Alternator field control circuit

[AndreyB]

I want to use this schematics - it says P-channel


while soldering it on this breakout board

with a P-channel FQB12P20TM I have already

I am just not sure which wire would go where because I am not super comfortable with S and D and etc.

[kb1gtt]

See attached schematic captured in Dave CAD (EEVblog's terminology) Does this make sense to you?

[rus084]

i draw at morning , just now came to PC .
kb1gtt , your schematic is better .

[AndreyB]

See attached schematic captured in Dave CAD

It works!



red wire is 12V
white wire goes to low-side driver on Frankenso
green wire is alternator field
there is a diode under the tape
blue wire is GND

with 35% duty @ 400Hz, I've got ~0.8 volts of charging (12.2 becomes 13v)
diode<>GND: 0.8A
field wire<>field wire: ~1.2A

I then proceed to a higher duty cycle - probably 80% (?), got 6A in the field wire<>field wire, a lot of charge, and melting solder with nasty smell in 10 seconds. Mental note: do not try this with duty above 50%.

PS: we need to add diode support into the DDPAK board. Could it be a surface-mount diode, or does it have to be a through-hole?

[kb1gtt]

Thanks for the data points. I'll have to re-spin the DPAK board to add some components.

Also good read found here, and attached in PDF form to prevent bit-rot.
http://www.embedded.com/design/real-time-and-performance/4006472/Designing-a-microcontroller-driven-alternator-voltage-regulator

[kb1gtt]

I see chips like this L9409F exist which includes pretty much all the parts, diode, MOSFET, ect in one chip. Makes for easy work and less parts to keep track of, as well it comes with thermal overload. http://octopart.com/l9409f-stmicroelectronics-18031937

Can we find a readily available chip like this to be used as the alt regulator?

[AndreyB]

google "Alternator Regulator chip" returns a lot of options, none of which are available for sale in low quantities.

[puff]

junk yard is your friend

[kb1gtt]

I found this one http://octopart.com/l9484-stmicroelectronics-13884940 but price is a bit of a yuck.

[kb1gtt]

Turns out that if you click that octopart link for newark, they don't have it. So that was blah. Here are some that seem to be actually available.

http://octopart.com/l9911p-stmicroelectronics-18034601 $8.93
http://octopart.com/l9911f-stmicroelectronics-18034875 $8.75
http://octopart.com/l9914b-stmicroelectronics-18034602 $10.50
http://octopart.com/l9914c-stmicroelectronics-20571333 $10.71

[AndreyB]

Here are some that seem to be actually available.

http://octopart.com/l9911p-stmicroelectronics-18034601 $8.93
http://octopart.com/l9911f-stmicroelectronics-18034875 $8.75
http://octopart.com/l9914b-stmicroelectronics-18034602 $10.50
http://octopart.com/l9914c-stmicroelectronics-20571333 $10.71

So which one should I order if I should order any? They are definitely not cheap I would not mind doing the PWM logic as long as we can find a protected P-channel chip.

Do we need some heatsinking huge-via areas around the proto areas on Frankenso?

[kb1gtt]

I would say go with the L9911F, it's the lower cost and it has a better datasheet. It looks like you'll have to PWM the P signal, as that will cause problems if the charging circuit if it does not have a proper signal. In your case the Neon doesn't have the P field wire so you can't get that from the alt. I'm not sure what it does if you are in a fault detected state. See page 16 of this datasheet http://datasheet.octopart.com/L9911F-STMicroelectronics-datasheet-10836994.pdf I don't know if it just triggers that DFM line or if it does something else.

If we can find a thermally protected P-MOSFET, I'd like that. I could mount the diode near the MOSFET such that if the diode gets to hot it triggers the MOSFET protection. However so far no dice in finding a good P-MOSFET with a 10 to 20A capability.

[AndreyB]

Why is it so hard to find protected MOSFET's?

I've been looking for a P-Channel, thermally protected, 10amp or more, enhanced mode MOSFET, DPAK or D2PAK, preferably with a voltage clamp protection.

You should not choose a discrete MOSFET as a driver for a signal or source going out from an ECU...

Example of ST's HSS-devices (you should find devices up to 100Amp without problems):
http://www.st.com/web/en/catalog/sense_power/FM1965/SC1037

OMG this link is amazing, is not that exactly what I need for my alternator control?

VN750PS-E is available in qty 1 from some suppliers I've never used - http://octopart.com/partsearch#!?q=VN750PS-E

VN750 Series Single 6 A 36 V 60 mOhm Surface Mount High-Side Driver - SOIC-8

and VN5160S-E is available from everywhere including digikey http://octopart.com/partsearch#!?q=VN5160S-E

if these use the same pinout on SO-8 that would be really nice.

[kb1gtt]

As noted above and here http://rusefi.com/forum/viewtopic.php?f=4&t=286&start=55 This is the list I think we should consider for a thermally protected P-MOSFET, apparently called a Hight Side Switch (HSS) by ST.

For above 1A I get concerned with using SO-8 packages as the heat will be hard to get out of the chip and it will be hard to uphold the rated amp of that chip. Much like the hi/low is only able to do 1/8 amps when it's technically rated for 2A. The thermal properties become important when you go higher in current.

$4.20 2ch http://octopart.com/vnd5e025aktr-e-stmicroelectronics-17417297 also has eval board http://octopart.com/ev-vnd5e025ak-stmicroelectronics-25737927 for $20
$3.25 2ch http://octopart.com/vnd7020ajtr-e-stmicroelectronics-30039249
$3.18 2ch http://octopart.com/vnd5e025mktr-e-stmicroelectronics-18186948
$3.95 2ch http://octopart.com/vnd5e025bktr-e-stmicroelectronics-21218838
$2.38 1ch http://octopart.com/vn7020ajtr-e-stmicroelectronics-31115209
$2.99 1ch http://octopart.com/vnd7030ajtr-e-stmicroelectronics-30039251

My favorite is VND5E025AK-E, as it has 3 potential drop in replacements so it should be good for long term and it has an off the shelf eval board so faster development cycle. Then my next runner up is VND7020AJ-E because it's cheapest per channel.

Code: Select all

$/1	  PT#	        ch  volts amps mohm Vmax '$/ch	pkg           notes
$3.25	VND7020AJ-E	 2	28	63	22	40	$1.63	PowerSSO-16	
$2.38	VN7020AJ-E	  1	28	63	20	40	$2.38	PowerSSO-16	
$2.99	VND703AJTR-E	1	28	56	31	40	$2.99	PowerSSO-16	
$3.95	VND5E025BK-E	2	28	60	25	41	$1.98	PowerSSO-24	
$4.20	VND5E025AK-E	2	28	60	25	41	$2.10	PowerSSO-24	Eval $20
$4.20	VND5E025MK-E	2	28	60	25	41	$2.10	PowerSSO-24

[AndreyB]

My favorite is VND5E025AK-E, as it has 3 potential drop in replacements so it should be good for long term

You are the boss. Need a board

2nd channel would be useful if someone needs VTEC with constant 2A current.

[AndreyB]

You are the boss. Need a board

just placed an order

[AndreyB]

PowerSSO-24 board has arrived. Need pin#1 marker and the larger heatsink vias would be easier to work with, but the main thing is that everything aligns right.

I do not have the diodes yet, how essensial are they to try the board?

[kb1gtt]

The diode is critical, needed for fly-back topology.

Added reminder in knownissues.txt about the pin1 indicator.

[AndreyB]

Looks like the hardware side of Dodge Neon high-side field control wire is taken care of now, next step the software.

Check the video - notice how voltage changes bit by bit while I am increasing PWM duty cycle, and then it suddenly runs away. Looks like PID would be needed?

[video][/video]

[kb1gtt]

Very nice, very nice. In terms of software you probably want to filter the analog signals some, to smooth out that needle. The PID will not like the noise either.

How hot did the chip get?

[AndreyB]

How hot did the chip get?

I think it was slightly warmer then the rest of the engine bay in that area.

[AndreyB]

Now configurable via Tuner Studio

[kb1gtt]

Great and thanks for that offset thing. One request, can you put a note on it some where that notes output % duty = Pterm + Iterm + Dterm + offset%

Hmmmm, it might also be good to also post the general algo some where on that screen. Such that those of us who want to know the details of what kind of PID we are dealing with (mostly how the I term is handled) it would help show that with out requiring jumping into the code.

As a note to others who might come across this thread. The offset is for a kind of manual tuning. For example, if the P, I and D set to 0, the output will equal the offset% which means manual control. You can then manually adjust the output duty until you reach 14V. At this point you know what the PID loop should be adjusting too, which can be handy when tuning. AKA if you know where you should be you'll have some insight to see if you are getting there correctly.

[gptech2444]

What was the end result of the alternator control?

I ended up getting a westfield alternator board for my 2003 MX5.

[AndreyB]

It just works

[gptech2444]

It just works

Ok, but where and how do I get hold of this circuit? And where is it documented to set it up?

edit

I've ordered the IC from Mouser, was $4.73aud

[gptech2444]

Is it as simple as this? Where do you place the diode for flyback?

[AndreyB]

Is not miata alternator controlled by just logic level pwm signal? I believe my 2001 miata controls the alternator with only frankenso as is.

On dodge neon yes we used that large mosfet with the small addon board. I need to document better when i get home on Saturday

[gptech2444]

Is not miata alternator controlled by just logic level pwm signal? I believe my 2001 miata controls the alternator with only frankenso as is.

On dodge neon yes we used that large mosfet with the small addon board. I need to document better when i get home on Saturday

Yes, you are correct, I thought we still needed some external chip before the alternator.
https://rusefi.com/wiki/index.php?title=Vehicle:Mazda_Miata_1999