Pre-regulator & post-regulator controll scheme.

[David_]

Hello.

I have designed a linear regulator(post-regulator) circuit currently using TPS7A4701 that is being controlled with a DAC to output between 0 and +30V, I also have designed a adjustable switching step-down regulator(pre-regulator) using LTC3864 which can output a voltage between 0.8V and the input voltage which in my case is 40V, but I plan to use it to output ~2-33V to allow the linear regulator stage to output 0-30V.

I don't know how I should control the pre-regulator though, one idea is to control it in the same way the linear regulator is controlled.
I am going to design in a current limit circuit of some kind and to st the limit I'll need another DAC output and I have a suitable quad DAC so I have outputs available to use for controlling the pre-regulator.

But the kinds I have seen have some circuits which by automation keeps the pre-regulator one or two volts above the post-regulator which sure does sound more convenient but I have not jet come up with a circuit to accomplish this. If I would use my first idea I also need to measure the output from that stage which amount to a unsuitable solution.

But I don't know if my circuit can be adapted to such a sheme, it works like this:


Might I use a op-amp to buffer the DAC control voltage and ether shift it up a little or just design the pre-regulators feedback resistors in such a way that it will output a higher voltage?
That does actually sound as a good idea, what do you all think?

 

[FvM]

A straightforward control scheme for the switched mode pre-regulator maintains constant voltage difference across the linear regulator. The topology has the advantage that it works independent of a particular setpoint, also tracks transistions between constant voltage and constant current etc.

 

[KlausST]

Hi David,

the post regulator has a given drop out voltage.
One way could be to for the pre-regulator to keep it maybe 2V (post regualtor drop out voltage + haedroom) above the post regulator´s output voltage.

This could ba achieved with some diodes or a zener and a PNP transistor.

The benifit is that even when the post regulator is in current limit (outut voltage is lower than set up with DAC) then the pre regulator voltage also drops and the power dissipataion is limited also.

The drawback is, that the risetime of the output voltage may be extended.

here a draft... please run a simulation to check loop stability.

(I expect good stablility when emitter resistor and collector resistor of PNP have the same value (unaltered loop stability) but then V(PO) = V(Out) + V(FB-pre) + 2.2V
Change count of Si diodes to change V(PO) )


Good luck

Klaus

 

[David_]

Re: Pre-regulator & post-regulator controll scheme.

Hello Klaus.

I am in the process of simulating the proposed circuit, however one aspect of my circuit raises a question.
The circuits surrounding the linear regulator is all driven by a 3.3V rail, and in the current configuration the pre-regulator can go below that, then I have the following options.

Ether drive a 3.3V rail in parallel with the pre-regulator which is feed by 39.6V or limit the pre-regulator to output 4-33V or there about so that the regulator driving the 3.3V always have enough voltage headroom.
There are regulators that can handle > 40V input and the DAC, ADC, reference and op amps don't draw much current so a parallel regulator for the control circuits might be best.
I should check so I know exactly how much current those circuits will need.

- - - Updated - - -

Unless I have missed something the control circuits will draw around 1mA. Even with 36.3V over a regulator the power dissipation will be minimal considering the 1mA load, something like 36mW. Then it sounds as a extra regulator will be the more suitable option.

 

[KlausST]

Hi David,

you could change the circuit, in a way that the output voltage of the pre-regulator allways is more than 3.5V (or so).

A low dropout 3.3V regulator could supply your circuitry.

Klaus

 

[David_]

I have spent a few hours simulating the purposed circuit but the end result is not clear to me, I have not been doing this long and I am not akin to any simulation software nor do I know what aspects might differ greatly from reality.
Though I would assume that pretty much anything could happen in a simulation.
Heres a picture of how the circuit looked like when I quite:


I had writen a long post discussing a few things but my browser scre**d me over and it all was lost leaving me in a very bad mood, where is the restore auto-saved content in advanced mode!
Anyway I'll re-write it later today cause my limit is reached.

Regards

 

[KlausST]

Hi,

from the feedback path i expect an output voltage of 1.5V more than V2 = 1.5V + 30.0V = 31.5V.
it should be stable. With the 10R load i expext an output current of 3.15A.

What does the simulation say?
I did not check the rest of the circuit.


Klaus

 

[David_]

Hello.
Sorry for the long delay i response, but it has been a long period of trying to learn about this stuff which has not been easy to say the least.

The thing was that during my simulation the results was just strange and I could not control the output at all, whatever I changed the output from the pre-regulator stayed within a 0,5V span.
During the simulation I keppt thinking of loop compensation, I knew that I needed to go into tht hole thing and find out if and how I would compensate over such a output span as 3-33V 0-3A and I was also thinking about the subject of output filter, a second output filter and I had read that problem with that might arise due to the control loop feedback which is something of a mystery for me so I endevoured to find out but there is a big gap in between my understanding of things and control-loop theory.
I find it unsuitable to proceed with the simulations and the control of the pre-regulator circuit until I have a stable ordinary regulator, I will start another thread concerning loop stability to solve that before continuing this topic.

Thanks all for you help.

Reagrds

 

[David_]

Hello.

I have changed the design in a rather big way to this:

from this app note:


And I think that this is a better way to do what I want, it is for sure cheaper and simpler.
I still need to work some more on it to be confident in its performans.

 

[FvM]

I would prefer a linear regulator with precise, load independent regulation.

 

[crutschow]

Your circuit will have poor regulation with load and temperature change due to the Darlington output stage, which is not inside the reglator feedback loop.
If you need more current you can parallel two or more LT3083's.

 

[David_]

Re: Pre-regulator & post-regulator controll scheme.

Okey, here is my hole circuit as it stands right now but I get the feeling that it might not work as good as hoped:


My other option is a real switching regulator as pre-regulator such as in the previews pictures in this thread, I have little experience and I do value your input.

- - - Updated - - -

The points just made about load regulation, should I have been able to work that out my self through LTspice?

- - - Updated - - -

My goal is to enable the noise specification of TPS7A4001 while drawing 3A, an alternative is TPS7A4701 that can output 1A unlike TPS7A4001 that can only handle 50mA.
And the output voltage range do demand a pre-regulator.