VC Buck Converter, will it work?

[Gasboss775]

I found this circuit online that appears to be an lm317 type adjustable regulator with thyristor based preregulator. It looks like it could work but there are no component values given, could anyone suggest possible values?

My input would be 24VAC, output perhaps 1.25 ~ + 20 volts if possible.

Thanks in advance.

 

[spec]

I found this circuit online that appears to be an lm317 type adjustable regulator with thyristor based preregulator. It looks like it could work but there are no component values given, could anyone suggest possible values?

My input would be 24VAC, output perhaps 1.25 ~ + 20 volts if possible.

Thanks in advance.

Looks interesting GB,

I will see if I can work out the circuit and suggest some values.

Would you really like the output to be 0V to 20V, rather than 1.25V to 20V?

spec

 

[ronsimpson]

Would you really like the output to be 0V to Vmax, rather than 1.25V to Vmax?

http://cds.linear.com/docs/en/datasheet/30801fc.pdf
I would change the LM317 to a LT3080. It is only good to 1A but you can parallel two to more. It can be adjusted to "0".
The LT3081 is for 1.5A.

 

[Gasboss775]

Thanks Spec & Ron, going down to zero isn't essential but could be useful.

 

[MrAl]

I seen a similar idea applied to another SMPS controller IC. Is there any reason why it won't work with an mc34063?

Hello,

I have a very similar circuit set up on my Li-ion charger, and have recommended this many times for adjusting the output voltage of a buck circuit that feeds a linear regulator. The whole idea is to keep the input to the linear (like an LM317) about 2v higher than the output. That keeps the whole thing much more efficient than using a linear alone.

Rather than use a resistor however, use a zener or string of diodes to set the voltage set point of the differential output of the buck, and monitor the output of the linear. When the linear voltage is turned up normally, the buck voltage then goes up automatically thus keeping the input to the linear 2 to 3 volts higher than the output which is required for normal operation of the linear.

Find a schematic drawing from year 2004 for the whole circuit in the attachment. Note the way the transistor is connected. Ignore the op amps and other regulators.

One thing you will find hard to do though is if you try to control the base drive manually with a pot or something, the set point will drift.

 

[ronsimpson]

going down to zero isn't essential but could be useful.

When you are powering something and it starts to smoke....it is nice to turn the power down to zero!

Look at #25 and see all the parts needed to get the LM317 to go to zero volts. (could be done better but)
That is why I like the LT part because SIMPLE.

I do PWMs every day and understand them so that is what I would do. Might even get one from ebay.com. I think the PWM from post #25 is on ebay.

 

[Gasboss775]

This application note from Linear Technology covers a number of ideas.

 

[MrAl]

This application note from Linear Technology covers a number of ideas.

Hi,

While those circuits are interesting, they are overly complicated. Take a look at the circuit in post #25 and compare, but ignore the op amps as they are for something else.

Basically all we need is the transistor and a few small parts because the input to the linear does not have to be super accurate. It is also the most stable because the transistor is much faster than anything else in the circuit so it has no effect on the dominant poles.

 

[ronsimpson]

MrAl on post #25.

I like the LM2576-adj. Good part and simple.
On the current limit:
http://cds.linear.com/docs/en/datasheet/3081fc.pdf
LT3081 has adjustable current limit and voltage can be set to zero.

Resistor on Ilim: 4.6k=1.5A, 3k=0.9A, 1.5k=0.4A (so a simple pot will work)
By putting a meter on Imon you can see the output current. (not a high current meter, just a simple meter) 1k resistor = 0.3V at 1.5A
You can also read the temperature in the same way on TEMP pin.
I would use the TO220-7 package.

 

[ronsimpson]

Vin to Vout on the linear regulator is 1.86V which is more than enough for this LDO regulator. (set by R2)
Current limit set by R8, Voltage set by R5. Voltage can be set to 0 volts.

 

[MrAl]

MrAl on post #25.

I like the LM2576-adj. Good part and simple.
On the current limit:
http://cds.linear.com/docs/en/datasheet/3081fc.pdf
LT3081 has adjustable current limit and voltage can be set to zero.
View attachment 102411

Resistor on Ilim: 4.6k=1.5A, 3k=0.9A, 1.5k=0.4A (so a simple pot will work)
By putting a meter on Imon you can see the output current. (not a high current meter, just a simple meter) 1k resistor = 0.3V at 1.5A
You can also read the temperature in the same way on TEMP pin.
I would use the TO220-7 package.

Hi,

Interesting linear regulator chip there, thanks for sharing.

Only drawback i guess is the cheapest one i can find is about 6 dollars USD. Compare to LM317 which can be had for about 50 cents. Special purpose circuits could benefit a lot from that newer chip though.

I generally stick with the lower priced parts but if i have to do something special i might make an exception. I might need a 5 dollar op amp for something i built for example.

 

[spec]

Hi all,

For about a year I have been meaning to do an article on 0V output standard three terminal regulators. A rough draft, which may be of interest, is now @ https://www.electro-tech-online.com...nal-regulator-adjusts-down-to-zero-volts.785/

spec

 

[spec]

View attachment 102413

Vin to Vout on the linear regulator is 1.86V which is more than enough for this LDO regulator. (set by R2)
Current limit set by R8, Voltage set by R5. Voltage can be set to 0 volts.

Pretty neat Ron.

I expect you know that you can get higher current versions in the LM3080 family:
Positive 3A: https://www.linear.com/product/LT3083
Negative 1.5A: https://www.linear.com/product/LT3091

spec

 

[ronsimpson]

I expect you know that you can get higher current versions in the LT3080 family:

I generally stick with the lower priced parts

Good points all.

For a bench power supply I think current limit is important and it should work to 0 volts. I like to charge 1.5V batteries or test the forward voltage drop of a diode.
Look at any $500.00 power supply.

I have not used the LT3081. Other members of the family can be paroled to increase current.
The current and temperature outputs work to Vout + 0.4V. So if you have a short, keep the Imon and Temp voltages under 0.4 volts. I have some little meter movements that work well with a full scale voltage of 0.3V.

Probably should have a reverse voltage protection diode on the output and maybe a small dummy load resistor.
I wish I had time to make a PCB. Maybe I could add one to my next project. I usually make break off boards and add them to something I am getting paid to do.

 

[MrAl]

Good points all.

For a bench power supply I think current limit is important and it should work to 0 volts. I like to charge 1.5V batteries or test the forward voltage drop of a diode.
Look at any $500.00 power supply.

I have not used the LT3081. Other members of the family can be paroled to increase current.
The current and temperature outputs work to Vout + 0.4V. So if you have a short, keep the Imon and Temp voltages under 0.4 volts. I have some little meter movements that work well with a full scale voltage of 0.3V.

Probably should have a reverse voltage protection diode on the output and maybe a small dummy load resistor.
I wish I had time to make a PCB. Maybe I could add one to my next project. I usually make break off boards and add them to something I am getting paid to do.

Hello again,

Yeah that temperature measurement output was a really good idea. That could tell the operator that it is overheating. It's about time somebody started thinking about how to make basic power regulator chips better.
The other thing i was hoping to see some day is a switching buck regulator with an internal overkill MOSFET. It seems most chips today use border line decent MOSFET's internally when they could make one with a very low Rds MOSFET even for lower currents like 1 amp. If we use a 1 amp MOSFET for a 1 amp design, the voltage drop might be almost that of a bipolar, so why not jump up to a 10 amp MOSFET for a 1 amp output and then be able to brag about having a much higher efficiency buck regulator. Yeah a little more money, but LT already doesnt skimp on design just for the sake of pricing, apparently

 

[dr pepper]

Mr Al, with regards to the circuit in post #25, do you think you could just have a diode feeding a cap and then use a zener to generate an approx -1.25v supply for the Lm317.
Would the reference within the Lm317 be good enough this way, or do you need the op amp and regs to generate a stable negative rail.
I ask as I put something toghter using a Lm723 and used a simple - rail using a diode / cap / zener and it works well, allthough I needed a trimmer to set the output 0v.

 

[MrAl]

Mr Al, with regards to the circuit in post #25, do you think you could just have a diode feeding a cap and then use a zener to generate an approx -1.25v supply for the Lm317.
Would the reference within the Lm317 be good enough this way, or do you need the op amp and regs to generate a stable negative rail.
I ask as I put something toghter using a Lm723 and used a simple - rail using a diode / cap / zener and it works well, allthough I needed a trimmer to set the output 0v.

Hi,

Well the most stable would be using a reference diode to generate the minus voltage reference.
You can get away with say two diodes in series or a zener or something, but the temperature drift will be higher. So it depends how temperature stable you need the adjustment to be. If you adjust it to say 5.00v and you want it to stay that way (within the spec of the LM317 anyway) then if you use diodes or a zener you'll see a little change in output with temperature, but with a reference diode you'll see less. Diodes are often assumed to have a -2.2mv per degree C temperature characteristic. That would mean over a 10 degree external temperature change you'd see roughly 22mv increase in output voltage. That might not be that bad really. That's of course with constant current.
The LM317 has a temperature characteristic that also affects the design anyway, and although it is not that much for light loads for heavier loads it will be more because the reference diode for the LM317 itself is INSIDE the same package as the pass transistor and as the pass transistor heats up so does the reference diode. That causes more output change than would be expected if the reference diode was external to the LM317 package. The data sheet does not mention this they just give the specs on the temperature characteristic as if the diode was outside the package which is a little misleading, because it also depends on load current and heat sink thermal resistance.
All this means the design should be checked for proper operation with the required min and max loads, in a real life breadboard as a simulation will not show this either.
I might be able to find some numbers for this i'll have to find my notes.

BTW with the negative reference R1 should be something like 150 ohms to ensure the minimum load current is always present even when adjusting down to 0v output. That also means the diode or zener or ref diode bias has to be able to handle something like 10ma.

 

[dr pepper]

Ok makes sense.

 

[spec]

Crude picture!
PWM is simple MC34063. Vin, Vout, ground, and FB shown. (some not shown)
The PWM will work hard of FB<1.25 volts.
The PWM will stop working if FB>1.25V.
To get 1.25V across the 1K resistor current must flow from the transistor. This same current will cause 2.50V across the 2K resistor.
To get the transistor turned on, the voltage across the LM317 must be about (2.50+0.65=3.15).
If you don't like the 3.15V number change either resistor.

View attachment 102211
The 100 ohms on the base is not critical. Could be 1k.

I hadn't appreciated how your circuit worked Ron- pretty neat converting the voltage drop across the LM317 into a current which can then be referenced to ground.

Another advantage is that the voltage drop is easily changed by changing a resistor and also the circuit does not introduce a whole load of voltage gain which could make the system difficult to stabilize in the frequency domain.

spec

 

[Gasboss775]

One thing that I've thought of but haven't tried yet is using an LM337 to provide the -1.25 V reference so the LM317 can go down to zero, two diodes would be cheaper, however I do happen to have a few LM337.