insulated. I think it is because it has a fan cooling the workpiece so the barrel needs covering to stop it being cooled as well.

---

[www.hydraraptor.blogspot.com]

I run a fan on mine, too. I just point the power supply exhaust fan at the work piece.

Hot air rather than cold then.

---

[www.hydraraptor.blogspot.com]

Lol! I don't draw enough power off of my ATX power supply for the air coming off of it to get warm.

Forrest Higgs Wrote:
-------------------------------------------------------
> Oh Lordy! You're right! I remember now. Struck
> me as a lot of bother to go through at the time so
> that you wouldn't have to change PIC chips. There
> are, after all, a lot of inexpensive 16F's that
> have A/D.

Actually thinking about this while measuring the thermistor englewood sent me it is probably not a bad way of doing it because the resistance varies over such a large range. You can easily measure time with as many bits as you want but ADCs with a large number of bits are expensive. Come to think of it I think that is basically how digital multimeters work.

---

[www.hydraraptor.blogspot.com]

Dirt cheap PIC's get you multichannel 10 bit A/D. How much resolution do you want?

Well my thermistor goes from about 2K at 20C to 20 ohms at 200C. To cover that range with a 10 bit A/D only gives you 10% accuracy at the 200C end. If you are measuring time, as long as your timer is fast enough to get good resolution at the high temperature it is no problem slowing it down with a prescaler to get similar resolution over a very wide range indeed.

I suppose in this case we are not interrested in temperatures below say 80C so if we let temperatures below that go off scale then 10 bits would be adequate. It would certainly simplify the software a bit and people would only have to change a resistor for different thermistors rather than a cap.

In my own case I am using TMS430 chips with A/D convertors so I will probably do it that way anyway.

---

[www.hydraraptor.blogspot.com]

Thinking out loud here but putting a resistor in parallel with the thermistor probably helps with making it a bit more linear. It would certainly tend to cramp up the high resistance low temperature end where the resistance is changing rapidly and have negligible effect at the high temperature end.

---

[www.hydraraptor.blogspot.com]

Yes putting a fixed resistor in parallel with a thermistor can make it almost a straight line over the range we are interested in. In my case I used 100 ohm across a thermistor that has a resistance of about 2K at room temperature and 20R at 200C.

I have put some graphs up on my blog [hydraraptor.blogspot.com]

---

[www.hydraraptor.blogspot.com]