LM 350 adjustable supply design issue

[krs]

Hi,

I need to build a LM 350 variable supply to deliver 0 to 24 volt at the output.

The standard circuits with R2 as the pot have a basic design problem in that the output voltage will jump to 24 volts should the wiper of the pot loose contact even momentarily.

I recall seeing a circuit configuration where the R1/R2 combination is modified, maybe another resistor added to prevent that situation - ie the output voltage will drop if the wiper of the pot looses contact.

Has anyone come across this on the Net?

 

[kchriste]

Connect the unused leg of the pot to the wiper. Then if the wiper goes open the output will go to the max setting and not beyond. A better solution is to make R1 the POT (With a suitable series resistor to limit Vmax) and R2 the fixed value. Then if the pot goes open, the voltage will go near zero instead of max. This will also prevent a noisy POT from spiking a circuit that you only want to run at say 10V with 24V spikes. The limitation with this configuration is that the min-max voltage adjustment is more restricted and you will dissipate more power in R1/R2. With either design, the output voltage cannot be adjusted to zero.

 

[krs]

Thanks for the reply.

Is your first option not the standard way to hook up the LM 350? One leg of the pot to the LM 350 adj terminal and the wiper and other leg of the pot both connected to ground.
**broken link removed**

That would allow the output voltage to jump to the maximum setting determined by R1 and R2 if the wiper would go open which is not what I was looking for..

I think the second option of making R1 the pot is more to my liking although I thought there was a way to keep R2 as the pot and design a simple resistor network around it so that the output voltage will drop when there is an open wiper.

I just can't get my head around this to figure out how that could possibly work.

 

[Hero999]

Don't worry about it, that circuit does work.

Keep the bottom resistor a pot, the top resistor can't be a pot as it will short circuit it when turned down to the lowest resistance setting.

 

[k7elp60]

If you are worried about the adjustment pot going bad and causing the output to go to high, you could use a circuit that is known as a crowbar.
In simple terms it has a fuse between the filter capacitor and the the regulator input. A SCR anode connected to the input of the regulator and it't cathode to the common of the regulator ckt. The gate is connected to a voltage sense circuit on the output. The voltage sense circuit is set to trigger the SCR when the output of the regulator goes to a predetermined upper limit. When this happens the SCR triggers and blows the fuse cutting off the input to the regulator.

 

[krs]

I looked at a crowbar circuit, but I was hoping for something simpler.

In a crowbar circuit, I would have connected the SCR across R2 to essentially bring adj down to ground. Reset could probably be done by shorting out the SCR or with a NO push button switch in series.
Or one could short the output of the regulator. That would really be fail safe as well.

The problem is that this is a controller for a Garden Railroad.
We have had wipers loosing contact, the output jumps to 24 volts, the train takes off at full speed and if that happens just before a curve the engine derails and usually falls over.
Since these engines can cost anywhere from $1000 to $5000 and up, the owner of the engine isn't too happy when that happens.

 

[Hero999]

Don't use a variable power supply.

Build a fixed 24V supply and use a PWM speed controller that shuts off if the wiper looses contact.

 

[RDL2004]

Is this what you were thinking of?

"Change improves regulator's reliability"

**broken link removed**

 

[aussiepoof]

The problem is that this is a controller for a Garden Railroad.
We have had wipers loosing contact, the output jumps to 24 volts, the train takes off at full speed and if that happens just before a curve the engine derails and usually falls over.
Since these engines can cost anywhere from $1000 to $5000 and up, the owner of the engine isn't too happy when that happens.

I got around the problem by using a digital pot and controlling it with a microcontroller... software lets you be a *lot* more creative in dealing with errors! Oh, and when badly programmed creates more creative failure modes too

The digital pot/LM350/PIC combo I used is the heart of an automated block controller for DC locos. I designed it for my local model railway club and its new layout. We'll be running 30-40 trains automatically for exhibitions with just 2 people overseeing operations. The system also uses RFID tags to cater for each loco's specific performance curves.

Like you, we've had problems with previous control systems that would result in trains racing, derailing and generally creating havoc. In one instance the motor of a particularly sensitive loco was burned out.

Let me know if I can be of any help.

Regards,
Paul

 

[audioguru]

Your unregulated input voltage to the LM350 will be 32VDC. The LM350 begins reducing its output current when it has more than 10V across it so the output current might start reducing at voltage settings of 22V and less. At a voltage setting of 2V the minimum current rating is only 250mA. It is a range of currents so you might be lucky to get one that limits the current to 1A. Then the LM350 will try to dissipate 30W and will melt. No, it will just shut down when it gets too hot.

Your main filter capacitor is small for 3A so it will produce a high ripple voltage. Then at a low voltage setting the LM350 will vary the output current with the ripple.