The first 5 subparagraphs of this paragraph deal with the electrical characteristics of the switchboard and only in the last paragraph is discussed how it can be built. This leads to the strange situation that in par. 4.9.3 and par. 4.9.4 everything is explained about calibrating the indicator and adjusting the voltage regulator while there is nothing yet to calibrate or adjust. Still it seems best to have the paragraph on building the switchboard at the end since it is easier to build it when one understands the way it works better. It only means that after building it, the reader will have to look back to the paragraphs on calibrating and adjusting it.

The electrical circuit of the switchboard of fig. 4.25 might discourage people who are not experienced with electricity. In a car, there is no special switchboard for the alternator. The voltage regulator is often built into it, there are no difficult adjustments, the fuse is integrated into a common fuse box or there is none at all and there is just the alternator control lamp instead of a complicated indicator. Such a simple circuit would be attractive for the firefly charger as well: Just two wires coming out of the charger to which the battery must be connected and that's all. Such a charger can be built and it would charge a battery. But compared to such a simple circuit, the switchboard circuit as given in fig. 4.25 performs the following extra functions:

• Field current control: Reducing the field current makes that the efficiency of the charger can be raised in two ways:
  • It makes it possible to adjust the speed of the charger so that it will be close to the optimum speed of the runner (at that head), so that turbine efficiency will be optimal.
  • It means that less of the gross electrical power output of the alternator will be used to power its own field, so that alternator efficiency will improve.

  See par. 4.9.2 for more details.

• The indicator: It is used for 2 main purposes:
  • Check whether a battery was discharged too deep (voltage measurement before charging has started).
  • Check whether battery is fully charged already (voltage and current measurements during charging).
• Re-adjusted voltage regulator. By readjusting the voltage regulator, the charging characteristic of the firefly charger will be such that batteries are charged completely in a reasonably short time. With the normal adjustment for cars, the charging characteristic is such that the last bit of charging takes exeptionally long.
• With a firefly charger, having a fuse is adviseable because with some luck, it will save the diodes inside the alternator in case of a battery being connected with the wrong polarity. In a car it is quite unlikely that a battery will be connected with the wrong polarity since batteries are not changed that often and if they are changed, it is probably done by a trained technician.
• A separate switchboard makes it possible to have the battery a bit furter away from the charger, at a safe, dry place. The cable between the voltage regulator and the battery should be short and thick since the voltage drop over this cable affects the charging characteristic negatively. So with an alternator with built-in voltage regulator, the battery must be connected with a short and thick cable. The cable between the alternator and the switchboard does not affect the charging characteristic so this cable can be made longer.
• If the switchboard can be disconnected easily from the charger, it can be taken home by the operator to prevent unauthorised use of the charger. Bringing home the charger itself is not practical because disconnecting the penstock is quite some work and because it is heavier.

Then there is the 100 µF 63 V `Elco' capacitor. This part protects other electronic components against voltage spikes that are created when a battery is switched off while it is being charged, see box 4.15

Fig. 4.25: Electrical circuit of the charger.

Addition to internet version:
Electronic switchboard.

There is also a design for a fully electronic switchboard, see http://www.microhydropower.net/mhp_group/portegijs/firefly_exp/Tech_issues.html#elect_switchb
It is more user-frendly and it can charge two batteries in one go. But it requires quite some electronics experience to build.

In the electric circuit of a car, the `DF' connection of the voltage regulator is connected directly to the `F' connection on the alternator. On the switchboard, there is a set of 4 lamps connected in between `DF' and `F' (see fig. 4.25). This set of lamps acts as a resistor connected in series with the field coil inside the alternator. Consequently, the total resistance of lamps + field coil will be higher and the field current will be lower than it would be without these lamps.

Reducing the field current in this way has two effects:

• The field current influences the speed at which the charger will run. To the turbine, the alternator acts as a brake and the field current determines how strongly. The turbine in its turn, tries to accelerate the alternator shaft. Then there is a speed, at which the two balance, it neither accelerates nor slows down. At a high field current, this speed will be slower (the alternator acts as a strong brake) while at a low field current, this speed will be relatively high.
• It reduces the power consumption of the field itself. Without such lamps, the field would continuously consume up to 40 W. This effect is especially important if you want to run the charger at half capacity and at a low head, so with a low electrical power. Suppose the charger would generate a gross electrical power of 100 W and 40 W of this is consumed again by the field.

The first lamp (the one to the left) is always `on' while the 3 others have switches. The lamps are connected parallel with respect to each other. So if more lamps are switched on, the total resistance of the set of lamps will be less and the field current will be higher. The last lamp is only 10 W so it is as if this lamp counts only for half (if 10 W lamps are hard to get, using four 20 W lamps will also do). This makes that with the 3 switches, already 6 different resistance values can be selected (with respectively 1, 1.5, 2, 2.5, 3 and 3.5 lamps switched on).

Trying to calculate what is the actual resistance of the set of lamps makes little sence because the resistance of each lamp varies strongly with the temperature of the filament inside, which in turn varies with the voltage over it.

The speed is not just determined by the field current as set by the number of lamps switched on:

• Also the power generated by the turbine influences the speed. When it is running at full capacity and at a high head, it is much stronger in trying to accelerate and the speed will be higher.
• The voltage regulator can also reduce the field current, which in turn will make the charger run at a higher speed. It does so when the voltage reaches the set point (see par. 4.9.4) and the battery is nearly or fully charged.

After a battery has been connected and charging has been started, the operator should try out what is the best number of lamps to be switched on. This can be done in the following way:

• Switching the indicator to current measurement. Then it measures the net charging current to the battery, which is related directly to the net power produced by the charger.
• Switching on more lamps or switching off some until the setting that gives the highest charging current.

The best number of lamps to be switched on, depends on the type of alternator fitted, the head and the size of the blocking timber (if any). So it will be the same as long as the same charger is used at the same site and trying out what is the best number has to be done only once.

If it was just for regulating the field current, a simpler circuit with normal resistors instead of lamps, would have been easier to build. But using lamps has some additional benefits:

• The lamps indicate whether there is a field current. The number of lamps that light up and their brightness even roughly tells whether the field current is large or small. This is helpful in troubleshooting.
• The fact that the lamps light up, indicates that the lamps (and the field coil in the alternator) consume energy. They will still consume energy when the charger is not running but the battery is still connected, so then the battery is discharged instead of charged. So hopefully the lamps will remind the operator that a battery must be disconnected when the charger is not running.

With the indicator on the switchboard, both voltage and current can be measured. The voltage that is measured is voltage V_i `at the indicator', see fig. 4.25. During battery charging, there are voltage drops over all cables that conduct this high charging current so then this voltage V_i differs slightly from voltage V_r at the regulator and voltage V_b over the battery poles. The current I that is being measured, is the net charging current to the battery. The gross current produced by the alternator is the net charging current plus the current consumed by the field inside the alternator.

The indicator circuit is designed such that it will measure voltages from 10 to 15 V, and currents from 0 to 20 A. Of course, the original scale of an indicator with 10 mA full scale current reads from 0 to 10 mA. For easy interpretation, the scale should be replaced or modified so that it reads directly from 10 to 15 V and from 0 to 20 A. The scale in fig. 4.26 has the right dimensions for a type of indicator that is commonly available in Holland and could be glued onto the original scale, if it fits. The small circle with cross outside the scale itself shows the right location for the shaft of the indicator needle. Normally, the plexiglass cover is clicked onto the indicator housing and it can be removed by carefully wriggling a knife in between. Take care not to damage the delicate needle and when putting the cover back, make sure that the pin of the zero adjustment screw falls in the slot of the indicator mechanism.

Measurements of voltages and currents are needed for the following purposes:

• To check whether a battery was discharged too deep. This can be done by measuring battery voltage before charging has started. If the operator notices that a battery was discharged too deep, he/she can warn the owner and maybe check the charge indicator of this owner, see par. 5.4.2.
  This check will also warn the operator in case a battery is connected with the wrong polarity: Then the needle will violently hit towards the extreme left.
• To check whether a battery is fully charged. For this, both battery voltage and charging current are measured. When the voltage has risen above 14.5 V and the current has dropped below 4 A, the battery is considered fully charged.
• To find best setting for field current control. For this, the charging current is measured, see par. 4.9.2.
• As a help in troubleshooting. Measurements of voltage and current might indicate what is wrong if the charger does not function well.
• To find out what is the power output of the charger (= voltage times current). This satisfies a very understandable curiosity and it might boast the confidence of the one who built it if the charger reaches the calculated power output or even exeeds this.

For serving these purposes well, the indicator should measure voltages quite accurately. A measuring error of only 0.1 V when measuring battery voltage before charging, would already mean that the state of charge of this battery is estimated 10 % too high or too low. Another thing about the voltage measurements is, that only the range between 10 V and 15 V is interesting. If a battery has a voltage below 10 V before charging, it is completely empty and there is no need to measure the voltage accurately in order to tell how empty it is. If the voltage during charging raises above 15 V, the voltage regulator is adjusted too high and again there is no need to know exactly how high this voltage is.

The way the voltage measurement circuit in fig. 4.25 works, is as follows:

• The two LM336 / 5V devices substract 10 V (5 V per device) from voltage V_i. So the voltage over the remaining part of the circuit (the panel indicator, 470 Ohm resistor and 100 Ohm trimmer) will be V_i minus 10 V.
  This also makes that voltages below 10 V can not be measured: Then the indicator needle hardly moves from its rest position at the extreme left of the scale.
• The remaining part of the circuit measures voltages of 0 to 5 V. The panel indicator is made to indicate `full scale' when the current through it is 10 mA (panel indicators with other full scale values are also availble, be sure to buy one with 10 mA full scale). So with 5 V over this part of the circuit, the current through it should be 10 mA, which means that according to the law of Ohm, the total resistance must be 5 V / 0.01 A = 500 Ohm. This is the resistance of the panel meter itself, the 470 Ohm resistor andpart of the resistance of the 100 Ohm trimmer.

The voltage measuring circuit in fig. 4.25 makes that the interesting range of 10 to 15 V covers (almost) the whole scale of the panel indicator. So compared to an ordinary indicator circuit that measures all voltages below a maximum voltage of 15 V, the interesting voltage range of 10 to 15 V is amplified 3 times on the indicator scale.

Fig. 4.26: A scale for the indicator.

This also improves the accuracy of the indicator considerably. It is as if the first 10 V is measured by the two LM 336 devices, which produce a very stable reference voltage. So errors in the panel indicator only affect the measurement of the remaining voltage. This makes that with an ordinary `class 2.5' panel indicator (with a maximum error of ± 2.5 %), the complete voltage measuring circuit will have an error of less than ± 1 %, when adjusted correctly.

Above is described how it works in theory. In practice however, it behaves slightly different:

• The LM336 voltage references need a minimum current of 0.4 to 0.6 mA before they will produce a stable, 5 V reading. The current through these devices will only become so high once voltage is above 10.2 to 10.3 V. Lower voltages can not be measured that accurately and that is why the indicator scale of fig. 4.26 shows no readings below 10.5 V.
  In fact, the indicator needle will move from its zero point already at voltages much lower than 10 V (meaning that the LM336 devices allow a small current to pass even though the voltage across each of them is way below 5 V). Then once voltage reaches 10.2 to 10.3 V, the indicator needle starts to react much stronger to slight voltage changes and a reliable reading can be taken.
• The voltage over these LM336 devices won't be exactly 5 V. To compensate for this, one of them has a 50 kOhm trimmer connected to its calibration connection, with which its voltage setting can be calibrated. There is no need to have trimmers connected to both of them since the calibration of the one with a trimmer can be set such that it compensates for any deviation of the other one.

The voltage measuring circuit should be calibrated. This could be done in the field using test voltages of batteries. Or it could be done in the workshop, using a stabilised voltage supply (see annex E *building your own testing and calibrating equipment) to provide test voltages. Make sure that the indicator is held upright during calibration since this will be the position it will be used in in the field.

1. First check the zero-adjustment of the indicator itself. Without any voltage being connected (and with the indicator being held upright), the needle should exactly read 0 V and 0 A. Correct with the zero adjustment screw on the indicator itself (just above the shaft around which the indicator needle rotates). Make sure that this setting won't be changed later on, maybe stick a layer of electrical tape over the zero adjustment screw.
2. Connect it to a voltage of about 11 V, but at least no less than 10.5 V. This could be the voltage of a battery that is already discharged way too much (if such a battery is not available, connect lamps to a discharged battery so that its voltage will drop a bit further). Measure the voltage with a reliable digital tester. Then adjust the 50 kOhm trimmer until the indicator shows the right reading.
3. Connect it to a voltage of about 14.5 V. This could be the voltage of a fully charged battery while it is still being charged further. When using a battery as voltage supply, make sure that the digital tester is connected to the right points in the circuit (see arrows with V_i and -V_i in fig. 4.25). Otherwise voltage drops over cables might make that the indicator experiences a different voltage than the tester that is used as a reference. Now adjust the 100 Ohm trimmer until the indicator reads the right voltage.
4. The setting of the trimmers influence one another. So if with step 3 the setting of the 100 Ohm trimmer was changed, the 50 kOhm trimmer needs to be checked. So repeat step 2 and 3 until the indicator produces correct readings at both 11 and 14.5 V.

The current measurements do not have to be as accurate as the voltage measurements. Consequently the circuit can be simpler. There is just the current shunt, which produces a voltage drop proportional to the current that flows through it. This voltage drop in turn is measured directly by the panel indicator.

For choosing the correct resistance R_s for the current shunt, the voltage V_f at which the panel indicator displays `full scale' should be known. This is not given with the indicator data but it can be calculated from the resistance R<sub>i</sub> of the panel indicator, which can be measured with a tester. The resistance of the copper wire inside increases with the temperature. So measure the resistance of the panel indicator at 20 ° or substract 0.43 % for each ° the temperature is above 20 °. Then the voltage V<sub>f</sub> at which the indicator displays `full scale' can be calculated using Ohm's law (the 0.01 stands for the 10 mA full scale current of the panel indicator):

V_f = 0.01 * R_i

Now the resistance R_s of the current shunt can be chosen. At 20 A charging current, the panel indicator should indicate `full scale' (it is also possible to choose a lower or higher full scale current).

V_f = 20 * R_s so: R_s = V_f / 20

Now can be calculated what size and length of cable will give the desired resistance value for the current shunt.

R_s = 0.0175 * l / A

with:

0.0175 = specific resistance of copper at 20 °C, Ohm/m/mm²
l = length of cable in m
A = cross-sectional area in mm²

Cables are sold by their cross-sectional area and sometimes it is even printed on the cable itself. In case cables are sold by the American Wire Gage classification, the cross-sectional area is as follows:

Probably a cable of 1.5 mm² (or #16 in the American classification) will result in a convenient lenght for the current shunt. Better not choose a cable that is thinner than this because it might get too hot.

If indicators with 10 mA full scale current are not available, the voltage measuring circuit needs some modification (the right value for the current shunt can still be found as explained above). Suppose only a 1 mA full scale indicator is available. Now one could replace the 470 Ohm resistor and 100 Ohm trimmmer by a 4.7 kOhm resistor and 1 kOhm trimmer. Then, after calibration, the indicator will give the correct reading for 14.5 V. But the LM336 devices will only reach their minimum current at a voltage of some 12 to 13 V so measuring lower voltages accurately is impossible. To increase the current through these LM336 devices without influencing the current through the indicator itself, fit an extra 470 Ohm resistor between the point indicated by the arrow `V<sub>i</sub>' and the `+' connection of the top LM336 device.

If this indicator circuit seems too difficult to make and calibrate, a tester could be used instead. Then it is adviseable to make connections for the tester on the switchboard:

• For the voltage measurements: Fit wires to the points in the circuit of fig. 4.25 indicated by the arrows with `V<sub>i</sub>' and `-V_i'. Now the voltage can be measured directly over these wires.
• For the current measurement: As a current shunt, fit a piece of cable that has a resistance of R_s = 0.01 Ohm.

  cross-sectional area of cable:	length between the wires:
  2.5 mm²	1.43 m
  1.5 mm²	0.86 m

  Fit a third wire to the end of the current shunt that is nearest to the battery switch (on the other end, there is the `V<sub>i</sub>' wire already. Now the voltage drop over this current shunt can be measured using the `mV' range on the tester. Each mV represents 0.1 A, so 100 mV means that the current is 10 A.
  Of course one could also measure the current directly using the current range on the tester, but this has some disadvantages:

  1. There will be a significant voltage drop over the tester, which affects the measurement itself.
  2. When accidentally the leads of a tester that is set to current range, are connected to the battery voltage, a short circuit will result. Hopefully only a fuse will blow but also the tester might be destroyed. A digital tester that is set to mV range will survive full battery voltage without problems.
  3. The maximum current that many testers can measure, is about 10 A or less. The current to be measured can be higher so this is not enough.

Normally, voltage regulators are adjusted to 14.2 to 14.4 V. This is safe from the point of protecting a battery against overcharging (for the effect of overcharging, see annex C *corrosion of grid in positive plates). At this voltage, a fully charged battery will draw such a low charging current that it does little harm. For a car, this is necessary because the battery will not be disconnected once it is fully charged.

For a firefly charger, such a low adjustment of the voltage regulator is undesirable:

• It makes that during the last bit of the charging process, the regulator limits the voltage to 14.2 V and consequently charging current will drop to a very low value. Therefor the whole charging process will take quite long. It is as if one wants to fill a bucket, but once it is filled for 90 %, the tap is nearly closed so that it only trickles. Then it still takes a long time to get the bucket filled completely.
• Any voltage drops in the cable between the battery itself and the point where the regulator is connected, will amplify the effect mentioned above. These voltage drops cause that the regulator senses a higher voltage V_r than the actual battery voltage V_b, see fig. 4.25. Consequently it starts limiting the charging current even before battery voltage V_b has reached 14.2 V. In a car, such voltage drops will be small because cables are short and thick and no other components are fitted between the alternator and the battery. With the firefly charger, such voltage drops are inevitable: There is the battery switch, current shunt, cable to the battery, connector and battery fuse.
• It increases the risc of batteries being disconnected before they are charged completely, which will make them wear out abnormally fast (see annex C: `active material falling out of positive plate' and `Sulphatizing'). The operator might be in a hurry and not take the time. Or he/she might feel that there is no sense in charging further since it goes so slow anyway: The indicator shows a very low charging current and the charger is practically running free since the alternator hardly slows it down any more...

Choosing a better value for the voltage regulator is a choice between evils. A low setting (say 14.2 V) means that charging takes long and there is a risc that batteries will wear out because they are not charged completely. A high setting (say 15.0 V) means that batteries will wear out because of overcharging if they are not disconnected once they are fully charged. This is especially likely to happen if the charger will be used at night. As a compromise, 14.7 V is chosen for the firefly charger.

Readjusting an adjustable electronic regulator is easy. You just have to find the trimmer that controls the voltage setting. If there is no cover that can be removed easily, look for holes where a small screwdriver fits through since the trimmer might be just behind there. However, most electronic regulators are not adjustable. Of those types, still the voltage setting can be influenced, see box 4.14.

Mechanical regulators work with an electromagnet that pulls a switch towards disconnecting the field current, and blade spring that pulls the switch in the other direction. When the voltage is above the set point, the electro magnet is stronger than the spring and the field current is disconnected. The easiest way to change the set point to 14.7 V is by increasing the force exerted by the spring a little so that only at a higher voltage, the electromagnet can pull the switch towards disconnecting the field current. Most types are calibrated in-factory and have no adjustment screw. Find the place where the blade spring touches a support on the frame. By placing one or two layers of electrical tape between the spring and the support, the spring is bent a little further and thus the spring force will be a little higher. An alternative way to change the set point is by connecting a 5 Ohm trimmer in series with the electromagnet. Due to the extra resistance, the current through the electromagnet will decrease a little and with it, the force it exerts to the switch.

A mechanical regulator could very well have 2 electromagnets operating 2 switches and some 6 connections to the outside. Then there is a problem of finding out which electromagnet acts as the voltage regulator and how it should be connected. This problem is dealt with in par. 4.9.6.

Readjusting the voltage regulator can best be done in the field:

• Connect a nearly or fully charged battery to the charger and check whether the voltage regulator starts reducing the field current:
  • The field current control lamps burn less bright.
  • The sound of the charger changes towards a higher pitch since it starts running faster.
  • The charging current is lower than its maximum value (as set by charger head, blocking timber, overall efficiency).
  • Switching on an extra field current lamp does not change the field current any more since the same current is just divided over more lamps so they all burn less bright. Now it is the regulator that controls the field current and not the lamps. This means that switching on an extra field current lamp has no more effect on the sound of the charger or on the charging current.
  If not, wait until the battery is charged a little more.
• Measure the voltage V_r over the regulator with a tester (this is not equal to the voltage V_i measured by the indicator on the switchboard itself since there is a voltage drop over the current shunt, see fig. 4.25).
• Readjust the voltage regulator until the voltage V_r at the regulator is 14.7 V. Check whether the regulator is still reducing the field current. If not, wait a little until the battery is charged some more and fine tune the regulator if the voltage is not 14.7 V any more.

Ideally, the regulator should provide full field current as long as voltage V_r is below 14.7, and once V_r reaches 14.7 V, it should reduce the field current in such a way that V_r remains at 14.7 V sharp irrespectively of the charging current. In practice however, it might not limit the voltage to 14.7 V sharp (the regulator will react to little peaks in the voltage V_r which are bigger when the charging current is higher, while the tester measures only the mean voltage). Therefor it is best to readjust the voltage regulator while the charging current is around 4 A. This is the charging current at which the battery is considered fully charged and at this current, a correct setting of the regulator matters most.

With most voltage regulators, also the temperature has a slight effect on the voltage setting, with the set value becoming lower as the temperature increases. This is to protect the battery since at a high temperature, overcharging happens already at a lower voltage. So don't think that something is wrong if you have adjusted the regulator carefully and the next day the setting has changed, it might just be the temperature that has changed. To complicate things further, regulators heat up themselves because of the power dissipated internally. So they change their own setting as they get warmer until they reach an equilibrium temperature. Preferably, the charger should run for say 15 minutes before the regulator is readjusted, and ambient temperature should be close to the mean daytime temperature in that area.

index page

next