This page is for any photos that may be of general interest. As well as supplying standard Mk2 units, there is plenty of other activity going on behind the scenes. Many of the systems that I build are 'specials' as agreed with the customer.
If you're interested in anything that you've seen here, or you have a suggestion for another way of doing something, do drop me a line.
This is my rig for testing an output board to which the triac has not yet been attached.
The pre-tested triac and PCB can be soldered together at later time when their mechanical 'fit' is known.
Without the triac in place, this board can be posted as a "Large Letter" :)
Otherwise it's a "Small Parcel" :(
The 'gate' signal for the triac is connected to the upper end of R2 by a flying wire.
To aid this process, I leave this leg slightly raised on any output boards that are to be tested in this way.
It is vital that all of the triac's connections are sound, otherwise ...
... all of the load current goes via the trigger chip rather than the triac.
Users of Mk2 Routers have extended their CT and output cables by several tens of metres without any problems.
Here, the 4-digit display is connected via 30m of standard ribbon cable, and is working perfectly.
Here, a complete system is being tested on the bench without an enclosure.
By fitting a 5V regulator instead of a 3.3V one, an SSR can be used without requiring a transistor stage to boost the control voltage.
This system is now in Australia.
Here, a triac-based system is being tested without an enclosure.
By attaching everything to a single board, my standard sequence of tests can be applied, including calibration.
It should be easier for the customer to test this assembly as a complete system too.
Here is an equivalent 3-phase system without an enclosure.
A simple heatsink which is shared between the three triacs will allow initial testing to take place on site.
Here, a 6-channel driver board has been added to allow up to six SSRs to be driven at 5V.
All the data and power lines have been routed via a standard ribbon cable.
This approach allows the 3.3V RF module to be fitted at a later stage if required.
This system is also now in Australia.
NB. The later (green) PCB can support both 5V SSR(s) and the 3.3V RF module without requiring any off-board components.
This system supports one local load and three remote loads via CAT5 cabling.
A larger enclosure has been used to ensure plenty of space.
Because the 4-digit display was not needed, I've modified that area of the PCB to provide the extra connections.
NB. This system (above and below) was assembled before the new (green) PCB became available. When using the later PCB, connection points for additional loads are available without the need for any modifications.
Here is the same system with all its wiring in place.
The CAT5 and power cables appear through the rear of the enclosure which provides a neat finished appearance.
Each of the remote loads has a local LED (the red/yellow wires) to show when the load is active.
The Mk2 system that is being planned here does not have a display.
Instead, the CT which monitors the diverted power will be connected to an adjacent emonTx. The data can then be sent to the user's emonBase.
The inter-unit cable is a standard 3.5mm M-M coax lead.
Here I am building a prototype version of my 3-phase router.
The two smaller boards just provide the V and I sensors for L2 and L3.
My 3-phase PCB is more compact and simpler to assemble, with no need for any inter-connecting cables
Checking the layout for my new "main" PCB.
To minimise its overall width, this system has been assembled in 'portrait orientation.
The main board has been moved sideways a few mm to allow clearance for the heatsink screw.
One corner of the output board has been filed away to allow more room between the two boards.
This system (above and below) is configured to drive multiple outputs. By fitting the pin-saving hardware (IC3 & IC4), the five pins which are intended for the RF module are directly available for this purpose.
With the multiLoad Mk2 sketch, the external switch is used to determine the load priorities.
This switch selects between the normal output (at the "trigger" port) and physical load 1.
In this photo, an LED + series resistor can be seen at the "physical load 1" port (in the RFM12B area).
In the system shown above, the main PCB is attached using the
standard mounting points.
Another option for a potrait layout is to rotate the PCB within
the box as shown here.
In the photo below, I am testing the operation of the sketch Mk2_withRemoteLoad_4.ino before release.
The board on which the sketch is running is on the left ...
the output stage for the 'local' load is at the centre ...
and the RF-linked output stage for the 'remote' load is on the right.
To provide a realistic test environment, both output stages are wired through CT1.
CT1 encloses the Neutral (blue) core of the PV simulation circuit, and also the Live (brown) core of each of the output stages.
This particular CT started life with an internal burden resistor, but it has since been removed.
At first glance, this Router looks much the same as many others that I've built.
But this board has an extra feature ...
Because this system does not include the RF module, one of the spare IO ports is being used to measure temperature. The photo below shows how the three connections for the Dallas sensor have been made. These connections are: any IO port with a 4K7 pullup resistor to Vcc; ground; and Vcc.
Here, D10 is being accessed at the mid-point of the space for R23. The adjacent ground pin is also being used.
Vcc is accessed at one of the C9 pads, a (pink) 0.1" wire link being required at the VR2 point.
The pull-up resistor should be 4K7 rather than 3K9 as shown here (at side of the RFM12B zone).
The sketch Mk2_RFdatalog4.ino includes the measurement of temperature for dataloggoing purposes. The ability to measure temperature could also be used to provide an automatic top-up facility for a water tank if insufficient surplus energy is available.
By mounting the burden resistors on DIY adapters like these, the scaling of the current sensors can be easily changed.
These adaptors, with their 220R burdens, are for use with a set of larger CTs.
With this setup, the working range for each phase is approx 10 kW.
For the system below, a simple internal aerial has been made from a standard SMA extender cable. In this case, the SMA socket is nearer to the top of the box so the aerial runs downwards from the socket rather than up. This kind of aerial seems to work just as well in either orientation. According to Wikipedia, it is a quarter-wave monopole.
This system uses 868 MHz for which the 1/4 wave aerial length is 82 mm.
For a 433 MHz system, the 164mm aerial would probably need to protrude outside the box.
Red heat-shrink sleeving for the 230V Line terminals
Heat-resistant sleeving for any wire that's near the triac
A "load on" LED in the lid is connected via the SIL pins on the output board
and one from July 2015:
Although this system runs at 5V, a 6V transformer is fine. To maintain the size of the voltage waveform, R4 is 18K rather than 10K.
The temperature sensor is again at D10, via R23, this time with the correct value of pullup (4K7).
A lid-mounted "load on" LED is driven from D12, at the top of the RFM12B area.
The above system has an extra control feature.
When the red button is pushed, the most recent value from the temperature sensor is shown at the display.
Although not for the same system, the cables for these lid-mounted LEDs have been clipped to the underside of the lid.
The other end of these two cables go onto the SIL pins of the output boards to show when each of the loads is "on".
The next four photos are for a 3-enclosure fully wired system. Because the behaviour of the supply meter was already known, this system can operate permanently in Anti-flicker mode. Hence there is no need for an "output mode" switch. Testing how the supply meter behaves in the presence of a burst-mode power diverter can be easily achieved by borrowing one on my Demo Units.
Two output stages are connected to the control unit by 4-core cables.
A pair of switches allow the loads to be operated singly with switchable priority, or in parallel.
One of the output stages has been inverted so that its upper face can be seen.
Although one side of this box looks rather empty, the other one is feature packed.
The two loads and their local LEDs are driven from four of the IO ports at IC3.
The pair of switches for selecting how the loads should be operated are at IC3 pin 7 (D5) and the "mode" port (D3).
A local load could still be fitted later.
IC3 provides access to 5 IO ports, so I should only have fitted 10 pins, not 12
Each of the remote units has a CT which is linked via the control cable. The signal from load #2 is connected to port A3 at one end of R11.
The burden resistor is at R12 with its legs bent around under the board.
When using the R11 location as a third current sensor, the Vref signal must be extended to this point (blue wire).
The bent legs of the associated burden resistor, and the track that connects R11 to A0 (pin 23), can also be seen.
The next pair of photos are for a bespoke system which uses many of my standard components. It's for controlling the heating of Domestic Hot Water in a house with an Economy 7 tariff. By means of dedicated wiring, storage radiators are only supplied with power during off-peak periods, but the DHW heater is fed from a supply that is permanently live. Using a pair of RF-linked boxes, the DHW heater is now only activated during periods of cheap power. Previously, a timer did a similar job but its setting was always problematic, particularly when the "hour" changed.
This next photo is from New Zealand where the power supply for heating the hot-water tank can be interrupted by the utility company at any time via a "ripple relay". To avoid the router being isolated on such occasions, a low-power feed from a wall socket has been used to provide a permanent supply for the router's control circuitry.
During periods of surplus power, energy from the interruptable supply is diverted to the hot water tank by the Router's output stage. This cable carries Line, Switched Line and Earth, hence only a single high-power cable can be seen.
By means of a "cooker" connector, the high-power wiring for the immersion heater can be easily returned to a
The switch at the bottom allows the hot water tank to be topped up using cheap overnight electricity.
Here is another way of adding a second output stage.
For this system in Denmark, I just supplied the PCBs and a few of the main components.
It's always interesting to see what can be achieved with a bit of ingenuity.
The triac for the output stage is under the expensive looking heatsink.
Here, a novel approach has been used to equip one of my original PCBs with dual power supplies.
Note the three pairs of resistors which reduce the signal levels from 5V to 3.3V, circuit details are here.
These original PCBs are available for the cost of postage. Let me know if you would like one to experiment with :-)
This Mk2 system is quietly diverting surplus power in New Zealand where a commercial equivalent is in the region of NZ$700.
More recently, this system has had an interesting upgrade, as shown below. The underlying Mk2 Router structure has not changed, but a Raspberry Pi Zero has been fitted above the main board to intercept data from the Serial port. The data is reformatted using a Python script, and sent via a Wi-Fi dongle to a broadband router. The data is forwarded to a conventional datalogging website where it can be viewed at any time. The live data for this system is at: http://pvoutput.org/intraday.jsp?id=49355&sid=34477
NB. The generated power and ambient temperature graphs that can be viewed for this location (June 2016) are not sourced by the Mk2 Router. But "Mk2" contributions for Diverted Energy and Cylinder Temperature are soon to be included ...
The large black box is a switched-mode power supply for the Raspberry Pi Zero.
A DIN-rail alternative to terminal block has been used for the power cables.