A first look at Pocket P.C., Booting PostmarketOS, Arch Linux and more

Hello again and welcome to another update!

After a shipping delay earlier in November caused by a DHL slow down due to the high volume of packages in their warehouses, reviewers and developers are now finally receiving their units.

We were planning to get this update out on December 1st but were waiting to hear back from the wireless certification testing house. We are still waiting for their results so instead of delaying this update any further, we will send another update later this month when the results arrive.

Initial Reviews

Chris Fisher over at Linux Unplugged received a prototype unit recently.

He talked a bit about it during the beginning of the latest Linux Unplugged episode.

Chris is going to dive deeper into reviewing Pocket P.C. next week.

A few other people have received review units and are waiting for more software images to be established so they can put the units through their paces.


Martijn Braam, a core developer of PostMarketOS, in less than 24 hours after receiving his unit was able to get PostMarketOS booting on Pocket P.C.

Martijn wrote a first look article on Pocket P.C. for Tuxphones.

IdleHax on Discord was able to get Arch Linux booting on the unit.


We have set up a Wiki so that we can document development and production progress, hardware revisions, various software quirks and a list of Linux distributions that work on our hardware. We are working on adding content daily.

Mass Production of Pocket P.C.

With the mold modified last month, we will start mass production of the plastics this month. We will be on site to take video of the first few plastics to be produced and inspect them. Expect to see that in an upcoming update later this month.

The last step before we start Mass Production is Production Validation Test (PVT). In order to kick off PVT, we need to get feedback on EMI testing and FCC certification from the wireless certification house, which as mentioned earlier, has still not arrived. It has taken longer than expected because we had to ship an SD card to Taiwan where the testing house is located. Based on feedback from them, they wanted to run a certain test that was not available on the SD card we originally sent them with the test unit. We expect the results shortly according to our communications with them.

We are making all efforts to ensure that all aspects of the project beside the PCBA, such as packaging, test fixtures and plastics, are in place so that when the PCBAs are ready they can be assembled and we can ship out the units to you as quickly as possible.

We wish you all a happy and safe holiday season and we will update you again shortly as we approach completion of our long journey.


Pocket P.C. Certifications and Original Popcorn Update

Welcome back to another Popcorn Computer update!

We know a lot of you have been patiently waiting for an update. We have been extremely busy working on the final development and production of the units and in navigating our way through the constantly-changing Shenzhen electronics environment.

First, we would like to clarify any confusion that may have arisen from our previous email. The units we started shipping at the beginning of October are going to software developers who are working on porting new operating systems to Pocket P.C. and working on the firmware.

PVT Fixes

Here’s what we learned from our last production run.

The footprint for the chokes used on the high-speed DSI signals lines had a fault in the pinout. DSI stands for Display Serial Interface which is where all the LCD data is transferred over from the SoC to the Display Controller IC.

This meant, we could not receive an FCC ID and certification using the PCBAs from this production run. The PCB used for certification has to be “electrically equivalent” and since we are making some changes to the routing, this could affect emissions in unexpected ways.

This has been corrected in the design.

FCC EMI Testing

In the meantime, we have been contacting multiple testing houses to inquire about quick turnaround times. Thanks again to our antenna vendor, Unictron, who suggested a few testing houses which could accommodate our needs. Once again, it comes down to who you know and being on the ground in Shenzhen and making face to face contact (with masks on!).

We are in the process of doing a preliminary EMI test to make sure that there are no major issues with the PCB that would force us to have to correct and to do another production run. This update is being sent later than anticipated because we had to create the necessary software and written materials to allow for this testing.

We will have the results within the next two weeks.

Mold Change

In the last production run, we moved the GPS module from the top of the PCB to the bottom. This was done to provide greater clearance for the LoRa antenna. Because of this, we need to modify the mold. Luckily, the mold change is straightforward and should be completed within 2-3 weeks. Once completed and verified, we can start mass production of the plastic enclosures.

Original Popcorn

Since we conceived of an improved derivative of C.H.I.P, which we called Original Popcorn in 2018, we have gone through four revisions. The last two were planned to be the production units since we believed that we had corrected all the issues found in the earlier versions.

USB Type-C Power delivery is really complicated. This is doubly so when you don’t select the right Power Delivery IC. We initially chose a Cypress component which turned out to be a PSoC4 internally instead of a custom ASIC. This meant it required a number of external support components. Further, Cypress would not provide any SDK or firmware source for us to modify and instead insisted on providing a firmware blob. We then chose an IC from a Chinese vendor which used a tool written in Visual Basic 6 to program the flash in the IC. Again, we were not allowed to view the source code and any development would take months to implement. Finally, we looked at and chose a Texas Instruments TPS65988 Power Delivery IC which has good support software that allows us to customize the settings of the device. A customized version of this IC is used in the latest Apple Macbooks.

All of this work on the Original Popcorn was not in vain. The power circuit we used on Original Popcorn is the same circuit we used on Pocket P.C. Along the way we learned which are the best manufacturers to use, where to purchase quality parts at low prices and we developed a process for verifying a design before ordering the PCBs.

For those who have preordered an Original Popcorn, some of whom have waited a very long time, we are offering a free pre-production unit. We have already begun shipping them out this month to those who said they would like one and will continue shipping them out within the next few weeks.

There are three errata with the pre-production units.

  1. We originally intended to use the GPIO on the TPS65988 Power Delivery IC for the 8 “XIO” GPIO pins on the pin header. We wanted to get rid of the IO Port Extender used on C.H.I.P. as it could not source power as well as it could sink. Unfortunately, when it came time to write the Linux driver for the GPIO we learned of significant limitations to the ability of the GPIOs in how it handles inputs. To resolve this issue, we added a more advanced port extender which is controlled over I2C which also doubles as an 8-bit ADC meaning all 8 “XIO” pins can read analog voltages. This is a feature that C.H.I.P. lacked.
  2. The other issue with the pre-production units is that we missed routing a trace for the enable pin on the power switch for the USB ports. (A hyphen instead of an underscore mixup caused the net to be left unconnected.) This power switch allows Original Popcorn to supply power to the USB ports when powered off the battery or through the CHG-IN pin on the pin header. We corrected this.
  3. The third issue is that we connected an active-low reset signal for the TPS65988 PD IC to a pin on the R8 CPU that is pulled-up on boot. This meant that if Original Popcorn is powered over the USB then it will infinitely reset. We solved this problem by putting a 0R pull-down resistor on this signal in the pre-production units. We have since moved this reset signal to another pin on the R8 CPU without the pull-up.

We are gearing up for production of the final version of Original Popcorn this month. We are planning to share video on our Youtube channel of the manufacturing and testing process. We will send another update when that video is ready.

Final Thoughts

With working units shipped to developers and all remaining issues identified, all that remains to be done is the corrections to those issues and obtaining certifications which are already in motion.

As more developers receive their units this month, we will share their developments with you. We plan on showing you different graphical environments running on Pocket P.C. in a separate video on our Youtube channel soon.

We are keeping a close eye on shipping logistics and the supply chain. We have had a shipment of developer units delayed because DHL postponed all package pickups due to the high-volume of packages in their warehouses. We are in discussions with multiple third-party logistics companies and will be making the decision soon on whether to ship packages individually from Shenzhen or Hong Kong or send bulk shipments to warehouses in Europe and the United States and ship from there.

Finally, while we try to send an update out each month the timing can vary depending on various information we are waiting to receive from manufacturers or suppliers. To ensure regular updates, we will adhere to sending an update on the first of every month. As we receive more information throughout the month, we will send a follow-up update. We hope this will allow you to have the latest information on a regular basis.

We can’t wait to get Pocket P.C. in your hands. Please be assured that we are working everyday to make that happen as soon as possible. If you have any questions or concerns, please feel free to reply to this email and we will try our best to get back to you as quickly as possible.


Developer Units are Shipping!

Many of you have been asking for an update and we have been working very hard on completing the final development of the Pocket P.C., so here it is:

After nearly a month of intensive work on the display, we reached out to Icenowy who found an overflow bug in the kernel driver for the Allwinner DSI interface. [see the fix here]

The display looks great:

We’re in the final stretch: The Pocket P.C. is almost fully tested!

We just need to verify HDMI with an adapter and Display Port.  In the schematic, we used the pinout of an HDMI A type connector instead of a HDMI C type connector.  This has been corrected.

We will also be replacing the U-blox CAM-M8Q GPS/GNSS module with a SIMCom SIM33ELA GPS/GNSS module. The U-blox component has been banned from import into China. As a result of the recent ban, this component has gone from $15/unit to $80+/unit. The SIMCom module is a drop-in pin-compatible replacement. The only difference is that it does not support the Chinese Beidou Navigation Satellite System which the U-blox module did.

Finalizing the software for the Pocket P.C. is ongoing and progressing rapidly.

We know everyone can’t wait to receive their Pocket P.C., which is why we have been working nonstop to get it perfected and ready to put into your hands.

People in our Discord chat state that they will use it just to SSH into some headless Raspberry Pis or tinker around with it or just use it as a portable terminal.  Whatever your plans for it, we think you will agree it was well worth the wait and your ongoing support of our efforts.

Lastly, great news!  We are now shipping developer units.

Remember, there is still time to upgrade to the LoRa version for $99.  Please email us to take advantage of this great offer.

Thanks again for your support and encouragement.  It means a great deal to us to have you be a part of our community.


🔁 Repeatability of Quality is Essential

Important Notice for Early Supporters:

We are still in the process of sending out invoices. If you have not received one yet and placed an order in November or December of 2019, do not worry, the process is half automated and half manual. This was intended so that we could review each shipping invoice before sending them out to ensure they are correct.

Here’s the update in brief:

  • PCB fabrication company not meeting our standards
  • Are you or do you know a Windows Driver Developer?
  • What’s next?

Welcome to the latest update in which we will take you along on our journey here in Shenzhen as we continue to move forward in completing and manufacturing the long-awaited Pocket P.C. Thanks as always for your continued support and encouragement!

We know you have been anxious to get the latest information on the status of the development and completion of Pocket P.C. This has been a challenging time in the electronics industry with significant supply chain and shipping problems. By being here in China, we have been able to face these issues directly and find ways to continue toward production despite the delays, component price increases and scarcity and other hurdles.

Pocket P.C. Production Update:

We were expecting to produce the final production samples two weeks ago. When our mass production (M.P.) factory went to perform a final inventory check, they found that 18 resistor and capacitor reels were missing. (Some of those reels cost over $200.00 USD each.) The factory we used in February did not return these reels after we had the last batch produced.

Sourcing some of these parts took some extra time because they were for tuning the WiFi antenna and of an extremely high tolerance. Our usual vendors did not have these components but, fortunately, we found the exact ones we need.

Our M.P. factory received the replacement parts we sent and prepared for production by doing one final quality check. In doing so, they found discrepancies in the PCB solder mask between the previous PCBs and these new ones. Through experience we learned that just because a factory did good work in the past does not mean they will continue to produce at that level. We, however, are always vigilant when it comes to quality control.

In the following pictures, you can see some of the discrepancies between the latest PCB on the right and the PCB previously produced by the same factory in February on the left.

At first, the factory which produced the PCBs stated that they would rework the PCBs by applying a new solder mask over the existing one. This was unacceptable to us and we told them they had to produce new PCBs for us at no additional charge. That’s where the trouble began.

They sent us “revised” manufacturing files to review before producing the new PCBs but the files were almost the identical to the previous version. They would have produced new boards with the same problems again! Frustrated, we demanded a refund and, of course, will never use them again.

We are now working with our main manufacturer which promises to meet our quality demands and work very quickly to produce our final samples. The new lead time on new PCB fabrication is 3 weeks from now with assembly to start promptly after completion.

Are you or do you know a Windows Driver Developer?

You can’t have great hardware without great software and we sure have some amazing software to share with you that we have been developing for the past few years. We are ready to open up access to beta testers but there is one thing holding us back. We need a signed Windows WinUSB driver. If you have been through the process of creating a signed driver package and submitting it to Microsoft for inclusion into their driver database, we would love to hear from you. Please reply to this email if you can provide any guidance or advice on creating driver packages or if you are or know of someone who is a software development consultant who can create a signed Windows driver. Thanks!

What’s next?

The events of the past few weeks added a month to our production timeline. Had the quality of the PCB fabricator not been poor and had they returned all our components then we would have production samples in our hands right now.

We really appreciate your patience while we go through the final stages of product development. We can’t wait to ship Pocket P.C. to you and get it in your hands. We really look forward to seeing and hearing about what you do with yours!


Important Notices and Custom Antenna Update

Important Notice for our Early Supporters: 

For anyone who ordered from our crowdfunding campaign early-on in November and December of 2019, we will soon be sending you an invoice for the shipping fee. 

When we started the crowdfunding campaign we used Stripe as our payment processor. Stripe did not allow us to charge shipping according to your location and we would have had to set a flat rate that would be higher than what it should have been for most countries. At the time of check out we had a notice that shipping would be charged at a later time. Now that we are getting close to shipping that time has come. When you receive the invoice make sure to take special note of your shipping address to make sure it is up to date. If not, please let us know by emailing us at

Important Notice for those who purchased the LoRa version of Pocket P.C.: 

Pocket P.C. with LoRa will ship a few weeks to a month after we start shipping the Wi-Fi / BT only version of Pocket P.C. 

Despite pushing our antenna vendor to shorten the lead-time for our custom antennas, they responded that they have limited factory capacity at this time due to a high demand for their products. We will keep those who purchased the LoRa version updated on the new timeline as soon as we are able to reschedule production with our factories based on this new information. See below for more information about the custom antennas.

We will be assembling the PCBs for all the units at the same time but will store the PCBAs for the LoRa version in our factories climate controlled warehouse until we receive the new antennas.

Community Update:

We have set up a Discord community for more real-time discussions and eventually AMAs with the team.

Everyone is welcome to join, ask questions and chat with other community members.

We only ask that you please avoid sending our team members DMs to ask about production or shipping status as our team has limited bandwidth to respond to these in the upcoming few weeks while we gear up for mass production.

Also, if anyone is having trouble with logging into or posting to this forum, you may message @sourceparts on Discord or email and we will get that sorted out for you.

Here’s the link to join our Discord:

New antenna provides increased efficiency for 868Mhz and 915MHz frequency bands:

This update will review how we obtained the best range for Pocket P.C.’s radios and how efficient the antennas are. To provide a quick definition: Antenna Efficiency is the ratio of power radiated by the antenna to the power supplied to the antenna. An ideal antenna has 100% antenna efficiency, i.e., it transmits all the power fed to it.  But in the real world, a good antenna radiates 40 to 60% of power supplied to it.

In early May, we spoke to our antenna partner, Unictron, which makes all the antennas for Pocket P.C.’s radios.  We requested that they improve on the antenna they had initially developed for us.  That is the antenna we showed in our last update. We were getting an efficiency around 20% in the 868MHz and 915MHz bands according to test reports.  While this was an improvement over our original chip antenna we were using in earlier versions of Pocket P.C., 20% did not meet our expectations for the 868MHz – 920MHz bands.

Unictron returned back to the drawing board and requested two weeks at the beginning of May to do further research and development. We had initially planned to order the latest DVT PCBs around that time and were told that the position of the pogo-pins connecting the antenna to the PCB would not be moved but to be safe we waited until they got back to us until we proceeded. We were glad we waited because after their response about the latest antenna they developed, they said that this antenna would only require one pogo-pin and the position had changed slightly.  If we had gone ahead with the PCB fabrication and assembly before waiting for the test results we would have had to order another batch of PCBs and have them assembled. This would have delayed mass production even further.

And the results are in:

Antenna for 868MHz

41.9% at 868MHz

Antenna for 915MHz

42.2% at 915MHz

Unictron engineers are now able to achieve over 40% for both 898MHz and 915MHz frequencies with the newly re-engineered antenna.

Wi-Fi Antenna Tests

While Unictron was developing the new LoRa antennas, we asked them to test the efficiency of our Wi-Fi antenna. The results show an efficiency of over 40% for all Wi-Fi channels. We were pleased with the result. 

Additional Information:

In the following image you can see the antenna location in Pocket P.C. as well as an image of the prototype antenna that was used to obtain the information presented above and below. The antenna that will be in all production LoRA Pocket P.C.s will look like this:

We can learn a lot about how an antenna will perform from its patterns. The following are 3D gain patterns of the separate 898MHz and 915MHz antennas. If you want to learn more about these patterns and how they are used, there is a good document that goes into more detail by Cisco Systems and can be found here:

To obtain these 3D Gain patterns, the following test setup was used in the Lab:

To obtain the efficiency of the antenna, the following test equipment was used in the Lab:

Final Thoughts

Overall, we hope that the real-world benefits for those who purchased the LoRa version of Pocket P.C. of waiting for an improved antenna design will be realizable in a longer effective range of the LoRa radio. For those who purchased the Wi-Fi / Bluetooth only version, the extra time spent this past May allowed Unictron to spend more time tuning and testing the antenna. Without this extra time, we would not have been able to achieve the efficiency that we received as a result of this test. 

This week our manufacturer will be producing the last DVT that rolls in everything we learned as a result of the previous two DVT revisions. Assuming everything goes as planned, we should be able to show a completed unit in the next week or so. Our software developers have already developed a significant part of the LCD drivers and as soon as they receive the sample units, they should be able to demonstrate Pocket P.C. booting Debian into a desktop environment on the LCD.

    I apologize for the longer than expected time between updates this past month.  You can expect more frequent updates this month and if you haven’t already, feel free to join us on Discord or in our community at  I greatly appreciate everyone who is patiently waiting for their Pocket P.C. and have shared their words of encouragement and support throughout the entire process. Thank you.


Making a Custom Pocket P.C. LoRa Antenna

Welcome to another update regarding Pocket P.C.’s development and production plan. We want to thank everyone for your patience and understanding. As you will read below, we have recently overcome a major hurdle regarding selecting and designing the right LoRa antenna.

There has been much testing of each and every component of the Pocket P.C. PCB and there have been many communications and in-person meetings with vendors to get support and clarification on product features.

This past month, we have also made some minor corrections to the PCB design and schematic. Once we have placed the order for the new sample PCBs, we will open source the latest design on GitHub.

The LoRa module communicates

We were able to successfully test the LoRa capabilities of Pocket P.C. w/ LoRa. Our tests showed that we were able to send data through multiple concrete walls in our office and from a 16-story window to the ground floor.

We have selected the RAK4200 LoRa module from RAKWireless. This module has the same specifications as the Murata module we originally selected however RAKWireless provides us with much better technical support and customer service that is invaluable during development and afterwards when you, the user, has questions about the module that we can’t answer.

The RAK4200 has a robust AT command set that allows you to easily join a LoRaWAN network or communicate peer-to-peer. You can read more about this module on their documentation page here:

By using the built-in Black Magic Probe, we were able to download the latest firmware available on their website and update the RAK4200 module immediately without any special tools besides the GDB software which will come pre-installed.

LoRa Antenna Efficiency Issues

Despite our initial successes with the LoRa module we found that during initial RF lab testing of the performance of the antenna the efficiency of the FR4 chip antenna that we selected was very low. This was determined to be due to the close proximity of the antenna to the LCD, which has a metal backing, and the structure of the enclosure itself. To resolve this issue a different type of antenna had to be selected and tested. We consulted with a few expert antenna engineers and performed software simulations with special software and found similar results to what we found in RF lab testing.

Fortunately, our antenna vendor proceeded to create a custom antenna for our device from scratch.

Here you can see that the antenna vendor’s engineers used copper foil to prototype various antenna designs. The antenna prototype pictured above provided the best efficiency.

After prototyping the antenna with copper foil, we had some samples made. On the left of the above photo is the sticker side of the sample antenna where you can see the antenna pattern and on the right is the connection side where there is an exposed copper area that connects to pogo-pins soldered on the PCB.

In order to test the new antenna our antenna vendor made some modifications to our existing PCB. Pictured above are the two pogo-pins that connect the main PCB to the antenna. One is the RF feed signal and the other one is ground. To design and develop a new antenna we removed the original antenna and scraped away the antenna pads and ground copper. We then drilled holes in the PCB so that the pogo-pins can be mounted. The RF feed pogo pin was then connected to the LoRa module via a coaxial cable with an ipex connector. In the final PCB, we will use the RF pad on the module and will not use a coaxial cable which should improve performance.

External LoRa connectors?

We’ve received some requests for the ability to connect an external LoRa antenna to Pocket P.C. w/ LoRa. We are evaluating how we can offer that capability. This would require a PCB mounted antenna and a hole in the case. We won’t be able to test the feasibility of an external antenna connector until after we enter mass production of the Wi-Fi/Bluetooth only version and the LoRa version with an internal antenna for those that opt for it.


  • We recently were able to get Ubuntu 18.04 booting on Pocket P.C.
  • The display backlight driver has recently been tested to work properly. (See below)
  • The touch driver has been tested but will require some more work to properly calibrate and improve responsiveness.
  • The display video driver development is put on hold until the next version of the PCB is assembled. It is estimated that it will require an additional 5-10 days to complete and work will resume once the new units are in our software developers hands.

PCB Updates

We added another communication bus (I2C) to the system controller from the main SoC so that it can have more insight into the operation of the entire system while performing its role as the keyboard controller over USB.

Shipping Update

As both versions of Pocket P.C., one with LoRa and one without both use the same PCB, the above issue with the antenna led to delays for both versions. As we have a clear path forward now, we will be ordering the next round of PCBs for another quick verification production run to prove the changes we made in the latest revision. This should take two additional weeks. With the most recent changes verified then we can start P.V.T. with mass production to follow soon after. Our current target for when we will be shipping is in June. We will be shipping the version without LoRa first to allow for any extra time that may be required for certifying the radios in the LoRa version.

Final Thoughts

With major development work being done, we will be focusing our efforts on the software offerings that will be available at launch. We will have a limited number of Pocket P.C. available in the next production batch that can be used for bring-up of new distributions. If anyone is interested in porting a Linux distribution that they are familiar with to Pocket P.C. please contact us. Also, please contact us if you are interested in using Pocket P.C. in a business or industrial setting, we can offer pre-production units for a limited-time product evaluation.

If anyone has contacted us and has not received a response, we apologize for the delay. We have been heavily involved in solving the remaining design issues so we can get Pocket P.C. into production. Please feel free to follow-up on your previous email to receive a quicker response.

Thank you everyone again for your patience and for being early supporters of Pocket P.C. We are excited to get your Pocket P.C. in your hands and see what you do with it.


Pocket P.C. Urgent Update Follow-up

You may have missed our previous email regarding the significant increase in the prices of components, plastics and packaging. If you did, you may miss the deadline for pre-ordering a Pocket P.C. before the prices increase. Not to worry! Because we did not have the opportunity to send a reminder update earlier, we are extending the deadline to order at the lower price until Sunday night 11:59 PM EST (UTC-5), March 21st, 2021.

If you want to read the original update, check out our previous blog post here:

Thank you for your continued support.


Pocket P.C. Urgent Update

Living in Shenzhen, China, has given me direct access to one of the largest electronic manufacturing hubs in the world.  Because of this, I have been able to find, negotiate and obtain the best components for our devices at the lowest prices.  Now, however, a major change is occurring which directly and greatly impacts the financial cost and time for development of all products, not just ours.

Since the 4th quarter of last year, we have seen indications that lead time and prices will be increasing dramatically.  At that time, and as we noted in a previous update, we started to source the major components which we believed would be in short supply and high demand.  We sourced enough components to fulfill all the existing preorders and a limited number of future orders.

For example, one component we use two of in Pocket P.C., an STM32F103 by STMicro, has increased from $2.30 to $8.30, a 400% increase in price. 

This striking increase in the cost of components, circuit boards and plastics leaves us no choice but to raise our prices on Pocket P.C. for future orders commencing on March 19, 2021.  As of that date, the new price for the purchase of a Pocket P.C. will be $299 and Pocket P.C. w/ LoRa will be $399. In addition, because of the new longer lead times, the newly-ordered units will not be available until the 4th quarter of this year.

For those of you who have already ordered, rest assured that you will be getting the product you ordered at the lower price.  

If anyone wishes to order any additional units at the lower price, you must do so before March 19th to take advantage of our special offer for our existing customers and supporters.

Jose Torres
CEO, Source Parts Inc.


Pocket P.C. is Alive!

Hello again and thanks for being a part of the Pocket P.C. community. We are excited to give you another update as we head into the final phase — the manufacture of the product you have been waiting for.

We had a production run in early February and a successful board bring-up for DVT2. We only had five PCBs assembled for this production run because had not worked with this factory before. We normally produce 25 units for a DVT.

Below are some photos from the assembly process:

Automatic Optical Inspection (AOI) of the PCBA after assembly.

Here you can see the LoRa antenna lettering being inspected. This is to ensure that the correct component has been placed. This is one of hundreds of points of analysis that the AOI machine does for each PCBA.

X-ray Inspection of the PCBA after AOI.

Here you can see again the LoRa antenna. The antenna is actually made of a special PCB and here you can see the traces on the various layers.

The factory owner personally helped us to test each device after all PCBAs went through AOI and X-ray analysis. After all boards were tested and powered on successfully, we carefully wrapped each board and headed home. When we finally left at around 10PM all the workers left for the holiday. We made it just in time!

The next day we shipped two units to Bootlin. They quickly got to work once they received the package. And the results are in!

First boot demonstrated in short-run of DVT2 PCBAs.

Taking Pocket P.C. out of the lab and into the wild.


We have started to post progress videos on YouTube which are being posted regularly. Make sure to subscribe to our channel to view our latest Pocket P.C. videos as they are released.

A video of flashing the keyboard / system controller with the built-in Black Magic Probe and a demonstration of a working keyboard will be uploaded next week.

Software Development

We will be working in the public and pushing changes to the following repositories for our U-boot, Linux and Buildroot changes respectively. Once the changes are thoroughly tested, we will be submitting patches upstream so that you don’t have to use our repositories.

Circuit Design

There is a lot of protection circuitry in Pocket P.C. — much more than most standard consumer electronics including most cellphones. We want Pocket P.C. to be a high quality and durable tool in your toolbox (or your pocket!)

As a result of studying this latest production run, we added five protection diodes in various parts of the PCB to prevent damage that can occur when using faulty chargers or cables. We also added a voltage regulator to the USB Serial port that can accept up to 30V. We always keep in mind that some defective USB-C chargers can output 20V by default without power negotiation which could cause some parts including the USB Serial converter to fail short, permanently ruining it. By adding the voltage regulator, the USB Serial port can handle these types of situations. We are doing everything possible to prevent cases of unintentional damage to Pocket P.C. to ensure that you will be able to continue using it for a long time.

During software development for the LoRa module and keyboard / system controller, we realized it would make things easier if we could use a UART to debug the firmware. The great thing about the built-in dedicated Black Magic Probe (BMP) device is that it offers us an additional UART. We have added an analog switch to share the UART on the BMP between the keyboard / system controller and the debug UART on the LoRa module. The LoRa module will now have both its UART1 and UART2 connected to a functional UART. It must be noted that the UART on the keyboard / system controller is shared with a ROW and COLUMN of the keyboard. If you are using the UART it will interfere with the operation of the keyboard. Just something to keep in mind.

We contracted a few engineers to do a full review of the design and outside of a few minor suggestions we gained their approval of the overall design. This gives us additional confidence that Pocket P.C. will be a reliable tool.

Next Steps

We will be doing another quick (10-15 day) production run to verify some of the changes we introduced into the latest version as a result of the lessons learned in the last production run.

After that, we will move straight into Production Validation Test (PVT) with Mass Production (M.P.) right after. The PVT should be less than 2 weeks to verify that our process to produce Pocket P.C. is sufficient to do so with high precision and without introducing defects.

Final Thoughts

Overall, at this point we are pleased with the design. This gives us confidence that Pocket P.C. will be a reliable tool you can depend on. We have about a month of development and additional testing before we can start PVT. This puts us into Q2 of 2021 as an anticipated ship date.

We will have a limited number of units available for evaluation for business use. If you have an urgent need to evaluate Pocket P.C. for your business, please reply to this email and we will try to accommodate you.

We appreciate your continued support and encouragement.

We promise to give you another detailed update as our exciting project continues toward completion.


Beating the Clock before Chinese New Year

We have been working nonstop this month and the last to get DVT2 into production and assembled before Chinese New Year. All factories take a holiday from February 5th to February 18th. Chinese New Year is on February 11th this year.

The process of verifying everything again was really worth it. We discovered an issue with the LCD pinout. Luckily, we had not yet ordered the PCBs and could change the pinout on the connector in less than half a day.

There were over a hundred changes and refinements made to the schematic, layout and BOM. Many were minor cosmetic changes to the schematic and some were major including moving and re-routing some components to accommodate the new antenna. We also swapped some data lines in the LPDDR3 Memory in DVT2 to make it function properly.

The latest schematic and layout has been open sourced and released on GitHub. You can click here to check out the latest files. We will release the design files after we start shipping Pocket P.C.

Getting a PCB Fabrication House to produce 10-Layer impedance controlled PCBs before Chinese New Year was a difficult challenge. We tried numerous PCB services and companies we know or have worked with in the past and none could start production until after the holiday. Many companies are trying to get existing orders out of the door and are not taking new orders.

Fortunately, one of our team members checked with one of his long term partner factory and they agreed to accommodate us.

As part of the normal process of auditing a new factory, we took a number of pictures. Here are a few.

PCB Copper Clad Laminate Storage Warehouse
Brown Oxide Machine
Manual Inspection Stations

There were a number of Engineering Questions (EQ) that were raised that required us to carefully check the potential issues before providing a response on how the fabrication house should proceed. This is part of the process of getting PCBs manufactured. Since our board uses some complex technologies such as a few Wafer Level Chip Scale Packages (WLCSP) that have very fine ball spacing, we had to get the manufacturing tolerance correct to avoid issues during assembly. One issue we had is that their processes can’t accommodate white soldermask for the WLCSP component footprints. Their solution is to use a green soldermask around these components while the rest of the PCB will be white. This will not be the case for Mass Production as the fabrication house we use for M.P. has more advanced machinery capability of higher tolerance.The color white was chosen for the PCB so that it can reflect the keyboard backlight LEDs thereby making them appear brighter. If we had chosen a different soldermask color, the true color of the LEDs would not come through.

Our perseverance is paying off. We are getting a quick-turn around on the PCBs and will be doing assembly of five pieces this week.We will be sending another update shortly after we have had some time to test the PCBs and let you know the results.