flyback diode aplication for linear power supply?

Thread Starter

Hamlet

Joined Jun 10, 2015
519
I'm finishing up a simple 2-30v, 15A linear adjustable power supply (LM723).


I am concerned about back-emf killing my circuit when I test motors, solenoids,
coils, etc. In addition to the LT10A04 shown in the schematic, I installed a 1N4007
across each power transistor. I subbed a BD139 for the BD243C, and used four TIP3055
for the output transistors. Other than that, everything else is same.

Do I need anything more at the final output? I'm thinking a large Shockley on the positive terminal as the ultimate emf
protection...

I am reassembling my test layout into an enclosure and I don't want the smoke
to get out after all my work.

Note, I omitted the mistake where the input is directly
shorted to the output. I think the author goofed up. Otherwise, it seemed to kick butt
nicely on the benchtop.
 

#12

Joined Nov 30, 2010
18,224
It never hurts to place a properly oriented diode on the output, but it works a little better when the diode is connected to the motor. If you made a test cable with the diode on the load end, that would work, but I guess you are trying to protect the supply against the day when you can't find (or simply forget) the test cable.:rolleyes:
 

Thread Starter

Hamlet

Joined Jun 10, 2015
519
I built it on the workbench, now I am finalizing it in an enclosure.

I think if I leave off an output diode, I can always install one on
any device under test that I am concerned about spurious emf.

Yes, I built a 1.5A supply with LM317, and even though I had
protection diodes, I killed it when I thoughtlessly tried powering
an antique electric fence charger. The load was probably way
too much, and emf didn't help matters. I need to rebuild that
project next.

I am learning the hard way.
 

AnalogKid

Joined Aug 1, 2013
10,987
A power Shottkey diode reverse connected from Vout to Vin should give you the protection you need. For the belt and suspenders approach, another thing to add is a tranzorb or power zener diode from Vout to GND. For a 30 V output, a 36 or 39 V diode will clip things nicely.

But - if you are worried about back-emf from a device connected from the supply positive output to GND, note that the voltage spike will be negative with respect to GND. Again, a 40 V tranzorb across the output will catch this much more quickly than a 400 V standard recovery diode.

ak
 
Last edited:

Thread Starter

Hamlet

Joined Jun 10, 2015
519
I don't have any tranzorb tv diodes. I do have a big 3-pin Shottkey I removed from a switcher.
I suppose that will be faster than the LT10A04 as shown in the schematic. My understanding
is that the LT10A04 that is called for in my schematic is really more for reverse polarity protection,
say, if I tried using the power supply as a battery charger. It isn't fast enough for multiple-rapid emf
event.

Looking at that diode again, I think a fuse is needed on the final, final output, least someday that diode
inadvertently becomes the fuse. :oops:
 

Thread Starter

Hamlet

Joined Jun 10, 2015
519
After I finish my project, should I post a pic, or schematic in the finished projects part
of the forum, or is that place for something different?
 

ian field

Joined Oct 27, 2012
6,536
A power Shottkey diode reverse connected from Vout to Vin should give you the protection you need. for the belt and suspenders approach, another thing to add is a tranzorb or power zener diode from Vout to GND. For a 30 V output, a 36 or 39 V diode will clip things nicely.

But - if your are worried about back-emf from a device connected from the supply positive output to GND, note that the voltage spike will be negative with respect to GND. Again, a 40 V tranzorb across the output will catch this much more quickly than a 400 V standard recovery diode.

ak
Shottky diodes with reverse voltage of only 20V are very common, 30 and 40V types are also common. Its rare to see any rated more than 60V, I assume the prices rise rapidly at about that point. In the junk box I have a couple of 90V SB diodes liberated from scrap LG VGA monitors. The highest voltage I've ever seen advertised was 135V.

SB rectifiers may not be the best choice for blocking back emf!
 

AnalogKid

Joined Aug 1, 2013
10,987
My understanding is that the LT10A04 that is called for in my schematic is really more for reverse polarity protection...
It is a common item in lab supplies, where multiple supplies might be connected in series, such as a +/- analog supply. In this case, if one supply comes up before the other it will appear reverse-connected to the later supply through the load. The diode keeps the reverse voltage down below anything that might do damage to the not-yet-awake output.

ak
 

Thread Starter

Hamlet

Joined Jun 10, 2015
519
I have 4 fans for cooling, a plexiglass greenhouse style cover, & brushed aluminum end caps (earth grounded.)
I have a 250va transformer, with two 13.5 volt secondaries. I have a DPDT switch installed before the bridge for
high & low range.

I definitely want to protect this creation, and I do anticipate placing it in series with other supplies. (I probably should
buy a variac.)
 

GopherT

Joined Nov 23, 2012
8,009
Then you may want to add fuse and a crowbar SRC protection. As soon as an over current is detected, s comparitor trips the SCR and that big scr and a small value, high wattage resistor suddenly draw a bunch of current to pop your fuse. As described it my beginning electronics book, it is like dropping a crowbar across the buss bars to save the sensitive bits downstream.
 

Thread Starter

Hamlet

Joined Jun 10, 2015
519
I tested a lashup of my circuit, and while using four 50w halogen bulbs wired in parallel, I was able to switch
the power on/off/on without blowing anything. A cold halogen have very low resistance, so I figger that the
current limiting/foldback part of my circuit is adequate to protect the transistors from a low-resistance load
or short, and this is the story I get from others who have built this circuit. As I anticipate using my supply
to test solenoids, big coils (not testla), motors, my concern is back emf from reactive loads wiping out the
TIP3055 or the LM723.

I like the the suggestions from everyone, yet those from AK seem closer to what I need to achieve. I have two
Schottky diodes in my junkbox; 45v/10A & 200v/10A. If I add "another thing" then I need to purchase parts.
I'm looking at power zeners and transorb on ebay. There are circuits where a "Power Zener" can be created
from a regular 1w zener & a power transistor. I might go this route, as I have those parts on hand. Perhaps
a transorb is the best solution. If there is no substitute, I will source and install this part. I have no previous
experience or understanding of the transorb device. What rating should I look for?


A power Shottkey diode reverse connected from Vout to Vin should give you the protection you need. For the belt and suspenders approach, another thing to add is a tranzorb or power zener diode from Vout to GND. For a 30 V output, a 36 or 39 V diode will clip things nicely.

But - if you are worried about back-emf from a device connected from the supply positive output to GND, note that the voltage spike will be negative with respect to GND. Again, a 40 V tranzorb across the output will catch this much more quickly than a 400 V standard recovery diode.

ak
 

ian field

Joined Oct 27, 2012
6,536
I tested a lashup of my circuit, and while using four 50w halogen bulbs wired in parallel, I was able to switch
the power on/off/on without blowing anything. A cold halogen have very low resistance, so I figger that the
current limiting/foldback part of my circuit is adequate to protect the transistors from a low-resistance load
or short, and this is the story I get from others who have built this circuit. As I anticipate using my supply
to test solenoids, big coils (not testla), motors, my concern is back emf from reactive loads wiping out the
TIP3055 or the LM723.

I like the the suggestions from everyone, yet those from AK seem closer to what I need to achieve. I have two
Schottky diodes in my junkbox; 45v/10A & 200v/10A. If I add "another thing" then I need to purchase parts.
I'm looking at power zeners and transorb on ebay. There are circuits where a "Power Zener" can be created
from a regular 1w zener & a power transistor. I might go this route, as I have those parts on hand. Perhaps
a transorb is the best solution. If there is no substitute, I will source and install this part. I have no previous
experience or understanding of the transorb device. What rating should I look for?
A transistor assisted zener can be imprecise - if you use the TL431 programmable zener, you can include the transistor in the nfb loop for much more accurate regulation.

There are dozens of manufacturers, but groups of them all seem to publish near identical appnotes - some exemplify the boosted TL431 configuration, but not all.
 
Top