EEVblog Electronics Community Forum
Electronics => Projects, Designs, and Technical Stuff => Topic started by: prosper on May 07, 2021, 02:29:31 am
-
I'm thinking about building an adjustable lab supply.
Ultimately, I want it to be a dual independent supply, so that I can connect each channel up in series (to get a bipolar supply), or as two isolated voltages.
I've found a design that I'm going to poach (at least as a starting point), which incorporates a microcontroller to monitor and adjust the voltage and current limit.
Is there some way I can use a single uC to control both channels? Given that they (each channel) don't share a common reference voltage, I don't see how I could use a single uC to interface with both sides of the supply. I *could* just throw in two uC's - they're cheap enough - and I could connect them together through an opto-coupler.
Attached is the design - at least for one of the channels. Second channel is basically a mirror of the first
-
I think its common for power supplies that have a tracking mode that tracking is only used when the supplies are used in series or parallel. For instance, my Sorenson LT 30-3 has the following modes of operation:
- both channels independent and floating wrt each other
- parallel mode, voltage and current set by the "master" channel
- series mode, voltage and current set by the "master" channel
That is, there is no "tracking but independent floating" mode.
-
Thanks - it's good to know what commercial units are doing.
-
...
Is there some way I can use a single uC to control both channels? Given that they (each channel) don't share a common reference voltage, I don't see how I could use a single uC to interface with both sides of the supply. I *could* just throw in two uC's - they're cheap enough - and I could connect them together through an opto-coupler.
...
Use a dual optocoupler and opamp in a feedback loop to "linearize" the phototransistor current.
You'll still need a voltage reference per isolated supply (zener enough?).
That is, adapt something like this: https://www.radiolocman.com/shem/schematics.html?di=514065 (https://www.radiolocman.com/shem/schematics.html?di=514065)
Another article about using linear optocouplers but the recommended ones are pricey.
https://www.digikey.com/en/articles/linear-optical-isolation-for-safe-sensor-operation (https://www.digikey.com/en/articles/linear-optical-isolation-for-safe-sensor-operation)
Or, make your own optocoupler with an LED+LDR in black heat-shrink tubing. Commercially known as a Vactrol. Popular with modular synths. It's not fast.
Otherwise, dual optocoupler, SPI DAC, and voltage reference, per isolated supply.
Oh, I just noticed current and voltage read-back. You might as well just use another uC then.
-
the current sense, at least, is done with a hall effect sensor. So that part should be isolated. The rest... maybe not.
I'd never thought of an optoisolator as a linear component, but, I'll definitely look into that
come to think of it, I'm planning to use filtered PWM to set the voltage and current levels, and an optocoupler can pass PWM. So, I think the threshold setting, and reading the output current is doable. It's just reading the voltage output that I need to think more about.
-
You could use an external ADC/DAC and isolate the digital interface, which gives you the option of better precision/accuracy vs the ones built-in to the MCU as well. Or keep the design as-is and move the UI to a separate MCU with isolated serial interface to the two power supply control MCUs.
-
I think its common for power supplies that have a tracking mode that tracking is only used when the supplies are used in series or parallel. For instance, my Sorenson LT 30-3 has the following modes of operation:
- both channels independent and floating wrt each other
- parallel mode, voltage and current set by the "master" channel
- series mode, voltage and current set by the "master" channel
That is, there is no "tracking but independent floating" mode.
It depends on the implementation/manufacturer. On both H24005 and BB3 we implemented tracking (https://www.envox.eu/eez-bench-box-3/bb3-user-manual/9-system-functions/#bb3_man_tracking) when output parameters can be set on two or more channels at the same time. In the case of coupling (https://www.envox.eu/eez-bench-box-3/bb3-user-manual/9-system-functions/#bb3_man_coupling) in series and parallel this is also done and then the firmware also does load balancing.
-
Is there some way I can use a single uC to control both channels? Given that they (each channel) don't share a common reference voltage, I don't see how I could use a single uC to interface with both sides of the supply. I *could* just throw in two uC's - they're cheap enough - and I could connect them together through an opto-coupler.
Well, that *is* the problem, isn't it?
The usual solution is to galvanically isolate the digital control circuits from the ADCs and DACs of each channel, and provide a separate voltage reference for each channel.
An acceptable alternative in my opinion is to implement a floating dual tracking supply with common ground, so no isolation is required and the control circuits operate on two outputs which allows series operation but not parallel or floating (from each other) operation.