Understanding Constant Current Supplies

I am trying to understand this basic Constant Current Circuit


The relays are 12V Omron G5v series rated at 41.7 mA

I am using a variable DC LAB PSU  to send the LM317 Vin and for GND.  I calculate 62.5 mA from this 1.25/20 as my constant current. Then I took measurements for both one and two relays in series.  Test leads of DMM were from the neg side of the relay to ground.

12V in gives 31mA
24V in gives 62 mA

Then with 2 relays, I saw I needed more voltage before the relays kicked in, about 15V.

15V gave 22 mA
30V gave 44 mA

I was under the impression that Current would be constant, but the more Volts I pump in the more current I seem to get.
I am a bit stumped. Can someone enlighten me?

This is a current limiter, not a current source.

> relays are 12V Omron G5v series rated at 41.7 mA

> the more Volts I pump in the more current I seem to get.

Well, sure. You have 24V worth of relay, plus that '317, so maybe 11V across each relay. They won't be pulling their nominal 41.7mA until you get 12V across each relay, 24V on two relays PLUS 2V or 3V for '317 losses.

You need voltage to make current in a resistor. The '317 won't "make voltage". It can only "waste voltage" to attempt to keep current to set-value.

And note that your observed 62mA is 0.5mA shy of the 62.5mA where the '317 should kick-in and start limiting the current.

Also note that you are at 148% of relay rated current and over twice relay rated power. Keep doing that and you'll smoke one.

Two 12V relays in series is gonna need 24V, plus '317 loss, the '317 is not doing any good until you get past 27V.

The relays will prefer 41.7mA, so you really need to get a 29.976 ohms (try 33R) resistor instead of 20R which is too small to "limit" current to a safe value for these relays.

I suggest you first try a 10-cent 100 ohm resistor instead of 5-buck relays. With 33R under the '317 you should get 38mA from 7V in all the way to ~~30V (a naked '317 may overheat at 26V 38mA) or heat-sunk to 44V in (where the 40V '317 is rated to die).

Your actual relays are more like 288R, so try again with a couple handy 270R resistors. Now for '317 set for 1.25/33= 38mA you will need ~~24V just to get the resistors up to 38mA. You can run from 24V to 50V and current should be pretty-solid 38mA. When you know you have not made any major boo-boo, that it DOES limit current, you can find a 29.976 ohm to get your nominal 41.7mA (or toss 330R across your 33R).

> I saw I needed more voltage before the relays kicked in, about 15V.

A relay on a quiet workbench should pull-in at around 2/3 of rated voltage. (If it is in a bouncy Jeep, don't bet on it staying On with 2/3 voltage). That says a "12V" relay may work with 8V, yours need 7.5V, entirely as expected. Omron only promises "must operate 80%", so yours more than beat their claims.

BTW: this series is rated for 200% over-voltage. So it won't smoke at 62.5mA. In fact I must wonder why you feel a need for "constant" relay power. A 12VAC or 24VAC winding with rectifier and minor filtering will be good-enough for any household or studio purpose. If your 117V wall outlet wanders toward 200%, it won't be your relays you will cry about. All your incandescent lamps and then half your electronics will fry-and-die before small relays will smolder.

Like PRR said, 12v relays are not ideal for this circuit. I would use the lowest voltage type.  This is a useful configuration to reduce disturbance in other circuits when relays energize (clicks). Short the relay coil with a switch when you want it off.  Neve drives all the LEDs in the V series like this, in series chains with ccs.

Umm change the resistor value for 40 ma?

About the only stupid electronic trick I recall doing with relays over the years was to provide more pull in current transiently to insure reliable switching, then relaxing back to a lower hold current. 

JR

Thanks.  I'll need some time to digest the math.  I had 12 and 24V relays because I was running 6 to 12 relays which off of a voltage supply, would pull way too much current and seemed like a waste of current.  In this setup, I will have somewhere between 5 and 8 relays.  This time around, i thought it would be better to use a current limiting setup.  I'll try 5V relays. I started with  12V since I had them and to keep current down.

The good thing is that I am planing on using a 24V transformer that is used to double up for phantom for these.

Does Constant Current only work in series?  If I had a parallel setup, the current draw would add per component draw, right?

With 33R under the '317 you should get 38mA from 7V in all the way to ~~30V (a naked '317 may overheat at 26V 38mA) or heat-sunk to 44V in (where the 40V '317 is rated to die).

Click to expand...

PRR, what calculation do you use to get 7V for minimum Volts?  This was a loose end in anything I read.

Driving from a current source only makes sense in series.

Your problem is a 6 of one half dozen of another.

The higher the nominal voltage the lower the nominal current for relay coils, so 2x 5v in series is similar power to 2x 12v in parallel. Do what makes sense for you. The higher voltage relays will be easier to switch electronically due to lower current.

If you run them all in parallel there is no need for current source.

JR

At the end of the day, what is the expected benefit of running these relays with constant-current?
At first sight, I see only disadvantages:
a)You'll have to run the relays at higher current than really needed, just to ensure positive switching
b) voltage drop in the regulator is another source of energy deperdition (heat)
c) cost
If your concern is noise, you have to make sure that the relay coil current doesn't share any audio ground or supply tracks, that's all.

> would pull way too much current and seemed like a waste of current.

POWER!!

The relay shown is 12V 41.7mA or 0.500 WATTS.

That is about the power needed to pull a clapper with that many contacts that big. Reed relays can work on less power, with fewer smaller contacts. Automobile starter relays have one BIG contact, need 10-20 Watts to pull that sucker up (and against the spring needed to bust it loose after 200-Amp load arc-welds the contact).

You can buy those Omron G5v in 5V or 48V, maybe custom in 1V or 100V. But the WATTS is what pulls the clapper(*). So you could get 1V 500mA or 100V 5mA.... it is still 0.5 Watts per relay.

> Your problem is a 6 of one half dozen of another

Or maybe: 12 of one and two half-dozens of the other.

> series?  If I had a parallel setup, the current draw would add per component draw, right?

Yes, but the voltages don't add in parallel!

The real thing is: how you gonna turn them off? Unless all relays are on all the time (why?), it is usually much easier to run the relays in parallel so the switch or transistor return may be common to all.

> use a current limiting setup.  ....to keep current down.

This works like putting an orifice in the shower head. If large, it has no effect; if small, it doesn't work right.

Eight relays at 0.5 Watts each is 4 Watts. You need a 4W power supply.

Relays do NOT need regulated power. The Omron G5v will work at 80%-200% of rated voltgae.

On a clean-sheet design, you could use "any" power voltage. But these are AC relays, we live in an AC world, we need a rectifier. Rectifiers have 1V-2V loss, We want to keep the system voltage well above 2V so we get more power on the relays than in the rectifier. That suggests going past 100V, but the fine wire needed to wind the relays raises the cost, also 2-cent transistors are rated 40V so even a 48V system costs more.

Use a 12VAC or a 24VAC winding with ample Power as determined by summing the loads. Rectify and filter slightly.

> 24V transformer that is used to double up for phantom

1) "Phantom" is usually about a Watt total. If you "add" a 4W burden to a 1W supply, it gets unhappy. You need to design for ALL the loads, and the relays are not negligible compared to Phantom.

2) most "doublers" will not happily give large output at half the double-voltage. 600V at 300mA with a 300V 30mA tap makes sense. 48V 20mA with a 24V 333mA tap does not make sense.

3) you can get 48V Omrons. If you do your switching with transistors, you must pay the 0.5-cent extra for 60V transistors. If you do it with switches, you are right AT the 50V rating of most cheap switches. And since relays are clacky, and Phantom should be clean, and Phantom should be regulated while relays don't care, it is wasteful to hang more than 1 or 2 relays on a Phantom line.


> what calculation do you use to get 7V
> I'll need some time to digest the math.

You are very weak in Ohms Law. I tell you, you MUST understand basic battery and resistor problems before you will be comfortable at any electric design or modification. If I tell you my battery is 12V and my O2 heater is 3 ohms, what is the Current? What is the Power? If I say I need this heater to give 20 Watts, what battery should I buy? Do a hundred of these pointless exercises every morning for a week, you can do them FAST which is the only way to move ahead to real problems.

---
(*) Yes, in a single turn, the pull is proportional to current. But we always use a many-turn coil. 100 turns at 10mA pulls the same as 200 turns at 5mA. The total coil size is limited by relay frame. When you figure out the smaller gauge to fit 200 turns, you find the resistance has quadrupled. So double turns needs double voltage and half current to give the same pull. It comes to the same Watts.