Adjustable Voltage Regulator Step Down Power Supply

[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.

 

 

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.

[bertiedog]

I have just been researching this point on LM 317 based regulators, and yes if the voltage in is reduced the max output should go down in proportion. input in theory has to be 3volts above output to work.

Sounds an interesting unit, I will order one myself to test out. They are on Ebay as well....

The non zero voltage should be no issue with most locos, but some sensitive coreless may try to run on the 1.25 volts, but this can be cured by adding two diodes in the output before the reversing switch, but without the item in front of me I cannot say it can be altered easily. Normal motors do not mind 1.25 volts, they stop running at about 2 volts.

Assuming the specs are the same the input can be AC or DC, which would help setting the max Vout required.

Stephen

Edited September 29, 2017 by bertiedog

[Crosland]

It sounds like your nominal 16V is actually a few volts higher, due to the regulation of the transformer. Did you measure the output with a load or just using a multimeter? A 24V output plus the required headroom across the LM317 plus the diode drops in the bridge rectifier put it nearer 19V AC. it will drop under load.

 

An 8V AC supply will be somewhat less than 12V peak, less the drop in the bridge rectifier and the required headroom across the LM317 you might only get 9 or 10V, again depending on the regulation of the transformer.

[cpman46]

Stephen

 

Reading John's post I think he means he has added an external centre off, forward, reverse switch to the output of the unit. There is no built in switch if you look at the photos.

 

Regards

Mike

[bertiedog]

Stephen

 

Reading John's post I think he means he has added an external centre off, forward, reverse switch to the output of the unit. There is no built in switch if you look at the photos.

 

Regards

Mike

I had guessed that already, and read several other descriptions of the type to confirm, from Ebay. What they do fit is dual pots to make sure there is no lost contact as the pot is turned, as this can cause spikes of full voltage on the output.

One on order from Hong Kong! It's not worth building one from bits at under £4 post free!

Edited September 29, 2017 by bertiedog

[bertiedog]

With the old Tri-ang unit the simple test would be to set the Knob to eight volts DC , with a meter, and connect the board up to it, or simply set the Tri-ang to zero, connect up, and read the output as the control on the Tri-ang is increased till you get your expected output, with the boards knob in the max position.

The input to these boards are AC/DC so it does not matter which way round the lead from the Tri-ang unit is fitted.

Stephen.

[cctransuk]

With the old Tri-ang unit the simple test would be to set the Knob to eight volts DC , with a meter, and connect the board up to it, or simply set the Tri-ang to zero, connect up, and read the output as the control on the Tri-ang is increased till you get your expected output, with the boards knob in the max position.

The input to these boards are AC/DC so it does not matter which way round the lead from the Tri-ang unit is fitted.

Stephen.

 

Thanks for all of the input - I knew there would be a simple solution; I will do just that.

 

From my, admittedly electronically-challenged perspective, these compact units have a lot of potential as walkabout controllers.

 

I'm thinking a project box, with the pot shaft protruding from the face, and a two-way, centre off rocker switch above it for directional control.

 

I currently use H&M Walkabout controllers, (superb), but they're getting long in the tooth and something a little more up-to-date does appeal.

 

Regards,

John Isherwood.

[sharris]

It sounds like your nominal 16V is actually a few volts higher, due to the regulation of the transformer. Did you measure the output with a load or just using a multimeter? A 24V output plus the required headroom across the LM317 plus the diode drops in the bridge rectifier put it nearer 19V AC. it will drop under load.

 

An 8V AC supply will be somewhat less than 12V peak, less the drop in the bridge rectifier and the required headroom across the LM317 you might only get 9 or 10V, again depending on the regulation of the transformer.

 

Looking at the photo in the original post, there appears to be a bridge rectifier at the input, so the 16V AC power is being rectified - remember your meter will show you RMS, so 16V AV RMS and will rectify to about 22.6V DC(I'd guess the 16V is actually slightly higher, maybe about 19V AC if you're getting  24V DC out of the regulator). 

 

The LM317 has a head-room of up to 3V - that's the difference between the input voltage (after rectifying in this case) and the maximum output voltage you can get out of it - you'll have to take this into account if you're scaling down the input.

 

Assuming you have enough input voltage you won't get less that 1.25V out - this is part of the design of the LM317 voltage - as was pointed out above, the easiest way to get to 0 if you need it is with 2 rectifier diodes in series. Remember this will knock about 1.2V off the maximum voltage too!

 

An 8V AC input will provide the regulator with about 11.3V DC after rectification and smoothing, so allowing for head-room you might only get a maximum of about 8V DC out of the regulator. 

 

Considering the head-room and RMS conversion, you would need an 11V AC RMS power supply to guarantee 1.25V - 12V at the output, or a 12V AC input if you want to use series diodes and to guarantee a 0-12V output. 

 

[Edit] 

I just realised I didn't take into account the voltage drop across the bridge rectifier (about 1.4V), so add an extra volt AC at the input! 

Edited September 29, 2017 by sharris

[Junctionmad]

 

 

I have just been researching this point on LM 317 based regulators, and yes if the voltage in is reduced the max output should go down in proportion. input in theory has to be 3volts above output to work.

Hold your horses here folks , its a voltage regulator 

 

You cant program the output range by adjusting the input voltage ( thats the whole point of a voltage regulator ) 

 

The LM317 has an internal 1.25 voltage ref, The output voltage is set as a function of that voltage ( and that voltage only ) , Wired conventionally  the Output voltage  is Vref/(1+r1/r2) where r1 is connected from ADJ to output and and R2 is from ADJ to GND. ( ignoring the ADJ current ) 

 

The input must be 3V above the output for the device to regulate , ie hold the voltage steady 

 

If you drop below 3V ( difference Vin Vout )  , the device will not function correctly 

so the simple choice ,  replace R1 with a value calculated to give you 12V ,  then you just need any AC input  ( approx 0.7 AC RMS  depending on capacitor , so for 12V , you need say 16 with some headroom , i.e. 11V RMS AC , MINIMUM. In practice with loads , a transformer outputting  16 VAC would be suitable 

 

Be careful about larger input voltages as high current low output voltages can exceed the power dissipation of the device easily , i.e. short circuits etc 

 

 

Technically , the output of a LM317 , is 1.25+ whatever voltage is applied to the ADJ pin , assuming the current flowing in the adjustment circuit is well above the 50uA offset current. 

 

Note if the aliexpress circuit is wired conventionally , the transfer function between rotating the pot and the output voltage will not be linear, but that may not be an issue in practice 

Edited September 29, 2017 by Junctionmad

[sharris]

I currently use H&M Walkabout controllers, (superb), but they're getting long in the tooth and something a little more up-to-date does appeal.

 

 

Amusingly the H&M Walkabout controllers were introduced in the 1980s, while the LM317 appeared in 1976. 

[bertiedog]

Hold your horses here folks , its a voltage regulator 

 

You cant program the output range by adjusting the input voltage ( thats the whole point of a voltage regulator ) 

 

The LM317 has an internal 1.25 voltage ref, The output voltage is set as a function of that voltage ( and that voltage only ) , Wired conventionally  the Output voltage  is Vref/(1+r1/r2) where r1 is connected from ADJ to output and and R2 is from ADJ to GND. ( ignoring the ADJ current ) 

 

The input must be 3V above the output for the device to regulate , ie hold the voltage steady 

 

If you drop below 3V ( difference Vin Vout )  , the device will not function correctly 

so the simple choice ,  replace R1 with a value calculated to give you 12V ,  then you just need any AC input  ( approx 0.7 AC RMS  depending on capacitor , so for 12V , you need say 16 with some headroom , i.e. 11V RMS AC , MINIMUM. In practice with loads , a transformer outputting  16 VAC would be suitable 

 

Be careful about larger input voltages as high current low output voltages can exceed the power dissipation of the device easily , i.e. short circuits etc 

 

 

Technically , the output of a LM317 , is 1.25+ whatever voltage is applied to the ADJ pin , assuming the current flowing in the adjustment circuit is well above the 50uA offset current. 

 

Note if the aliexpress circuit is wired conventionally , the transfer function between rotating the pot and the output voltage will not be linear, but that may not be an issue in practice 

But the lm is quoted as a variable input device, lowering the input (above the minimums), must surely reduce the Vout in proportion.

If say six volts in is used then six volts out is the max, ( less the 1.25 loss)

[bertiedog]

Thanks for all of the input - I knew there would be a simple solution; I will do just that.

 

From my, admittedly electronically-challenged perspective, these compact units have a lot of potential as walkabout controllers.

 

I'm thinking a project box, with the pot shaft protruding from the face, and a two-way, centre off rocker switch above it for directional control.

 

I currently use H&M Walkabout controllers, (superb), but they're getting long in the tooth and something a little more up-to-date does appeal.

 

Regards,

John Isherwood.

Don't forget if you feed the unit from the Tri-ang DC variable, you will get no output at all under approx 3 volts in, and the lm 317 kicks in,

Stephen.

Edited September 29, 2017 by bertiedog

[Junctionmad]

But the lm is quoted as a variable input device, lowering the input (above the minimums), must surely reduce the Vout in proportion.

If say six volts in is used then six volts out is the max, ( less the 1.25 loss)

 

Stephen , its a voltage regulator with a fixed reference voltage in the feedback loop. Lowering the input voltage HAS NO EFFECT on the output voltage UNTIL  you force the regulation into dropout , and that point both line and load regulation are lost , in effect the device simply acts as a series pass unit, ( but the transfer function is not documented as far as I can see ) . What I mean is the Vout=Vin-2V in dropout, but the voltage will not vary because of the resistors , i.e. its not controllable  in the top 3V or so ( i.e. its not regulating ) 

 

I mean if you want that , just get a 2n3055 !

 

The advantage of using the LM317 is that when operating within its regulated region , the output voltage will not fall depending on output current 

Edited September 29, 2017 by Junctionmad

[cctransuk]

Hold your horses here folks , its a voltage regulator 

 

You cant program the output range by adjusting the input voltage ( thats the whole point of a voltage regulator ) 

 

The LM317 has an internal 1.25 voltage ref, The output voltage is set as a function of that voltage ( and that voltage only ) , Wired conventionally  the Output voltage  is Vref/(1+r1/r2) where r1 is connected from ADJ to output and and R2 is from ADJ to GND. ( ignoring the ADJ current ) 

 

The input must be 3V above the output for the device to regulate , ie hold the voltage steady 

 

If you drop below 3V ( difference Vin Vout )  , the device will not function correctly 

so the simple choice ,  replace R1 with a value calculated to give you 12V ,  then you just need any AC input  ( approx 0.7 AC RMS  depending on capacitor , so for 12V , you need say 16 with some headroom , i.e. 11V RMS AC , MINIMUM. In practice with loads , a transformer outputting  16 VAC would be suitable 

 

Be careful about larger input voltages as high current low output voltages can exceed the power dissipation of the device easily , i.e. short circuits etc 

 

I knew that, as soon as you guys got into this, I'd be out of my depth !!

 

The Tri-ang power pack outputs 19v AC at the nominal 16v AC terminals; 17v DC at the nominal 12v DC uncontrolled terminals; and 7v - 17v (reversible) at the controlled 12v DC terminals.

 

I can tell you that, from having tried it, powered from the 19v AC terminals, the voltage regulator's on-board digital output indicator shows 1.12v - 22.2v DC as the regulator pot is rotated.

 

Using the 17v DC terminals on the Tri-ang unit, exactly the same output is indicated on the voltage regulator.

 

Using the 7v - 17v DC controlled terminals on the Tri-ang unit, the voltage regulator indicates a maximum output range of 14.6v - 22.0v DC; the minimum output in all cases is 1.12v DC.

 

So, it would appear that, using the lowest 7v DC controlled output from the Tri-ang unit, a usable output range of 1.12v - 14.6v DC can be achieved with the voltage regulator.

 

I will test this out - but, on the face of it, a lower AC output transformer of, say 8v - 9v, may well provide the desired DC output range of 1.12 - 12v DC.

 

As to the advice that series diodes can produce a 0v minimum output from the voltage regulator, can someone provide a VERY BASIC circuit diagram that will indicate were the diodes should be placed, and what their spec. should be, please?

 

Regards,

John Isherwood.

[Junctionmad]

Thanks for all of the input - I knew there would be a simple solution; I will do just that.

 

From my, admittedly electronically-challenged perspective, these compact units have a lot of potential as walkabout controllers.

 

I'm thinking a project box, with the pot shaft protruding from the face, and a two-way, centre off rocker switch above it for directional control.

 

I currently use H&M Walkabout controllers, (superb), but they're getting long in the tooth and something a little more up-to-date does appeal.

 

Regards,

John Isherwood.

This solution will not work well  , once Vin to the LM317 falls below around 2V for small loads, Vout - VIN-2V, but the pot will in essence have no effect until turned down to allow the device to operate with the lowered VIN, i.e. there is a 1.25V discontinuity until the device regulates again. 

Edited September 29, 2017 by Junctionmad

[Junctionmad]

I knew that, as soon as you guys got into this, I'd be out of my depth !!

 

The Tri-ang power pack outputs 19v AC at the nominal 16v AC terminals; 17v DC at the nominal 12v DC uncontrolled terminals; and 7v - 17v (reversible) at the controlled 12v DC terminals.

 

I can tell you that, from having tried it, powered from the 19v AC terminals, the voltage regulator's on-board digital output indicator shows 1.12v - 22.2v DC as the regulator pot is rotated.

 

Using the 17v DC terminals on the Tri-ang unit, exactly the same output is indicated on the voltage regulator.

 

Using the 7v - 17v DC controlled terminals on the Tri-ang unit, the voltage regulator indicates a maximum output range of 14.6v - 22.0v DC; the minimum output in all cases is 1.12v DC.

 

So, it would appear that, using the lowest 7v DC controlled output from the Tri-ang unit, a usable output range of 1.12v - 14.6v DC can be achieved with the voltage regulator.

 

I will test this out - but, on the face of it, a lower AC output transformer of, say 8v - 9v, may well provide the desired DC output range of 1.12 - 12v DC.

 

As to the advice that series diodes can produce a 0v minimum output from the voltage regulator, can someone provide a VERY BASIC circuit diagram that will indicate were the diodes should be placed, and what their spec. should be, please?

 

Regards,

John Isherwood.

Sorry . John but 

 

So, it would appear that, using the lowest 7v DC controlled output from the Tri-ang unit, a usable output range of 1.12v - 14.6v DC can be achieved with the voltage regulator.

 

 

are you saying fixing the voltage at 7vdc gives you 14.6vdc output 

 

sorry, too many words, 

 

The answer is that it will work for the purposes that you want , i.e. controlling a motor , but the top 1.25 volts will not be regulated , which may make little difference in practice

 

Note at high load current and low output voltage , the heatsink may be entirely deficient 

 

personally Id adjust the resistors and have a regulated output from 1.25 to 12V ( i.e. using the full dynamic range of the pot)

 

will the short circuit and temp sensing work in drop out  ?, is that characterised anywhere 

Edited September 29, 2017 by Junctionmad

[cctransuk]

Sorry . John but 

 

are you saying fixing the voltage at 7vdc gives you 14.6vdc output 

 

sorry, too many words, 

 

The answer is that it will work for the purposes that you want , i.e. controlling a motor , but the top 1.25 volts will not be regulated , which may make little difference in practice

 

Note at high load current and low output voltage , the heatsink may be entirely deficient 

 

personally Id adjust the resistors and have a regulated output from 1.25 to 12V 

 

will the short circuit and temp sensing work in drop out  ?, is that characterised anywhere 

 

Yes - with a 7v DC input the output is 1.12v - 14.6v DC, depending on the setting of the regulator's pot.

 

Can you indicate which resistors should be adjust, and what their revised value should be, please? Remember, you're talking to a numpty as far as electonics is concerned.

 

What diodes should be fitted where to produce a minimum output of 0v DC?

 

Regards,

John Isherwood.

[kevinlms]

Yes - with a 7v DC input the output is 1.12v - 14.6v DC, depending on the setting of the regulator's pot.

 

Can you indicate which resistors should be adjust, and what their revised value should be, please? Remember, you're talking to a numpty as far as electonics is concerned.

 

What diodes should be fitted where to produce a minimum output of 0v DC?

 

Regards,

John Isherwood.

Don't do this. Using a voltage input from the output from an old Tri-ang power supply (or any such poor quality source) isn't going to make a reliable/good controller. Why, because the Tri-ang's output will be all over the place, as its a current limiting device. Its how all resistance controllers work.

 

You might as well just use the old Tri-ang controller & chuck your new bit of kit in the bin!

[Junctionmad]

 

Yes - with a 7v DC input the output is 1.12v - 14.6v DC, depending on the setting of the regulator's pot.

 

unless Im misreading the english, thats impossible 

Edited September 29, 2017 by Junctionmad

[bertiedog]

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.

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.

My own use is to run 3 to 4 volt motors, and the lm317 with reduced input at approx 6 volts does just that, and if the diodes are added to get nearer zero, then about 9 volts does the trick to run Busch Feldbahn locos..

Edited September 29, 2017 by bertiedog

[cctransuk]

unless Im misreading the english, thats impossible 

 

Nonetheless, the output of the Tri-ang unit, at it's lowest setting on the pot, is read as 7v DC using a multi-meter.

 

At that input setting, the output readout on the voltage regulator's own digital read-out is 1.12v to 14.6v, depending on the setting of the regulator's pot.

 

This is confirmed by connecting the regulator's output to track, and a loco is controlled within the full range of the regulator pot, without excessive speed at the top end of the pot range.

 

I have NO idea of the processes going on - I can only report what I observe.

 

Regards,

John Isherwood.

Edited September 29, 2017 by cctransuk

[Junctionmad]

Nonetheless, the output of the Tri-ang unit, at it's lowest setting on the pot, is read as 7v DC using a multi-meter.

 

At that input setting, the output readout on the voltage regulator's own digital read-out is 1.12v to 14.6v, depending on the setting of the regulator's pot.

 

This is confirmed by connecting the regulator's output to track, and a loco is controlled within the full range of the regulator pot, without excessive speed at the top end of the pot range.

 

I have NO idea of the processes going on - I can only report what I observe.

 

Regards,

John Isherwood.

Something is wrong , what you report can't be happening , the lm317 can't boost the voltage.

 

What's the output of the tri-ang, is it unsmoothed DC. I'm not familiar with it.

 

Note that the top of the range , i.e. From 13.0 approx to 14.6 that would require 2V above that as Vin ,

 

Be interested to read the input to the lm317 rather then the triang ouput

 

And that voltage 13.0 to 14.6 is unregulated

 

Ps. It struck me this is all at zero load , the dropout at zero load is actually very small on the lm327 , be interesting to see it with a xo4 motor attached ?

 

Correction I see you have done that

 

I have a bunch of lm317s in a drawer ( hasn't everyone ) , I'm going to breadboard it up Sunday , and put a scope on it. It's tweaked my curiosity. I have a controller using a lt3081 which is a much improved device but this is rather weird.

Edited September 29, 2017 by Junctionmad

[bertiedog]

The Tri-ang unit sounds a bit iffy in operation, it may be the 16 vac is not independent of the feed to the Tri-ang controller and I wonder if the rectifiers are at fault, is it as old as using selenium rectifiers, (they will have gone with age but silicon diodes would replace them)

 

I would have thought that there were two tappings on the same transformer or two independent windings, one for the 16 VAC and one to feed the controller.

 

If it reports 7 volts DC at the lowest setting it must be something wrong with the rectifiers. By all means put a load of an old motor on and check the voltage, and see if the motor runs, if it does, again the rectifiers must be replaced.

 

It all depends on how old the unit is, they were well sealed up, but can be opened. In theory the unit can be left as is an used for 16 VAC, but the old selenium rectifiers may get worse and failure creep in, I have seen over heating due to this.It only takes a couple of quid to fit a bridge rectifier, or if Tri-ang used two rectifiers replace with two Diodes.

 

Once identified does any body on here know for certain the wiring diagram of an older Tri-ang unit? Transformers rarely go wrong , unless grossly over loaded or damaged physically, or attacked by damp.

 

Stephen

[sharris]

Likewise I don't know what comes out of the Triang 'DC' output - full wave rectified, half wave rectified?

 

I suspect that whatever it is, it's not smoothed, so the peak voltage may be enough to give the measured output from the regulator circuit - there appears to be a large capacitor in the regulator circuit that might be 'filling in the gaps' between peaks - at least for low current loads. An unsmoothed half or full wave rectified output may be confusing things - your meter may be showing the RMS of the signal rather than the peak voltage.

[cctransuk]

Likewise I don't know what comes out of the Triang 'DC' output - full wave rectified, half wave rectified?

 

I suspect that whatever it is, it's not smoothed, so the peak voltage may be enough to give the measured output from the regulator circuit - there appears to be a large capacitor in the regulator circuit that might be 'filling in the gaps' between peaks - at least for low current loads. An unsmoothed half or full wave rectified output may be confusing things - your meter may be showing the RMS of the signal rather than the peak voltage.

 

All the above is way above my head.

 

The Tri-ang unit has, prior to the latest 'playing' with the voltage regulator, been used to provide power to my test track via two, switchable alternative supplies.

 

Supply 1 is from the nominal 16v AC (actually 19v AC) terminals of the Tri-ang unit to a H&M Walkabout controller, the output from which gives excellent control of mechanically sophisticated loco mechanisms.

 

Supply 2 is from the nominally 12v DC controlled terminals of the Tri-ang unit direct to track; this gives less precise control but more 'umph' for less mechanically sophisticated loco mechanisms.

 

I have seen no evidence whatsoever that the Tri-ang unit is malfunctioning - it is actually branded Tri-ang Hornby RP14.

 

Perhaps I should explain that I am in the process of upgrading my older locos from old-fashioned (ie. Tri-ang X04 / Hornby Dublo) mechanisms to more up-to-date Mashima / Mitsumi plus Ultrascale / High Level mechanical  systems - hence the 'before and after' power / control arrangements. All of my loco stock - approximately 100 units, operate entirely satisfactorily via either Supply 1 or Supply 2 - though the former gives far more sophisticated control at lower speeds.

 

My interest in the voltage regulator unit is purely academic. It is my (very basic) understanding that a low voltage DC motor can be controlled either via a fixed voltage and a variable current limiter (resistor), or via a variable voltage - is this correct?

 

IF that is the case, I was interested to see how a variable voltage supply would perform vis-a-vis a traditional variable current-limited supply.

 

I am probably way off course - but nothing ventured, nothing gained (learned).

 

Regards,

John Isherwood.