Honey Run Apiaries-A Beekeeper’s Blog

One critical part in the honey stick machine is the optical sensor.  They will be used in several locations in the honey stick machine to determine if a straw is in position and when honey is in the straw at a particular location.  it consists of a very simple circuit containing an IR LED and transistor.  The output voltage is detected by the Phidgets UBS interface.

The circuit is setup so that the output voltage increases when the IR is blocked.  Testing showed that the circuit was plenty sensitive enough to determine the varying levels of light when a empty or full straw was placed between the LED and transistor.   I did observe one very interesting thing when the straw with honey (or water) was centered between the LED and Transistor.    In this case the IR intensity increased instead of decreased as one might expect.  I suspect in this case the straw is acting as a lens, focusing light on the transistor.  Offsetting the IR LED and transistor so that it lines up with the bottom half of the straw eliminates this problem and actually increases the voltage difference between honey and no honey in the straw.

In keeping with the modular theme for the prototype, I’ve assembled the AC Driver board as a separate circuit.  It contains 4 of the very simple AC driver circuits designed earlier.  Assembly and testing was uneventful.  There is a lot of wasted space on this board that can be eliminated later with a custom printed circuit board, but I definitely wanted to keep the 120VAC circuit separate from the others for the prototype.

There is also be portions of the honey stick machine, including heaters and pump, that will need to be run on 120VAC and sill be switched on and off by the Phidgets interface.  Relays would be the simplest option here, but again these mechanical relays have a far less than desirable rated lifespan in number of cycles.   So I’ve opted to use solid state optically isolated AC relays.  They typically have a far lower amperage rating, but I only expect to drive a maximum of 0.5 amps presently on any given AC device.  The circuit pictured shows the very simply NPN driver circuit triggering an LED and the solid state relay.

The local association’s meeting was canceled tonight due to snow.  This gave me a little bit of time to work on the honey stick machine that’s been left sitting for awhile while I caught up on other things.  I was able to finish and test the final 4 to 8 decoder previously built on a prototype board.  For testing I connected to the DC driver board simply to have a display of the output on the LEDs.  I was a bit worried at first because it didn’t work at all, but then I found some insulation had melted allowing two wires to come into contact with each other.  After fixing that issue the circuit worked perfectly while running it though some 20,000 on/off cycles.

One thing I learned building this circuit was how slow and tedious it can be building these circuits on prototype PCBs.  The boards have no circuit traces and every wire needs carefully positioned and soldered by hand.  Should I get to the point where I can produce even a small number of machines, I’ll definitely be looking to design a printed circuit board.  Fortunately that’s almost as easy as desktop publishing now with software like PBC123 that lets you design and order custom circuit boards in small numbers.

One of the challenges in the circuit design that I was worried about was obtaining the additional outputs I expect to need for the honey stick machine.  The Phidgets USB controller only has 8 outputs and while I could simply add a 2nd Phidgets module, I’d really rather avoid the extra expense in the final model.  I had proposed using a 4 to 8 decoder previously to add the extra outputs.  The good news is that the 74259 addressable 8 bit latch works quite well and replaces the entire circuit proposed.  The bad news is that the 4 outputs from the USB controller aren’t guaranteed to update at the same time.  This means that random signals can be generated on the output.  Not a good thing.

The solution? An edge detection circuit on all 4 inputs to generate a pulse any time the input changes, a delay timer to debounce the multiple pulses sent when the inputs change at different times and a 2nd timer to send the all clear to the 8 bit latch signaling that it’s ok to look at the input signals now that they are stable.  The Phidgets controller updates the output signals approximately every 8mS, so it’s quite possible that when multiple inputs are changed that the change will occur over two consecutive updates.  So if you wait and look at what signals changed over 12mS (1 1/2 update cycles), you should catch any output that occurs over two cycles.  Or at least that is the theory.

It’s been about 12 years since college and since I’ve done any circuit design and I really don’t have access to the oscilloscopes , etc. needed to see really what is happening with the signals in the circuit.  Though I’m  happy to say that something must have stuck after 4 years of college, the circuit works flawlessly.  As long as I limit the software to sending out signals every 16mS , I can selectively turn on and off each individual output using just 4 outputs from the Phidgets controller.  This does have the downside that it will be slower and only one  output can be addressed at a time,  but many of the outputs needed are not  so time sensitive  so this should work well for them.  Those that are time sensitive can be driven by the remaining 4 outputs of the Phidgets controller directly.

Given the unknowns  developing a machine such as this, I’ve opted to build much of it in smaller functional blocks.  This certainly won’t be a cost effective way to produce machines if we end up marketing it, due to the increased cost of parts.  And assembling circuits on these generic prototype boards is quite time consuming.  But it is just about ideal for this stage of development.  I’m happy to say the DC Solenoid driver board is completed and working.  It’s basically just 8 of the mosfet driver circuits wired on a single board.  It is fairly simple but should be much more reliable than mechanical relays, and it’s one step closer.

It’s winter again, or at least it looks like it out the window.  We had several inches of snowfall Tuesday night and it looks like it’s going to stick around for awhile.  So it’s time to get back to the honey stick machine.  It’s been over a year now since I designed the simple MOSFET solenoid driver circuit, and I’ve finally constructed it on the prototype board and tested it. I’m happy to say it worked flawlessly.  It’s probably overkill for the lower power solenoids, but keeping them all the same (the project calls for 7 drivers) will allow significant flexibility.  This is good since I don’t know what I’ll get into once I get into the mechanical side of things.

The picture, from left to right, shows the power supply, Phidgets USB interface, and the prototype board with solenoid on the right side of it.

Now you can own your own ‘non working’ orignal Sticky Machine.  Only $430 right now on ebay, a fraction of the original $3000 price tag.  Unfortunately it sounds like the seller had as many problems with the machine as others I’ve talked to about it. 

Having filled quite a few honey sticks by hand, it’s not really surprising there are some problems with it.  The viscosity of the honey can vary significantly from jar to jar, at least enough to throw off any sort of timing for filling and sealing.  Temperature of the heating elements plays a big part in sealing time.  Too cool and it seals slowly, too hot and it may melt though the plastic breaking the seal.  It’s really a pretty tricky operation to get right, even by hand.  So with apparently little or no feedback in the system I’m not sure how one would get it to consistently work.

With a change of jobs and being in the middle of bee season, I haven’t had much time to devote to much else, including working on the honey stick machine.

I have noticed however that Dunbar Honey Farm is now advertizing their honey stick machine in the American Bee Journal for $1900  (Phone 586-770-9953).  I suspect this is the same machine he had a year ago when my wife talked to him.  If so, it is a relatively manual system  that uses a specially made pump, a holder for a dozen or so straws and an impulse sealer.  It certainly beats filling them individually with a syringe, but isn’t yet the automated machine I’m hoping for (though I’m sure it’s far cheaper).  If anyone finds out differently please let me know.

I recently was informed of another Honey Stick Machine it is on the market. It is not automatic, but sounds like it may work a bit faster than how I’ve currently been filling straws. It consists of a special pump that is made special in india that fills the straws with raw, unheated honey and special trasy that hold 14 straws for sealing. The gentleman selling them says that with two people he can fill and seal 1000 straws per hour.

The machine costs $1000 and is available from Dunbar Honey Farm, Jerry Dunbar, 586-770-9953. I do not know if that price includes the sealer or not (I’m assuming he uses an impulse sealer as I do). If anyone gets to see this machine or has one, please let me know what you think of it.

Update: The gentleman who is offering this machin says he knows the the person who purchased the Patent for the Sticky Machine and that that it isn’t being produced yet because he has been unable to get the design to work. This seems to confirm the rumors I’ve heard elsewhere. Dissapointing as I would have loved to save all the time and money of designing my own. I had some hope that the Sticky Machine would soon be on the market again as I have know some people who have seen the Sticky Machine’s Inventer with a working version of the machine, but it appears it won’t be any time soon.