PWM-PSU Voltage Adjustment

[Kerrowman]

Hi all,

I have a PWM power supply that has a nominal output voltage of 12V and a current capacity of 25A continuous. While there is about 0.5V adjustment on the PSU itself, I need to be able to adjust this output to supply 7-12V so I need to drop a few volts after the output. I will lose 1-2V due to a current flow of 10A or more but still need some further adjustment.

I attach my suggestion based on an approach I used with a lower powered supply that used an MJE3055 and so am asking:

1. Would this arrangement work with the high power MOSFET (acting as a sort of valve like on a hose pipe)

2. Is this MOSFET the most efficient device to use as I've heard that it can waste a lot of heat if not fully switched on?

Thanks

 

[HarryA]

It may work but as you say it is not  very efficient. You can stack voltage regulators to get the current you need if you wish to roll our own. See the 4-20 volt 20 ampere circuit here:

https://www.eleccircuit.com/high-power-supply-regulater-0-30v-20a-by-lm338/

Also you can buy a motor controller that may work for you - rated 30 amperes. With luck you may get one that works - read the reviews.

https://www.amazon.com/RioRand-7-80V-Motor-Controller-Switch/dp/B071NQ5G71/ref=sr_1_4?dchild=1&keywords=adjustable+voltage+regulator&qid=1610728630&sr=8-4

Now I see " it is a DC Motor Speed Controller,Not a Voltage Controller "

 

[Kerrowman]

Hi Harry,

Thanks for the suggestions. The voltage regulator seems a good idea but I don't need the whole power supply but just an add on set up to adjust down the 12V nominal output I already have. Would I need then item 4 in the list? If so how is the voltage energy dissipated and how efficient is it overall?

The motor controller uses a PWM approach it seems (12kHz) and I already have a PWM in my setup.

Jules

[HarryA]

You do  not need the bridge diode nor the 4700 mfd capacitor but keep the 0.1 mfd as it is typically in voltage regulators of that type.  You may need heat sinks and heat sink compound; look up "lm388 pdf" files for more information.

This illustration is set up for 5 volts:

[Kerrowman]

Thanks but my particular need is to take 12V, 25A max and make adjustable from 6-12V and a max current of 25A so would the attached from the previous link be suitable? I find the text doesn't make easy reading and is hard to understand.

J

[HarryA]

The max for the LM388 is 12 amperes:

Look at the the Buck Converter  here; one reviewer is pushing it to 25 amperes with a fan/blower cooling.

https://www.amazon.com/Anmbest-Converter-Adjustable-Regulator-Protection/dp/B07R832BRX/ref=sr_1_41?crid=3ETDJKMX4K9ZO&dchild=1&keywords=dc+12v+converter+step+down+module&qid=1610919636&sprefix=dc+12v+step+down+converter+12v%2Caps%2C407&sr=8-41

I should get a kick-back from Amazon.com for all the times I reference them.😉

 

[Kerrowman]

Hi Harry,

I have used this buck converter before but its never supplied even 20A consistently.

In the first link you gave it talks about using the LM338 in parallel, as in the attached. Can I not stack 3 of them in parallel to do the job? If so I will assemble a circuit and run it by you?

Thanks

 

 

[Kerrowman]

Hi Harry,

As suggested above I have drawn my own version of the power supply on the first link you gave using 5 LM338s. I could try and simulate this in LTSpice but that will take me a while. Any thoughts on whether it would do the job - probably better than the Buck converter?

Jules

[HarryA]

Bear in mind the headroom for the LM388 is 3 volts. The headroom voltage refers to the actual voltage drop across the regulator that must occur during operation. So that means that the most you can get out of the above circuit is 9 volts. Minus the drop across the 0.3 ohm resistor.

You maybe better off using your one transistor circuit if you do not need the power/current for extended periods. If so perhaps there is a spice model for the IRFB4110? You could run it in the LTspice simulator and/or I could run it in the TINA-TI simulator. Also there may be a similar transistor that there is a spice model for; for simulation if you decide to go that way.

[Kerrowman]

Hi Harry,

 

The good point you raise about voltage headroom is presumably why 18V is supplied by the transformer in the full power supply circuit shown on that first link. However, the note nearby says that it has not been built so is it a bit of a gamble that it will work as indicated?

 

The FET approach to dropping volts will work for the time periods I’m using for certain tests but for long term running I will likely need something more efficient.

 

For the FET approach, in Spice simulation the FDS6699S mosfet works well but, given the potential for high heat dissipation, I’m not sure how a heat sink can fit to the SMD SO8 package in a way that I can use.

 

Perhaps the answer is to buy or build something originally designed to deliver 8-12V at 30A max rather than add on extra devices to adjust an output.

 

J

[JohnKlein94]

On 1/19/2021 at 11:31 AM, Kerrowman said:

Hi Harry,

 

The good point you raise about voltage headroom is presumably why 18V is supplied by the transformer in the full power supply circuit shown on that first link. However, the note nearby says that it has not been built so is it a bit of a gamble that it will work as indicated?

 

The FET approach to dropping volts will work for the time periods I’m using for certain tests but for long term running I will likely need something more efficient.I think you should still play Vulkan Vegas  https://5vulkanvegas.com/ro

 

For the FET approach, in Spice simulation the FDS6699S mosfet works well but, given the potential for high heat dissipation, I’m not sure how a heat sink can fit to the SMD SO8 package in a way that I can use.

 

Perhaps the answer is to buy or build something originally designed to deliver 8-12V at 30A max rather than add on extra devices to adjust an output.

 

I completely agree with your opinion.