Friday, December 27, 2013

CryoBUG: Running Load Tests

Having zeroed in on the refrigerant charge I would like to use, I thought it was time to create a heat load versus temperature chart. Since the Demo unit has a vacuum insulated cold head, it doesn't lend itself very easily to attaching a load resistor. So I pulled out of storage my original bread-board prototype and recharged it with the same refrigerant as I presently have in my Demo unit (HC/Argon blend).

Heat Load Test Set-Up

Load Resistor
For the load heater I used a 6 ohm 50 watt resistor suspended directly in the center of my 6 foot long 1/4" OD tubing evaporator coil. The evaporator was then wrapped in multiple layers of aluminum foil to help contain the heat emanating from the resistor within the interior space of the evaporator coil. The entire thing was then insulated with fiberglass.

My heat load power supply consisted of a 0-18 VDC 3 Amp adjustable unit with digital readouts for current and voltage. So basically this gave me a top range of nearly 55 watts.

Monitoring of all the temperature data was via a USB-TEMP module made by Measurement Computing, and running a DasyLab charting application on my laptop computer.

Test data was collected at zero applied load, and then at 5 watt increments ranging from 5 to 45 watts applied. Estimating the added static heat load at 10 watts from insulation losses, I would say that the total heat load with 45 watts applied, is in reality 55 watts (10 W static + 45 W applied). This is pretty much the absolute maximum heat load that CryoBUG is capable of handling with a -150 °C refrigerant charge. And to be conservative, I would rate it for 50 watts total, because it starts to lose it after that (see chart below).

CryoBUG Temperature versus Heat Load Chart

From some of my earlier refrigerant charge tests, I can see the possibilities for taking on more heat load, assuming that the ultimate temperature requirements were to be relaxed.

Thursday, December 19, 2013

CryoBUG Rev 2 Controller Boards on Order

The printed circuit boards got ordered from ExpressPCB a couple of days ago, and should be here by tomorrow. I went for the Mini-Board special, which gives you three 2.5"x3.8" boards for as little as $51. However I went for the enhanced version, which includes the soldermask and silkscreen, giving it that production quality look.
(Click on Image to Enlarge)

This time around I was able to go with a smaller transformer, which along with a reduced component count (thanks to the ECIO-28 micro-controller), allowed for the smaller sized board while still retaining the dual voltage aspect (115/230 VAC).

I had some left over I/O on the ECIO-28, so I squeezed in an expansion port (J4), which has an LCD ready pin-out for an optional display.

I am rather pleased with the design, and its compact size.

(Click on Image to Enlarge)
As can be seen in the schematic, the design is very simple compared to the previous version. This of course is due to the micro-controller taking over much of what was done with discrete circuitry, and moving the complexity over to the firmware.

As I mentioned in my previous post, the firmware development is being done with Matrix Multimedia's FlowCode product, which greatly reduces the effort of writing the code and programming the chip.

Since the ECIO-28P has a built-in USB connector and a boot-loader, it is a one button operation within FlowCode to transfer the code to the actual hardware. So writing, testing, and updating firmware is actually a fun process.

As I mentioned earlier, there is also an expansion connector as part of my design suitable for connecting an LCD display (or an OLED, VFD). Although I wont be using the display aspect on my current project, I thought it might be nice to have the flexibility to do so down the road.

Since LCD back lights require a fair amount of power, a VFD or OLED display might be more within the power range of my controller board. So just for fun, I placed an order for a small OLED 8x2 character display from CrystalFontz to see how that might look.

Possible uses could be for an operational status display, or a way to display how many faults, and of what type had occurred. I'm sure it'll be fun playing around with this to see what I can come up with.

Can't wait until tomorrow!

Update 12/22/13: Tomorrow has come and gone. The boards I got back were of excellent quality. The problem is I made a few mistakes, that when all added up were far too many to live with. First problem was that the hole size for the machine pin socket for the ECIO-28P, was simply too tight. Then I noticed that the expansion connector for the external display was flipped right-to-left from what was needed. And then to top it off, I was tied into the wrong four data lines for the display (D0-D3 instead of D4-D7 as it was suppose to be). This is what happens when I skip bread boarding every single aspect, and rush the design out to get made. Anyway I have corrected all the problems and ordered a 2nd set of boards. Also the schematic and PCB layout shown in this post have been updated to reflect these changes (Rev 2.1).

Friday, December 13, 2013

CryoBUG Project Resumed

Well its been a long time since my previous posting, and a lot has transpired over the last few months. A truly Green refrigerant charge has been developed, utilizing only hydrocarbons (HC) and Argon, which promises to give me even better cool-down times and stability.

The chart above reflects a heat load of approximately 10 Watts on a copper High-Mass Cold Head that is utilized on the CryoBUG Demo unit.

Since I was already using some HC refrigerants in my blend to begin with, it always kinda of bothered me that I was still reliant on R-14 an HFC. Besides being a high GWP gas, R-14 also carries a very hefty price tag. So after experimenting with some other HC gases, I was able to zero in on a combination of 4 gases which not only worked in place of any HFC's, but also offered superior capacity and stability. 

During my experiments I did have a series of mishaps. First I fried my DAQ temperature test & monitoring hardware, and then I started to notice erratic behavior in my unit control electronics. The DAQ problem was easily, and luckily, cheaply fixed by  the manufacturer Measurement Computing. The control electronics unfortunately wasn't so easily fixed. It is a design flaw, and something attributable to using components outside of their intended specs. So back to the drawing board.

On my original control system I was intentionally avoiding the use of a microcontroller, because of concerns with electrical noise issues and just being a bit lazy about programming. Anyway I decided to bite the bullet, and have opted to design a new control board around an embedded microcontroller chip. To make things a bit easier, the initial design will utilize a ECIO-28P from Matrix Multimedia as my controller base.

  1. USB connection
  2. PIC18FX455 microcontroller
  3. 4Mhz ceramic resonator
  4. Power/Programming LED
  5. Reset switch
  6. Power selection jumper
  7. Device pins - 0.6" DIL
Picture shows ECIO-40P. ECIO-28P is similar.

To make the programming easier, Matrix's FlowCode application is being used, which as the name implies sports a graphical flow chart approach to coding. In fact it's so easy to use, that I was able to write my entire code, simulate it, and then write it to the ECIO and test it on a protoboard all in one day. It was also nice to see the simulation mimicking the hardware test results.

A piece of the CryoBug controller's FlowCode showing simulation panel...

I really enjoyed this aspect of the project, and wished I had just done it this way in the beginning. Using a microcontroller also allowed me to easily accommodate some extra features that were very much needed.
  • Remote now overrides local ON/OFF control for the compressor, and works in a steady state mode, unlike local, where the switch action is momentary toggled ON/OFF.
  • Delayed restart following a momentary shutdown, in order to allow compressor pressure equalization, for easier start-up and less amperage draw.
  • A method of differentiating between a Low Pressure and High Pressure fault while still only using one status indicator (flashes when the fault condition = High Pressure).
Next up; redesign of controller PCB.

Wednesday, January 30, 2013

CryoBUG Demo Unit Achieves -156°C

I ran a few different charges and kept hitting a ceiling around -148°C. I began to suspect that not only was the static heat load greater than the previous prototype (actually this was a no brainer), but that it was probably more attributable to radiant heat coming off of my vacuum vessel wall, and not conduction as I had originally suspected. This became more clear when I noticed that the bellows section of my vacuum vessel was also getting very cold (see image below).

Click image to Enlarge
In fact the entire bellows appeared to be evenly cold on both ends, something that seemed a bit odd. So I started thinking that perhaps I should give some super-insulation a try, in order to minimize the influence of radiation.
For those not familiar with super-insulation, it simply refers to a method of shielding or blocking the radiant energy emanating from an external heat source. In most cases this involves using multiple layers of a metallic foil like sheeting separated by an insulating material such as mylar or nylon.

Luckily a place that I subcontract to, had some old vacuum jacketed lines laying around that had been disassembled, which just happened to have an insulation material very similar to what is pictured to the left. Since my cold head is fairly small, I only needed a few feet to test out my theory. As it happened one of the lines had begun to unravel at one end, so I snipped off what I needed and wrapped this material around my CryoBUG's cold head, and then re-installed and evacuated the stainless steel bellows.

I then restarted the CryoBUG's compressor, and allowed it to cool down. Within 50-55 minutes I was seeing temperatures drop down below -150°C, and in just barely over an hour, it leveled off at -156°C. finally I was back in my target temperature range, and this was only with an 80-90 micron vacuum. Probably with a few more layers of super-insulation and a better vacuum, -160°C would be possible.

Mytek Controls Prototype Heat Load Controller
Knowing that I was now ready to run some heat load tests, resurrecting an old load controller project of mine seemed like a good idea.

This was a prototype of a 6000+ watt radiant heat load controller that I had stopped development work on about 2 years ago when I ran into problems with the linearity and sampling speed of the power transducer. I have since gotten a much better (as in also more expensive) watt transducer, but just hadn't gotten around to swapping it out for the bad unit, or finishing the embedded software code.

Now seemed like as good a time as ever, so I spent about two days making the hardware changes, and then embarked on setting up my laptop to communicate with the on-board processor chip. This of course took way longer than it should have, but eventually I got everybody talking to each other.

Next I started tweaking with the code, and was able to get a fairly stable output at the low wattage levels that I would be needing (10-100 watts). The only problem was that the new transducer required a divide-by-four circuit to adjust it's 0-10 Vdc output signal down to the A/D converter's 2.5 Vdc analog input limit. I used some 1% resistors to create a voltage divider, but with the transducer's 1.5 mv per watt output, the stability and drift became an issue. So I am presently waiting for some instrumentation op amps to arrive with which I can fashion a precise and stable voltage divider from.

Other things on my agenda, will be to come up with a vacuum electrical feed-through for connection to a load resistor attached to the cold head. Once this is done, I can run a heat load capacity curve on the CryoBUG.