开云体育

ctrl + shift + ? for shortcuts
© 2025 开云体育
  1. Duncanampspsud
  2. Topics

Modelling PSUD2 into a voltage regulator

 

I have a design I want to explore to power a 5V device that consumes up to 3A on start up as the system boots up and consumes more power, and then settles and takes a steady state at 240mA
?
How is this best modelled to make sense of the war DC design before the regulator
?
Any ideas/experience?
?
Thanks
 

There's a couple of things I can think of.
?
First is to ensure that the voltage feeding the regulator doesn't drop below the dropout voltage. You'd have to check the data sheet, but say the dropout for your 5V regulator is 6V then ensure that it doesn't drop below that. You probably want to add safety margin for component tolerance, lower than normal mains voltage, etc.
?
Second thing is you can simulate the environment with a stepped load. This could be 3A for N seconds then change to 240mA after that.
?
Set up stepped load
?
  1. Right-click in the load box (outside of the resistor itself), select Change then select type of load as Constant current.
  2. Right-click on the constant current symbol click edit, then click the stepped load box
  3. Value 2 and After will light up, so you can key in value 1 of 3A, value 2 of 240mA and the delay between the two you would have to decide for yourself as you know best what the timings are
?
Regards,
Duncan
?
 

Duncan,
?
Thanks for this; for some reason I had to really play with the circuit it get it to produce reasonable voltages in both scenarios.

This is the regulator I was looking at https://www.ti.com/lit/ds/symlink/lm1084.pdf?ts=1721541195139&ref_url=https%253A%252F%252Fwww.mouser.com%252F
?
The start up current is 2.5>3A and running 240-260mA so this is what I ended up with.
?
Rich

Screenshot 2024-07-22 at 07.05.17.png

 

Thanks Rich.
?
For the load, is it a motor or something like that? Just trying to get my head round why the startup current is 10x the run current.?
?
Some feedback on the PSU, not trying to pick it apart, just some thoughts:
?
  • There's quite a bit of resistance in the chain. Adding it up in my head it's over 3 ohms. A 3 amp draw will cause a minimum of 9 volts drop, and the resistance in the earlier part will cause much more. The chart bears this out - there's a drop of 12 or 13 volts between the low current and the high current draw
  • As you are feeding into a regulator, some ripple is OK on the output of your PSU. The solid state regulator from your datasheet has a line regulation of 0.015%. That means it will cut the ripple going in by 6000-7000 times. So some ripple is OK as long as the "trough" of the cycle doesn't drop below the dropout value of the regulator
  • The RC sections will give a small amount of ripple reduction but will mess up the regulation. Again, think about how much ripple is OK and drop the RC sections. And probably the LC too
  • If I was having a go at this, I would probably just have transformer, bridge, capacitor. And that's it. Nice and easy, compact and should give better regulation
  • The 16.5V seems high for a 5V regulator, would need to keep an eye on power dissipation as it's a linear regulator
?
Putting all the above into practice, I've clipped an example where I've used a 9V / 3A (10% regulation) transformer into a simple capacitor. On the high current draw there is about 3V peak to peak of ripple. That may look bad, but the regulator will get it down to 0.0005V which may or may not be OK for your application. The light load rises to 12V which means your regulator will drop 7V @ 0.25A which is 1.75W, but worse is the 3A draw where it will lose 12-15W. I skimmed the data sheet for a power envelope but couldn't quickly find anything, just some stuff about how it limits the power by shutting down automatically.
?
Hope this is some help to you.
?
?
Regards,
Duncan
?
?
 

Duncan,
?
This is great! thanks so much
?
I have only designed power supplies that are not regulated before :-)
?
That approach makes life very simple!
?
I can also add a capacitor after the Regulator, when I looked at this unit (an alternative one I might consider) the application of a 100uF seemed useful.
?
https://www.mouser.co.uk/datasheet/2/268/MIC29510_12_5A_Fast_Response_LDO_Regulators_DS2000-3443948.pdf
 

开云体育

I don't know what the powered circuit is, but all regulators are not created equal.
They can have enormous phase & impedance non-linearities that really screw up the sonics.
I can assure you that attention to PS linearity pays huge sonic dividends.

This thread from 2009 goes into the detail.


As far as the PSU, why not parallel a high current winding feeding a simple transistor regulator with a lower current precision one.
After the appropriate time, the high current can be shut off.
To avoid a transient, kill the high current winding input at the line zero crossing

More complicated in some aspects for sure.
As simple as possible in others

Cheers,
- Ian
On 7/22/2024 03:07, Tonescout wrote:

Duncan,
?
This is great! thanks so much
?
I have only designed power supplies that are not regulated before :-)
?
That approach makes life very simple!
?
I can also add a capacitor after the Regulator, when I looked at this unit (an alternative one I might consider) the application of a 100uF seemed useful.
?
 

Ian,
?
Thanks for the guidance/idea.

So the concepts is a dirty regulator circuit for the start up and a clean linear non-regulated for the in-use period. = sounds like a great idea :-)

What I am unclear on is the process of transition, and the way to manage the balance of current draw to the dirty circuit and then to only the clean circuit?
?
Rich
?
?

1 - 7 of 7