Jump to content
Sign in to follow this  
Automate

AC power usage Booster Pack

Recommended Posts

There are some good design ideas in that Instructable, but the implementation is weak. For example the current sensors are biased to allow the ADC to read the AC signal they produce - that is OK. The problem is that the bias voltage is not measured, it is assumed to be 1/2 of supply, so there can be a DC offset in the current measurement. There is also no anti-aliasing filter on the ADC. The way AC voltage is measured also has a few problems.

 

Give me a few days and I'll draw up a schematic that fixes these design problems.

Share this post


Link to post
Share on other sites

That would be great oPossum!

 

A few things I have been thinking about.

 

1. I like how the design will handle either 120V or 240V split-phase but there is no guarantee that the voltage is the same on both legs of the 120V. Would it be possible to monitor AC voltage on both legs and then in software specify which of the two voltages to use for each current sensor input?

 

2. Similarly, it would be nice if you could monitor any mix of 240V and 120V circuits and not use dedicated inputs like the Instructables design. Then just use software to add two channels together for the 240V circuits.

 

3. Since the board only deals with 60Hz analog signals and to maximize flexablility, use a cable to connect the Lanchpad to the booster pack. This would allow the booster to more easily be used with other boards such as a Raspberry Pi just by using a different custom cable.

 

4. Does it make sense to remove the transformer from the board so there is no 120V on or near the booster pack?

Share this post


Link to post
Share on other sites

TI has the great appnotes and code library for building AC usage monitoring and measurement systems.

http://www.ti.com/lit/an/slaa494/slaa494.pdf

http://www.ti.com/tool/msp430-energy-library

They are based on low cost analog front-end for electricity metering MSP430AFE2xx. I'm alredy used it in my own smart plugmeter and highly recommend.

 

But actually I don't think what LP + AC usage Booster Pack is suitable for that. Because there have a deal with high voltages and there is a risk of voltage damage without the proper enclosure and isolation.

Share this post


Link to post
Share on other sites

@Sasha, thanks for posting the links they have some good info. If you compare the TI design to the Instructables, a big difference is the voltage transformer used in the Instructales design. With this transformer the only high voltage is on the primary side of the transformer, nothing else needs high voltage. This voltage transformer does not need to be mounted on the booster pack. It could be a stand alone transformer like many other electronic projects.

 

The other big difference is the TI design is suitable for power companies that are using certified power meters that are required to have accuracies in the 0.5 to 0.2 percent range. For the Booster Pack I am thinking of 1 to 2 percent accuracies after calibration.

Share this post


Link to post
Share on other sites

If someone wants higher accuracy they could go with the MSP430F2013 DIP in the LaunchPad to get 16-bit resolution instead of 10 but then you only get 4 channels instead of 8.

Share this post


Link to post
Share on other sites

@Automate, look on this. It's safe, accuracy and has low cost BOM. Isn't this device is a beautiful? :)

 

post-34929-135135556917_thumb.jpg

post-34929-135135556926_thumb.jpg

post-34929-135135557196_thumb.jpg

Seriously, why you think it should be a booster pack? Surface area of booster pack less then Arduino shields, MSP430 value series litle bit weaker than ATmega328 used by guys from openenergymonitor.org and in last, I doubt that sketches from emonTx would work now with Energia and LP (may be I'm wrong). Where is a profit?

Share this post


Link to post
Share on other sites

@Automate

 

1. I like how the design will handle either 120V or 240V split-phase but there is no guarantee that the voltage is the same on both legs of the 120V. Would it be possible to monitor AC voltage on both legs and then in software specify which of the two voltages to use for each current sensor input?

While it is possible to monitor them separately, it is generally not necessary because the electric company provides the guarantee to provide 120VAC and 240VAC steady-state, and the fact that they are split phase ensures that they are in phase (or rather, 180 deg. out of phase). However, it may be more beneficial to use per-phase voltage measurement on a three phase system, where the alternative is to ensure that you have the phases ordered correctly and just calculate the phase shift mathematically.

 

2. Similarly, it would be nice if you could monitor any mix of 240V and 120V circuits and not use dedicated inputs like the Instructables design. Then just use software to add two channels together for the 240V circuits.
This is certainly possible; the only drawback is a slight reduction in the attainable resolution (by half) of the 120V measurements. A circuit set up to measure 240VAC can also measure 120VAC.

 

3. Since the board only deals with 60Hz analog signals and to maximize flexibility, use a cable to connect the Lanchpad to the booster pack. This would allow the booster to more easily be used with other boards such as a Raspberry Pi just by using a different custom cable.
I don't see any issue with this.

 

4. Does it make sense to remove the transformer from the board so there is no 120V on or near the booster pack?

It depends on the application. In general, it is preferred to avoid letting 120V near the Boosterpack/Launchpad whenever possible to mitigate the potential shock hazards. However, if a board is designed specifically for use/interaction with AC power (e.g. an AC relay board, or the watt hour meter shown in the link to TI's app note earlier in this thread), then there is a necessary risk that can be evaluated and appropriately mitigated. Certainly there are several commercially available Arduino AC relay shields that interact with AC power, so AC power on such a board is not unheard of. Many of these provide isolation between the high voltage and low voltage sections which help to mitigate risks. The TI app note does not provide isolation between the high voltage and the MSP430, but there is an implicit assumption that the end-user product will be appropriately packaged to avoid high voltage exposure (either in a metal enclosure that is tied to the safety/chassis ground or in a plastic housing) The images of the metering device alsenin created uses a plastic housing around a microcontroller powered transformerlessly which uses wireless for communications, so the end result has no exposed high voltages. It all boils down to making sure you think about safety and making the result as safe as possible for the intended application.

Share this post


Link to post
Share on other sites
The TI app note does not provide isolation between the high voltage and the MSP430, but there is an implicit assumption that the end-user product will be appropriately packaged to avoid high voltage exposure (either in a metal enclosure that is tied to the safety/chassis ground or in a plastic housing)

I'm sorry, but that appnote provide isolation. If you mean isolation of metering IC, then yes, currently only Maxim (Teridian) has suitable technology isolating of metering IC from the sensors (sensors as a current shunts and voltage sensors).

 

post-34929-135135557219_thumb.jpg

Share this post


Link to post
Share on other sites
The TI app note does not provide isolation between the high voltage and the MSP430
I'm sorry, but that appnote provide isolation.

 

http://www.ti.com/lit/an/slaa494/slaa494.pdf

 

Looking at Fig. 1, there is definitely an MSP430 on the left side of the electrical isolation boundary, right alongside the AC mains. While there is isolation though digital isolators going between that MSP430 and another MSP430 on the right side of Fig. 1, the point I was trying to make is that there are times when the MSP430 or Boosterpack may to be non-isolated from high voltage, and that safety precautions should be used anytime this is done.

 

In the case of the board picture you included (wherein the MSP430 on the high voltage side is toward the bottom-center of the board), they've clearly marked the bottom half of the board with labels like "HIGH VOLTAGE AREA", "CAUTION", and "CAUTION UNSAFE AREA". If I were to prototype a circuit like this using two Launchpads and two custom Boosterpacks, I would designate one as a high voltage "do not touch" Boosterpack and one as the low voltage, electrically-isolated Boosterpack that is safe to touch.

Share this post


Link to post
Share on other sites
There are some good design ideas in that Instructable, but the implementation is weak. For example the current sensors are biased to allow the ADC to read the AC signal they produce - that is OK. The problem is that the bias voltage is not measured, it is assumed to be 1/2 of supply, so there can be a DC offset in the current measurement. There is also no anti-aliasing filter on the ADC. The way AC voltage is measured also has a few problems.

 

Give me a few days and I'll draw up a schematic that fixes these design problems.

The Instructables design in my OP is based upon the OpenEnergyMonitor design. They seem to have some accuracy problems http://openenergymonitor.org/emon/emontx/accuracy

Hopefully your design would improve on this.

Share this post


Link to post
Share on other sites

Here are two design proposals. Up to 6 of each can be stacked to increase channel count.

 

The first uses a 12 bit ADC and has 2 channels that can be used for voltage or current and 5 channels that are current only. This could be used for single or split phase applications. Accuracy would be poor at low load due to limited ADC resolution.

 

A 2.5 V reference voltage is supplied by the MSP430 and buffered by IC1A. The buffered reference is used by the ADC and also divided by 2 and buffered by IC1B to create a bias for the sensors. The first ADC channel reads the bias voltage so it can be subtracted from the other 7 channels and produce a signed result representing voltages above and below 'zero'.

 

Each input has a simple passive first order low pass filter to reduce aliasing. A second or higher order filter would be better, but require many more components.

 

post-2341-135135558086_thumb.png

 

The second uses a 24 bit (max) ADC and has 3 channels that can be used for voltage or current and 3 channels that are current only. It could be used for single, split, or 3 phase applications. Accuracy would probably be limited by the voltage and current transformers used. It could approach 'utility grade' precision with careful hardware and firmware design.

 

Each input is balanced and ground referenced. TP or STP cable would be used to minimize pickup of unwanted signals in the wiring. Passive first order low pass filter to reduce aliasing.

 

The MCP3903 ADC is made for power monitoring applications and has hardware assist for phase lead/lag compensation. This would be a SMD design to ensure maximum performance - ground planes and guarding are critical for high resolution measurement.

 

post-2341-135135558096_thumb.png

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.

Sign in to follow this  

×
×
  • Create New...