LM317 (Again!)

[muso52]

Hi Gents,
Well I might be considered lazy but would welcome your assistance with the following.
I need a 3.5v supply from a variable DC Voltage nom 5 -15v full wave rectified but unsmoothed. The LM317 would do the job but I would consider it feasible to put a smoothing capacitor on the output of the bridge prior to the LM317. Its many many years since I knew how to calculate value of Cap and also taking into consideration the variable input voltage.

Would very much appreciate any thoughts on the subject.

Bill

 

[ericgibbs]

hi Bill,
The smoothing cap value is determined by the current you want to draw from the bridge/LM317.

You may find 1.5V in/out difference for the LM317 just too low.

Was the 5/15V measured without any existing cap on the bridge.?

 

[muso52]

Hi Eric,
Thanks for replying.
With respect to current draw, I am guessing at 10ma.
Aprreciate 1.5v in/out may be to low for LM317 but can adjust low volts (5v) to 6 or 7 if needed.
The 5/15v is an estimate allowing for Vdrop in circuit sourcing the variable voltage + bridge.
This is fed from 18v transformer, the unmoothed voltage going to a variable voltage controller for a nominal 12v DC motor.

Hope this helps.
Regards, Bill

 

[ericgibbs]

Hi Eric,
Thanks for replying.
With respect to current draw, I am guessing at 10ma.
Aprreciate 1.5v in/out may be to low for LM317 but can adjust low volts (5v) to 6 or 7 if needed.
The 5/15v is an estimate allowing for Vdrop in circuit sourcing the variable voltage + bridge.
This is fed from 18v transformer, the unmoothed voltage going to a variable voltage controller for a nominal 12v DC motor.

Hope this helps.
Regards, Bill

hi,
The 18V transformer give approx 18V * 1.4 =~25V when smoothed.

Is this for a train speed controller.? that uses a full wave rectified voltage controller.??

If yes, to isolate the smoothing capacitor for the LM317 input from the bridge, connect a 1amp diode from the un-smoothed 18V rectified to the LM317 input and add say a 220uF or higher from the LM317 input to 0V.
On the output side of the L317 I would add a 47uF and 100nF, assuming a 15mA load

 

[muso52]

Hi Eric,

Yes, you are quite correct, it is part of a train speed controller!!

I will explain.

It is considered better to use full wave rectified but not smoothed DC for motor control.However, I wish to derive a smoothed voltage from this pulsed voltage for use of onboard (model train) electronics, this being a nominal 5v.

Now, a range of 6 -18v = 12v, the dc motor voltage max. Now by the simple use of parrellel inverse diodes I can 'drop' the first 6v (or more) to the motor. The idea being that at this voltage, the model train is stationary although approx 6v is across the track This thus being used for onboard electronics.
This can be in the form of led lights and indeed in my own engines sound units.

Unless I have completely forgoten or my knowledge of principles I believe it possible to place a bridge and associated circuit (LM317) in parallel with the motor across the track (supply) and derive a smooth supply for electronics and still maintain a pulsed supply(rectified full wave ) to the dc motor.

I think!!

Bill

 

[ericgibbs]

hi Bill,
You could use a bridge across the tracks, that taps into the motor controller output voltage, but wont the voltage you get across the bridge input be dependent upon the motor speed setting.??

 

[muso52]

Hi Eric,
'You could use a bridge across the tracks, that taps into the motor controller output voltage, but wont the voltage you get across the bridge input be dependent upon the motor speed setting.?? '

Yes exactly. The Voltage across the the bridge input will be dependent on the motor speed setting ie the variable voltage.
The idea is to use a LM317 post bridge to source additional circuits at a fixed voltage from a variable one. The bridge being necessary because voltage polarity is reversed to change direction of motor. Hence the use of a smoothing cap. Since the cap is now between the bridge and the track voltage no smoothing of the 'pulsed' dc will take place.

Now I am assuming that the LM317 can be used in this way. That is, a varying voltage in for a designated Vout allowing for internal losses (1.4v?)

Again, would welcome your comments Eric.

Thanks,
Bill

 

[ericgibbs]

hi Bill,
Just drawing a circuit, few minutes...

EDIT:
Look at this option.

Its with the motor at full speed ie:12V

I need to see your circuit for reversing and then I will complete a drawing.

 

[Mr RB]

...
Now, a range of 6 -18v = 12v, the dc motor voltage max. Now by the simple use of parrellel inverse diodes I can 'drop' the first 6v (or more) to the motor. The idea being that at this voltage, the model train is stationary although approx 6v is across the track This thus being used for onboard electronics.
...

Very clever! Thank you I have never seen that done before!

If I can make a suggestion for your (Eric's) circuit; 6v, the diode drop and ripple voltage leaves you no real overhead for the LM317, which is a regulator that needs quite a bit of overhead.

As an alternative you could use a low dropout 3.3v regulator, which will give 3.3v out even from as little as 3.5v in. Then you probably need to add a zener on the regulator input as most of these regulators are limited to a Vin of 9v or 16v etc.

Basically it's Eric's circuit plus a resistor and zener, and using a different 3pin regulator.

 

[ericgibbs]

hi Bill,
This circuit shows a full wave rectifier for the LM317 connected across the motor drive, for a full speed condition.

Gif1 shows the 3.5V using the 3.5V output as reference, NOTE: the 3.5V ground 0V, will be above the motor ground.
You cannot connect the LM317 0V, to the Motor Controller 0V

Gif2 shows the 3.5V with the motor 0V as reference.

The circuit inside the rectangular box is simulating the controller fwd/rev.

If your circuit uses the back EMF from the motor, the LM317 smoothing cap could load the back EMF, its a point to watch out for.

Also the input voltage to the LM317 will have two diode voltage drops, ~1.4V.
So this 1.4V should be added to the 1.5V low current drop out voltage of the LM317. so for 3.5Vout you will need ~6.5Vinp.

 

[muso52]

Hi Eric (and MrRB!)

Appologies for not writing earlier, have been playing on big steam trains with sons and grandchildren!

' Very clever! Thank you I have never seen that done before!

And me neither Mr RB! There are many circuits designed for Constant Lighting on DC Train Layouts. I prefer mine (above), simples. I use constant current devices to source leds in engine and trains. These being connected to the track prior to the dropping diodes to the motor.Of course in carriages, some form of circuit for changed polarity has to be implemented. Again I cheat!
As the LZ334 (low current device) has a reverse breakdown voltage rating somewhat higher than the max applied dc motor V, these are connected in a sort of inverse parallel again with no bridge but 2 diodes. This giving a Vdrop of only approx .7v whatever polarity is applied.

'As an alternative you could use a low dropout 3.3v regulator, which will give 3.3v out even from as little as 3.5v in. Then you probably need to add a zener on the regulator input as most of these regulators are limited to a Vin of 9v or 16v etc.'

Interesting MrRB. I must confess I have never used a Zener in anything, although it was part of a Btec unit I did many moons ago. But, if I recall, arn't Zeners used in fixed V circuits i.e. when the input and output voltages are knowns and not variable? The use of a 3.5v fixed reg is an option to consider thou.

Eric.

Many thanks for your time and effort.


'If your circuit uses the back EMF from the motor, the LM317 smoothing cap could load the back EMF, its a point to watch out for.'

Back emf is normal in closed loop PWM controllers. As we are using pulses that rise and fall approx 120 hz there is no where to sample and hold (nominally using a cap) the back emf. Normally this is monitored in PWM controllers when the pulse is off. Ie , if no pulse voltage is present , then any voltage detected between pulses is back emf. So no back emf.


You cannot connect the LM317 0V, to the Motor Controller 0V

Can you explain that to me please? The bridge is the only means of isolating the LM317 from the Motor Controller 0v( sort of). Is this not exceptable? I would also be very gratefull if you could explain reasons for R1 (10meg) and R 2 (10k). Essentially I believe I am looking at the circuit from the bridge D5 -D8 onwards.

Again thanks and regards,
Bill

 

[ericgibbs]

You cannot connect the LM317 0V, to the Motor Controller 0V

Can you explain that to me please? The bridge is the only means of isolating the LM317 from the Motor Controller 0v( sort of). Is this not exceptable? I would also be very gratefull if you could explain reasons for R1 (10meg) and R 2 (10k). Essentially I believe I am looking at the circuit from the bridge D5 -D8 onwards.

Bill,
Due to the full wave bridge, the output of the LM317 will not be directly connected to the 0V of the controller, you have the diode voltage drops to consider.

If you did directly connect the LM317, 0V and Controller 0V, you would effectively short out part of the diode bridge.

The R1 and R2 are just for the Simulation in LTspice.

The transformer and bridge are only to provide a test wave form for the Sim
and the switches so that I can show the effect of forward and reverse.

 

[muso52]

Bill,
Due to the full wave bridge, the output of the LM317 will not be directly connected to the 0V of the controller, you have the diode voltage drops to consider.

If you did directly connect the LM317, 0V and Controller 0V, you would effectively short out part of the diode bridge.

The R1 and R2 are just for the Simulation in LTspice.

The transformer and bridge are only to provide a test wave form for the Sim
and the switches so that I can show the effect of forward and reverse.

Hi Eric and good morning,Many thanks as allways.


OK fully understand all of above. One final aspect to consider. I may have to raise the output of the LM317 to 4.5v. Should this be the case then changing the ratio of the resistors is no problem. However it may then be considered a better option just to use a low drop out 5v fixed reg.Does such a thing exist in a TO220 case as size (inside model engines) is a factor.
Regards,
Bill

 

[ericgibbs]

Bill,
For loads upto 100mA, look at the TS2950CT 5V [ and 3.3v version]

 

[Mr RB]

MIC5209 in 5v is good for 500mA and has very low dropout and some nice features like reverse voltage protect etc. It's SMD but is large and has same 0.1" pin spacing as a 7805 so it is easy enough to use;
https://www.electro-tech-online.com/custompdfs/2011/09/mic5209.pdf

You will also need a 12v 3W zener and a resistor about 1W 33 ohm. That will limit the voltage at the input of the regulator to 12v DC.

I don't like Eric's idea of using a full bridge rectifier as it adds complexity and has a major disadvantage that the 2 grounds can't be allowed to touch. But it does mean you can use a smaller cap. Personally I would stick with the one diode on his orig circuit, and just use a larger cap 470uF or so.

 

[ericgibbs]

hi Roman,
Problem is that the two lines, +Vrect[unsmoothed] and 0V are swapped over for Reversing the engine.

So the 'rectifier' has to be active for forward and reverse.

 

[muso52]

Hi Eric ( and Roman)

Thanks again for info.
Having read the TS2950CT 5V .pdf it seems that this indeed would be preferable to the LM317. As space is a premium, having no Vset resistors are a bonus. Also it's low dropout properties.
I would imagine that the 220uf smoothing capacitor would be suitable (electrolytic?) and also just 1 additional cap, a 1uf across the output for stability. Would any additional components be required?

Regards,
Bill

 

[ericgibbs]

Hi Eric ( and Roman)

Thanks again for info.
Having read the TS2950CT 5V .pdf it seems that this indeed would be preferable to the LM317. As space is a premium, having no Vset resistors are a bonus. Also it's low dropout properties.
I would imagine that the 220uf smoothing capacitor would be suitable (electrolytic?) and also just 1 additional cap, a 1uf across the output for stability. Would any additional components be required?

Regards,
Bill

hi,
If I recall correctly you are only going draw low mA's from the 5V.??
If so I would have a 220uF 35Vwkg electrolytic on the input, with a 47uF 6Vwkg electrolytic and 100nF [0.1uF] on the output.

Whats the 5V powered circuit going to do.?? [curious]

 

[muso52]

Hi Eric,

Thats it then, sussed!
Many thanks!

Whats the 5V powered circuit going to do.?? [curious]

I'll let you know.........

Bill

 

[Mr RB]

hi Roman,
Problem is that the two lines, +Vrect[unsmoothed] and 0V are swapped over for Reversing the engine.

So the 'rectifier' has to be active for forward and reverse.

My apologies! I missed that point.