Author Topic: Subtract 2-4VDC at 0.1-10A+ from a DC line  (Read 1209 times)

0 Members and 1 Guest are viewing this topic.

Offline digsysTopic starter

  • Supporter
  • ****
  • Posts: 2209
  • Country: au
    • DIGSYS
Subtract 2-4VDC at 0.1-10A+ from a DC line
« on: March 04, 2020, 09:28:17 am »
Got an unusual one - I have a lot of "stuff" out there designed for 48VDC. It can tolerate 51VDC continuous, and will crowbar at 52V (app 1 sec) - all fine.
So everyone I deal with is moving to a SYSTEM 48V BUS and doing away with converters, both 12V etc up and 240VAC down. Some of these BUSes are 1,000s Amps. They are also directly connecting to 48V vehicles - with 4x12V batteries - with flat > Full alternator charging.  PoE cameras, switches etc etc. This now sees anything from 48-56VDC on the BUS. Re-designing 100s devices is not an option.
I could just "Insert" a string of high current diodes in the DC line (in a nice 2 terminal Al case), for 6-7V drop, and call it quits, but that is a lot of often UN-necessary heat and loss, especially for installs that are running at 48-51V.
I need to have that much higher Vloss, as damn diodes Vf vary from 0.5+V at 0.1A > 2.0V at 10+A. When the systems are in idle, they draw very little current, and once the crowbar fires, they'd have to restart the machinery, some of it is huge :-)
I had planned on a series of diodes with "taps", so the Installer could check the system voltage and select the appropriate "drop-out voltage" required ... perfect
Anyone know of a device with a junction that has ~ 0.6-2.0V loss at all currents? Even a zener assisted PNP varies too much .. or other ideas? Has to be DC line only and simple.
Hello <tap> <tap> .. is this thing on?
 

Online Zero999

  • Super Contributor
  • ***
  • Posts: 20362
  • Country: gb
  • 0999
Re: Subtract 2-4VDC at 0.1-10A+ from a DC line
« Reply #1 on: March 04, 2020, 10:29:07 am »
What's the minimum voltage it will work from?

How about using a buck converter to get 40V? It will have much lower losses.
 

Offline digsysTopic starter

  • Supporter
  • ****
  • Posts: 2209
  • Country: au
    • DIGSYS
Re: Subtract 2-4VDC at 0.1-10A+ from a DC line
« Reply #2 on: March 04, 2020, 10:55:02 am »
These are mostly existing installations that "basic" electricians will go out to and I have no access to the 0V rail. I can easily afford to waste 0-4V, but don't want to do a blanket install, as many installations will be fine, and need nothing. A diode would be perfect, except for it's damn variation with current.
Hello <tap> <tap> .. is this thing on?
 

Offline nali

  • Frequent Contributor
  • **
  • Posts: 732
  • Country: gb
Re: Subtract 2-4VDC at 0.1-10A+ from a DC line
« Reply #3 on: March 04, 2020, 11:00:44 am »
I needed something similar albeit much lower power, I used a simple NPN emitter follower with a zener reference which worked fine.
 

Offline Mazo

  • Regular Contributor
  • *
  • Posts: 66
  • Country: bg
Re: Subtract 2-4VDC at 0.1-10A+ from a DC line
« Reply #4 on: March 04, 2020, 11:05:31 am »
A TL431 style reference+the amount of transistors needed to dissipate the power?All you will need to ensure is that the minimum current on the line is >quiescent current of the "TL431"
 

Offline Rerouter

  • Super Contributor
  • ***
  • Posts: 4705
  • Country: au
  • Question Everything... Except This Statement
Re: Subtract 2-4VDC at 0.1-10A+ from a DC line
« Reply #5 on: March 04, 2020, 11:11:55 am »
you will need a beefy pass transistor, but this will work for what your asking,

Input on left, output on right, change zener to adjust drop, remove zener for minimum drop,

edit: would help if I remembered my transistor polarities, in reality a darlinton should be enough to make a constant 2V drop for yourself, lets go with zero's suggestion, probably with a pass element.
« Last Edit: March 04, 2020, 11:20:58 am by Rerouter »
 

Online Zero999

  • Super Contributor
  • ***
  • Posts: 20362
  • Country: gb
  • 0999
Re: Subtract 2-4VDC at 0.1-10A+ from a DC line
« Reply #6 on: March 04, 2020, 11:16:10 am »
you will need a beefy pass transistor, but this will work for what your asking,

Input on left, output on right, change zener to adjust drop, remove zener for minimum drop,
You've got the resistor and zener the wrong way round.

The voltage drop will still vary considerably with the current. The TL431 will be more stable. Refer to page 27, figure 31 of the data sheet.
http://www.ti.com/lit/ds/symlink/tl431.pdf?
 

Offline digsysTopic starter

  • Supporter
  • ****
  • Posts: 2209
  • Country: au
    • DIGSYS
Re: Subtract 2-4VDC at 0.1-10A+ from a DC line
« Reply #7 on: March 04, 2020, 11:25:48 am »
Quote from: Mazo
A TL431 style reference+the amount of transistors needed to dissipate the power?All you will need to ensure is that the minimum current on the line is >quiescent current of the "TL431"
I have thought of those, but was hoping there'd be a silicon junction that didn't vary so much. Definitely a possible fallback.
Hello <tap> <tap> .. is this thing on?
 

Offline digsysTopic starter

  • Supporter
  • ****
  • Posts: 2209
  • Country: au
    • DIGSYS
Re: Subtract 2-4VDC at 0.1-10A+ from a DC line
« Reply #8 on: March 04, 2020, 11:32:02 am »
Quote from: Rerouter
  you will need a beefy pass transistor, but this will work for what your asking ...
Played with various zener assisted circuits as well ... just seeing if there are any other clever options out there :-) ... If it wasn't for my fail-safe crowbar, I'd just use cr@ppy diodes :-) .. TL431 (and similar) is head of line.
Hello <tap> <tap> .. is this thing on?
 

Offline Mazo

  • Regular Contributor
  • *
  • Posts: 66
  • Country: bg
Re: Subtract 2-4VDC at 0.1-10A+ from a DC line
« Reply #9 on: March 04, 2020, 11:51:46 am »
A Sziklai pair with the "outer" transistor being itself a current boosted pair(darlington or again sziklai doesn't matter) will vary very little in voltage drop as the "inner" transistor is setting the "reference" Vbe and its current is changing very little.There will still be a temperature drift.
Still the TL431 solution seems superior by far.
 

Online Zero999

  • Super Contributor
  • ***
  • Posts: 20362
  • Country: gb
  • 0999
Re: Subtract 2-4VDC at 0.1-10A+ from a DC line
« Reply #10 on: March 04, 2020, 11:55:34 am »
Quote from: Mazo
A TL431 style reference+the amount of transistors needed to dissipate the power?All you will need to ensure is that the minimum current on the line is >quiescent current of the "TL431"
I have thought of those, but was hoping there'd be a silicon junction that didn't vary so much. Definitely a possible fallback.
Well there are high power LEDs of course, but they will have a similar V/I curve as any diode, just with a higher voltage. Diodes also have a negative temperature coefficient, which is bad.

I think the TL431+a power transistor is the only sane way. If you don't want a big PNP, then use a small one, with a big NPN, in a Sziklai configuration, as mentioned above. There are also lower voltage variants of the TL431, such as the TLVH431, which can go down to 1.24V.
http://www.ti.com/lit/ds/symlink/tlvh431a.pdf
 

Offline digsysTopic starter

  • Supporter
  • ****
  • Posts: 2209
  • Country: au
    • DIGSYS
Re: Subtract 2-4VDC at 0.1-10A+ from a DC line
« Reply #11 on: March 04, 2020, 12:29:51 pm »
Quote from: Zero999
Well there are high power LEDs of course, but they will have a similar V/I curve as any diode, just with a higher voltage. Diodes also have a negative temperature coefficient, which is bad.
I think the TL431+a power transistor is the only sane way. If you don't want a big PNP, then use a small one, with a big NPN, in a Sziklai configuration, as mentioned above. There are also lower voltage variants of the TL431, such as the TLVH431, which can go down to 1.24V.
Looking that way ... the diode option is just so clean and simple. No active junctions, no possible gain / feedback issues whatsoever. Mount several TO247s to a 4x4oz PCB, heat is spread out nicely. Taps ready and easy etc etc Just the stoopid Vf / current issue, and the tempco you mention. Oh well.
Hello <tap> <tap> .. is this thing on?
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf