08/20/2019 at 00:41 •
This is it! While not exactly ground-breaking news... it is newsworthy!
The 1.0 spec is a culmination of intermittent experimentation since the genesis of the project in 2015. It includes support for up to 400 watts via the average DB-25 connector, along with support for various common communication protocols -- both legacy and cutting-edge. Stability has been the goal, and now it is here for the most common connectors in the specification. Contacts have been optimized to reduce signal interference, separate high voltage supplies from low-voltage logic & to generally make design/assembly easier. These connectors support tiny, efficient low-voltage modules & big power modules alike!
Now that the spec is stable, large-scale production runs are on the horizon. Stay tuned!
06/12/2019 at 04:14 •
It's right here!
This release is pretty significant. I am preparing to shift from the oft-repeated arduous process of hand-soldering modules... to creating printed circuit boards. I wanted to have the specs for the major connectors (DE-9, DA-15, DB-25, etc.) firm before moving forward with any production runs.
The key things that have changed:
- DA-15 is now an Enhanced Battery Connector.
- LiPo batteries already have a balance connector. Why use an EBC instead?
- Pretty much all balance connectors in the RC industry merely allow a battery charger to monitor two or more cells of a battery pack with an unknown chemical composition. The EBC spec supports battery packs with 6 cells (e.g. 6S) & includes one id pin. This pin has a resistor connected which can be used to identify both how many cells are in the pack, and what their chemical composition is. For instance, a 4S LiPo could have a simple 1.3Kohm resistor embedded within. Off-the-shelf battery chargers could include new functionality that analyzes the resistor value of the battery pack, compares it to the matrix & pre-loads the best settings for charging the pack.
- LiPo batteries are already too heavy. Why add a heavy old-school DA-15 connector?
- You don't have to! You can use a 16 contact socket IDC header instead. All 16 contacts are only really necessary on a fully-equipped 6S battery pack. A 3S pack would still use the 16 contact socket IDC header, but could terminate as few as four wires on said header.
- Why not smaller?
- There's nothing that would prevent use of a smaller connector. It seems, however, that good-ol' 2.54mm headers are best suited for external battery management, and 16 contacts seems to provide just enough functionality & flexibility.
- Why no digital communication?
- There is! I was able to squeeze in `one-wire-data` on contact 1. This digital bus allows one to embed inexpensive digital temperature sensors & similar into the pack.
- Why support only LiPo?
- They support more! Enhanced Battery Connectors support Pb (lead acid), NiMH, LiIo & LiFe batteries as well! As seen on the matrix.
- LiPo batteries already have a balance connector. Why use an EBC instead?
- DC-37 is now 'The High-Def D-Sub'.
- The connector now is spec'd to support HDBaseT (1080, 4K, etc), USB3, & standard BaseT connections. Example: "6x Raspi's and/or 6x Nvidia Jetsons sharing a backplane."
- Most connectors now support up to 40VDC.
- Why so high?
- 24V battery systems (such as those in some RVs) may reach voltages as high as 36V while charging. I wanted Retro Modules to be compatible.
- A grid of 9x 18650's (a 3x3 9S battery pack) has a max safe voltage of 37.8VDC.
- While the average D-Sub contact can support up to 5A, D-Subs with IDC termination use wires that can only support about 1.2A each. An increase in the maximum supported voltage safely allows for more wattage on those same wires.
- Some module arrangements may have a common backplane, while some module arrangements may feature daisy-chaining. The higher voltage allowance helps reduce risk of connection-related power delivery issues.
- Why not higher?
- 40VDC is the maximum voltage supported by many inexpensive DC-DC buck converters that use the LM2596 (or similar) regulator.
- Voltages over 50V can be considered dangerous (hat tip to: @Njkotzur via https://hackaday.io/project/45999-dewalt-flexvolt-60v-battery-adapter/discussion-100764)
- Why so high?
- Many connectors now support (on a dedicated contact) voltages up to 15VDC.
- Why up to 15VDC?
- Smaller modules may not have any room for a beefy DC-DC buck converter. The lower voltage limit allows for simple linear regulators (etc) if needed.
- Smaller modules may not need much power at all.
- Many 12VDC automotive accessories are able to tolerate up to 15VDC.
- Why up to 15VDC?
- Some connectors now support 12VDC with the ability to request more.
- Why the complexity?
- There are many 12VDC automotive accessories (and similar) that prefer just that: about 12VDC. The connectors (such as `din5-240`, `din6-240` & `din8`) will provide about 12VDC by default if a suitable power source is available. Some devices, however, make do with 12VDC -- but work better with higher voltages if they are available. These connectors (much like functionality described in USB-C PD documentation) are able to provide more power if a source can provide it & the client device requests it.
- Why the complexity?
This spec is becoming easier & easier to work with. I'm now able to manually make one or two modules in an evening with great ease. I'll be posting a few new (related) projects soon!
- DA-15 is now an Enhanced Battery Connector.
05/04/2019 at 16:54 •
Work is continuing in earnest for the 1.0 spec release. This spec release is expected to be firm & as such, an audit of many connectors has been performed. The proposed changes include new flexibility on the power-related contacts & optimizations to reduce interference.
There have been some exciting developments concerning multi-cell battery support, as well as support for some more exotic modules. Looking forward to sharing more with you soon!
06/23/2018 at 18:55 •
This latest pull request introduces a few modules (and 3D-printable files) designed to solve power-related issues. One of the main concerns I've had over the past several months was how readily-available power supplies & batteries might fit into the specification. One of the goals of the specification is to encourage adopters to use parts they already have in their shop. Power tool batteries, for instance.
Presently, most connectors noted in the spec at http://retrospec.cc have pins which can accommodate up to 20VDC. Most power tool batteries have a nominal voltage of around 18VDC, but can exceed 20VDC at full charge. For these types of batteries to be used, the average max power rating for connector pins needed to be higher. How high, though? An RC-industry 6S LiPo battery can exceed 25V at max-rated charge. Most golf cart batteries (and newer fancy RV batteries) have a 24VDC nominal voltage. In the end, I decided to go with 30VDC max voltage on most pins in the new proposal. Many inexpensive DC-DC buck voltage regulators are rated to work with voltages up to 30VDC. These regulators can easily fit into most modules, run efficient & run cool. This higher voltage allowance permits more power-hungry modules on the same relatively thin wires.
The new pull request (http://retrospec.cc/new) includes STL files for the two leading power tool manufacturers & reference information. Files for other leading power tool manufacturers will become available within the coming months. Support for LiPo batteries of varying form factors will follow. For lower-power projects, the good ol' 9V Battery module will certainly suffice.
The larger batteries presented a new opportunity for a standardized modular footprint. After messing around with a few different approaches I settled on the idea that one should be able to just print (or CNC/laser) thin plates... or full enclosures. The hole pattern accommodates up to 2X DA connectors, up to 2X DB connectors, up to 2X DC connectors & up to 4x DE connectors. These connectors, when mounted, should ideally be arranged in a palindromic fashion for greater ease of use.
All power-supplying modules using this footprint should have all female connectors. Each pin on a typical D-Sub connector can accommodate about 5A. As such, resistor-style fuses (or auto-resetting fuses) should be between each power pin and the actual power source. The connectors should ideally have all contacts properly connected so a small module plugged in on one side of the footprint can talk to a small module on the other side. If you're in a rush... just wire up the power contacts. You can always come back later & upgrade.
The footprint will allow for all sorts of small-form-factor computing devices to run atop a power tool battery. The new pull request includes STL files for Raspberry Pi devices & full-size Arduino devices... with support for more devices coming soon. Each of the holes in the STL file designs are compatible with M3 screws or Imperial #6 screws. Screws & standoffs can be purchased at many PC or RC part supply stores.
Since there are 2X 30VDC pins on DB-25 connectors & 1X 30VDC pins on DE-9 connectors, the configuration shown above can accommodate up to 30VDC at up to 30A. Please make sure to check the new PR (again at http://retrospec.cc/new) for new pin specs for high-current D-Sub connectors (with some accommodating up to 60VDC at up to 80A & some soon which will accommodate 60VDC at up to 160A).
This PR is a big bold step for this little free modular framework.
06/13/2018 at 04:49 •
Putting the finishing touches on a new pull request. This one is going to be fun. The dust is settling. Some of the big blockers are being resolved.
Stay tuned. New post next week.
05/18/2017 at 04:57 •
This pull request sums up several months of occasional planning & research work. The projects listed are projects that interest me; projects that I intend to build and/or contribute to in the future. Given present time constraints, I've simply chosen to compile notes about each project in the repo with the hope that others might find the reference material useful -- and that I can reference that material when I'm ready to do some serious building again.
The general idea of the repository is to provide a common electrical specification for maker-friendly projects. Adding known modular projects into the mix helps provide some extra inspiration. Modules & modular project reference materials are expected to be the complete scope of the repository.
More to come!
07/18/2016 at 04:52 •
Just finished up editing module ideas I've been dwelling on since February. Since the specification is now fairly stable, the focus now is: what can be built? Many big custom builds are just that: big & custom. Once demoed... they may be purposefully destroyed or fall into disrepair. With a basic open spec, modular parts of a big build can be salvaged & saved for later. Most of the ideas in this latest proposal are bigger ideas -- broken up. I'm looking forward to building a bunch of these. Since I do not (nor will ever have) the time to build all the things, I have included many references to other components & in some cases, other projects.
08/17/2015 at 06:57 •
After the Maker Fair 2015 revisions were adopted, I went back to work updating a bunch of Retro Modules which were no longer in compliance. This is one of the reasons I'm striving toward a stable pin specification in short order. It is not fun looking at a box full of no-longer-compliant modules. I've had to sift through existing modules to find the few which were not affected by the spec revisions. Thankfully, I am gaining momentum again.
When I built out the hardware for my 'Christmas Truck' project, I used the Retro Modules spec extensively. Those modules will need slight updates now, but even the basic modularity of the install proved useful. I recently traded my truck in for another, and the uninstall process was a breeze. That is -- everything except some of the oldest modifications I made to the truck. The 2012 Tacoma features a LCD screen in the rear-view mirror. It, at times, displays composite video from the stock rear-view camera. I wanted a way to send my own video to the mirror without making a big mess of the mirror housing. I was able to locate the video cable & make a splice quite some time ago. To undo this splice, I had to break out the soldering iron & solder the short segment of wire back together. If I decide to do a similar modification today, I would make it easy to remove any mods. How? I would add connectors (blade, etc) which allow the stock cable to be reconnected to itself with ease. This can increase the cost of the mod, but not by much.
I have spent some more time fleshing out pin header specs & documentation. The very common 26 pin header is hardware compatible with a 25 pin DB-25 connector. The 10 pin header is compatible with the DE-9 connector. This allows the use of old-fashioned header-to-connector ribbon cables. You may have had some of these cables in an older desktop tower.
Another common staple of desktop computing is the "AC'97"/"HD Audio" header & breakout modules. Many modern single-board computers feature analog & digital audio, but are often broken out via consumer audio jacks. These small computers could leverage the AC'97 header with great ease -- reducing circuit-board complexity & increasing design flexibility. A source of confusion for many, however, is S/PDIF headers on motherboards. Legacy graphics cards often shipped with a tiny cable for routing digital audio. S/PDIF headers are not well standardized & may consist of two to eight pins. Given this problem, I opted to append five pins to the end of the "AC'97"/"HD Audio" header. This keeps compatibility with off-the-shelf desktop audio modules yet allows S/PDIF in/out on a custom header if so desired.
Pro audio is also an important area to focus on.
Since part of this spec features analog audio next to other digital signals, there is a greater risk of noise. Many environments do not require pristine audio conditions. Some do. This is why I have included support -- in certain connectors -- for AES Digital Audio. S/PDIF is the consumer version of AES. AES42 has some really great features that quite a few sound engineers enjoy. These signals are generally sent on a standard 3-pin XLR connector. There is risk, though, in sending digital signals on a connector primarily used for analog audio. Audio mixers may get damaged if a AES signal is connected to an analog input. I have proposed a spec for 6-pin IEC XLR connectors to hopefully address this issue -- and add a feature or two. The 3-pin connector only allows for one-way transmission of digital audio signals (generally). A 6-pin connector allows for power & bi-directional AES transmission. This opens up doors for chainable digital effects processors, digital audio pedals, mics & remote monitoring. For instance, an artist performing on stage generally needs a mic, guitar input & monitor mix. This could be accomplished via this single XLR cable. The monitor mix would arrive on the `digital-audio-aes-host-out-client-in` pins, and the audio source(s) on-stage would be sent back to the mixer via the `digital-audio-aes-client-out-host-in` pins. These on-stage modules could be powered via the `fourty-eight-volts-max` pin.
This latest work can be seen on the incoming branch of retrospec.cc.
08/17/2015 at 05:35 •
Maker Faire Bay Area 2015 and the events leading up to it really helped refine the specification in a few very important ways. First, I found many of the features packed into the DB-25 spec should be moved to DA-15. This was in an effort to make the spec more accessible as one gets started. The DA-15 spec is not the best candidate for a plethora of chained modules. If chainability is your main focus, shift over to DB-25 and/or DE-9.
The old version of the spec featured USB and lacked CAN. USB is difficult to support when chaining is considered. USB is also a riskier thing to mess around with, since it probably has an expensive computer connected to it. CAN, on the other hand, chains easily and can be run over long distances. It is not as fast, but is sufficient for many electronics experiments. It is used extensively in automotive, automation & robotics environments.
Finally, after taking a look at some other multi-pin connectors, I found it wouldn't be to difficult to support 1080i HD video. While not the cleanest means of transmitting video, it is an ideal transmission method for those who like to tinker. The other connectors I mentiond combined composite, 'super' video & component video onto three pins. This pattern has been followed with the DB-25 connector pin spec.
When I decided to make an appearance at Maker Faire, I knew I would need a set of modules to help attract interest. I chose to embed a number of modules into an old flatbed scanner -- along with an Arduino Uno to run everything. The control surface featured all sorts of buttons & fiber-optic-backlit DE-9 & DB-25 connectors. The buttons & LEDs were driven by I2C port expanders & the NeoPixels via the `pixel-bus`. When one (or a few) of the buttons were pressed, the pin holes corresponding to the button label(s) would illuminate. Since I already had a 16x8 I2C LED display, I programmed the Arduino to display button label text there as well. The LED display was behind the scanner glass, while the control surface was mounted on a board well below. The scanner 'display case' and other parts of the display were linked by a single 25 connector ribbon cable. There were other means of input & output as well. Important to the design: I needed certain modules to protected, while others could be messed with. Those goals were well achieved.
08/17/2015 at 02:09 •
Here are a number of the Retro Modules I've been working on as I've been fleshing out the spec. This video demonstrates some of the possibilities provided via this framework. In building the modules, I discovered pros & cons of varying connector specifications. As things became more stable, I found I was able to deploy quite a few modules around the house, shop & truck. Just the beginning...