Open Source Arduino Alternator Regulator

[CaptainRivet]

Originally Posted by Andrew Lippman

Is it posible to connect to the Victron smart shunt so you can measure actual current into the battery? That way you can control for the effect of running electrical gear while charging. My understanding is that Wakespeed does that. Link also did it once. When I asked Balmar about that, the reply was that it was too expensive. Likely less true now.

He adds a hall effect sensor that actually meassured the current already, not necessary to add another 130Euro for a smartshunt.

[Bellinghamster]

Originally Posted by markxengineerin

Another hardware question: for the power supply that feeds everything, I was planning to use LT8210 from Analog Devices. It takes in a wide range of dirty/bad input voltages and makes a clean 12V output to protect everything downstream, but it also does a lot of things we don't need, and it's expensive at $10 to $15 each in the quantities involved here.

Does anyone have ideas on how to more cheaply turn the sometimes highly variable battery voltage into ~12V, and additionally provide reverse polarity protection, protection from transient spikes, etc?

Ideally this part would work the same for 12, 24, and 48V inputs. Downstream of this is a TPS562202DRLR and then a LDO to produce a clean 3.3V or 5V, so the output doesn't need to be a perfect 12V, mainly looking to protect against brown-outs and transient voltage spikes.

I'm not sure why you want to stack so many regulators, and don't know the current needed (i.e. are going to to send field current off the bulk-regulated rail, or directly off the input?). Something like a LM5012 (6-100V input, 2.5A buck converter) would get you all the way to your LDO in one ~$3 part. Adding $10-15 (plus additional components) to "boost" if voltage is <6V (cranking? bow thruster?) doesn't seem necessary, but would love to hear your reasoning.

For reverse polarity protection for low-current applications, a simple diode is hard to beat as long as you're not up against voltage drop concerns or striving for high efficiency - neither should be much of a concern in this application.
If you're running 10A of field current through it, you have to consider power dissipation/heat, then probably better to use a MOSFET as an "ideal" diode: https://www.onsemi.com/download/appl...and90146-d.pdf

[maine-cruiser]

Originally Posted by Bellinghamster

I'm not sure why you want to stack so many regulators, and don't know the current needed (i.e. are going to to send field current off the bulk-regulated rail, or directly off the input?). Something like a LM5012 (6-100V input, 2.5A buck converter) would get you all the way to your LDO in one ~$3 part. Adding $10-15 (plus additional components) to "boost" if voltage is <6V (cranking? bow thruster?) doesn't seem necessary, but would love to hear your reasoning.

For reverse polarity protection for low-current applications, a simple diode is hard to beat as long as you're not up against voltage drop concerns or striving for high efficiency - neither should be much of a concern in this application.
If you're running 10A of field current through it, you have to consider power dissipation/heat, then probably better to use a MOSFET as an "ideal" diode: https://www.onsemi.com/download/appl...and90146-d.pdf

My thoughts exactly on the LT8210. Seems to need lots of external components. Would you ever need boost functionality? In addition to the LM5012, TI has a bunch of buck converters you might be able to use depending on your power requirements:
https://www.ti.com/power-management/...8max=48%3B100&

Bill

[nfbr]

I have a WS500 running a 9kw APS160 at 56v. My tips based on its weaknesses.
A onewire temp sensor rather than thermistor
Dual current sensors, alternator and battery
LONG damping of input signals. Seconds at least.
A real time clock so that adsorb can be limited to once daily or less.
A real real hard think on wiring for catamarans. Generally the battery and alternators are remote from each other. The WS wiring sucks for that. Think real hard about what connections are needed and what can be inferred instead.
Focus on LiFEPO4. Thats 99% of refit use cases today.

[becrux]

Following…

[markxengineerin]

Is it posible to connect to the Victron smart shunt so you can measure actual current into the battery?

Yes it will do that

I'm not sure I appreciate the need for real-time display.

Display is mounted on the enclosure, it's a tiny OLED unit mainly useful for troubleshooting. I agree, there's nothing interesting going on during normal operation.

The display should have a simple on/off switch for the alternator

It will include an on/off switch, and where to mount it will be up to the user

I would add presets for different battery profiles with the option to override bulk/abs/float voltages and time.

All these sorts of things will be user adjustable, but I'm sure the community will come up with some popular presets for those who prefer. The programming adjustments are currently done in the Arduino IDE, but I like the idea of using a web browser instead. I've never done that, but the hardware will be there to support it.

Sort of an aside here, but can a self-excited alternator be converted to be used with an external regulator? Can such an external regulator stop the alternator charge?

I don't know much about these, sorry, might try to google it.

[lateral]

Following with interest.
I have currently LiFePo4 and a MC718. The Balmar BMS SG200 randomly goes offline and reinitiates without recognising the 718. Losing faith in the system even though it is still charging as programmed.

[markxengineerin]

A onewire temp sensor rather than thermistor

I need to figure out the best way to attach these to the variety of alternators out there. What is the most "universal" way? I was planning a m5 (or other) washer style thermistor, but open to other suggestions.

LONG damping of input signals. Seconds at least.

No problem- software adjustment

A real time clock so that adsorb can be limited to once daily or less.

Looking into it

Would you ever need boost functionality?

No. Someone suggested this: https://www.ti.com/lit/an/snva717/snva717.pdf I haven't read yet, but hopefully "protections" can be done more cheaply, if not, $15 chip it is : )

Something like a LM5012 (6-100V input, 2.5A buck converter)

Thank you, research in progress

[CapnMarden]

Following

[nfbr]

Re alternator temp lcn. The jury is out on that. The three real limits are bearing temp / life, insulation rating and rectifier diodes.
There is no one point gives all of these and no manufacturer publishes the insulation temp ratings (which are standardised for electric machinery)

The stator laminations run hot due to eddy currents, so thats not a good spot. Most look for a stud on the rear of the case.
My WS probe is currently measuring exhaust air temp as that seems to give the most stable behavior.
A digital onewire sensor is much less affected by wiring and noise than thermistors. I had big WS problems there.

Big gotchya warning for 48v. Rotors are 12v or 25% pwm energized. Caution needed to not cook them.

Oh, also, limit rotor to 20% of peak if no tach signal. Saves heating a stopped alternator.

[markxengineerin]

A digital onewire sensor is much less affected by wiring and noise than thermistors. I had big WS problems there.

What is WS? I don't have much preference for what precise location is measured, I think any temp measurement can be made useful, but I want the solution to be universal and easy to install. I tried an IR sensor because that seemed to fit the bill, but it was a fail, it they're affected too much by ambient temperatures. When mounted close enough to an alternator to target them (in engine bay), it's not going to be at a stable temp very often.

Big gotchya warning for 48v. Rotors are 12v or 25% pwm energized. Caution needed to not cook them.

I have never had anything but a 12V alternator. Could you elaborate? I think you're saying that at 25% duty cycle 48V, I will be feeding the equivalent of 100% duty cycle at 12V to the field? I have the opportunity to feed any voltage and any duty cycle to a 48V alternator (or any other), but I don't know what's best. If I understand you correctly, I should only feed 12V, that way I effectively have finer resolution on the PWM? The smallest step change I can make is limited by the steps in a digital potentiometer, in the current design. Do most higher voltage (24, 36, 48V) output alternators run with a 12V field? That would actually make the design simpler.

"Oh, also, limit rotor to 20% of peak if no tach signal. Saves heating a stopped alternator."
I think that one is easy even without a tach signal. If field command is >X and output amps <Y, engine is not spinning, stop trying to charge.

Currently, I have no good plan for measuring RPM, because the analog inputs are using ADS1115's and not fast enough for that (but they're very accurate). I don't know if a raw analog input on the microcontroller is going to be good enough, I've found them to be terribly unreliable and noisy in the past, so if anyone has a suggestion for a fast + good I2C or SPI analog input chip, that would be a helpful addition.

[nfbr]

WS is Wakespeed 500

Yes, most 48v alts are 12v rotor.

[HankOnthewater]

Hi MarkEngineering, I see lots of people requesting additional features. There must come a time that you (as project leader) saying "OK, that is enough, I will now work with the suggestions I have received".
One needs to keep withing the scope of the intended project, and no, one can not please everyone all the time.
In regards to cost. I would not worry about an additional costs, as the end product will be (I hope) still much cheaper than a commercial product, and better than that product. In the big picture, it is a very small expense for an expensive hobby (sailing)

[markxengineerin]

Thanks, I'm cautious of it. So far, the additional features have added single digit dollars to the cost, so they're worth including. I think the forum is very valuable for getting more eyes on the design. Some ideas should have been obvious to me (ability to read a shunt, adding a CAN module, etc.) but were not at the time. And it's never going to be a more convenient time to add features than now!

[Jon Hacking]

Originally Posted by markxengineerin

...Are there any other features that people want? Now is the easiest time to add things before finalizing the first PCB order.

Excellent project, Mark! But just to be clear, you don't actually talk about your intended charging algorithm.

I started looking into algorithms after I once blew up a $3000 bank of Sonnenschein Gels that went into thermal runaway at just 13.8v. I ran into Al Thomason, the mother of the (Arduino, open-source) Very Smart Regulator (VSR) & now the main designer at WakeSpeed. BTW, I can point you to the VSR designs if you want. I was one of Al's testers, as I'm an EE, HW & SW.

I'm now pretty convinced that a "tail-current" charge is the best way to fully charge just about any boat battery, lead/acid or LiFePO4.

In the initial Bulk stage, you push the alternator hard until the battery reaches the Acceptance voltage. Then, in Acceptance mode, you keep the battery voltage stable while monitoring the current into the battery. When the current into the battery required to maintain that Acceptance voltage drops to ~1% of the Ah capacity of the battery, the battery is full & the system should drop down to a Float voltage, where the battery can sit all day.

Are you planning for your regulator to follow this sort of charging profile?

The main reason I ask is that you don't talk about sensing the current into the battery, which is crucial for determining when the battery is full. Many regulators just rely on timing for this, which is sub-optimal at best. You only seem to talk about sensing the output of the alternator, which is a very different thing (& not really necessary for the regulator to work correctly).

As for features:

• I love the display idea, & would like to know more about that.
• All lead/acid batteries need a battery temperature sensor, as the charge curves are temperature dependent (as I found out, the hard & $$ way)
• LiFePO4 batteries need a temp sensor for the alternator, not the batteries
• The current sensor should be for current going into the battery, not alternator output (or it needs 2 current sensors)
• LiFePO4 batteries prefer to be charged to only ~85%, but it's hard to know when this is reached