Author Topic: Multiple Power Supplies  (Read 11292 times)

0 Members and 1 Guest are viewing this topic.

Offline Paul Price

  • Super Contributor
  • ***
  • Posts: 1419
Re: Multiple Power Supplies
« Reply #25 on: July 02, 2013, 08:55:16 pm »
fpliuzzi,

This guy hasn't got it right either, there is no need for wasting parts and power with those extra diodes on the output and common point, these diodes just serve to add unneeded complexity as well as ensuring poor regulation..a poorly thought out  circuit design.
 

Offline fpliuzzi

  • Regular Contributor
  • *
  • Posts: 157
  • Country: us
Re: Multiple Power Supplies
« Reply #26 on: July 02, 2013, 10:15:23 pm »
I was always curious about the extra 1N4007 isolation diodes in EDN's implementation.

Regards,
Frank
« Last Edit: July 03, 2013, 12:09:15 pm by fpliuzzi »
 

alm

  • Guest
Re: Multiple Power Supplies
« Reply #27 on: July 03, 2013, 09:50:25 am »
Given the huge amounts of bulk capacitance that might be present at these current levels, I would add the datasheet-recommended reverse-biased diode from output to input instead. The 7812 might be unhappy when a huge cap gets discharged through the regulator if the input gets shorted.
 

Offline Paul Price

  • Super Contributor
  • ***
  • Posts: 1419
Re: Multiple Power Supplies
« Reply #28 on: July 03, 2013, 12:05:01 pm »
alm, I don't think you quite understand how the 7805 works or the EDN circuit functions.

There would not be any great discharge current through any individual 7805 regulator and the diodes would not help to protect if there was such an event.

The diodes are a poor design idea.
 

Offline mrflibble

  • Super Contributor
  • ***
  • Posts: 2051
  • Country: nl
Re: Multiple Power Supplies
« Reply #29 on: July 03, 2013, 12:20:45 pm »
I should hope all the 7805 opinions are rather academic, since the OP will of course have taken the pragmatic path and use buck regulation. I mean ... he has 30V and 60V supplies available and needs 12V and 5V... Seems buck is a better match than linear.
 

Offline Paul Price

  • Super Contributor
  • ***
  • Posts: 1419
Re: Multiple Power Supplies
« Reply #30 on: July 03, 2013, 01:29:02 pm »
mrflibble, You are quite right, buck is better than linear, 

but it is not simpler, it is not the simple, "I need to get this job done and get to the beach!"  solution.

Remember what the OP asked for,  "..a simple solution?"
 

alm

  • Guest
Re: Multiple Power Supplies
« Reply #31 on: July 03, 2013, 01:41:48 pm »
alm, I don't think you quite understand how the 7805 works or the EDN circuit functions.

There would not be any great discharge current through any individual 7805 regulator and the diodes would not help to protect if there was such an event.

Do you think that claiming everyone who disagrees with you doesn't understand the rather trivial circuit is an intelligent response?

Assume you are designing the 100A supply the EDN article suggests. You might have local bypassing after every regulator and maybe some bulk capacitance to improve transient response. Say you have 1000 uF of capacitance in total, charged up to 12 V. Now the input gets shorted. Each 7812 contains an internal diode from output to input. The TI datasheet recommends a reverse-biased diode with over 10 uF of output capacitance. So we have a cap charged to 12 V, then 100 diodes in parallel, and then a short. Silicon diodes have a negative tempco, so the load sharing will be extremely poor. That's not a robust design.

The series diodes also prevent this problem, but will affect regulation, as you stated.
 

Offline johndon2000Topic starter

  • Contributor
  • Posts: 28
  • Country: gb
Re: Multiple Power Supplies
« Reply #32 on: July 03, 2013, 02:07:44 pm »
Thanks all, ended up using a DELL PSU off an old PC for the -12V, and then 2 power supplies for the +5V and +12V... much appreciated :)
 

Offline Paul Price

  • Super Contributor
  • ***
  • Posts: 1419
Re: Multiple Power Supplies
« Reply #33 on: July 03, 2013, 02:16:45 pm »
alm,

Firstly, I think that anyone that doesn't understand this rather complicated argument about this "trivial" circuit is welcome to seek clarity and correct apparent confusion about what is actually happening in this circuit.

I don't think you quite understand the problem,

If any single diode attempts to take some more of its share of the input-shorted discharge of an output capacitor, then it's forward voltage rises causing all other diodes to be biased forward to share the surge current.
The forward voltage of the protection diode rises in a few nSecs in respect to current, while the thermal 2.2mv/degC effect takes some much more time, making all the other protection diodes the shortest path that electricity likes to take.
« Last Edit: July 03, 2013, 03:15:19 pm by Paul Price »
 

Offline mrflibble

  • Super Contributor
  • ***
  • Posts: 2051
  • Country: nl
Re: Multiple Power Supplies
« Reply #34 on: July 03, 2013, 03:11:01 pm »
mrflibble, You are quite right, buck is better than linear, 

but it is not simpler, it is not the simple, "I need to get this job done and get to the beach!"  solution.

Sure it is. *click* [Buy It Now] on ebay, go to beach. Continue going to beach for a few weeks. On the way back from beach grab cheapo buck regulator from the mailbox and integrate into design. Problem solved. After solving said problem put the rest of those LM2596 modules you bought in the parts box for future projects.

Although to be fair, if that 3A current is sustained that might be pushing it with those ebay LM2596 modules without some additional cooling.
 

Offline Paul Price

  • Super Contributor
  • ***
  • Posts: 1419
Re: Multiple Power Supplies
« Reply #35 on: July 03, 2013, 03:19:07 pm »
mrgibble, You seem to have somehow moved to an universe where ebay transactions take place along with delivery at someting near the speed of light.

You might take notice that the OP's post on this page shows he had the wisdom to find a simple, quicker way to the beach.
 

duskglow

  • Guest
Re: Multiple Power Supplies
« Reply #36 on: July 03, 2013, 03:23:50 pm »
I'm not sure how "the speed of light" and "a few weeks" got conflated in your head.  Ah well.
 

Offline mrflibble

  • Super Contributor
  • ***
  • Posts: 2051
  • Country: nl
Re: Multiple Power Supplies
« Reply #37 on: July 03, 2013, 03:36:56 pm »
mrgibble
Time for a new keyboard?

Quote
You seem to have somehow moved to an universe where ebay transactions take place along with delivery at someting near the speed of light.
I reside in the universe where cheap electronics from China typically takes about 2-3 weeks to arrive, people have a clue how component tempco affects current sharing, and where people ground their scope.

Quote
You might take notice that the OP's post on this page shows he had the wisdom to find a simple, quicker way to the beach.
I took notice of all the posts by the OP in this thread. He uses ready made switchers, which sounds like a good plan. :)

 

duskglow

  • Guest
Re: Multiple Power Supplies
« Reply #38 on: July 03, 2013, 03:38:20 pm »
I reside in the universe where cheap electronics from China typically takes about 2-3 weeks to arrive, people have a clue how component tempco affects current sharing, and where people ground their scope.

Reality is a pretty good place to live in. :)
 

Offline c4757p

  • Super Contributor
  • ***
  • Posts: 7799
  • Country: us
  • adieu
Re: Multiple Power Supplies
« Reply #39 on: July 03, 2013, 03:42:57 pm »
I'm not sure how "simple" and "guessing whether it works contrary to how it's supposed to - gee, I hope there are no unforeseen problems with this" got conflated in his head. Oh well.

Using a linear regulator to get 5V from 12V at 3A is retarded. You're not making a bench power supply here. Using them at the edge of their rating contrary to the way the datasheet instructs is also retarded. And going along with "just trust me, it will work, I tried it" when you present completely incoherent arguments as to why this is would be supremely retarded on the part of the OP.

It's hard to beat the long-term simplicity of doing it right the first damn time, and it's hard to beat the short-term simplicity of an LM2576 switcher module or similar.
« Last Edit: July 03, 2013, 03:45:41 pm by c4757p »
No longer active here - try the IRC channel if you just can't be without me :)
 

duskglow

  • Guest
Re: Multiple Power Supplies
« Reply #40 on: July 03, 2013, 03:48:47 pm »
Well, this bad advice would have only resulted in the popping of a few regulators, and not in bodily injury, so thank goodness for small blessings?
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf