Questions about a linear power supply design

Hi @kd9wzz,

   You will appreciate these are just comments from a quick search on the web .. I haven't built the supply Bill (@dronebot-workshop) described, or used the parts I am going to mention.

Linear regulators for 8+Amps are rather unusual, but I spotted Analog Devices LT1581 is rated to 10 Amps, so might be suitable. I haven't read the data sheet,

https://www.analog.com/media/en/technical-documentation/data-sheets/158125fa.pdf

so please take care to check it out thoroughly, to see if it meets your needs. Things to consider, include:

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

Heat dissipation ... linear regulators are essentially 'magic resistors' ..   

  To a first approximation:

  power dissipation = (Input V - Output V) * Output current

   e.g Vin = 24V Vout = 5V Output current = 8 A (roughly based on Bill's 24Vdc example)

     ==> (24 - 5) * 8 = 19 * 8 = 152W

But the thermal resistance from power output transistor is 2.5 DegC/W, so even if it was bolted to a perfect, infinitely large heatsink, just 40W {40 * 2.5 = 100 degC increase} would result in junction temperature of 125 degC, when ambient  temperature is 25 degC

Of course, if the input voltage is much lower, then this figure falls. The LT1581 data sheet gives examples of the full calculation involved.

It isn't clear what you expect your supply to achieve, but at this high current, the usage is limited to cases in which the input voltage never exceeds the output voltage by more than a few volts. This is quite different to the aims of Bill's design, which was for a flexible voltage output, but much more modest current.

-----------

Single chip regulators do not generally provide adjustable current limiting .. they may provide some internal 'self-preservation' limit, but for a high current regulator, that will allow a very high current to flow. It is possible to add further circuitry, to provide limiting, but I suggest you carefully evaluate the power dissipation question posed above, to see if your approach meets your needs, before worrying about this problem.

----------

If the 'battery' question means can I switch to a battery instead of the transformer-rectifier-smoothing capacitor power source, then in general the answer will probably be yes, but bear in mind the switch must be suitably rated, break-before-make action, and protection against excessive current demand must be addressed.

Also, some care when switching is needed if the load has a large capacitance that could result in the output voltage exceeding the input voltage, when switching. I expect that is discussed in the data sheet.

---------

Note the resistor values for voltage adjustment will be different for this regulator .. again the data sheet will advise.

----------

So, sorry this is not a simple or short or even comprehensive reply, but I hope it gives some things to check out, and maybe ask more specific questions.

In particular, please specify not only the output current maximum, but also the input and output voltages to the regulator, to give some guidance as to how much heat will be dissipated.

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

Of course, switch mode power supplies are much more efficient when the input and output voltages are significantly different, but they bring different challenges, particularly if you wish to design and build your own. Alternately, there are 'modules' avalible from the usual sources that might meet your requirements.

Best wishes and good luck, Dave

My main purpose is learning, so this is all really good information, so I appreciate it. I intend to follow up here and make sure I'm not trying to do something that will result in... tragedy. Heh.

Second, specifically want a linear supply for the purpose of radio. Linear supplies are quite a bit quieter in terms of creating a bunch of RF.

The convenience goal would be two knobs for voltage (one of them selecting between several set voltages), one knob for current limiting, a toggle for switching between the battery and power supply, and a switch to turn it off all together along with a place on the back to plug in to thew all and also plug into my load. Then I can power either of my radios. My main radio takes 12v and can consume up to 8 amps when transmitting, although I have empirical data that it really only consumes 6A. Then my other radio uses 9v, and while I don't know how much it will draw, it won't be a lot because it's only a receiver.

The video I mentioned is a REALLY solid start on all of this. I could actually do away with the continuous adjustment for the voltage and just have two selections for 9 and 12, but future proofing isn't bad either (and that part of the circuit is well demonstrated in the video so there is no reason not to use it that I can see.

It's getting a little late for me, but if I get some time tomorrow I will get a schematic together and do a little more research.

That whole thing of taking an idea from crackpot plan to actual, workable solution is a always exciting 🙂

Hi @kd9wzz,

  I think I have given you plenty to ponder and puzzle, so I'll just throw an extra point. I already guessed why you preferred linear, but the heat dissipation can be a real pain!

One trick is to have both switch-mode and linear .. the switch mode does the 'big volt drop', to about 1.5 Volts (say) above the desired value, then a linear trims off the remaining 1.5V, hopefully trimming most of the switch noise at the same time.

Then dissipation is 1.5 V * 8 Amps = 12 W ... a lot more reasonable to handle than 150W!

Of course some luck and care is needed to make sure the noise doesn't go straight through, but it is widely used.

Best wishes, Dave