[PIC] Car Cooling System PSU Design Related

[BradtheRad]

purpose of D6 diode

The batteries produce 24V total. The diodes prevent backfeed into neighboring 12V circuitry.

If you wish to ground reference the batteries, then a different charging arrangement is needed.

Since you're pressed for time, it may be easier to make a 12V system.

 

[Easy peasy]

Are you talking about D1, D3, ? they could be replaced with P -fets, 15V zener gate to source and then 100k to gnd, then the internal fet turns on to assist the internal diode, for 6 milli-ohm at 10A the losses are now 0.6 watt, manageable on a pcb with 1 square inch or larger ...

- - - Updated - - -

for 28 amps, you will need to parallel mosfets to get approx 1 watt overall ( watts = I^2 R ) so, 1.25 milli-ohm, so 4 x 6 milli-ohm in // giving 0.3 watts each on a common pcb heatsink area ...

 

[baileychic]

This is the MOSFET I am using for all tasks in the circuit like Battery charging, Peltier driving and Fan driving.

The Battery charging and Peltier voltages are max +15V and Fan voltage is +12V @3A.

The Battery charging current is max 4A.
The Peltier current is max 20A.

I have considered 30A as max for calculation of power dissipated in MOSFET.

So, Pd(mosfet) = I * I * Rds(on) = 30A * 30A * 2.1m Ohms = 1.89W. I am using small TO220 heatsinks for them.

For the diodes I am using this one. Vf = 0.51V and so Pd(diode) = Vf * I = 0.51V * 30A = 15.3W. I am using large heatsinks for them.

I added one more diode to block flow of Solar Panel voltage directly into Battery.

I have designed the lo-power PCB layout. Here are its screenshots.

I still have to design the Hi-power PCB layout.



Edit:

I will try to replace the 3 power diodes with P-fets, zener, 100k combination. That will bring down the power diddipation in the diodes to less than 5W per diode for 30A. Thanks for the suggestion.

- - - Updated - - -

Please provide a link for the 6m Ohms Rds(on) P-fet.

- - - Updated - - -

Can I use this P-fet with Rds(on) = 5m Ohms?

If yes, please mention the Zener value for each diode circuit and how to connect them.

https://www.mouser.in/ProductDetail...=sGAEpiMZZMshyDBzk1/Wi5vJ6PzONUYjMwFU9cVWIB0=

This will reduce the Pd(p-fet) very much.

Pd(pfet) = I * I * Rds(on) = 30A * 30A * 5m Ohms = 4.5W per P-fet.

I can then use a small TO-220 heatsinks for them and this saves the cost of big heatsinks used for the diodes.

- - - Updated - - -

Can this P-fet be used? It has max D to S voltage of 30V.

SQP50P03-07


Why 3x Parallel P-fet sneeded to replace 1 power diode? P-fets's internal diode cannot conduct 30A current?

D1 max current is 4A as it only used for Battery charging for Solar Panel.

D8 is also 5A as it is in Battery charging circuit.

D8 is 30A as it is in load circuit that is 30A from Battery goes to DC-DC Converter through D8.

 

[baileychic]

@ Easy Peasy

In post #42 were you telling me to use this circuit, of course with below P-fet?

See, in the 2nd image you can see that there is no reverse flow of voltage/current from drain to source that is from +30V to +28V side. How much reverse voltage can internal diode withstand before breakdown and shorting drain to source?

Can I also use this P-fet?

https://www.mouser.in/ProductDetail...=sGAEpiMZZMshyDBzk1/Wi4lI00SCTc047yeVAvMQSdk=

Mention which of the above two mosfets will be a better choice. I prefer the above one in the link.

- - - Updated - - -

Are you talking about D1, D3, ? they could be replaced with P -fets, 15V zener gate to source and then 100k to gnd, then the internal fet turns on to assist the internal diode, for 6 milli-ohm at 10A the

for 28 amps, you will need to parallel mosfets to get approx 1 watt overall ( watts = I^2 R ) so, 1.25 milli-ohm, so 4 x 6 milli-ohm in // giving 0.3 watts each on a common pcb heatsink area ...

I don't understand this. Why P-fets in parallel for each diode? Check this post. I am using a P-fet with 2.9m Ohms Rds(on) and it can handle 40V.

So, at 30A current the Pd per P-fet is 30A * 30A * 2.9m Ohms = 2.61W. The Pd is approximately that of a 3W resistor. The P-fet can handle max 50A and my current will not exceed 28A.

So, now the question is can I use one P-fet to replace each diode?

 

[Easy peasy]

the internal mosfet diode is rated at the same voltage as the fet itself - as it is integral to the fet

paralleling p-fets is used when the total dissipation must be low - usually needs to be low for SMD, i.e. if you need <= 0.5W per device

 

[baileychic]

Okay, thanks. Then I will omit parallel P-fets per diode replacement and use just one single P-fet per diode with small TO220 heatsink.

At 30A current the Pd per P-fet is 30A * 30A * 2.9m Ohms = 2.61W.

Only one P-fet will conduct max 28A in my design. The other 2 P-fets in the Solar Panel + battery charging circuits conduct max 4A because battery charging current will never exceed 4A in my design.

So, actually I need small heatsink only for the 28A conducting P-fet but I will put for all the 3.

All N-Mosfets in my design are PSMN2R0-30PL

All P-Mosfets in my design are IPB120P04P4L-03

https://www.mouser.in/datasheet/2/196/Infineon-IPP_B_I120P04P4L_03-DataSheet-v01_02-EN-1227285.pdf

This is my latest circuit. Is this circuit okay? Just check the 3 P-fet circuits. Others parts of the cirtcuit are okay.

SPV max is 30V.

BATV max is 29V or 30V while charging.

DC-DC Converter output is +15V fixed. It might be 15.5V max.

Q5 is used to block flow of battery voltage into Solar Panel. Q5 should conduct into F1/CS1- that is DC-DC Converter input.

Q7 is to block Solar Panel voltage flowing into Battery but it should conduct into CS1+

Q6 is to block flow of battery voltage into DC-DC Converter output which is +15V. But it should conduct +15V (+14.7V) charging voltage into battery.

 

[Easy peasy]

2.61 watts is simplistic, as soon as the fet heats, the Rds-on goes up, the watts go up, the temp goes up, the Rds-on goes up - see the spiral happening....?

a small heatsink will not cut it - less than 1 W may well be manageable - hence P-fets in parallel ...

 

[baileychic]

Okay, I will parallel 2x P-fets for the 30A line that is Q6. That will divide the 30A max to 15A per P-fet and so the new Pd will be

15A * 15A * 0.0031 Ohms = 0.6975W = 697.5mW.

I will then put small heatsinks for each parallel P-fet.

Regarding Zeners can I use a single Zener and 100k for two P-fets in Parallel?

Did you look at my latest circuit in my previous post? Is that okay, that is the P-fet connections?

 

[Easy peasy]

I did not look at your ckt in detail - unfortunately advice is worth exactly what you pay for it - any free advice should be taken with some large grains of salt ...

also - if I had a hurried look and said it was OK - you might take that as gospel - when in fact I may have made a mistake ...

 

[baileychic]

If anybody can answer about my P-fet connections then I appreciate it else I will take the risk of using it. It is for a student project. At max the student will fail and has to take his final year student project next year because he is not paying me additional to get advices.

 

[baileychic]

Can anybody give confirmation about the P-fet circuit?

This is the simulation result with max SPV and BATV. +15V output of DC-DC Converter is fixed. DC-DC Converter output will be present only if SPV or BATV or both are connected to it.

The simulation result show the current flow path.

 

[ZASto]

If anybody can answer about my P-fet connections then I appreciate it else I will take the risk of using it. It is for a student project. At max the student will fail and has to take his final year student project next year because he is not paying me additional to get advices.

OK, how much are YOU paying this community for solving problems that are out of your knowledge.
From the beginning I was suspicious on your "projects", are they commercial or home works. Now it's obvious.

Students that have this kind of projects should be here on forum, not you.

 

[baileychic]

I am not going to pay anything as the student is not paying additionally for that. I very well know how to solve the Mosfet equations as they are just basic Mosfet biasing equations. My issue is because the red arrows in the simulation doesn't look fine because reverse voltages/currents are flowing. I don't know if that is just the Proteus issue.

With diodes the voltages look perfect.

 

[ZASto]

You did not understand what I wrote?
You are asking help on the forum, because you, obviously, lack a little bit of knowledge, to be able to get some money from poor students that should not skip classes.
If you have some problems with Proteus, there is a support forum.

 

[baileychic]

Yes, I know that labcenter has a forum but they are eon holidays till 6th Jan. I have submitted my simulation file to them but I can't wait till 6th Jan because the PCB layout is pending and the student has to get the PCB manufactured and assembled and it takes some time. The student already got the deadline extended once.

 

[Easy peasy]

Crikey, in any event you can simulate in LTspice (or other ) if it works in sim it is likely to work in the real world ( layout aside )

 

[baileychic]

I modified the circuit a lot. I totally eliminated the old Battery charging circuit and the 3 power diodes.

I have used 40V max input 28.8V adjustable output @ 5A to charge the batteries using a simple switcher. The Solar Panel Voltage may go up to 35V max and so the switcher can handle it. Battery charging current for a 12V x 2 40Ah batteries will be 4A max. So, the switcher can provide sufficient charging current.

The 3 nasty power eating power diodes and 2 Mosfets in the Battery charging circuit has been totally eliminated.

Now the 2 Batteries charge in series.

14.4V charging voltage per battery in series = 28.8V total charging series voltage.

Is this okay now?

The D1 diode at the switcher output blocks the reverse flow of battery voltage into switcher output.

When battery is fully charged that is 28.8V series voltage then no current flows through the switcher.

SPV is totally isolated from Battery voltage.



Should I add a N-Mosfet Logic Level between GND pin of Switcher and GND and control it with a PWM to control battery charging? Is that necessary?

Or is it better to feed Logic Level PWM into ON/OFF pin of switcher to control the charging?

 

[Easy peasy]

say the eff is 95% ( data sheet best case for 35Vin, 28.8Vout ) then at 144 watts out, you need to dissipate 7.2 watts and keep the case temp below 70degC

this will require a fair sized heatsink.

What happens when the Vin from the solar panel, goes below 28V? you will not get maximal charging to the batteries ...

 

[baileychic]

Then shall I use two switchers whose output is fixed to 14.4V and feed each 14.4V to one Battery and so the 2 switchers will be charging the batteries in parallel. Will this work? Then SPV can go below 20V.

 

[Easy peasy]

that will be better assuming useful current can be got from the PV panel in the 15 - 28V range ...