Another Engine Control and Regulator project - DC Alternator

[thomasonw]

Here are a couple of You-tubes,  the 1st one showing a trial run of the controller with the UNO based generator simulator:  http://www.youtube.com/watch?v=4zDpfDg4biY


And here is walk-through of the system at this point:  http://www.youtube.com/watch?v=80CcIWkqnqM

-al-

[RJ]

To operate on 24v would be a small change - the main inductor in the switching power supply should be changed, and the I2C regulators would need to be watched for heat - though so little current is used in them I suspect there would be little issue.   Then of course the resister divider ratios and crow-bar would need to be adjusted.

This would allow for full operation on 24v - controller, remote, powering 24v devices (starters, field, throttle control motor, etc..)

Directly driving LOW side would be more difficult - though mounting some FETs on an external heat sink could be used to convert the high-side drive to a low-side drive.  (Interesting there was a lot of early talk about High side vs. Low side drive early on...   )


Have also looked at using this controller with a 48v system, and that brings in different issues.  Many of the logic devices have Vmax limits around 40v.  However, if one is able to power the controller from a 12/24v source and drive 12/24v devices (throttle controller, etc) and a 12/24v Field this will work well.  And one could also look to add an external Field FET pack to drive a 48v field if needed.

Have been working with some one looking to do a 48v alternator that uses 12v for the 'system' (starter, throttle, etc.) as well as a 12v Field.    If you are interested in 24v verion  I can look into that in more details.  Let me know.  And into the idea of an external FET pack, to get the LOW side drive . . . .  Thinking perhaps an opto-isolated connection would allow for a 'universal' high or low side driver...

-al-

My engine controls will operate off of 12v, starter, throttle control, etc. I have  attached a diagram on how my alternator is hooked up. It's a large brushless 6 phase alternator rated at 400a at 28v and 50a at 12v, however the 12v is derived from the regulator and not directly from the alternator.

Take a peak and see what you think. Talking with C.E. Niehoff the max field current draw is nearly 28a although speaking with them typically it would be half that.

-Randy

[thomasonw]

My engine controls will operate off of 12v, starter, throttle control, etc. I have  attached a diagram on how my alternator is hooked up. It's a large brushless 6 phase alternator rated at 400a at 28v and 50a at 12v, however the 12v is derived from the regulator and not directly from the alternator.

Take a peak and see what you think. Talking with C.E. Niehoff the max field current draw is nearly 28a although speaking with them typically it would be half that.

-Randy

Randy,  Yes - that is quite the beast of an alternator!  In thinking about 'other voltages' over the past couple of days, I think it can be framed in three broad catagories:

1) Voltage of accessories:  Fuel Pump, Throttle Control, etc.
2) Voltage of alternator Output and Battery
3) Voltage of Field Drive.

With regards to this controller in its current form:  In a 12v world all three above are the same and all is good.  In a 24v world, this would still be true - though some small change to the +5v regulator is needed  as well as resister values in the voltage dividers.


However, in a pure 48v world there are issue in that many components (ala the H-Bridge for the throttle controller, VBat input to INA-220's etc) are limited to a V-Max of around 35v-40v.   In a PURE 48v world this design would not work as is.

------------------------------------------------------------

Now, if all the above voltages are NOT the same, for example if the Battery and alternator output is 48v (Item 2 above) while all the accessories are 12v (or 24v) - this design will work with simple resister modifications to pre-scale the 48v sampling lines down.   AND if by chance the Field is also 12v (or 24v) (Item 3 above) all is well!  (This is how "yellowhead" is looking to run his system).


BUT:   If the Field Voltage (Item #3 above) needs to be 48v, or the Field needs to be driven from the LOW side- the current layout will not work.  Either an external FET daughter board would be needed, or the design modified to accommodation a low side drive.  I am planning on revising the artwork once I bring up this board and have it running on the actual Kubota/DC generator we have, and can look into making those design changes.  So, THAT issue can be solved.

-----------------------------------------------------------

NOW:  To your alternator / system:  Sounds like you would be running in a 12v Accessory standpoint (Item #1 above), so that is no problem.  24v output (Item #2) is easly handled via resister changes, and item #3 (The LOW side drive) could be handled either with the current PCB artwork via an external FET driver, or with revised artwork.

But I am wondering what you are looking to do about the 12v charging output?  I located this doc from Ce Niehoff: http://www.ceniehoff.com/Documents/Ctrl_Hyperlink/TG0014D_uid12172009224052.pdf  which describes how an SCR pack is used to reflect the 12v output and at the same time regulate it to 12v.  If you wanted to retain that feature, I think it would be a bit much to support with this project:  managing both the 24v and 12v outputs of the alternator.  However, if you are looking to only bring out 24v from this alternator, and get the 12v replenished via some other means - a 2nd small alternator or perhaps a 12v 'battery charger' then with something to handle the LOW side field issue I think this project would work fine.  (And I think the 28A Field current can be easily handled with FETs, it is just a case of getting them to drive from the LOW side)

[RJ]

I'm not terribly concerned with retaining the 12v feature of the regulator. I can charge the starter battery with either a traditional charger or simply leave the system manual start, I could run the controls through a DC/DC converter to bring the voltage back down to 12v. Everything is still in the planning and collecting parts stage at this point. The only real reason I wanted to maintain 12v for the controls was the more abundant supply or 12v parts. Things can be changed to 24v as long as I can get the parts needed. Ideally I would like to go to 48v, but I have not had the time to try my hand at taking the alternator apart to see if that is something that can be done. I'll need Bob's help with that. The phases are currently wired in delta, connecting them in WYE would get my voltage up there right around 48v. Testing would then need to be done to see how fast I would need to spin it, and also what my field requirements would be. I'm working on building a test bed to drive the alternator off a small motor for testing purposes.

The only need for the 24/48v would be to charge the batter bank. So it sounds like this system would work for my application with some minor tweaks.

[thomasonw]

<snip>
The only need for the 24/48v would be to charge the batter bank. So it sounds like this system would work for my application with some minor tweaks.

Sounds like it would.  I have been revising the artwork and moved to a LOW feed to that will simplify supporting field voltages that are different then the controller voltage.  Still bothered by why almost 100% of other controllers I have seen drive High. . .   But at this point low simplifies support for the mixed voltages so no choice.

<SNIP>

Another thought would be to consider elimination of the buck converter for 5V supply.  The board use of power is so low that a linear regulator might end up saving power, especially if a low noise LDO regulator was used.  (I've found this to be true on my battery bank controller.) Secondary benefit would be less noise problems on current and voltage conversions.  I know for your own use you have other 5V things planned but for public use the linear approach is smaller, cheaper, bullet proof.

I have been looking into using cascaded liner regulators (to spread the heat load).  1st regulating down to 9v and then to 5.  Doing the math I come up with a Tj on the 9v regulator of 75c while operating in a Ta of 120f.  Reasonable, however if going to a 24v 'system' supply voltage (and using 15v intermediate) we get a Tj of 108c.  The spec calls for Tj-max 150c, but I always worked under the thought of Tj <100c as a magic number.  Would be very interested in hearing others opinions:  Is it worth / reasonable to run at Tj = 108c?  Or is that getting too much on the edge?

Thanks
-al-

(detail math here: http://smartdcgenerator.blogspot.com/2013/01/looking-at-5v-regulator.html)
(And BTW:  I have been posting links for the TMI level of stuff as opposed to placing it directly in this forum, is this an acceptable practice?  Or, is the custom to Fill The Posts Up and not link to an outside forum?? )

[BruceM]

Hi William,
I reviewed your note on linear vs PWM regulation.  How did the controller draw get to 450ma? My PIC processors draw about 2ma at 5V (I'm assuming Atmel is competative), your ina's are about 2 ma each, so where's the 5V power hog?

I was assuming that your board was drawing about 50ma on average, not 10x that. At 450 ma I'd stick with the buck converter.

I'd be looking carefully to find out where all that power is going.  Something must be wrong, or I'm out to lunch, again.

Best Wishes,
Bruce

[thomasonw]

Hi William,
I reviewed your note on linear vs PWM regulation.  How did the controller draw get to 450ma? My PIC processors draw about 2ma at 5V (I'm assuming Atmel is competative), your ina's are about 2 ma each, so where's the 5V power hog?

I was assuming that your board was drawing about 50ma on average, not 10x that. At 450 ma I'd stick with the buck converter.

I'd be looking carefully to find out where all that power is going.  Something must be wrong, or I'm out to lunch, again.

Best Wishes,
Bruce

Well Bruce, the answer is simple:  I was being lazy, and am a bit embarrassed.  Being lazy in that with all the wires associated with the test jig, I just did not want to risk disturbing things.  So instead of unsoldering L2 and taking the 5V current draw there I measured the overall "+12v" feed. Used an assumed efficiency ratio for the switching regulator.  Then subtracted out the spec sheet draw of the LCD.   Going back it seems the LCD draws more then the Spec sheet.  And there is more more:  I had unplugged the USB cable to the simulator UNO - and as a result it was drawing +5 from the controller as well....  So there you are:  Lazy, and embarrassed.

Today I did the work I should have done in the 1st place - the controllers +5 draw is  22mA at idle when measured via L2.  Peaking at 26mA while 'under way'.  Redoing the calcs gives a max Tj of 53c in a 24v system with a big heat-sink, and only 77c if no heat-sink is used.   Looks like a simple PCB based heat-pad area (like on the LCD controller) will be just fine.

Thank you Bruce for getting me out of the heat. . .

[BruceM]

William, glad your power consumption is where it should be. 

26ma is more like what I expected, and there's no buck converter that's going to be efficient at that current, so you might as well keep it simple and use a linear regulator for your public version.

[mike90045]

I have been looking into using cascaded liner regulators (to spread the heat load).  1st regulating down to 9v and then to 5.  Doing the math I come up with a Tj on the 9v regulator of 75c while operating in a Ta of 120f.  Reasonable, however if going to a 24v 'system' supply voltage (and using 15v intermediate) we get a Tj of 108c.  The spec calls for Tj-max 150c, but I always worked under the thought of Tj <100c as a magic number.  Would be very interested in hearing others opinions:  Is it worth / reasonable to run at Tj = 108c?  Or is that getting too much on the edge? 

Unless you need ultra clean DC (sn -120dbc) don't stack regulators, look into a switcher that is clean enough, or use a big honking resistor to dissipate some of the heat, near the expected current level.
Or look into the LDO (low drop out) regulators, they drop less voltage across them and dissipate less heat.

Tj of 108c is fine, as long as you make it easy to replace the fried regulators. (and whatever else downstream fries with it)

[BruceM]

The LDO linear regulators do have about .5 ma or less to ground idle current, vs about 3ma+ for the standard linear regulators.  In William's case of starting with a 12V battery source, with generator power, either would be fine, he only needs one and won't need a heatsink, as he has stated. 

The only problem with series linear regulators (with plenty of voltage drop) is stability, but if you're willing to watch that re: matching the regulator's specs with caps you can do it as William proposed (when he thought his current was much higher than it is.)

If the starting voltage is too high for a 78xx regulator or LDO type I use a PNP darlington transistor with a small zener to base for a pre-regulator, you only need about 0.5 ma of current to ground to adequately stabilize a 1/4 watt zener.

[thomasonw]

Thank you all for the continued comments on the regulators, still have time to revise it as I am not looking to 'release' a revised design until after the current controller is tested on an actual generator.  One thing I keep in mind:  the 'source' voltage for this project is really related to the voltage being used by the peripherals, ala the Fuel pump, Starter key-switch, throttle control module.   Even if the driven alternator is 24v, or 48v -   I suspect many deployment will end up being 12v, just because 12v 'parts' are so really available.  And in any case - supplying the controller with more than 24v nominal will cause issues with other components, like the H-Bridge and the FET drivers - all of which have 30-35v Vmax ratings.

On other news,today  have reached a kind of mile-stone.  I posted the next release of the Source (v0.1.5) with has major edits as a result of bench testing using the simulator.  And have come to a transition point - I think I have gone as far as I can with bench-testing.  Need to get this thing connected to actual hardware!  Given our generator is outside the weather will play a little in there.  And we have guests coming this weekend, so I suspect progress will take a short breather.  (Plus, I think my Wife is getting a bit tired of seeing me hunched over the computer staring into the screen and muttering things to myself).   (Look under the Source tab on: http://smartdcgenerator.blogspot.com/)

As an overview, here is a run down of the key features of this project (It has grown a little on scope - after all many things are 'just a few lines of code'!)

Overall Features:

• Controller with separate remote display and operating switches.
• Support for local operating switches on controller.
• Supports system voltages (Starter, throttle control, etc) of 12 or 24v.
• Support of Alternator field and battery voltages up to 48v systems.
• Alternator voltage support is independent of System Voltage - as 12v system can manage a 48v alternator.

Engine Management:

• Selectable Auto Generator Start-up (based on battery voltage)
• Auto stop after charging is completed
• Warm-up and cool-down periods w/o alternator load.
• Active throttle speed control, adjusting engine speed to match current loading.

Alternator Management:

• Fully configurable 3-stage Alternator regulation with soft ramp of initial power on.
• 'Over Charge' (or 4th stage) support for AGM batteries.
• Battery Temperature Compensation and adjustment for all charging voltages.
• Precise remote measurement of battery voltage,  including  remote negative lead to accommodation any common ground line voltage drops.
• Alternator charging actively managed by measured battery voltage, charging current, and total load placed on driving engine.
• Charging rate reduced under excessive EGT conditions.
• Adaptive alternator reduced power mode to support simultaneous driving of other loads, ala water pumps, hydraulic pumps.

Fault Detection:

• System fault monitoring:  Over temperatures (Engine, EGT, alternator, battery), voltage, amps.
• Broken sender / wire check for Oil pressure and Cooling Water Flow Senders.
• Fail safe protection:  Watchdog timer, hardware crow-bar for over-voltage
• Software self-checks for incorrect configurations, and internal operating errors.

And some likely future enhancements will include:

• Auto Start quite period holds off (with optional RTC added)
• Reducing charging rate due to high alternator temperature.
• Lifetime recording of total amps /  watts produced by alternator, hours of operation, etc.

There is one PCB board left if anyone is interested in it, two of the other ones have been spoken for.  Just drop me an Email or Message via the board.

And again, thank you to everyone who has given input and comments, there will still be opportunity for more  - but I want all to know how much I really do appreciate the inputs and ideas I have received here. 
Next step will be connecting this to our existing generator / watemaker: http://mvvikingstar.blogspot.com/search/label/Kubota%20DC%20Generator

[thomasonw]

It occurred to me last night that I am not being true to this forum.  Yes, am using a small engine to drive a generator (DC in this case) as many people here seem to do.  And the inclusion of a Watermaker pump does bring in more then one purpose, but to call that co-gen is perhaps stretching it a bit.
.
It also occurred to me that to repurposing the Watermaker clutch output to drive a water circulation pump instead would just be a "small matter of programming".   So a question:  Aside from engine coolant temperate, would be be good to take anything else into account when deciding to enable a heat extraction water circulation pump/valve?   Hold-off periods during warm up? (Let the engine run at full temp for some time to heat-soak, as opposed to just look at current coolant temp?)    Current load?  (Ala, on reduced load, stop circulating in anticipation of engine cooling?)    Anything else?

With so much being monitored in this design (Engine load, Coolant temp, EGT, engine RPMs) there are opportunities to go beyond simple engine temp. But to be honest I am not sure if there is really a need to, or  just look at coolant temperature with perhaps some type of hysteresis between the On and Off triggers and be done with it.

Thanks
-al-

[veggie]

One thing to consider is an auxiliary set of contacts for some form of a load dump.

For instance:
If the system is to be used for co-gen, then there should be a way of producing heat (from a loaded engine) even after the batteries are charged.
If the battery charging cycle is short due to minimal discharge, then the heating portion of the CHP may be useless (too short of a run cycle to do any good).
However, if the load can be transferred to something like a heating element (to assist heating) once the batteries reach full charge, then the full HP and cooling BTU output of the system can be utilized.

Adding CHP capability does open a can of worms on the programming side because there are so many variables and ways to achieve the same result.
Such as... how long do I dump the load into the heating element after the batteries are charged to "x" voltage.

just a thought,
veggie

[Lloyd]

One thing to consider about feature creep. If turns out all systems are controlled with a single controller with no ability for manual over ride for each of the subsystems, then a controller failure can render the whole system caput until a new controller can be had/installed.

Don't mean to be negative, but you must allow for limp home.

Lloyd

[BruceM]

Since this controller is doing alternator regulation, I don't think any sort of manual reversion is practical.  Instead, I think having a couple of fully tested spares on hand would be best, after the design is fully tested.  My various PIC based boards have proven to be so reliable, even the non-pcb prototypes, that I wouldn't mind an integrated solution like William's. I expect the AVR processors will also be robust.

To combat feature creep, I think I'd just have some spare I/O and perhaps bring the 2- I2C serial buss pins to a connector or header. Then someone using his open design hardware and software could always add on almost anything externally.