diy solar

diy solar

How much Amperage is compromised by a 4mV drop.

Even if we assume (must add I think that's not a good idea) the SCC has both negatives tied, what will happen when the BMS disconnect the battery?

Regarding thermal protection you can add some but if they are correctly sized they will never overheat thanks to the PVs being unable to supply more current than their Isc. So given it should never trip a thermal fuse can be used even if it's one use only.
 
My reason for suggesting thermal protection was not excessive current but rather excessive voltage drop. If the gate isn't turned on hard enough the MOSFET burns more power, overheats. Alternatively if it gets too hot then a small part of its active area can go into runaway and burn up. Pulling gate voltage to zero at some temperature may prevent that.

Mouse builds a nest, insulating the heatsink?



I found a birds nest on my truck's exhaust manifold. Might have looked like a safe place, but since I also saw a rat creep off it wasn't really.

Isolated supply - you've already got one, PV.
Just connect 32V (what is worst-case Voc, maybe 50V?) to a megohm resistor in series with a zener. That'll draw 50 uA, 2.5 mW drain.
Put a capacitor across the zener. Use a SPDT switch to either pull gate up to zener voltage or clamp it to GND.
 
Ok I see, not a bad idea against mices and co ;)

Yea, I thought about the zener solution but it wasn't compatible with an optocoupler; a relay will work nicely with it however (but relays can be power hungry, that's why I avoided them here) :)
 
Maybe you don't.
Are those two negatives of SCC actually separate nets, which might be at different voltage? Or is it just a wire straight through?
I'm trying to keep a common ground reference for all components, because multiple modular functions are expected - which would go haywire if voltage reference is not identical. However having said that I'm no expert either on circuit design so would need support to filter out blunder (if any) on the way lest - all goes up in smoke later :LOL:.

If straight through so your MOSFETs are referenced to battery negative, then you should be able to drive them with a non-isolated signal referenced to battery negative.
That's the idea, but I liked your brilliant suggestion of pulling +ive voltage from PV panels for driver as this would ensure Mosfets are closed only when the sun is out and powerful enough to begin charging the system. Can we put 2 Zener in parallel for redundancy?

Review data sheet, determine what gate voltage is necessary to turn MOSFET on hard enough for the current flow it will see, minimizing power dissipation.
VGS(th) is 4V, going by most circuits 10-12V should be good enough to put them into completely active state.

To make it safer, consider a thermostat, thermistor, PTC fuse, or other way to turn gate off if MOSFETs get too hot. It should snap off and remain off with hysteresis.
Absolutely, working on it.
 
Even if we assume (must add I think that's not a good idea) the SCC has both negatives tied, what will happen when the BMS disconnect the battery?
I think the schematic is not so clear, would try to get a better one soon.
 
That's the idea, but I liked your brilliant suggestion of pulling +ive voltage from PV panels for driver as this would ensure Mosfets are closed only when the sun is out and powerful enough to begin charging the system. Can we put 2 Zener in parallel for redundancy?

But then I thought some more, and it seems to me that after MOSFET opens the negative, it breaks the negative connection to zener.
That's the problem when I draw schematics in my head instead of with pencil or pixels.
Gotta consider all intended modes of operation, and all faults too.
Make sure you figure out how to wire it, because I got confused.
 
Here's a better schematic of things we discussed yet, however more modules are in process.
This would help translate things better I suppose.
Please have a look and let know what you think.

1597202394092.png
Regards
 

Attachments

  • Schematic.pdf
    133.9 KB · Views: 1
Last edited:
1k resistor in series with 12V zener - what's that supposed to do?
And I hope it's at least a 1 watt resistor.

Also, I see a body diode in the MOSFET symbol. With the MOSFET oriented the way it is, what is it supposed to do?
 
1k resistor in series with 12V zener - what's that supposed to do?
And I hope it's at least a 1 watt resistor.
I believe we need at least 100mA current to drive those Power Mosfets - Each takes about 250nC charge to get active.
But then we're capped by that Zener. So may be with 500mW limits, it is supposed to be 10K (For a parallel config) - would hardly drive 10mA though.
Any comments on Parallel Zeners - are they suggestible.

1597211855101.png

Also, I see a body diode in the MOSFET symbol. With the MOSFET oriented the way it is, what is it supposed to do?
That is how Mosfets are modeled in component library of the software I'm using.
Body diode is not an explicit part of the Mosfet as such, but denotes resistance to reverse leakage current in the mosfet I suppose.
Checking it more on the internet found following description:
"Even though the body diode in MOSFETs is often drawn as a physical diode next to the MOSFET, it's not actually a discrete building block of a MOSFET. It is a 'parasitic' element that exists inherently because of the silicon build-up of a MOSFET. There are no MOSFETs without body diodes, it's a fact of life.
That also means that this body diode isn't actually a purpose-made diode. It doesn't quite behave like either a signal or power diode and it has fairly poorly defined characteristics. However, it also means that it doesn't have reverse leakage like a traditional power diode. It's part of the entire package, so the datasheet usually defines an overall reverse leakage figure."


Well this board has 2 modules as of now:
A) First one detects voltage buildup at Panels and activates Mosfets to delivery charge to SCC.
B) SCC's output towards Battery (Charging) is regulated with Mosfets, using control signal from micro processor.
This is to control charge limit for the batteries. Switching Off SCC Input from panels could also have done it. Which one does looks better?

Thanks for your sanity check and response.
 
Last edited:
I believe we need at least 100mA current to drive those Power Mosfets - Each takes about 250nC charge to get active.
So for 45v/0.01A it was supposed to be 10K (For a parallel config). Typo. Any comments on Parallel Zeners - are they suggestible.

View attachment 19679


That is how Mosfets are modeled in component library of the software I'm using.
Body diode is not an explicit part of the Mosfet as such, but denotes resistance to reverse leakage current in the mosfet I suppose.
Checking it more on the internet found following description:
"Even though the body diode in MOSFETs is often drawn as a physical diode next to the MOSFET, it's not actually a discrete building block of a MOSFET. It is a 'parasitic' element that exists inherently because of the silicon build-up of a MOSFET. There are no MOSFETs without body diodes, it's a fact of life.
That also means that this body diode isn't actually a purpose-made diode. It doesn't quite behave like either a signal or power diode and it has fairly poorly defined characteristics. However, it also means that it doesn't have reverse leakage like a traditional power diode. It's part of the entire package, so the datasheet usually defines an overall reverse leakage figure."


Well this board has 2 modules as of now:
A) First one detects voltage buildup at Panels and activates Mosfets to delivery charge to SCC.
B) SCC's output towards Battery (Charging) is regulated with Mosfets, using control signal from micro processor.
This is to control charge limit for the batteries. Switching Off SCC Input from panels could also have done it. Which one does looks better?

Thanks for your sanity check and response.

My point was the body diode appears to be in the conducting direction. I think MOSFET should be oriented so body diode is reverse-biased. Check data sheet for relative voltages that belong on each terminal.

As for zener in series with resistor, the zener drops 12V and the resistor sees 45 - 12 = 33V or something like that.
To make a regulator, put the zener in parallel with the load. That way the load gets 12V.

With the circuit set up in a way that could work, as the sun comes up it would turn the MOSFET on. If slow, MOSFET could be in linear region. It ought to have a circuit that snaps it from off to fully on. But really, PV puts out near Voc while it still has very little current capability.

MOSFETs don't need current to hold on, just for the transition, charging gate capacitance. So I suggested 1 Meg ohm charging a cap. When a switch closes, cap supplies an instantaneous high current needed to switch.
 
Yep, all mosfets but the 2N7000 have their drain and source reversed on your schematic.

And that's not how you use zeners for what you want to do, check this search results for more info: https://www.google.com/search?q=zener+shunt+regulator ;)

And there should be a resistor between the gates and sources of the IRF1404 to turn them off when the 2N7000 is open.

And no series resistor on the gates is a bad idea.

I don't see what the purpose of the mosfets on the PV side, they are simply off when the PV doesn't provide power and on when it does, what's the purpose of that?
 
Yep, all mosfets but the 2N7000 have their drain and source reversed on your schematic.

And that's not how you use zeners for what you want to do, check this search results for more info: https://www.google.com/search?q=zener+shunt+regulator ;)

And there should be a resistor between the gates and sources of the IRF1404 to turn them off when the 2N7000 is open.

And no series resistor on the gates is a bad idea.

I don't see what the purpose of the mosfets on the PV side, they are simply off when the PV doesn't provide power and on when it does, what's the purpose of that?

I think IFR1404 is oriented correctly.

Any time +5V is present it draws current though 1.2k, 4.2 mA and 21 mW
Voltage on gate of IFR1404 will be pulled up to (5V - Vt). Vt of 2N7000 is 0.8V min, 3.0V max so you get 2.0V to 4.2V on gate.
Vt of IFR2404 is 2V min, 4V max, no guarantee it is on. Even if it is, will be in linear region, drops voltage and burns lots of power.
I was looking for a graph of Rds(on) vs. Vgs but don't see that. Use: Fig 3. Typical Transfer Characteristics
Vgs(max) is 20V. You'd be better off using a P-channel FET to pull it all the way up to 12V.
That should never reach 20V, but could add a clamping component to eat any inductive spikes for complete robustness (not really needed for DIY, good for a commercial product.)

Instead of P-channel you could have gate of IFR2404 pulled up to 12V and use your 2N7000 N-channel FET to ground it.
Much bigger resistor plus cap would let gate transition quickly, but reduce steady-state power dissipation.

Oh by the way, IFR2404 has RDSon of 0.004 ohms, but if it ever overheats and fails the resistance would be much higher and it would dissipate a lot of power. Then it gets really hot. Make sure nothing catastrophic could happen beyond your electronics in that situation.

" That is how Mosfets are modeled in component library of the software I'm using."
Did you run SPICE? What waveforms did it give?

Now that you're switching battery side of charge controller, is that one which says, "Do not connect PV before battery"?

Stick with us, and we'll give you an EE education more useful than from a university. Minus the calculus and device physics.
 
Last edited:
:ROFLMAO::ROFLMAO::ROFLMAO: Looks like those components are few but the mess they have created is the exponential square of probability for incorrect combinations possible in the entire universe. Literally can't stop laughing at the critique's review analysis:LOL:
I value your time gentlemen and was just about to call it quits before I read your last line "Stick with us, and we'll give you an EE education more useful than from a university. Minus the calculus and device physics."
If you are ready to teach, I am here to learn.
Let me come up with a better version of the board(y)

And Thank You!
 
Here's a simple SSR I have operational in my VW Westfalia.


I have a transistor driving 6 fet gates through 10 ohm resistors. Very simple circuit.
 
I think IFR1404 is oriented correctly.

If the SCC is the power source then nop, they are in the wrong direction. If you consider the battery is the power source then, yes, but usually the SCC charge the battery, not the reverse... ^^
Voltage on gate of IFR1404 will be pulled up to (5V - Vt). Vt of 2N7000 is 0.8V min, 3.0V max so you get 2.0V to 4.2V on gate.
https://www.onsemi.com/pub/Collateral/2N7000-D.PDF Vt of IFR2404 is 2V min, 4V max, no guarantee it is on. Even if it is, will be in linear region, drops voltage and burns lots of power.
https://www.infineon.com/dgdl/irf1404pbf.pdf?fileId=5546d462533600a4015355dae92618b0

Uh??? they're not BJTs, the Vth isn't applied to the drain/source path in any way. If the 2N7000 is turned on the IFR 1404 will have the full 12 V on their gate ;)
 
If the SCC is the power source then nop, they are in the wrong direction. If you consider the battery is the power source then, yes, but usually the SCC charge the battery, not the reverse... ^^


Uh??? they're not BJTs, the Vth isn't applied to the drain/source path in any way. If the 2N7000 is turned on the IFR 1404 will have the full 12 V on their gate ;)

Hmm. If Charge controller is trying to put power into the battery, charge controller "battery negative" terminal is more negative than battery negative terminal. I think that pushes current though the body diode.

If no PV is available and battery is powering charge controller (which it has to do for charge controller is going to recognize the battery), then battery negative terminal is more negative than charge controller "battery negative" terminal, and MOSFET is capable of isolating charge controller, preventing drain into it.

What was this circuit supposed to do again? Would a bidirectional MOSFET solid state relay be needed?


Looks to me like 12V is applied to drain of 2N7000. 5V is applied to gate.
Vgs(th) is between 0.8V and 3.0V, so 2N7000 is only on if its source is below 4.2V, possibly only if below 2.0V.
Source follows voltage on gate, hanging below it by Vt. At least, gate gets pulled up to that. Nothing to pull it down so drifts when (previously 5V) signal on gate is pulled below source. Other than that, there's a bit of leakage and capacitive coupling that boots voltages up or down.
So that's all the voltage you get on gate of IFR 1404. May or may ot turn it on at all, definitely not hard enough.

(Excuse us @Bhupinder while @BiduleOhm and @Hedges compete to see who is the most confused.)
 
Hmm. If Charge controller is trying to put power into the battery, charge controller "battery negative" terminal is more negative than battery negative terminal. I think that pushes current though the body diode.

Yep, that's why I say they are reversed.

If no PV is available and battery is powering charge controller (which it has to do for charge controller is going to recognize the battery), then battery negative terminal is more negative than charge controller "battery negative" terminal, and MOSFET is capable of isolating charge controller, preventing drain into it.

Hum... maybe not a bug but a feature then... ^^

What was this circuit supposed to do again? Would a bidirectional MOSFET solid state relay be needed?

Not sure, at first it was a disconnect on the PV side... now I don't know what he has planned anymore.

ooks to me like 12V is applied to drain of 2N7000. 5V is applied to gate.
Vgs(th) is between 0.8V and 3.0V, so 2N7000 is only on if its source is below 4.2V, possibly only if below 2.0V.

Vth is referenced to the source, not the drain, so we don't even care about the voltage applied to the drain. His schematic is confusing (not following the highest potential at the top and signals left to right conventions) but if you redraw it the usual way it'll probably be clearer ;)

Source follows voltage on gate, hanging below it by Vt. At least, gate gets pulled up to that. Nothing to pull it down so drifts when (previously 5V) signal on gate is pulled below source. Other than that, there's a bit of leakage and capacitive coupling that boots voltages up or down.
So that's all the voltage you get on gate of IFR 1404. May or may ot turn it on at all, definitely not hard enough.

Hum no, voltage is applied between the gate and source so the source doesn't follow anything as it's considered the reference (0 V).

BTW I'm always talking steady state, ignoring dynamic things.

I'm 100 % positive the IFR1404 would receive the full 12 V on the gates (and turn-on if their direction is corrected from the schematic as they are reversed, a priori) if the 2N7000 is turned on.

Edit: wait a second, I'm wrong. The 2N7000 is floating, so not even sure what would happen when you put 5 V on it's gate (and I'm too lazy to think right now, you'd need to take into account dynamic stuff and capacities). Didn't saw that at first because of the weird schematic. Well, in any case it'll not work, at best it'll work unreliably.
 
Last edited:
Certainly 5V on gate of 2N7000, relative to source, would turn it on and could pull almost 0.4A according to curves in the data sheet.
But unless drawn as a battery or other voltages source referenced to source terminal of MOSFET, I take "+5V" in his schematic as referenced to GND.
And that would barely pull the source up a couple of volts, not to 12V.

I don't call it floating, just an N channel MOSFET used as a high-side switch (or source-follower? meaning source follows gate voltage, in this case). I've seen 'em and used 'em, and pointed out the voltage/speed needed for different applications. Charge pumps can sustain a voltage of 10V above source but area slow. Boot circuits are fast, good for chopping, but can't hold up forever.
 
Certainly 5V on gate of 2N7000, relative to source, would turn it on and could pull almost 0.4A according to curves in the data sheet.
But unless drawn as a battery or other voltages source referenced to source terminal of MOSFET, I take "+5V" in his schematic as referenced to GND.
And that would barely pull the source up a couple of volts, not to 12V.

Yea, of course. I made a mistake, see my edit ;)

don't call it floating, just an N channel MOSFET used as a high-side switch (or source-follower? meaning source follows gate voltage, in this case). I've seen 'em and used 'em, and pointed out the voltage/speed needed for different applications. Charge pumps can sustain a voltage of 10V above source but area slow. Boot circuits are fast, good for chopping, but can't hold up forever.

Yeah. But that's not going to work properly here.

The simplest would to use the 2N7000 as a low side switch with a pull-up on the gates of the IFR and inverted logic in software.
 
What was this circuit supposed to do again? Would a bidirectional MOSFET solid state relay be needed?

Originally circuit was suppose to Isolate PVs from SCC when they were not needed by the system.
This was bases on following requirements:
1) As a check to verify batteries were in place "before" any charge from PVs could hit SCC.
(Idea was to drive relay and then later SSR Mosfets from Battery source - that got forgotten somewhere down the lane - one of the BIG disadvantage
of not documenting things).
2) Safeguard SCC components from unnecessary stress arising out of bypassing current when not active.
3) Better control over Battery Charging Control logic.

But things got haywire as clearly visible and am apologetic for that.
So I am thinking of starting a new thread with an appropriate name to document and trace things better here on.
Current thread may be left to its original intent of finding Amperage degradation with change in IR conditions of conductors.

Would send the link to new thread and we shall continue there.
 
Last edited:
Back
Top