cupoftea
Advanced Member level 6
Hi,
We need to make a PSU for our application..
Vout 22-38V, (it can be anywhere in that range and may vary within that range though preferably remain pretty "flat" over say any 10 second period)
Iout = 2.2A max.
Isolated from mains. Mains will be primary power source.
No capacitance can be "seen" from the load(!!!)
So the output caps of the PSU must be "hidden" behind say an output linear regulator or whatever.
We would like the voltage to be as "flat" and unchanging as possible.
Most of the time, the resistive load will be fixed and unchanging.
In fact, a change in the load is only likely to ever happen once per week or so.
We were thinking of something like a current source which is used to give the voltage on the cable.
So perhaps a Current that is regulated so as to regulate the Vout. {i suppose thats kind of like a PFC stage, which regulates a current (input current),
in order to regulate the Vout}
So, We have a number of small resistive heaters which are connected at various points across the wires of a long 2-core cable. (up to 200m)
We must power these heaters. The cable voltage can be between 22V and 38V or so.
Each heater takes only some 3mA. But for 2 minutes each week, each heater will draw 10mA.
It will suddenly go from 3mA to 10mA in some few microseconds or so.
Any heater can at any time be switched on or off the cable.
But this isnt likely to happen more than once per day.
The maximum current that the cable has to deliver is 2.2A. It has to deliver this no matter if the voltage is anywhere between 22V to 38V.
I say 2.2A but this is the general average.....if regulation cant cope for intervals of a few seconds at a time then thats fine...as long as the long term average is 2.2A delivered.
There may be anything from zero to 220 heaters along the cable.
So any ideas? We will finish our current job then get started on this. It sounds easy? Just an iteratively adjusted current source?
The one thing we are certain of is that a standard frequency compensated PSU is out of the question.....the lack of output capacitance at the load would make such a thing challenging to control.
Would you agree?
And after all, the load is generally fixed and constant and there is no requirement for dramatic transient response etc etc.
The company that wants this PSU is saying that due to the "uniqueness and complexity" that any suitably designed PSU would possess, that we must therefore pay a £30k "NDA insurance" fee to the company,
and that they will subtract this from our fee. This is because they say that this design will be a "unique gem" and that they are in danger of us taking it elsewhere when we have completed the design.....
but surely this is just a bog standard iterative-set-and-check current source?......as simple as "bread and butter"?
What's with all the IP paranoia with folks these days? We are talking pretty well fixed load with no special transient response requirements.
Was thinking of a current source which can be incremented by a microcontroller in response to the micro reading the output voltage.
("iterative set and check").
It wouldnt be considered "cheating" to have say a 40V TVS in every heater so that if the whole cable's loads suddenly went from 10mA to 3mA simultaneously,
then the TVS's could "catch" the skyrocketing Vout before the micro got it back under control.
We need to make a PSU for our application..
Vout 22-38V, (it can be anywhere in that range and may vary within that range though preferably remain pretty "flat" over say any 10 second period)
Iout = 2.2A max.
Isolated from mains. Mains will be primary power source.
No capacitance can be "seen" from the load(!!!)
So the output caps of the PSU must be "hidden" behind say an output linear regulator or whatever.
We would like the voltage to be as "flat" and unchanging as possible.
Most of the time, the resistive load will be fixed and unchanging.
In fact, a change in the load is only likely to ever happen once per week or so.
We were thinking of something like a current source which is used to give the voltage on the cable.
So perhaps a Current that is regulated so as to regulate the Vout. {i suppose thats kind of like a PFC stage, which regulates a current (input current),
in order to regulate the Vout}
So, We have a number of small resistive heaters which are connected at various points across the wires of a long 2-core cable. (up to 200m)
We must power these heaters. The cable voltage can be between 22V and 38V or so.
Each heater takes only some 3mA. But for 2 minutes each week, each heater will draw 10mA.
It will suddenly go from 3mA to 10mA in some few microseconds or so.
Any heater can at any time be switched on or off the cable.
But this isnt likely to happen more than once per day.
The maximum current that the cable has to deliver is 2.2A. It has to deliver this no matter if the voltage is anywhere between 22V to 38V.
I say 2.2A but this is the general average.....if regulation cant cope for intervals of a few seconds at a time then thats fine...as long as the long term average is 2.2A delivered.
There may be anything from zero to 220 heaters along the cable.
So any ideas? We will finish our current job then get started on this. It sounds easy? Just an iteratively adjusted current source?
The one thing we are certain of is that a standard frequency compensated PSU is out of the question.....the lack of output capacitance at the load would make such a thing challenging to control.
Would you agree?
And after all, the load is generally fixed and constant and there is no requirement for dramatic transient response etc etc.
The company that wants this PSU is saying that due to the "uniqueness and complexity" that any suitably designed PSU would possess, that we must therefore pay a £30k "NDA insurance" fee to the company,
and that they will subtract this from our fee. This is because they say that this design will be a "unique gem" and that they are in danger of us taking it elsewhere when we have completed the design.....
but surely this is just a bog standard iterative-set-and-check current source?......as simple as "bread and butter"?
What's with all the IP paranoia with folks these days? We are talking pretty well fixed load with no special transient response requirements.
Was thinking of a current source which can be incremented by a microcontroller in response to the micro reading the output voltage.
("iterative set and check").
It wouldnt be considered "cheating" to have say a 40V TVS in every heater so that if the whole cable's loads suddenly went from 10mA to 3mA simultaneously,
then the TVS's could "catch" the skyrocketing Vout before the micro got it back under control.
