PDA

View Full Version : Computer-Controlled Variable Voltage Source?



roblad
02-04-2008, 01:48 PM
So I have a strange question. I have been looking around online for a while, and having no luck. Who knows, maybe you guys can help me? Here goes:

I am attempting to make a temperature controlling unit. I know I could go buy a PID temperature controller and relay and all of that, but I'm trying to do one that is computer-controlled, and have as-of-yet been unable to find one. I was looking through the Phidgets and Trossen catalogs, and found almost everything that I would need. The temperature sensor Phidget, the 8/8/8 I/O board. I even wrote software for a PID when I was finishing my thesis in college. The problem is, that software belongs to my alma mater, and was written in LabVIEW using National Instruments hardware. What I'm looking for is a computer-controllable variable voltage source, preferably USB powered, but that's not necessary. With that, I'd then just run the PID algorithm to determine what voltage to send.

I'd like to have a voltage range of -6V to 6V or so. It doesn't need that much current, 2A or so would be fine. Unfortunately, I just don't really know where to look for this sort of thing. Hope somebody knows where I can get something like this.

Matt
02-04-2008, 04:23 PM
Just an FYI, if you haven't seen it yet, check out the USB-Q "The Automated BBQ Grill Temperature Control. This might give you some more ideas to work with also.

http://forums.trossenrobotics.com/showthread.php?t=1141

roblad
02-04-2008, 05:08 PM
I actually did see that. As a matter of fact, that's how I found this site. :happy: The difficulty is that I'm looking to build a solid state version. Meaning I'm planning on running the variable voltage through a thermoelectric cooler. Wonderful little things, but rather than being able to drive a servo motor to open the air vent on the barbeque a certain amount, I need to run varying amounts of voltage through the TEC. Other than that, the system will be remarkably similar. Though I'm also going to try to make the whole thing battery powered so it can actually be mobile.

Dave
02-04-2008, 05:16 PM
I'd like to have a voltage range of -6V to 6V or so. It doesn't need that much current, 2A or so would be fine.

One quick and dirty way to accomplish this would be to filter the output of a DC motor controller. The motor controller generates a PWM signal, so all you'd need is a lowpass filter circuit to produce a variable analog voltage. The output wouldn't be pretty, but as long as you don't need a perfectly stable DC signal, it should suffice.

fitchett
02-04-2008, 10:20 PM
Just use a motor controller to direct drive power resistors. A product like the PhidgetMotorControl HC has two outputs, and each output would give you variable control up to 200 Watts, depending on how you design your resistor bank. Digikey sells lots of power resistors.

Chester

roblad
02-05-2008, 01:22 PM
Hrm. Well thanks for the ideas, I think I'll just have to get some components and fiddle a bit. My electronics chops are a bit rusty, so I'll have to hone those again. If anybody has any more ideas, keep 'em coming. I'll probably be working on this for a while. :happy:

Dave
02-05-2008, 03:49 PM
Chester's right, concerning power resistors. Although I've never seen a Peltier cooler driven directly by PWM, you should get the same result. Let us know how it works out!

fitchett
02-05-2008, 08:39 PM
That's an interesting enough idea that I just bought a couple peltier cells to try it out.

Chester

tom_chang79
02-05-2008, 09:12 PM
roblad,

To accomplish this, you need two components. A digital potentiometer and an adjustable voltage regulator. You need to select digital potentiometer's interface depending on how you want to control it, I2C, SPI, etc... You then need to select an adjustable voltage regulator. Then you need to look at that particular datasheet to see what the R1/R2 feedback would be to determine the digital potentiometer's R1/R2 range.

A simple digital potentiometer will work as a variable voltage source, but it depends on the load. If you have a high impedance load, it should be fine, but once your impedance starts to drop, there is nothing there to regulate the voltage since it's a simple divider, and will follow the basic Ohm's law.

roblad
02-06-2008, 12:07 PM
Well Chester, let me know how your attempts turn out, I've got about 6 projects going right now, so you may beat me to the punch in trying to make this work.

Rob

fitchett
02-06-2008, 08:27 PM
It's a piece of cake. I managed to break one Peltier Cell, so I was only able to produce 10C to 78C within the 8"x6"x4" insuated temperature chamber with the remaining cell. Treat the wires on the cells very gently - they are fragile. So, in conclusion, the 1064 will control Peltiers, no problem. Of course, the response of the temperature is so slow, that you could just use a 3051 with your 888 to build a simple H-bridge, and run a very slow PWM period - 30 seconds would likely be fine, especially after you reach a steady state.

Chester

Alex
02-07-2008, 03:35 PM
Thanks Chester!

Just an FYI for everyone reading this thread:

1064 (http://www.trossenrobotics.com/PhidgetMotorControl-HC.aspx)

3051 (http://www.trossenrobotics.com/store/p/3204-Dual-Relay-Board.aspx)

888 (http://www.trossenrobotics.com/store/p/5341-InterfaceKit-8-8-8.aspx)

:)

robot maker
09-22-2008, 03:17 AM
what voltage and current do you need i can design a circuit with a DAC and a micro processor and on the DAC output goes to a op amp to adjustable regulator,i made a design close to it for a project ,but i used a BCD DAC with switches

6 volts to -6volts at 2 amps not to hard to design,coding the micro can be a problem for me
but with a basic stamp could be easy

kiwikungfu
01-19-2009, 07:50 PM
It's a piece of cake. I managed to break one Peltier Cell, so I was only able to produce 10C to 78C within the 8"x6"x4" insuated temperature chamber with the remaining cell. Treat the wires on the cells very gently - they are fragile. So, in conclusion, the 1064 will control Peltiers, no problem. Of course, the response of the temperature is so slow, that you could just use a 3051 with your 888 to build a simple H-bridge, and run a very slow PWM period - 30 seconds would likely be fine, especially after you reach a steady state.

Chester

Pardon the beginner question, but this thread is very similar to my project.

I need to replace a potentiometer (two 0-5vdc synched outputs) with a computer controlled output. The phidget 3051 or 1060 might be only piece of hardware I need, but I don't see any data on their outputs (in volts DC) with, say, a 12V power supply. Would either of these units be able to output 0-5vdc controlled by the software run on the hosting computer?

If possible, a battery powered or USB powered (versus household AC like the 12V power supply I see suggested for use with the 1060 LV motor controller) would ensure portability, as I'm hoping to drive this from a laptop.

Thanks for the answer to such a beginner question,
Joe

Adrenalynn
01-19-2009, 07:58 PM
Welcome to the forum! How much current do you need to sink?

kiwikungfu
01-20-2009, 12:48 PM
I will be measuring the amperage, impedance, and capacitance of the potentiometer to be replaced and posting them to this thread shortly. The is potentiometer to be replaced with a phidget is powered from a standard automotive battery. Much appreciated,

Joe

Adrenalynn
01-20-2009, 01:50 PM
Oh - you're needing analog out. I was thinking digital out. I wasn't aware that any of the phidgets boards were analog output boards.

http://futurlec.com/DC_Opto_Output_4.shtml

4 channel analog ouput board, TTL interface (would require an adapter to serial or USB for PC interface), 12v powered. You would also need to buy or make a small voltage divider to get the 12v down to 5v.

robologist
01-20-2009, 04:07 PM
Would a digital pot (http://www.maxim-ic.com/products/digi_pot/) combined with some sort of Phidgets GPIO or serial interface work?

kiwikungfu
01-21-2009, 11:25 AM
Oh - you're needing analog out. I was thinking digital out. I wasn't aware that any of the phidgets boards were analog output boards.

http://futurlec.com/DC_Opto_Output_4.shtml

4 channel analog ouput board, TTL interface (would require an adapter to serial or USB for PC interface), 12v powered. You would also need to buy or make a small voltage divider to get the 12v down to 5v.

Ok, I've taken some measurements of the potentiometer to be replaced. The inputs are 2 5vdc leads, and the outputs are 2 synched values ranging from >.2vdc to <4.9vdc. My fluke 70 III doesn't have amperage, but I'm anticipating the load is relatively low (car battery range).

Just when I knew I didn't know very much I realise there is so much more to know- a digital out can range from 0-5vdc? Wouldn't the output, if we had digital output, be at a constant voltage but toggle on and off to simulate binary states? Oh, I see, I wrote previously "ranging from 0-5vdc" whcih does sound like an on-off relationship, my apologies!

So yes, looks like I need an anolog output that can take 2 5vdc inputs and output 2 (synched) .2 to 4.9vdc variable outputs based on a C# algorithm, hopefully connected to a host laptop by USB (or RS232 or any standard interface).

Aren't the Phidget 3051 or 1060 effectively anolog outputs of DC voltage, mimicking an analog potentiometer? Again, please pardon the beginner question!

jes1510
01-21-2009, 11:32 AM
You could use an op-amp with a digital pot setting the gain which would in turn be controlled by a Microcontroller.