Fullscreen HTML5Med Res (??MB)HTML4

High Current Solid State 12V Battery Isolator This device connects an auxiliary battery to the main vehicle battery/ alternator while the engine is running, charging that extra battery. But it disconnects it once the engine shuts down, so that the vehicle battery can’t accidentally go flat. It’s cheap and easy to build but also very robust. It’s ideal for RVs, campers, offroad vehicles and boats. I When the voltage drops, it detects that the engine has have had ongoing problems with the battery systems on my 4WD vehicles. My car is fitted with an auxiliary been stopped and breaks that connection. Not being at all happy with the commercial units I tried, 12V battery system that I use to run a fridge, some raI decided to design my own. dios, camping lighting etc. My design criteria were: I tried using a commercial battery isolator to connect it to the main vehicle electrical system but found that this • Low current drain from the main battery when the engine is off. had two major shortcomings. • Fully solid-state operation (no relays). Firstly, its case offered little protection from the elements, • A low forward voltage drop when switched on, minimisand it occasionally filled with water – not good. ing heating and power loss. Secondly, it uses two open-frame style relays to connect the batteries in parallel. The contacts in these relays • Must not interfere with radios (ie, no RFI/EMI). are nothing special and occasionally weld together, leav- • Must use commonly available parts. ing the batteries permanently connected. That can lead to • Must handle very high currents without damage (>100A). • A completely waterproof and dustproof housing. both batteries going flat. Also not good! These made the first design decision easy: Mosfets are an The idea of these isolators is to parallel the batteries when the engine is running and remove this connection ideal solid-state switching device for large direct currents. While P-channel Mosfets are easier to drive for highwhen the engine is off. So when you are camped overnight and you discharge side switching, N-channel Mosfets offer lower losses at the the auxiliary battery, you can still start the engine in the same price thanks to a vanishingly small ‘on-resistance’. So I decided upon six Infineon IRFS7434TRL7PP Mosmorning. It works by measuring the vehicle battery voltage, which fets, which have an on-resistance of less than 1mΩ (0.001Ω) is usually below 13V with the engine off and around 13.5- and are each rated at 40V and 362A. (I initially used similar IRFS3004-7PPBF devices in my 14.5V when the engine is running. prototype, but these have now been disSo when the voltage is high enough, it continued). determines that the alternator is chargby Bruce Boardman The S7434TRL7PP Mosfets come in ing the battery and connects the auxila 7-pin D2PAK (TO-263) SMD package iary battery. (VK4MQ) 24 Silicon Chip Australia’s electronics magazine siliconchip.com.au Shown here without its connecting leads (with their insulating covers, they’d hide half the panel!) use of the isolator is simplicity itself: connect the “main” terminal to the “main” battery positive and the “aux” terminal to the “aux” battery positive, with a chassis connection provided through the diecast metal case secured to the vehicle. That’s it! The LED will glow when the main battery voltage is high enough to charge the auxiliary battery. with a large mounting tab, which serves as both the drain and thermal contact for the device, allowing heat to dissipate into the PCB. Despite the impressive specifications, these devices cost under $4 each. Circuit description The circuit is shown in Fig.1. You can see the six power Mosfets (Q1-Q6) at the top, between the two battery positive terminals. They are not all connected in parallel, for an important reason. All power Mosfets have an internal ‘body diode’ (also known as a parasitic diode or internal diode) which is an inherent part of their construction, and this allows current to flow in one direction even when the FET is switched off. So to prevent unwanted current flow in either direction, the six Mosfets are arranged as three pairs (Q1-Q3 & Q4Q6), which are connected in ‘inverse series’. This way, the body diodes of each set of three Mosfets are connected anode-to-anode and so block current flow in both directions, unless both sets of Mosfets are switched on. In this case, all the body diodes are effectively shorted out. Despite the FETs having very high current ratings, three have been paralleled in each set as cheap insurance against failure. For example, the isolator could happen to be switched on during engine starting and starter motor currents can be siliconchip.com.au very high, and high currents can also flow when the auxiliary battery is first connected to the vehicle electrical system after being fully discharged. A single LM339 quad comparator (IC1) is used for all control functions. This contains four standard comparators with open collector outputs, which go low when the voltage at the inverting (-) input is higher than the voltage at the non-inverting (+) input, and are high impedance the rest of the time. That turns out to be quite useful in this circuit. I chose a switch-on threshold of 13.4V and a switch-off threshold of 12.6V. The main battery voltage is applied to pin 4 of CON1 and to a string of resistors to ground, which forms a voltage divider. The top part of the divider is 11.5kΩ [4.7kΩ + 6.8kΩ] and the bottom part is 6.8kΩ. This gives a division ratio of 2.69 [(11.5kΩ + 6.8kΩ) ÷ 6.8kΩ]. So at the switch-on battery voltage threshold of 13.4V, that means that 4.98V is applied to pin 6 of comparator IC1b (very close to 5V), and at the switch-off threshold of 12.6V, pin 6 of IC1b sees 4.68V [12.6 ÷ 2.69]. A 5V reference voltage is supplied by linear regulator REG1, powered from the main battery via a 100Ω resistor, and this voltage is applied to pin 7 of IC1b, the non-inverting input. Initially, output pin 1 of IC1b is high but once the main battery voltage rises above about 13.4V, the pin 6 input voltage exceeds that of in 7 (ie, 5V) and so output pin 1 goes low. This pulls current through the 4.7kΩ resistor and LED1, Australia’s electronics magazine July 2019  25 Fig.1: the circuit is basically a comparator which senses when the main battery voltage is high enough to charge the auxiliary battery and turns Mosfets 1-6 (or 1-12) on to do so. When the main battery voltage drops the Mosfets turn off. so LED1 lights up. In this condition, diode D4 is forward-biased and so the voltage divider formed by the 100Ω and 1.5kΩ resistors comes into play, reducing the voltage at pin 7 of IC1b from 5V down to about 4.69V (ie, 5V x 1.5kΩ ÷ [1.5kΩ + 100Ω]). That has the effect of reducing the switch-off threshold to 12.6V (4.69V x 2.69) as desired. That prevents the unit from switching on and off rapidly if the battery voltage is near either threshold. The output voltage from pin 1 of IC1b is also fed to the pin 8 inverting input of IC1c, which has its pin 9 non-inverting input connected to the 5V rail, so it acts as an inverter. So when the main battery voltage rises and IC1b’s output goes low, IC1c’s output goes high allowing the gates of the FET’s to be pulled up via the 10kΩ resistor, switching them 26 Silicon Chip on (as described below) and connecting the two batteries. REG1 is a micropower regulator, both to minimise the quiescent current but also (and most importantly) because it has an excellent initial tolerance of ±0.5%. This, along with the 1% resistor tolerances, determines how accurate the switch-on and switch-off voltage thresholds will be. Note that if you change the battery sense voltage divider resistors, you can calculate the new switching thresholds by calculating the divider ratio, then multiplying 5V and 4.7V by this ratio. To change the hysteresis (ie, the spread of these two thresholds), you would need to change the value of the 1.5kΩ resistor at pin 7 of IC1b; a lower value gives more hysteresis, and a higher value, less hysteresis. Australia’s electronics magazine siliconchip.com.au Mosfet gate drive To switch on an N-channel Mosfet, the gate needs to be driven several volts above the source. In this circuit, all the Mosfet sources are connected together and when the Mosfets are switched on, they will all rise to the battery voltage – ie, around 12V. Therefore, the gates need to be driven to at least 17V and ideally higher, to 20V or more, to ensure that they switch on fully and have the lowest possible resistance and dissipation. This voltage is generated by comparator IC1a, which is configured as an astable oscillator and drives a charge pump. The frequency of this oscillator is set to around 15kHz by the combination of the 22kΩ feedback resistor and 3.3nF timing capacitor. Output pin 2 of IC1a is pulled high by a 4.7kΩ resistor, and the resulting square wave causes the 100nF capacitor to charge up to around 12V, via diode D2, when output pin 2 goes low. When that pin goes high, to around 12V, the anode of diode D3 is lifted up to around 22V and this voltage in turn charges the following 100nF capacitor which supplies the Mosfet gates with about 20V via the following 10kΩ resistor. That is, as long as output pin 14 of inverter IC1c is not being held low. If it is, this shunts any current flowing through that 10kΩ resistor to ground, holding the gates low. At the same time, to save power, when pin 14 goes low, diode D1 becomes forward-biased and this discharges the 3.3nF timing capacitor, disabling the oscillator which generates the gate drive voltage. Zener diode ZD1 protects IC1 from supply spikes, in combination with the 100Ω series resistor from the main battery, which limits the current through ZD1 should it conduct. Zener diode ZD2 protects Mosfets Q1-Q6 from damage due to excessive gate voltages. This is important as when the ~20V gate drive is initially applied, their sources are at 0V, and this could otherwise exceed their maximum ±20V VGS ratings. However, ZD2 will not conduct for long, as the source voltage will quickly rise, reducing VGS to around 7-8V under steady state conditions. Features & specifications • • • • • • • • Suits most 12V batteries Waterproof Silent Solid-state (no relays) Easy construction and installation Switch-on voltage: 13.4V (13.13-13.67V*) Switch-off voltage: 12.6V (12.35-12.85V*) Quiescent current: approximately 3mA when off, 7mA when on • High current handling (>100A peak, >40A continuous) • Low voltage drop: typically <1mV/A Low dissipation: typically <1W <at> 30A *if some ±0.1% resistors are used (see parts list) TVS1 and TVS2 are transient voltage suppressors, similar to zener diodes but more robust. These protect the unit and especially the Mosfets from high-voltage transients which are common in the automotive environment. Construction The prototype was built on two boards, with the control circuitry on a piece of stripboard and the Mosfets, TVSs and battery connectors soldered to a double-sided ‘blank’ PCB which was manually cut into large, isolated sections of copper that the components were then soldered to. You can also build it this way, and we will give some information later on how to do so. However, to make your life easier, we have produced two commercial double-sided PCB designs. Again, one is for the control circuitry and the other for the larger components. You then just need to solder the components to these two boards, join them and mount them in the case. Fig.2 shows the control board while Fig.3 is the Mosfet board overlay diagram. Use these and their matching photos as a guide during construction. While the prototype had all six Mosfets on the same side Fig.2: one of two PCBs in this project, the control board, with matching photo alongside. You could also build this on stripboard if you wished (see page 30) but PCBs make a much neater job and minimise the chance of errors. siliconchip.com.au Australia’s electronics magazine July 2019  27 Here’s the top side of the completed Mosfet PCB. It’s fitted with six Mosfets as shown in Fig.3a (top). But if you wish, another six Mosfets can be soldered to the underside of the PCB for even better current handling (Fig.3b, lower) of the board, our Mosfet PCB (shown in Fig.3a) actually has twelve possible Mosfet mounting locations; six on the top and six on the bottom, with each pair of Mosfets directly above and below each other (Q1 and Q1’, Q2 and Q2’ etc). Fig.3b shows where the Mosfets can be mounted on the underside of the board. This gives you the option to mount three or four Mosfets on one side of the board and the remainder on the other side, which will help to more evenly distribute what little heat is generated in the device, and may also make slightly better use of the copper, reducing losses slightly. But it’s a minor advantage, and you could just as easily fit them all one side, which is what we did. For the control board, install the resistors where shown, then the 1N4148 diodes, ensuring that in each case, the cathode stripe faces as indicated. You can then fit the single zener diode, with its cathode stripe facing to the left. Next, solder IC1 to the board, ensuring that its pin 1 dot/notch face towards the top as shown. We don’t recommend that you use a socket as these can cause failures over time. Now fit the non-polarised capacitors, which can be either ceramic or MKT types, followed by the single electrolytic capacitor, with its longer positive lead through the righthand pad (marked with a “+” symbol). That leaves REG1 and CON1. Gently bend REG1’s leads to fit the PCB pads, then solder it in place. CON1 is a regular 5-pin header that’s soldered to the top side of the board. You can then move on to the Mosfet board. Building the Mosfet board This board has eight SMDs (six Mosfets and two TVS diodes) plus three through-hole components, not including the battery connections, which we’ll explain below. Start by soldering the Mosfets. These are quite large and are soldered to large, thick copper planes so you will need a hot iron to solder them. In each case, start by spreading a thin layer of flux paste over all the pads, especially the large one for the tab. Then locate the Mosfet in position and solder its pin 1 (near the 28 Silicon Chip dot). This is the gate connection so should be the easiest to solder. Check that all the pins and the tab are lined up correctly. If not, re-heat that solder joint and nudge the device slightly. Solder the remaining five small pins next. It doesn’t matter if you accidentally bridge them to each other, as long as they don’t bridge to the middle stub pin (which is not connected on this board) or pin 1 (the gate drive). Finally, flow solder onto the junction of the tab and its large mounting pad underneath. You will need to apply heat and feed in solder until the solder flows to form a smooth fillet between the two. It’s OK to add a little extra solder until it covers the tab. The flux you added earlier should aid in this process. Repeat the above for the other five Mosfets. Then solder the two TVS diodes in place using a similar procedure, ie, applying flux paste to both pads, tacking the part down on one side, soldering the other side, then refreshing the first solder joint to ensure it is reliable. Next, solder ZD2 and LED1 in place on the top side, with the orientations shown. It’s a good idea to fit LED1 with some space between its lens and the PCB, so that it can poke through a hole in the case. The base of its lens should be a little bit more than the thickness of one M8 nut above the board. Having done that, fit 5-pin header socket CON2 on the Australia’s electronics magazine siliconchip.com.au Make sure any added wires do not project above the board any higher than the bodies of the Mosfets; otherwise, they could potentially short to the metal lid of the case later. Testing The two PCBs are stacked as shown, with the 8mm brass battery connection posts fitted firmly in place with washers ensuring good contact with the PCB tracks. underside of the board. The easiest way to do this is to plug CON2 into CON1 on the control board, attach the two boards using the four corner mounting holes, 12mm tapped spacers and short machine screws and then solder CON2 to the Mosfet board. That ensures the two headers line up properly. The M8 brass screws that will be used as the battery terminals can now be fed through the Mosfet PCB, with a shakeproof or crinkle washer under the screw head (which goes on the bottom side of the board) and another under the nut which is done up tightly on the top side of the board. This should give good electrical contact to the PCB and means that you don’t need to solder the screws and nuts to the boards, which is difficult and makes disassembly impossible. (You can see that this was done on the prototype in the photos below.) While the Mosfet board is now complete, you could consider adding some tinned copper wires in parallel with the copper on the board. This will reduce the voltage drop across the device, as well as its dissipation, and make it more robust. However, we do not feel that this is strictly necessary due to the use of extrathick 2oz copper on this board. If you do want to run some extra wire, you can solder lengths of tinned copper wire from between pins 2 & 3 of each Mosfet to between pins 5 & 6 on the Mosfet on the other side of the board. You can then solder wires from the tabs of each Mosfet to the nearby battery terminal. You may be able to solder these to the exposed copper around the nuts, or directly to the nuts themselves, with a very hot iron. Ideally, you should use an adjustable bench supply with current limiting for testing. Set it to 12V and around 50mA, then apply power between the main battery terminal and the ground pad on the Mosfet PCB (or pin 5 of CON1 or CON2). You should observe a current flow which settles at around 8mA. LED1 should remain off. Measure the voltage at the auxiliary battery terminal relative to GND. It should be low, close to 0V. Now increase the supply to around 14V. You should observe LED1 switch on. The current draw should increase slightly. The voltage at the auxiliary battery terminal should now have risen to the supply voltage. Reduce the supply voltage back to 12V and confirm that LED1 switches off and the voltage at the aux battery terminal drops back to 0V within a few seconds. This verifies that everything is working as intended and you can now proceed to finish construction. Adding a bypass switch There may be times where the vehicle battery is low, but you still want to connect it to the auxiliary battery. One example would be if the vehicle battery is flat but the auxiliary battery is charged, and you want to ‘jump start’ the vehicle using the aux battery. While you could do this with a screwdriver across the terminals, it’s much nicer to have a switch which forces the unit to operate. This is quite easy to do, but it does have one limitation in that this won’t work if the vehicle battery is dead flat, since the unit is powered from it. But it should work down to at least 10V, or possibly even less. The easiest way to achieve this is to connect a switch between pin 7 of IC1b and GND. When this switch is closed, Fig.4: the front panel can either be photocopied or even better, downloaded from siliconchip.com.au/shop/11/5059 Ideally, it should be laminated before glueing in place. siliconchip.com.au Australia’s electronics magazine July 2019  29 Alternative construction method using stripboard and hand-cut PCBs Instead of using the PCBs that we designed, you could copy the approach used for the prototype and build the control system on a piece of stripboard (Veroboard, for example) and handmake your own PCB to host the Mosfets and related components. My suggested stripboard layout is shown at right. This requires a board with at least 13 strips and 21 rows of holes. The diagram is drawn looking from the top of the board (ie, from the non-copper side). The copper tracks are shown as a visual aid, as if you can see them through the board. You may want to use a larger piece of stripboard so that you have space to drill some mounting holes later. Before fitting the components, cut the tracks in the sixteen locations shown (including all seven tracks under IC1). It’s often easier to cut the tracks with a 3mm twist drill, just removing the copper around the hole. Having soldered the components in place, fit the wire links. The shorter links can be made using component lead off-cuts, or in some cases, by merely bridging adjacent tracks with solder. Longer links are best made with solid-core insulated wire (eg, Bell wire). For the Mosfet board, you will need a piece of double-sided copper laminate around 100 x 100mm (slightly smaller, if you’re planning to fit it into the specified box; check it fits before proceeding). Ideally, this should have thicker-than-normal copper (eg, “2oz” which is double normal PCB copper thickness). The required layout is shown clearly in the photos below. On the top of the board, you will need to make three straight cuts (eg, using a rotary cutting disc) to separate the copper into four islands. The central islands should be around 25mm wide. Be careful not to cut through the fibreglass substrate; just the copper. Ensure the cuts are wide enough to guarantee electrical isolation. The underside requires just one cut down the middle, separating the copper on either side. Next, drill two 8mm holes for the battery terminals and eight 2mm diameter holes (around the locations where the Mosfet tabs will be soldered) for wire vias to pass through later. Now is also a good time to drill four 3mm holes which the control board will be mounted to later (lining up with holes on that board). Bend pin 1 (the gate) of each Mosfet up, then solder the remaining five small pins to the central island. Be careful to place the Mosfet so that the body does not bridge the cut in the copper plane. Then, using a hot iron, solder the tabs in place. Join the gates with light-duty wire; it’s easier to use stiff b ell wire, but you could use Kynar or multi-strand wire. The small copper island at the bottom is the ground connection point. Solder the anodes of the two TVSs to this island, with the cathodes to the large planes on either side. You can now add the zener diode, with its anode to the large central copper area and its cathode to the Mosfet gate wire. Stripboard prototype with matching layout below. Don’t forget to cut the tracks where indicated – you’ll have a massive short circuit otherwise! Next, run a strip of thick copper wire down the central island, soldered near every pair of Mosfets, plus wires on the underside fed through each of the 2mm holes you drilled earlier and bent over to touch the battery terminals. Solder them near the terminals and on both sides of the 2mm holes to form vias. If you can’t easily get thick copper wire, you can use a bundle with multiple pieces of 0.71mm or 1mm diameter tinned copper wire. Solder four wires to this PCB: one to the main battery terminal side, to supply 12V to the control board; one to the small ground area, to connect to GND on the control board; one to the cathode of the zener diode, which goes to the gate drive pin on the control board; and one to the central copper island (or zener diode anode), which goes to the control board Mosfet source terminal. Note, though, that this source terminal only connects to a 10kΩ resistor with the other end connected to GND. So you could make your life slightly easier by simply soldering a 10kΩ resistor between the two central copper islands on the Mosfet board and then you won’t have to run this fourth wire. The only part that’s left now is LED1, which can be chassismounted to your box, with its anode connected to pin 4 of CON1 on the control board, and its cathode to pin 1. Make the three other connections from your Mosfet board to CON1 on the control board, as described above, and you are ready for testing. The photo at left shows the original (hand made) prototype “Mosfet PCB” with its hand-cut breaks between the copper sections. Note how the gate pins here are all connected to (the red) insulated wire, not to the PCB. At right is the opposite side, with 8mm brass bolts soldered firmly in place, with heavy copper wires which pass through the board and are soldered to the top copper as well. 30 Silicon Chip Australia’s electronics magazine siliconchip.com.au it will pull that pin down to 0V, which means that the voltage at pin 6 will always be higher than pin 7, so output pin 1 will go low, switching on Mosfets Q1-Q6. This switch is shown with dotted connections in Fig.1. We’ve also shown the most convenient points to solder wires to go to the switch in Figs.2 & 3. Simply solder a wire here, to the COM terminal of an SPST switch, then a wire from the NO terminal of that switch to a convenient ground point. When you activate this switch, you need to remember to switch it back into its normal position later, for the unit to go back to doing its job! Case assembly There are only four holes to drill: two in the lid for the battery terminals (main and auxiliary), plus one for the LED, and one 3mm hole in the side of the case for the ground eyelet. If you’re installing the optional bypass switch (S1), then you may wish to mount it on the lid, in which case you will need to drill an extra hole. Make sure that the switch won’t foul the Mosfet board once it’s mounted. You will probably find that you have more room if you mount it low on the side of the case, and that may also make it harder to trigger the bypass function accidentally. If you’re using a metal case, ground is connected to the case internally and then externally, to the vehicle chassis or one of the battery terminals. You will also need to find a way insulate the two 8mm bolts from the lid of the case. With a plastic case, the easiest way to provide a GND terminal is to feed a long M3 screw through the GND terminal on the Mosfet board, attaching it to the PCB in a similar manner as the two large 8mm screws (ie, with shakeproof washers and nuts). This can then project up through a fourth hole in the lid. Or you could connect the ground eyelet to a screw which is externally accessible elsewhere. There’s no need to provide any insulation for the 8mm screws when using a plastic case; however, you will need to seal all the exit holes with neutral cure clear silicone, to ensure that the case remains watertight. Download the panel label artwork from the SILICON CHIP website and print it at actual size. You can then cut it out and use it to mark out the hole positions in the lid. Drill them all to 3mm, then enlarge the two battery terminal holes to 8mm with larger drills, a stepped drill bit or a tapered reamer. Laminate the label and cut out the holes using a sharp hobby knife. You can then stick it to the lid using contact adhesive or a thin smear of neutral-cure silicone. Other options for creating adhesive panel labels are described on our website at siliconchip.com.au/Help/FrontPanels Now plug the two boards together and join them using Nylon tapped spacers and machine screws. Mount the whole assembly on the underside of the lid, remembering to use insulators for the 8mm screw shafts if the lid is metal. Attach the assembly to the lid using a flat washer and nut, then another flat washer and nut, which can later be used to clamp the battery wires or terminals. Seal any possible water entry points (eg, around the LED lens) with neutral cure silicone, then, if using a metal case, drill a hole in the side of the case for the ground eyelet siliconchip.com.au Parts list – Solid State Dual Battery Isolator 1 double-sided PCB coded 05106191, 98 x 71mm 1 double-sided PCB with 2oz copper, coded 05106192, 98 x 71mm 1 IP65 diecast aluminium box, 115 x 90 x 55mm [Jaycar HB5042/HB5044, Altronics H0423] OR 1 IP65 polycarbonate box, 115 x 90 x 55mm [Jaycar HB6216/HB6217] 1 panel label, 115 x 90mm 2 35mm long M8 brass screws 6 M8 brass hex nuts 6 8mm ID brass flat washers 4 8mm ID brass or beryllium copper star/crinkle washers 4 8mm ID Nylon screw insulators (if using a metal case) 4 12mm long M3 tapped Nylon spacers 8 M3 x 6mm panhead machine screws 2 small eyelet quick connectors 1 M3 x 10mm panhead machine screw, shakeproof washer and two hex nuts Semiconductors 1 LM339 quad comparator, DIP-14 (IC1) 1 LP2950ACZ-5.0 5V low-dropout linear regulator, TO-92 (REG1) 6 40V 100A+ N-channel Mosfets, TO-263-7 (Q1-Q6) [eg Infineon IRFS7434TRL7PP*] 1 5mm LED (LED1) 2 15V 1W zener diodes (ZD1,ZD2) 2 5kW 15-18V transient voltage suppressors, DO-214AB/ SMC (TVS1,TVS2) [eg, Bourns 5.0SMDJ15CA-H*] 4 1N4148 small signal diodes (D1-D4) 1 5-pin SIL socket (CON1) 1 5-pin header (CON2) Capacitors 1 4.7µF 50V electrolytic 4 100nF 50V ceramic or MKT 1 3.3nF 50V ceramic or MKT * available from Mouser or Digi-Key Resistors (all 1/4W 1% metal film) 1 22kW 3 10kW 2 6.8kW# 3 4.7kW# 1 2.7kW 1 1.5kW 2 100W # use ±0.1% tolerance resistors for the tighter threshold ranges mentioned in the text and attach it using a machine screw, shakeproof washer and two nuts. You can then insert the sealing gasket into the channel in the underside of the lid, cutting it to size so that it fits around the full circumference. With that in place, lower the lid onto the case and attach it using the supplied screws. Don’t forget to attach the case (if metal) or ground screw to the vehicle’s ground, either via the chassis or to one of the battery negative terminals. You can then wire up the two battery positive wires to the unit and verify that LED1 lights and the auxiliary battery begins to charge when you switch on the engine. Don’t forget to use heavy automotive cable with a sufficiently high current rating (25A+) to handle the high charging currents which can occur. The prototype used 35mm2 SC automotive starter motor cable. Australia’s electronics magazine July 2019  31