Adjustable Voltage Regulator Step Down Power Supply

[Crosland]

For a start I do not quite understand where the 24volts is coming from in the original post about these unless the transformer is delivering more than 16 vac rms. The bridge rectifier will increase it but not by quite that much.

 

You have all the answer in your post. The 16V is RMS. When rectified and smoothed it will be 16 X 1.414 less the voltage drop in the rectifier.

 

Anyway I breadboarded several versions of the regulator and reducing the input does reduce the max V out as expected, It is non linear though.

If you mean the output varies as you vary the input then the regulator is not regulating anything.

 

A regulator produces a fixed output regardless of the input until the input is less than the output plus the "drop out voltage" (a datasheet spec for every regulator). If the input is reduced further then the output will indeed reduce but it is not being regulated.

 

Edited September 30, 2017 by Crosland

[cctransuk]

 

You have all the answer in your post. The 16V is RMS. When rectified and smoothed it will be 16 X 1.414 less the voltage drop in the rectifier.

 

If you mean the output varies as you vary the input then the regulator is not regulating anything.

 

A regulator produces a fixed output regardless of the input until the input is less than the output plus the "drop out voltage" (a datasheet spec for every regulator). If the input is reduced further then the output will indeed reduce but it is not being regulated.

 

 

Then why is this voltage regular provided with both a pot adjuster and a digital output voltage indicator?

 

I am assuming that the device is intended to produce aa adjustable, regulated output voltage at the indicated voltage.

 

Or am I just too much of an ignoramus to understand even that fairly basic premise.

 

Anyway - me as a layman the device does appear to be able to function as a model train controller, and I will source a 240v / 9v transformer to see if an input from this will produce an adjustable output from the device of 1.12v - 12v (or thereabouts) DC.

 

I really would be most grateful if someone could indicate in diagram for where and what series diodes can be added to the circuit in order to produce a 0v DC minimum output; (as suggested above).

 

Regards,

John Isherwood.

[Brian]

Each diode drops approx. 0.7volts. So two together in series drops approx. 1.4volts, 3 in series drops 2.1volts etc.

Use any diodes from the 1N540x range e.g. 1N5401

 

When measuring the output volts at the full output setting it will always be more than the input volts when an AC voltage is inputted, as the large smoothing capacitor on the regulator rises the AC input volts to 1.41 times that read on a standard AC volt meter. In other words what your reading on the multimeter is the AC RMS input voltage, but the capacitor gives the Peak voltage. Minus the loss occurring in the bridge rectifier which is converting the AC to DC, so 1.4volts is lost before reaching the capacitor.

Lets assume a 16v AC input voltage is read, then the Bridge rectifier will drop this by around 1.4volts and the capacitor increase it by 1.41 which is to say 16 -1.4 times 1.41 = 20.58 volts at maximum output setting on the pot. Reducing the pot will in turn reduce the output volts.

 

Drawing show diodes and a reversing switch in the output of the regulator.

 

 

Edit to correct drawing.

2nd edit to correct maths!

 

Edited September 30, 2017 by Brian

[sharris]

1N4001 through 1N4007 have a maximum forward current of 1A - I'd be tempted to use something a bit more beefy - specially if you're using old-style motors. The 1N5401 series have a forward current of 3A.

 

The LM317's over-current protection won't kick in until at least 1.5A, so might well be able to blow 1N400x diodes if there is a short or heavy load.

[Brian]

Yes a good point. The 1N5401 or any from the 1N540x range would be a better diode to use in this instance as they are 3.0Amp rated.

Drawing corrected to reflect this :-)

[cctransuk]

Each diode drops approx. 0.7volts. So two together in series drops approx. 1.2volts, 3 in series drops 2.1volts etc. 

Use any diodes from the 1N540x range e.g. 1N5401

 

When measuring the output volts at the full output setting it will always be more than the input volts when an AC voltage is inputted, as the large smoothing capacitor on the regulator rises the AC input volts to 1.41 times that read on a standard AC volt meter. In other words what your reading on the multimeter is the AC RMS input voltage, but the capacitor gives the Peak voltage.  Minus the loss occurring in the bridge rectifier which is converting the AC to DC, so 1.2volts is lost before reaching the capacitor.   

Lets assume a 16v AC input voltage is read, then the Bridge rectifier will drop this by around 1.2volts and the capacitor increase it by 1.41 which is to say 16 -1.2 times 1.41 = 20.8 volts at maximum output setting on the pot. Reducing the pot will in turn reduce the output volts. 

 

Drawing show diodes and a reversing switch in the output of the regulator.

 

 

Edit to correct drawing.

 

Now that is exactly what I need - thanks very much indeed.

 

I have ordered an https://www.rapidonline.com/vigortronix-vtx-126-020-212-chassis-transformer-230v-20va-12v-12v-88-3922 , which should be able to drive a couple of these adjustable voltage regulators, with an output variable voltage which I calculate to be :-

 

input 12v AC - 1.2v (bridge rectifier) = 10.8v DC  x 1.41 (capacitor) = 15.2v DC - 1.4v (2 No. diodes) = 13.8 v DC.

 

So, with the diodes in place the voltage regulator should output 0 - 14v DC, whic I would have thought should be OK; (my H&M Walkabouts output 0.1 - 14.6v DC).

 

I will report back once the transformer is to hand.

 

Regards,

John Isherwood.

[Brian]

My maths is getting worse! 2 x 0.7 is not 1.2! Its 1.4.

However my error isnt going to make that much difference to what you're after.

Of now to my man cave whilst chanting "2 times 0.7 is 1.4".

[cctransuk]

My maths is getting worse! 2 x 0.7 is not 1.2! Its 1.4.

However my error isnt going to make that much difference to what you're after.

Of now to my man cave whilst chanting "2 times 0.7 is 1.4".

 

I have also ordered a vented PSU case, power supply jacks / sockets and grommets for the transformer, plus a hand-held enclosure, control knob, diodes, and a DPDT centre-off rocker switch for the regulator.

 

If this project works, it should produce quite a neat and light, hand-held walkabout controller.

 

Regards,

John Isherwood.

[DCB]

It looks good system but I wonder about overload protection.  I don't know which Triang controller you use but I have several different and most have a 1 amp nominal cut out on the 12 volt DC with controlled and uncontrolled DC off the same cut out. This is a push button reset.  I don't remember if the AC has a cut out but the unit can't deliver that many VA anyway probably 12VA hence the 12 volt 1 amp branding.  Max voltage is actually around 19 V DC some are more.

A feature of using a voltage regulator as a speed controller is very good speed stability up hill and down dale, and the ability to provide a lot of Amps at low speed. very different to resistance controllers like Duettes and that ik.   I don't know what sort of PSU you have acquired but plenty of PSUs happily deliver 5 plus Amps. so you need a good overload cut out to protect your locos, especially coreless motor ones and a good heat sink/ airflow Maybe a small fan in the handheld enclosure?.  

[cctransuk]

It looks good system but I wonder about overload protection.  I don't know which Triang controller you use but I have several different and most have a 1 amp nominal cut out on the 12 volt DC with controlled and uncontrolled DC off the same cut out. This is a push button reset.  I don't remember if the AC has a cut out but the unit can't deliver that many VA anyway probably 12VA hence the 12 volt 1 amp branding.  Max voltage is actually around 19 V DC some are more.

A feature of using a voltage regulator as a speed controller is very good speed stability up hill and down dale, and the ability to provide a lot of Amps at low speed. very different to resistance controllers like Duettes and that ik.   I don't know what sort of PSU you have acquired but plenty of PSUs happily deliver 5 plus Amps. so you need a good overload cut out to protect your locos, especially coreless motor ones and a good heat sink/ airflow Maybe a small fan in the handheld enclosure?.  

 

See my posts #14 & #25 re the Tri-ang Hornby power unit that I have been using for breadboarding; it does not have an overload cut-out - it just buzzes!

 

See post #31 re the transformer that I have ordered; the vented enclosure that I have also ordered is rated as adequate for the va rating of the transformer.

 

The hand-held unit will house only the voltage regulator and DPDT centre-off switch, plus the diodes; for testing purposes it will be fed from the transformer unit via a lead that plugs into an output socket on the transformer enclosure. Ultimately, it will connect to the baseboards via a 5 pin DIN plug / socket, which will provide 12v AC input to the voltage regulator and variable regulated 0 - 12v DC output to the track.

 

The regulator has a heat sink - do you feel that the enclosed hand-held unit will require fan venting?

 

Where and by what do you suggest the current overload should be provided - I am a total novice at this kind of thing.

 

The transformer is rated at 20va with two separate 12v AC outputs. I did wonder whether a 50va rated transformer would be better, but it would have needed a much larger vented enclosure, which I had difficulty sourcing. If experience shows a need for a higher rated transformer, we are only talking £8.00 or so, so this shouldn't be a problem.

 

Any other advice will be gratefully received.

 

Regards,

John Isherwood.

[Junctionmad]

A few things to bear in mind as full wave bridge rectifier with smoothing capacitor is quite a bit more complex then the layman understands

 

Firstly the Traffo is 12vac , this is the rms voltage at full rated load , the regulation of that transformer is 12% , which indicates the no load voltage will be about 14.5 VAC RMS

 

The peak voltage as specified by Brian , is the theorectical peak , note this is not the DC voltage , and n fact is determined by the load current and the smoothing capacitor , your actual dc output could be much lower

 

Personally the whole things is a design running outside its envelope. It's fine for coarse motors etc which can largely handle significant variations and where ripple is actually sometimes useful.

 

As a controller for small low current motors , it's simply a bad idea

[cctransuk]

A few things to bear in mind as full wave bridge rectifier with smoothing capacitor is quite a bit more complex then the layman understands

 

Firstly the Traffo is 12vac , this is the rms voltage at full rated load , the regulation of that transformer is 12% , which indicates the no load voltage will be about 14.5 VAC RMS

 

The peak voltage as specified by Brian , is the theorectical peak , note this is not the DC voltage , and n fact is determined by the load current and the smoothing capacitor , your actual dc output could be much lower

 

Personally the whole things is a design running outside its envelope. It's fine for coarse motors etc which can largely handle significant variations and where ripple is actually sometimes useful.

 

As a controller for small low current motors , it's simply a bad idea

 

I'm afraid the tech. speak goes over my head - I did say that I am a complete electronic novice.

 

My stock of around 100 locos are powered by anything from the Hornby Dublo / Tri-ang X-04 era mechanisms (with five pole armatures), through Mashima and Mitsumi motors on Highlevel gearboxes or Ultrascale gearsets, Pittman motors, Romford motors, MRRC motors, and many RTR mechanisms from 30+ years ago up to the present day, including current production with coreless motors.

 

This experiment is merely that - in order to see if voltage regulator control can at least match the performance of my aging H&M Walkabouts - which, IMHO, offer the ultimate in control sensitivity.

 

If it fails, the component parts will find other uses in time.

 

Regards,

John Isherwood.

[bertiedog]

 

You have all the answer in your post. The 16V is RMS. When rectified and smoothed it will be 16 X 1.414 less the voltage drop in the rectifier.

 

If you mean the output varies as you vary the input then the regulator is not regulating anything.

 

A regulator produces a fixed output regardless of the input until the input is less than the output plus the "drop out voltage" (a datasheet spec for every regulator). If the input is reduced further then the output will indeed reduce but it is not being regulated.

 

But the LM series is not a fixed regulator, its regulation is set by the components values on the output, and reducing the input max voltage to a lower figure, (fully covered in specs), just alters the regulated output Max. It just works at the lower input levels (above the min allowed) to regulate the voltage output within the range set by the max input.

A fixed value regulator is different, you must be above the output or it does not work, but LM is OK as long as the input, at whatever value is above the output.

If the input has to be at the full allowed value to get a regulated output, then why do the makers specify an input range?

Forgetting the drop out and loss, if you feed 6 volts in you can get 0 to 6 volts out fully regulated at every set value, thats the whole point of a variable regulated voltage source. The LM317 works on 6 volts in, and gives a fully controlled range of voltages out..

[dhjgreen]

John,

 

Can you tell us the value of the potentiometer. If they have used the design recommended in the data sheet it should be 5k ohms.

[Crosland]

Then why is this voltage regular provided with both a pot adjuster and a digital output voltage indicator?

 

I am assuming that the device is intended to produce aa adjustable, regulated output voltage at the indicated voltage.

 

Or am I just too much of an ignoramus to understand even that fairly basic premise.

 

No, I wasn't totally clear. Once you set the output with the pot then the regulator will regulate the output to that voltage, so long as the input is higher by a few volts.

 

If you want to use it as a variable supply to drive trains, for example, then the input voltage must be greater than the highest voltage you want to get out of it.

 

If the input has to be at the full allowed value to get a regulated output, then why do the makers specify an input range?

Forgetting the drop out and loss, if you feed 6 volts in you can get 0 to 6 volts out fully regulated at every set value, thats the whole point of a variable regulated voltage source. The LM317 works on 6 volts in, and gives a fully controlled range of voltages out..

 

If the makers did not specify a range then some idiot would feed in 50V and wonder why they let the smoke out. It does not mean it is designed to work with a variable input voltage.

 

You cannot forget the drop out. A linear regulator can never provide a regulated output equal to the input.

 

So, to get 0 - 6 out you will need a fixed 9V in to satisfy the LM317 drop out voltage of 3V.

 

To get 0 - 24V out you will need 27V in, etc, ...

[dhjgreen]

John,

 

Can you tell us the value of the potentiometer. If they have used the design recommended in the data sheet it should be 5k ohms.

I can actually read the the values in the pictures. The pot is a dual 10k which they describe as a design feature, giving 5 k as they are in parallel. The design does use the datasheet recommendation and the reference resisitor is 240 ohm (can't read the colours, but the value is screen printed on).

 

If you replace the dual 10 k with a dual 5k you will have a regulated 14 V supply ( provided you meet Crosland's head room suggestions)

 

Edit to add: before you embark on building your transformer box, do read this

Edited October 1, 2017 by dhjgreen

[bertiedog]

The reason the dual is used is that a single track could suffer a minor glitch as it urns, like crackle on a wireless volume control, but in the regulator circuit this is very problematic, as the gliich can send the regulator output to full voltage for the period of the glitch, dependant on exact circuit used. So they double the value and in parallel it gives a virtually glitch free potentiometer.

This trick was used widely in High end Hi-Fi, with dual or treble wipers built in to the pot, plus in Marantz they dualled the pots as well.

[Junctionmad]

But the LM series is not a fixed regulator, its regulation is set by the components values on the output, and reducing the input max voltage to a lower figure, (fully covered in specs), just alters the regulated output Max. It just works at the lower input levels (above the min allowed) to regulate the voltage output within the range set by the max input.

A fixed value regulator is different, you must be above the output or it does not work, but LM is OK as long as the input, at whatever value is above the output.

If the input has to be at the full allowed value to get a regulated output, then why do the makers specify an input range?

Forgetting the drop out and loss, if you feed 6 volts in you can get 0 to 6 volts out fully regulated at every set value, thats the whole point of a variable regulated voltage source. The LM317 works on 6 volts in, and gives a fully controlled range of voltages out..

You cannot feed in 6v and get Ov to 6v out regulated.

 

Just to repeat , when in regulation, the output voltage IS NOT defined by the input voltage. Again Vout = vref( 1+ R2/R1 and for regulation to function Vin must be about 3V above vout , ( due to drops across the series pass circuit )

 

Again Stephen where is Vin referred to in that formula

 

To see what's really happening . Let's start with Vin at say 20vdc and the resistor divider programmed to provide 12vdc

 

As you reduce Vin , Vout WILL NOT reduce, until the regulator drops out of regulation at around 15v approx . The output voltage will then fall in proportion to the input voltage , accordingly to Vout ='Vin -3v ( less dropout for lower currents ) and if the resistors aren't reprogrammed the device will remain non functional as Vin continues to fall.

 

But at any given Vin , changing the pot to specify a valid Vout below the dropout threshold will again cause the device to function as a regulator for that programmed voltage.

 

So Vout only tracks Vin when the device isnt actually working at all

 

What confuses me is as Vin falls, the dynamic range of the pot should reduce , i.e. You end up with less angle to play with , at 2.5k the output is around 13v , and this is around half the pot , hence reducing the Vin to 15v , would mean that half the pot is lost , because any resistance above 2.5k would cause the device to drop out. Hence as this circuit set up for 20V , would suggest that, at 13v output the top 50%. of the pot range should be lost , ie once you get to 50% of the pot angle the output is saturated

 

The proper way to use this board, would be to replace R1 with a 560 ohm , with a 5k pot this would give 12.5 approx Vout at max pot travel ( assuming a 5k pot ) . Note that also the pot angle to Vout isn't linear , i.e. at 2500, ie mid way , the output voltage is 7v , but probably good enough for model trains !

 

Oh course with R1 at 560 doesn't allow the minimum current to pass to ensure regulation, so a simple expedient of a 2k resistor from output to gnd will ensure that. The series diodes can be added to produce near zero output as mentioned earlier

 

The dc voltage need to be simply 3 volts above 12.5 to produce a fully functional system , for safety 5V above would be better assuming there will be ripple current.

 

Just be aware at low voltages say 3V , at 500mA, with say 18v dc input , the power dissipation is 7.5 watts , no way that heat sink will handle that

 

That's one of the issues with series pass regulations , a simple solution is to,use a second LM317 as a tracking regulator in essence splitting the power dissipation over two devices ( or whack on a big heat sink )

 

Note that as far as I can tell , both short circuit and chip temperature limits do not work when the device is in dropout , hence shorts/over currents would be passed back to the transformer and potentially chip junction temps could be exceeded resulting in magic smoke

Edited October 1, 2017 by Junctionmad

[bertiedog]

But the point is 12 volts out is not required by me, just 6 volts with about 9 volts in, with appropriate resistors, it works! This range is fully within spec.

[Junctionmad]

But the point is 12 volts out is not required by me, just 6 volts with about 9 volts in, with appropriate resistors, it works! This range is fully within spec.

Sure , I was referring to the use of the OP board, with its resistors programmed for 20v

[Crosland]

But the point is 12 volts out is not required by me, just 6 volts with about 9 volts in, with appropriate resistors, it works! This range is fully within spec.

 

So why didn't you just say that, instead of claiming 6V in would give 0 - 6V regulated out? Maybe the 6V in was a typo?

[sharris]

Mine goes all the way up to 11

[cctransuk]

Having read some mention of adjustable voltage regulator step down power supplies, I ordered and have taken delivery of one of these - https://www.aliexpress.com/item/LM317-Adjustable-Voltage-Regulator-Step-Down-Power-Supply-Module-With-LED-Meter/32618386190.html?spm=a2g0s.9042311.0.0.svpAoZ.

 

Module.jpg

 

Testing it, I found that with a nominal 16v AC supply from an old Tri-ang control unit, the minimum output voltage is 1.25v DC, and the maximum is 24v DC.

 

The inability of the unit to output 0v does not seem to affect my locos and, combined with a cente-off, forward / reverse switch, the unit seems to have potential as a fairly sensitive control unit for 4mm. scale.

 

However, the maximum voltage output of 24v is excessive - approximately twice what I need.

 

Knowing nothing about electronics, my instincts suggest that  an 8v AC supply should halve the maximum output to 12v - but I suspect that not to be the case.

 

Can anyone with some electronics knowledge advise whether the lack of a 0v output will cause unforseen problems, and whether it is possible to adjust the maximum output to 12v or thereabouts, without making major amendements to the unit?

 

Many thanks in anticipation.

 

Regards,

John Isherwood.

 

I am pleased to report that I have completed and tested my controller based on the adjustable voltage regulator https://www.aliexpre...2311.0.0.svpAoZ

 

 

To power it, I installed a 240 V AC to 2 x 12V AC transformer https://www.rapidonline.com/vigortronix-vtx-126-020-209-chassis-transformer-230v-20va-9v-9v-88-3921 in a snug enclosure https://www.rapidonline.com/evatron-fe2-ltl-case-psu-ventilated-small-black-30-3180 , and fitted two power jack outlets https://www.rapidonline.com/truconnect-cn-dcj-002c-dc-power-socket-2-1mm-20-1094 . A mains plug and lead completed the job.

 

The controller comprised an enclosure https://www.rapidonline.com/hammond-rl6035-deep-lid-abs-hand-held-instrument-enclosure-144-x-65-x-40-grey-30-0026 used side-on; ie. with the split between the two halves sandwiching the control shaft of the regulator. The regulator was bolted into the centre of the enclosure using the fixing holes provided on the regulator circuit board; the control shaft protruding through a hole drilled on the joint between the two halved of the enclosure.

 

I installed a two-way / centre-off rocker switch at a reverser / isolator, and wired up the components, including the two diodes https://www.rapidonline.com/dc-components-1n5401-3a-100v-silicon-rectifier-diode-47-3144 , as per Brian's excellent diagram.

 

 

For testing purposes, I installed two two-core leads - a supply lead terminating in a power plug https://www.rapidonline.com/truconnect-cn-dcp-001a-2-1mm-standard-dc-power-plug-20-0905 , for connecting to the power supply unit; and a track supply lead terminating in a five-pin DIN plug - the standard controller connection to my test track. I intend to retrofit a coiled extending lead to combine the power input and track supply functions via the five pin DIN interface.

 

Testing the variable voltage regulator controller has proved to be very interesting. As the variable voltage adjustment operates over around 300 degrees of rotation, I had expected that using the control knob would be somewhat laborious; not so, each loco that I have tested seems to use around 90 - 120 degrees of rotation, depending upon its individual voltage requirements.

 

For instance, a Heljan GWR 1361 0-6-0ST starts to move at around the 9 o'clock, and reaches its desirable maximum speed at around the 12 o'clock position.

 

A Dapol BR Class 121 single car DMU starts to move at around the 11 o'clock position and reaches its desirable maximum speed at around the 2 o'clock position.

 

A pair of Mainline GWR 2251 0-6-0s, fitted with Perseverance / High Level / Mitsumi chassis start to move at around the 10 o'clock position and reach their desirable maximum speed at around the 1 o'clock position.

 

Thus each individual loco can be controlled very precisely within a specific sector of the control knob rotation.

 

A major benefit of this method of control is that, without exception, all locos tested have run MUCH more quietly. Quite a few of my locos have a decided 'growl' when controlled by my Hammant & Morgan Walkabout controllers; the 'growl' disappears when the variable voltage regulator controller is used. I can only assume that the 'growl' is generated by the particular wave form produced by the H&M controllers.

 

All in all, I'm very pleased with the results obtained from the variable voltage regulator controller, and would encourage others to try one.

 

Regards,

John Isherwood.

[bertiedog]

All solved very neatly then, the noise you were getting is the mains frequency left after rectification by the HM units.

Stephen

[Junctionmad]

As I outlined a single resistor change will restore 300 degree pot rotation and hence even finer control, should that be desired