Next steps for the Alternator Regulator effort

[Jd1]

Originally Posted by CarinaPDX

But one question that jumps out is why use ethernet connectors for a CANBUS? At least for marine applications, N2K CANBUS should have DeviceNet connectors (proprietary SimNet and SeaTalk connectors notwithstanding), or perhaps screw terminal strips. It is unclear to me at first pass what other interfaces are available on the current model (v3) that could be practically used to configure and monitor. An (optional?) Bluetooth connection for smartphones would at least allow a quick set-up. CANBUS is a great solution in the long term, but requires more work and assumes a programmable device and display (most N2K systems use MFDs, not PCs). Most N2K interfaces for PCs use a USB connector - perhaps a USB header would be useful and eliminate the need for N2K-USB conversion (or even an N2K bus at all). However you decide to go, each approach should be specified with all that is required for each one, end-to-end, to help us understand the options. For instance, the regulator to N2K pigtail, N2K bus, N2K-USB or N2K-RPi hat adapter, PC or RPi, OS support, and the applications required would be one example. One of the options should be very simple to get going, like a BT to Android phone app, and not require more hardware or expertise.

I am very interested in your project. I will be needing a smart regulator for the new engine, which comes with an N-type 115A Mitsubishi alternator. The US-made regulators seem to only support P-type alternators, which seems to me pretty short-sighted. So please keep up the good work.

Greg

All valid points but I thought I would clarify a couple of your questions:

Ethernet connectors are available for pennies all over the place. Pre-assembled cables are also readily available at a fraction of the price of an N2K setup (as far as I know). In any case, it can't get much cheaper/convenient,

CANBUS is (IMHO) not workable yet but could be interesting in the future. I see no purpose for it in the current version.

Programming and monitoring on the current model happens via USB. There is a micro USB connector on the board and also a place to install pins for a more permanent connection. A simple terminal program will do the talking but you will need to dig through the (very complete) documentation. The average consumer is unlikely to have enough patience to figure out things.

[CarinaPDX]

I appreciate the info.

Originally Posted by Jd1

Ethernet connectors are available for pennies all over the place. Pre-assembled cables are also readily available at a fraction of the price of an N2K setup (as far as I know). In any case, it can't get much cheaper/convenient,

Cheaper I get. Convenient not so much - few users have a RJ-45 crimper. Admittedly it is easy to enough to cut a patch cable, then chase down a DeviceNet field-installable connector.But using ethernet cable seems like a bad choice due to the small gauge wire used. It would be better to just put a 5-terminal screw block along the edge, then cut a DeviceNet cable if needed. A second connector isn't necessary; the original N2K spec did not support daisy-chaining, although I think they have extended the spec to include it now. That is just unnecessary complexity.

Programming and monitoring on the current model happens via USB. There is a micro USB connector on the board and also a place to install pins for a more permanent connection. A simple terminal program will do the talking but you will need to dig through the (very complete) documentation. The average consumer is unlikely to have enough patience to figure out things.

OK, that would work for me. But it does sound a bit user-hostile. Better to write a small cross-platform program to put a GUI front end onto the USB text interface. I was just looking into it for another reason, and found that Python has a library for displaying gauges so it might be a good choice.

Greg

[thomasow]

All,


Thanks for the thoughts, really welcome more than one simple head noodling on things!

One somewhat common thread I see is documentation / communications: This project offers a lot of configuration and flexibility – chiefly via the USB port and I know often folks (myself included) get focused around that and perhaps loose the sight that in its simplest form only 4 wires, a few jumpers, and the DIP switches are needed. Adding extra sensing wires enables more capability, but still in the end I suspect 90% of installs would get by just using the DIP switches for configuration. It seems to me that the Quick Start guide could use more focus. Maybe even a new expanded ‘Users’ guide with the present document really reserved as an advanced Reference Manual.

The original vision, 5 years ago, for this project was to have an App connecting via Bluetooth for config / monitoring (hence the ASCII commands / status string which were developed for M2M comms, not primarily M2H), but that never got done. There is a person working on a PC based app to do this – progress is coming, slowly of course. Interesting note about DC Generators, this winter I am refreshing my DC Generator controller design, including the remote display. Seems like a great direction would be to generalize that display to include the Alternator Regulator as well as any future offerings. Though the HUI would be simple (compared to say an app), it at least would be available w/o any other devices.

Oh, CAT-5 – I am following the CiA-303 standard for CAN connection, the DMX standard is close to it, but not quite. I went this way as CAT-5 cables are simple to locate worldwide, reliable, and low cost. Even so, the PCB is also laid out to support a screw connector if one wishes instead of the RJ45’s. Is how a couple of folks are wiring things into their NMEA2000 system to display status. (Just let me know when ordering and I can make one up that way). There are other options: could go with Device Net standard (what NMEA2000 uses) – does bring up costs, and creates an issue of access to cabling / connectors outside of the 1st world countries. Would folks like to see a different cabling system used? There are a LOT of simple low-cost solutions used in the transport sector (auto/industrial/trucking). Again, one concern is access to components. Easy for Audi to source nice water-tight connector molded onto cables, not so much for me, and even less so for someone in Fiji.

Name, ya. I can see how Alternator Regulator is way too generic. I had wanted to use NextGen Alternator Regulator or Smart Alternator Regulator – those both were already in use. Originally I was using Arduino Alternator Regulator (hence the Blog name), but that is not in line with Arduino’s marks directions. So back to the placeholder Alternator Regulator. Plus I am looking to migrate CPUs in future designs. Any creative Ideas for a good name?

Case; One critical decision is: How water tight should the packaging be? IP67? Is possible, comes at a cost. Esp in modest (under 1,000 / year) volumes. But there are tradeoffs which could be made. Example, instead of screw connectors – use hard wires and route them out through a cable gland. Perhaps pot the whole thing? (DIP switches becomes an issue – and back to the app / display above, and I am sure some other issues). But the basic question comes down to: IP67 or just dust / splash tight and install it inside one of those gray NEMA boxes if one wants true water-proof solution?

Thank you all for taking the time to type your thoughts, I really do appreciate them and in some ways am hoping this effort becomes more of a community effort (though I would say the past year has seen great progress on that front!) Please keep thoughts / comments coming!

[hellosailor]

Software can be "open source" but unless I missed an important new memo, HARDWARE CANNOT BE OPEN SOURCE.

That's a contradiction in terms. Anyone can cut the legs off a stool to make it shorter, that does not make the stool "open source".

So perhaps what you want is a regulator that controls many power sources (as opposed to an alternator regulator, which controls only an alternator) and then in addition to having a flexible piece of hardware, you want that hardware to be controllable by an active computer, with open source code?

It helps to define the question.

The hardware would be incredibly expensive to design, especially with a modular "add more of these" construction that allowed the one unit to regulate anywhere from say, 100-200-400-600 amps at both 12 and 24 volts. Or perhaps up to 150 volts, to allow for solar panel inputs as well.

So, first tackle the hardware problems, then worry about open sourcing the control software. Once you see what the hardware production is going to cost you, I don't think you're going to be wanting to give away control of any part of it, but maybe I'm wrong on that.

[Jd1]

Originally Posted by hellosailor

Software can be "open source" but unless I missed an important new memo, HARDWARE CANNOT BE OPEN SOURCE.

Yes, you missed the memo. When a schematic is published and especially when there are only off-the-shelf components used, the result is considered 'open source'.
Oh, the operating computer code must also be available.
Basically when you can build everything yourself without buying anything from the designer then it is open source.

[Jd1]

Originally Posted by thomasow

Maybe even a new expanded ‘Users’ guide with the present document really reserved as an advanced Reference Manual.

Exactly !

Would folks like to see a different cabling system used?

No!

Case; One critical decision is: How water tight should the packaging be?

just dust / splash tight and install it inside one of those gray NEMA boxes if one wants true water-proof solution.

It is amazing how much cost is added by packaging (maybe excepting potting). I do not think people in general appreciate the cost of packaging.
Potting obviously eliminates any fixing which makes it undesirable for most situations.

[hellosailor]

Sorry, jd, but "open source" refers to "open source code" as opposed to the normal compiled, encrypted, inaccessible trade secret source code used in most computer electronics.

Hobbyist electronics mags and amateur radio mags all over the world have been publishing "kits built with common parts" and encouraging readers to go out and build them and modify them for a hundred years now, and that's never called "open source" as far as I can tell.

There's no "source code" to be openly accessible. No source code at all.

A project kit is a project kit. Source code means an alleged, attempted, or sometimes real programmer is going to be involved.

Even if some Makers' Faire in the Great White North has corrupted the phrase.

[Jd1]

Originally Posted by hellosailor

Sorry, jd, but "open source" refers to "open source code" as opposed to the normal compiled, encrypted, inaccessible trade secret source code used in most computer electronics.

Hobbyist electronics mags and amateur radio mags all over the world have been publishing "kits built with common parts" and encouraging readers to go out and build them and modify them for a hundred years now, and that's never called "open source" as far as I can tell.

There's no "source code" to be openly accessible. No source code at all.

A project kit is a project kit. Source code means an alleged, attempted, or sometimes real programmer is going to be involved.

Even if some Makers' Faire in the Great White North has corrupted the phrase.

I guess we will have to agree to disagree ....

[john61ct]

Open hardware is a huge very well established movement, closely related to Makers and co-working, hackerspaces etc

Anyone not clued into it is just out of touch needs to educate themselves.

Arduino itself is one of its best examples.

Here's an example from over five years ago

Made In China: Eric Pan and open source hardware
http://boingboing.net/2013/06/11/mad...an-and-op.html

Like trying to say feminism isn't really a thing 8-)

[thomasow]

Guys,

Lets not get bogged down on "Open Source" and its nuances.

And also, the architecture of using a common high-voltage bus with point down converters, and/or using a single massive control / power conversion device is not the architecture of this project - for very good reasons. So, if we can also not get bogged down on that trail would also be great (or start a new thread and folks can work through the ++ and -- of it)!

And JD1, you are SO right as how costly packaging can be, unless one is on 1K+ / mo volumes, then it becomes a bit more cost effective. . .

-al-

[Charente]

Few thoughts from a satisfied AltReg user:

Case - definitely step 1, something simple/cheap using a standard case - maybe with the pcb extended so the screw terminals are outside the case for ease ?

Remote control/display - this would be a good addition esp if you can reuse from the Generator project. Doesn't need to be complex - just 2 line lcd and some generic buttons (eg AltReg on/off, force float, etc).

Alternator control via voltage sensing terminal - this is the biggie which makes installation possible on existing internally regulated alts without modification/replacement. Maybe this would compromise the overall operation of the AltReg ? But it would open up a much bigger market for the device.

Remote battery sensing - you've talked about this and it sounds good, makes sensing more accurate and wiring simpler - but it's another board/case ?

RJ45 - good for simplicity but imho poor in marine air - I've had the socket contacts corrode and break. Best to stick with screw terminals - and maybe offer a pigtail for Devicenet.

See SeaLITHIUM VRC-100 Advanced Alternator Controller | | Nordkyn Design for some ideas - though I've not seen/used one.

Jeremy

[rickbase1]

Guys, I am the guy "slowly" building a GUI for the "Alternator Regulator for Al.

I was hoping to have an Alpha release for xmas, but we are back in the snow with our family for a couple of weeks and have a bottom job waiting for us back at the boat so it now looks like it will be another month or two.

It will be a cross platform tool designed to work with Windows, Android and Linux. To build it on Apple requires licenses and a Mac computer, neither of which I have.

Al has kindly hammered a bunch of info into a database for me, so things like descriptions and help that match the manuals will all be part of the experience.

The initial release will be serial USB with Bluetooth and CANBus to follow.

If you have a bit of spare time, a linux computer, a spare Alt-Reg and would like to help with some Alpha testing please PM me.

I also want to add some quick config wizards and would like some input for developing those.

As to Al's original question about enhancements, I would like to see:


1) DeviceNet connectors
2) bigger terminals that are spaced out
3) an option to have it potted
4) an IP67 case option
5) jumpers instead of solder connections for the commoning the DC on the back of the board
6) spare memory for other tasks, possibly by upgrading the current AtMel processor
7) onboard min blade fusing, so fuse pigtails are not required

And for Christmas... a Solar Controller, an active LFP BMS, and bottom anti foul that lasts forever !

Merry Christmas all!

Sailor Rick

[CarinaPDX]

Originally Posted by Charente

Alternator control via voltage sensing terminal - this is the biggie which makes installation possible on existing internally regulated alts without modification/replacement. Maybe this would compromise the overall operation of the AltReg ? But it would open up a much bigger market for the device.

I suppose it would be a quite different product, but that would be a huge winner for some of us. As I stated earlier I have an N-type 115A Mitsubishi alternator. It has a remote sensor attachment point which as shipped just connects locally to +DC on the engine but can be used to work around diode isolators (or in my case through the ProIsoCharge to sense whichever battery(ies) is/are charging). While AFAIK there is no way to alter the OEM regulator set point, it should be easy to manipulate the alternator output by controlling the voltage that is sensed by the existing regulator. If there was a way to do that I would be one very happy camper. One of the benefits is that it might not invalidate the warranty (no internal modification required) and works with a replacement alternator without modification as well. And if the external regulator were to fail it would be easy to remove it and be up and running in stock mode.

Greg

[thomasow]

All,

Thanks - Lots of good ideas / comments here. At this point I would like to focus in on the case options, as it impacts other topics

After 35+ years in electronics development, I am still always surprised how much cases / connectors contribute to a products cost. 50% is not unknown! For this effort, am also a bit limited due to the volumes and to be honest up front tooling costs. So, a few ideas that I think would be workable:

• Dust / splash resistant casing. Something like this: https://www.mouser.com/productdetail...5qwsGPGYf6Q%3D

PCB could extend past case sides with screw terminals exposed for easy access. A nice stick-on label on the top to ID the connectors, and one inside to guide on how to set up the DIP switches.

Option #1, the Mounting Option

• Like case, but configured to provide some level of water / dust proofing starting with something like this: https://www.mouser.com/productdetail...T0C5OTT07K4%3D

And here it gets tricky. Can go down the path of water-proof connectors passing through the side of the case – costly, I would expect costs to be $50 just for connectors!

Another approach could be instead of using bulkhead connectors, attach pig-tails to the PCB and route them out through a waterproof gland. Make the wires marine-grade, about 5' or so long. Folks would cut and crimp the ends as needed and/or extensions if required.

Option #2: Waterproof w/connectors
Option #3: Waterproof w/pigtails

(And if folks want screw terminals to attach their own wires and use a bulkhead gland, well - that is really option #1 mounted in a nice gray plastic NEMA box. . .)

• Potting. Kind of a slight variation to option #3. Use pigtails, but instead of a bulkhead gland, pot the whole think up. Still have to consider the DIP switches and how to protect them. Not sure how much this approach would bring vs. #3, but in any case:

Option #4: Potting

Would like to hear folk’s thoughts on these, as well if there are any other ideas? After I get some direction on the case I want to come back to the CAN next.

Thanks again,
-al-

[Jd1]

Originally Posted by thomasow

Another approach could be instead of using bulkhead connectors, attach pig-tails to the PCB and route them out through a waterproof gland. Make the wires marine-grade, about 5' or so long. Folks would cut and crimp the ends as needed and/or extensions if required.

Unfortunately I can't contribute much to the packaging idea as that has always been my Achilles heel. I can however suggest that it is not wise to assume people have the ability to crimp anything. If pigtails are decided upon, it would be prudent to have them terminated with connectors for plug-and-play operation.

Is it possible to have a waterproof connection if you route more than one cable through a waterproof gland? I was always under the impression that you could not seal the inter-cable spaces properly. I would think that waterproof connectors would be the only way to go but as you mentioned, cost is immense. That is why, IMHO, splash proof makes more sense. It should be easy to find a mounting spot for the electronics that just requires a splash proof box.

If this would be a large production run item, I would suggest the pcb extend into a sub compartment that contains the connections (screw connections) but this is not cost effective in this case. One could forego the sub compartment and extend the pcb outside the case with the acceptance of the possibility of some corrosion.
This is the only cost effective and user friendly method that I can think of.

As mentioned before, potting is not my preferred method of going but if better moisture protection is deemed desirable, conformal coating could possibly provide that additional protection without having to do full potting. The board would be non repairable and I am not sure how much protection the dip switch mechanism (the internal contacts) would have from corrosion.
Potting or coating would also remove the options of attaching extra wires for things like hard wired USB, solder bridges for configurations etc etc

The more I think about it, extending the pcb makes the most sense and is the most cost effective. It also allows an off the shelf case to be used by simply modifying (cutting) one side so the case and lid have a gap to allow the pcb to extend through it.

I will be watching with great interest to see what other people come up with!