PK90 Questions / Pill Dispenser

[tlo57]

Hi,

For my senior project we are suppose to design a pill dispenser that receives schedule information from a website and dispenses the specific pill accordingly to the time of the schedule. The Pill Dispenser would have to connect to the internet using Ethernet, and receive the schedules and somehow control the dispenser unit to dispense the right pill at the right time. Also it would need to incorporate a variety of functions such as an audio alert and visual alert to tell the patient when to take the pills. And dispensing the pills it would need to send a signal to the motors of the dispenser unit. We've decided to use a stepper motor to dispense the pills.

I am very interested in the PK70 development kit to control the pill dispensing unit. I guess I would have to ask if the PK70 could do these functions:
Control a Stepper Motor (I'm guess an 1/0 input into the controller chip of the motor)
Control a Display LCD (Display customizable caretaker messages to the patient)
Control a Speaker (Emit an audible alert when the pill has to be taken)
Receive input from the dispenser (Dispenser will tell the controller if the pill has been taken or missed so it can be logged)
Maintain a website (The caretaker must be able to log on and customize the pill schedule for the dispenser)
Power (I think stepper motors are in the range of 5V-12V) - is it compatible with the motor voltages?

Thanks for any help or suggestions,
Ted

[rnixon]

Hi Ted,

What electrical interfaces do you need for each of the items in your list? Do you have any h/w design capability, or are you looking for a finished set of h/w that has all the features in the list. For example, lcd, keypad, etc. Is the goal of your project to design with an embedded device (hardware), or implement software?

[Ran_Talbott]

Ted,

The first question to ask is what your course goals are. If you're supposed to be building something "realistic", you should probably be looking at something more like a MOD5270 or MOD5272, because real-world products will have cost minimization as a major design concern. Especially if (as it appears) you're desiging a "per-patient" dispenser.

If text messages are sufficient, do a google search for modules built around the long-established Hitachi HD44780 controller chip: you'll find dozens of examples of projects interfacing it to almost every processor ever made. You can also find the modules very cheap from electronic surplus dealers.

You can also google up scores of examples (especially hobby robotics) of projects driving steppers from micros. I personally like the L293D chip, because I've been using them for decades (literally), they're cheap, and they're simple. They also take care of the fact that your motor(s) will almost certainly want to run off a different voltage from your CPU.

Keep in mind that, if you're supposed to make a realistic machine, you'll need to support multiple kinds of pills, of wildly-varying sizes. You'll want to make sure that it doesn't wind up dispensing extra pills.

If you only need simple beeps, you can easily do that using an on-chip timer. You can make some more-complex sounds, too, with more software work, before you reach the point of needing to add something like a DAC and .WAV playback.

After you get back to us with the answer to "the first question", I may be able to offer you some more ideas that won't undercut the learning value of the porject and/or piss off your professor

Ran

[tlo57]

Hi,

Thanks for replying, and I have looked into the project more closely as the semester has progressed and my team has met up more often.

For the electrical interfaces, we are looking the modules, such as a speaker, display, power supply, stepper motors to buy and then connect them to the micro-controller. So our goal is to both design an embedded hardware, with all the capabilities needed to dispense the pills and alert our patient, web interface, etc... but also to program the micro-controller to display the message from the caretaker, get the schedule from the web, dispense the pills by controlling the stepper motors, enabling the speaker..etc.

The course goals is to build a device for our 'employer' that fits under a budget of $1000, that meets all the requirements specified from our 'employer'. So, cost is an issue with this project as there are a lot of modules that we need to buy. I looked at the MOD5270 and it is significantly cheaper than the PK70, but I'm not sure why - the features are very similar. Is it harder to programming the MOD5270 rather than the PK70, for the functions of (controller stepper motors, display, speaker,manage a website, get data from the website)? I guess that is a concern too, since at the end of the year in May, we need a working project. I'm just thinking if the PK70 gives us an edge in ease of implementation that could save us a lot of time on the programming aspect of the Dispenser and would allow us to focus more on the electrical, mechanical aspects. But I suppose if the MOD5270 can do everything required by the dispenser and is cheaper then that seems like the ideal choice for a micro-controller.

I attached our concept report below.
Thanks so much for the help,
Ted

[Ridgeglider]

Ted: My suggestion is to prioritize prototyping speed and flexibility for now, ie the PK70. If your 1st prototypes actually meet a market need, you'll have the opportunity to iterate, cost reduce, and customize later. By then you'll have a much better idea of the real-world (rather than imagined) requirements and your 2nd iteration can therefore focus on answering real problems. If you try to do that now by using a less-flexible platform, you may find yourself painted into a corner and lacking the features needed to quickly deliver the inevitable, suddenly-realized, make-or-break requirements that always come with the new product development process. You may never use all those features, but in the beginnng it's nice to be able to have some options that you can shed later when you know they're not needed. The 5270 would be a reasonable migration path for a streamlined 2nd generation effort. As long as you can tell that story you should be OK.

[tlo57]

Okay thanks, I think I'll buy the PK70. I'm thinking the 32 i/o pins should be enough the control the 12 stepper motors, and the serial can be used to control the display, the speaker will take one enable output from the pk70. Thanks, for you help.

[Ran_Talbott]

I'd also lean toward the PK70 for rapid prototyping, but probably not as a base for the finished product: in the final product, you don't need the enclosure, you'd probably want to put the SD card socket in a different spot, and you'd want to build all your electronics on a a single board you'd plug the MOD5270 into.

It looks like you did a good job of thinking things through for your preliminary design document, and did a very nice job writing it.

Suggestion: since you're doing quite a few steppers, "multiplex" them much like an LED display to save I/O lines and driver circuitry. Use high-side drivers to switch the positive supply among the steppers, and one common set of low-side drivers for all of them. And be sure to design your driver circuit so that no motor is powered on accidentally: during power-on-reset, the GPIO pins will probably be set as inputs (double-check the datasheet for whatever CPU you wind up using). It's common for I/O pins to have a passive pullup in such a configuration, which may be strong enough for your power control circuitry to think it's getting a high output from the CPU. You don't want that default state turning on motors. In the real world, you'd probably have the FDA and/or product liability lawyers pouncing on such a weakness.

Ran