The DropBot system is described in detail in "Fobel et al., Appl. Phys. Lett. 102, 193513 (2013); doi: 10.1063/1.4807118". If you use this information in work that you publish, please cite as appropriate.

Measles and Rubella field trial in Kenya


Last week, four members of the Wheeler Lab departed for Kenya to perform a field-validation study for measles and rubella using a new version of the DropBot. The new system (which we refer to as the Measles/Rubella-BOX or MR-BOX) is capable of performing four simultaneous sandwich ELISAs in about 40 minutes using a chemiluminescent readout. It detects light by measuring the current from a photomultiplier tube (PMT) using another piece of open-hardware developed in the Wheeler lab called the DStat. It also features an integrated amplifier, a magnet for performing bead-based separation, a webcam, LED lighting, and a humidity and temperature sensor. The assay is based on earlier work with Rubella and we’ve since added the capability to test for measles at the same time. We are using 100% low-cost, inkjet-printed DMF devices.

In preparation for this project, we’ve made significant updates to the Microdrop software, including path routing, new quality control tests, and lots of usability improvements. We hope to preview some of these features here on the blog in the coming weeks and we will release updated installers later this summer.

We’ll be testing 150 clinical samples at the Kakuma refugee camp over the next 3 weeks. Check out the video below for an introduction to the project, and if interested, follow our progress on twitter.

Posted in Uncategorized | Leave a comment

Sciex Analyst Control app and Microdrop plugin

We just released a Windows app called “AnalystControl” that can connect to the Sciex Analyst application (used to control auto samplers, mass spec, etc.) to start/stop queued acquisition batches.

The binary release of the “AnalystControl” application is available at:

The source code of the “AnalystControl” application is available at:

See below for instructions on how a) directly control start/stop of acquisitions in Analyst queue, and b) coordinate acquisition in a Microdrop protocol using “AnalystControl” and the analyst_remote_plugin Microdrop plugin.

Direct Queue control  through AnalystControl

The AnalystControl app can be used to connect to the Sciex Analyst application to start/stop queued acquisition batches.

  1. Download and unzip AnalystControl release.
  2. Run “AnalystControl.exe” directly from unzipped directory (no installation necessary).
  3. Click “Connect” to connect to the “Analyst” application.
  4. Click “Connect” in the “Queue” area to connect to the queue manager.
  5. Click one of the following buttons in the “Queue” area:
  • “Ready”: Set queue state to “Ready”
  • “Start”: Start queued acquisition
  • “Stop acquisition”: Stop a running acquisition
  • “Stop”: Stop queue (must call ”Stop acquisition” first if acquisition is in progress)



Microdrop Plugin (analyst_remote_plugin)

To use the analyst_remote_plugin plugin for Microdrop

  1. Install analyst_remote_plugin plugin by selecting File > Manage plugins..., then selecting Download plugin...
  2. Create a Microdrop protocol and check the “Acquire” checkbox in at least one step.
  3. Queue a batch in Analyst.
  4. Start the “AnalystControl” program if it is not already running.
  5. Click enable in “Remote control” area of “AnalystControl” app to enable remote connection from Microdrop plugin.
  6. Start the protocol in Microdrop.

The Microdrop protocol will pause execution upon reaching the step where “Acquire” is checked.  The Microdrop protocol will remained paused until the acquisition has completed, after which the protocol will be resumed.


Posted in Announcements | Tagged , , | Leave a comment

DropBot, meet OpenDrop


Earlier this Spring, we were contacted by Urs Gaudenz from GaudiLabs and he told us about a low-cost, DIY-friendly DMF platform he was developing along with collaborators from the Bioflux project, called the OpenDrop.

The OpenDrop is a fully integrated device that is not much bigger than an Arduino Uno. It does not require any external high-voltage amplifier and uses MosFET transistors to control the state of each electrode (instead of the more expensive PhotoMOS chips used in the DropBot). These modifications reduce the system cost considerably, from ~$5,000USD down to a couple of hundred dollars, a level where many more people (e.g., DIY hobbyists) would be willing to build one just to play around with, which we view as a very exciting development!

Urs graciously sent us an OpenDrop, and we just finished writing firmware and a plugin that allow it to be controlled from Microdrop. We’ve added instructions on the Microdrop wiki.

The OpenDrop uses DC electric fields instead of AC (a simplification that drives a lot of cost savings) and does not perform any impedance-based sensing. It also uses a PCB-based electrode array that is covered by a removable film (Saran wrap), instead of removable devices. While this certainly works for moving drops around in an open plate configuration, it is unlikely to work in the more useful two-plate configuration (necessary for dispensing and splitting) unless an oil-filler is used. That being said, we think it should be possible to achieve many of the cost savings while maintaining most of the added functionality of the DropBot system, which is the focus of our current development efforts. In any case, we see the OpenDrop as a very exciting development that will bring more people into the open-source DMF community.

Posted in Uncategorized | Leave a comment

IPython notebook integration

We just published an update for Microdrop (v1.0.18) that adds an exciting new feature: IPython notebook integration. This provides a quick and easy way to dive in and explore data logged during the course of your experiments. This could be useful for custom analysis of one-off experiments, or you can save notebook templates so that you can quickly apply a common set of post-processing steps to future experiments. If you need tighter integration with Microdrop, it’s straightforward to transition these notebooks into Microdrop plugins.

Experiment log window

To access this feature, open the experiment log window from the menu item View/Experiment Logs and choose a log id from the drop down menu. There should be three new buttons : 1. Manage sessions..., 2. Open.. and 3. New... If you press the New... button, a file selection dialog will open in the Microdrop/notebooks folder. From here, you can choose a notebook template. We’ve provided two examples to get you started: “dump feedback results to csv.ipynb” and “Experiment log explorer.ipynb”.

If you choose one one of these templates, that file will be copied to your experiment log folder (i.e., “Microdrop/devices/device name/logs/log id/”) and an IPython notebook session will be launched in your default browser. To see an example of what one of these notebooks looks like, see this link. It demonstrates how you can dump feedback results to a *.csv file and plot drop velocity for a selection of steps in your protocol. After you’ve created a new notebook by copying a template (and optionally editing it), you can later open that notebook with the Open... button.

Notebook session manager

Note that each time you launch a new IPython notebook, it creates a new system process. These processes remain open until you exit from Microdrop, so if you create too many notebooks, it can cause a drain on your system due to high memory usage. The Manage sessions... button provides you with the ability to close down notebook sessions that you are no longer using by clicking on the Stop button next to the notebook you want to kill.

If you have questions/comments about this feature, please send them to the dev list.

Posted in Announcements | Leave a comment

Microdrop v1.0

We’re happy to announce the release of Microdrop v1.0 (Windows installer). A list of changes can be found here. Note that if you have previously installed a pre 1.0 version, you should uninstall it before installing the latest version.

This release is also available as a self-extracting portable version which you can run from a folder on your computer or a USB thumb drive. This option does not require Administrator privileges and allows you to easily switch between multiple versions of the software on the same computer.

We’ve also released an update to the DMF control board plugin (which should update automatically, or you can manually update from the menu item “File/Manage plugins”). To see what’s new, check out the change log here.

If you experience any problems, please send a message to the dev list.

Posted in Uncategorized | Leave a comment

Always make sure that the amplifier is off when you launch Microdrop!

We’ve recently noticed strange things happening when connecting/re-connecting to the control board during an experiment (e.g., drops randomly moving on the device, electrolysis, etc.), and while testing out the new hardware calibration routines (e.g., several fried op-amps in the feedback circuit). We hooked up an oscilloscope to the system and ran some tests to figure out what was going on.

When you initiate a connection to the DropBot (e.g., by launching the Microdrop application), the control board pulls a pin on the ATX power supply low to turn it on. This provides power to the daughter boards (e.g., the signal-generator board and high-voltage switching boards), which then power-up and initialize. During the power-up of the signal-generator board, transient voltages appear on the output pin within the first ~200 ms; if the amplifier is on during this time, it can output voltages in excess of +/- 200 V.

Oscilloscope screen showing high-voltage transients measured on the amplifier output during DropBot initialization. Each vertical line represents 100V.

Oscilloscope screen showing high-voltage transients measured on the amplifier output during DropBot initialization. Each vertical line represents 100V.

Because the high-voltage switching boards are uninitialized at this point, high-voltage transients also appear on the outputs of all channels (though they are attenuated by roughly 50%).

Oscilloscope screen showing high-voltage transients measured on channel 0 during DropBot initialization. Each vertical line represents 100V.

Oscilloscope screen showing high-voltage transients measured on channel 0 during DropBot initialization. Each vertical line represents 100V.

These transients can cause damage to a connected DMF device (e.g., dielectric breakdown, electrolysis, etc.) if it is loaded with liquid. They can also damage the control board‘s impedance measurement (feedback) circuit if the test board is connected, since this board contains many large capacitors which provide a low-impedance path to the feedback circuit. In fact, this scenario caused us to fry several op-amps on our control board, which helped us to identify the problem.

To summarize, if the amplifier is on when you connect to the control board, high voltages may be applied to all channels in the system for a brief period (< 200 ms). To prevent damage, always make sure that the amplifier is off when you launch the Microdrop application. As an added layer of safety, you should avoid loading liquids onto your DMF device until after you have launched Microdrop. This way, even if the amplifier is on, it will not cause any damage to your DMF device or to the feedback circuit.

This problem can be addressed in future designs through hardware modifications to the control board, signal-generator board and or/the high-voltage switching boards.

Posted in Announcements | Leave a comment

Preview of the Microdrop 1.0 release

Update 2015/03/18/: link to the portable v1.0rc0 instead of the earlier developer preview. This release candidate is considered stable for routine use; the only major change expected prior to v1.0 is the addition of a Windows installer.

We are getting set to release version 1.0 of the Micrdrop software as well as a couple of new hardware add-ons: (1) a test board and (2) an anti-aliasing filter shield. In combination, these improvements provide a simple method for users to quickly calibrate and benchmark their DropBot systems and to diagnose hardware problems. In addition, they significantly improve the accuracy of impedance-based measurements. We are in the process of upgrading our systems here to perform internal testing, but we thought we would share a preview of what’s in store.

This first set of screenshots shows a wizard for calibrating the reference load resistors. Once calibrated, this allows the DropBot to monitor the high-voltage output from the amplifier.

This next wizard is for calibrating the circuit responsible for measuring device impedance. This process makes use of the previously mentioned test board, which is essentially just a bank of known reference capacitors connected between one of the switching boards and the feedback cable (i.e., in place of a DMF device).

This final set of screenshots shows a test routine for checking the high-voltage switching boards. This can be useful for quickly identifying faulty solid-state relays.

One of the other changes with the upcoming software release makes it possible to run Microdrop as a portable application (e.g., from a folder or USB thumb drive) without installing the software. We still plan to provide an installer, but this will make it possible to run multiple versions of Microdrop on the same computer (e.g., one “stable” version and one “experimental” version).

If you are feeling brave and want to try out the release preview, download and double-click on the self-extracting zip file , then double-click on the Microdrop.bat file. Be warned that this version may not be compatible with your old protocols/plugins (we are working on upgrade tools, etc.). You will also need to download the new control board plugin (change log); click on the menu item File/Manage plugins, click the Download plugin... button, select dmf_control_board from the list and click OK. Once you restart Microdrop, it should give you the option to flash the upgraded firmware. Note that if you want to go back to the previous firmware version, you will need to re-flash the control board and recalibrate.

In the new year, we will post an official release (with a Windows installer) and provide some additional details on the new hardware and software features. If you have any comments or questions about the release, please direct them to the development email list.

Posted in Announcements | Leave a comment

Panelizing PCBs with gerbmerge

When you’re ordering small batches of PCBs, it is often cheaper to merged the gerber files for several board designs onto a single panel. We do this using a tool called gerbmerge. If you download the electronics zip file from the wiki, it contains a sub-directory called merged gerbers which contains gerber files for a multi-project panel composed of several copies of the various DropBot PCBs. If you want to change the contents of this panel or it’s size, you can edit the file merged gerbers/layout.cfg.

The default panel size is 14.5″ and 10″, but you can change this by editing the lines:

PanelWidth = 14.5
PanelHeight = 10

Each board design is represented by a section in the layout.cfg file. For example, the section for the control board is included below. To change the number of copies of any of the boards in the panel, you can simply modify it’s respective Repeat = x line, where x can be any number (including 0).


# You can set any options you like to make generating filenames easier, like
# Prefix. This is just a helper option, not a reserved name. Note, however,
# that you must write %(prefix)s below, in ALL LOWERCASE.
# Note how we are making use of the ‘projdir’ string defined way up at the top
# in the [DEFAULT] section to save some typing. By setting ‘projdir=somedir’
# the expression ‘%(projdir)s/proj1′ expands to ‘somedir/proj1′.

# List all the layers that participate in this job. Required layers are Drills
# and BoardOutline and have no ‘*’ at the beginning. Optional layers have
# names chosen by you and begin with ‘*’. You should choose consistent layer
# names across all jobs.

# If this job does not have drill tool sizes embedded in the Excellon file, it
# needs to have a separate tool list file that maps tool names (e.g., ‘T01′) to
# tool diameter. This may be the global tool list specified in the [Options]
# section with the ToolList parameter. If this job doesn’t have embedded tool
# sizes, and uses a different tool list than the global one, you can specify it
# here.

# If this job has a different ExcellonDecimals setting than the global setting
# in the [Options] section above, it can be overridden here.
#ExcellonDecimals = 3

# You can set a ‘Repeat’ parameter for this job when using automatic placement
# (i.e., no *.def file) to indicate how many times this job should appear in
# the final panel. When using manual placement, this option is ignored.
Repeat = 2

Once you’ve made your modifications to the layout.cfg file, open a command terminal (press the Windows Key + R, type “cmd”, and press ENTER) and navigate to the merged gerbers directory. Now type “gerbmerge layout.cfg” and press ENTER.

Follow the instructions on the screen (i.e., type “y” and press ENTER). Then the gerbmerge program will read in the gerber files for each of the projects and attempt to place them on the panel. It will keep a running count of the number of placement attempts and the panel area. Once you are happy, you can press CTRL + C to stop the process. If the program cannot fit the designs onto the defined panel size, you will need to go back and modify the layout.cfg file (either reduce the number of designs or increase the panel’s dimensions). For more information on viewing the resulting gerber files and ordering PCBs, refer to the BuildInstructions.

Posted in Uncategorized | Leave a comment

DropBot v2.1 hardware designs released

We just pushed up the v2.1 hardware designs to the git server. You can find instructions for ordering PCBs and a zip file containing all of the KiCAD, and gerber files on the wiki.

This new design is a minor update to the v2.0 series. The biggest changes are:

  • The new high-voltage switching boards are 2-layer (previously they were 4-layer). This means that they are cheaper to fabricate and can now be ordered on a single panel with the rest of the PCBs.
  • The new high-voltage switching boards have their own microcontroller. Previously they used a general purpose input output (GPIO) chip, which was one of the most difficult chips in the system to solder because it was so tiny! In the future, having a microcontroller on these boards will make it possible to add new functionality.
  • All surface mount capacitors/resistors now have a minimum size of 1206 (previously, some of the capacitors were much smaller and more difficult to solder manually).
  • All of the boards now have hardware support for in-system programming (meaning that in the future, it will be possible to flash the firmware for all of the microcontrollers over a single USB cable automatically).

If you have any questions/comments, post them to the dev list. If anyone gets boards assembled using the stencil reference numbers, let us know; we’re curious to see how that works out.

Posted in Announcements | 7 Comments

DropBots in the wild

About a month ago, we received an email from Markus Haapala from the University of Helsinki with some photos of their newly built DropBot system. This represents an exciting milestone for us as the first DropBot built outside of the Wheeler Lab! We’ve started a wiki page where we are hoping that people can link to websites/blogs/photos documenting their builds. We’ve also started a map to keep track of where other DropBots systems are out there.

Posted in Uncategorized | Leave a comment