ALTERNATOR GEN LAMP CIRCUIT MODIFICATION
© Copyright, 2014, R. Fleischer
Purpose of this modification: After this modification, the Bosch BMW Airhead motorcycle alternator will produce electricity even if the GEN lamp burns out ... or; if the GEN lamp printed circuit material (it is common for it to crack at the lamps area) does not connect to the lamp properly. This modification can be adapted for use where there is no normal BMW instrument pod. This modification, while referencing the 'instrument pod', applies to all types of the instrument pods BMW used, as well as the /5 bikes, where the GEN lamp is mounted on the headlight bucket, and there is no pod.
NOTE!...This modification has been done by many in past decades...and the 'accepted value' has generally been 470 ohms. I will expand upon the that resistor value determination. I will show that a slightly lower value is slightly better and makes an improvement in the rpm at which charging begins. I will also show why a much lower value is appropriate if you do not have a GEN lamp.
General comments about 'modifications': As a general rule, the 'factory knows best' is a fairly accurate statement. However, few of us own UNmodified motorcycles. Our bikes are a reflection of our personal
desires. I am NOT in favor of many of
the modifications that we all see or hear about. Some of these modifications, which I have
done myself, are not very economical; or, do less than is often
believed; or, are not good ideas in general; especially for the average rider/owner. I try to be
honest, describe all sides of any controversy or what can happen
with any modification.
What about this alternator GEN LAMP modification? If done INcorrectly, it will reduce reliability. You could damage the instrument 'printed' wiring connection board or produce a short circuit under the gas tank...etc. Note that these lamps seldom fail. Unfortunately, most folks never inspect them for sagging filaments, a sign of impending lamp failure; I don't recommend that for the function lamps anyway, due to the possibility of damaging the flexible printed board material. Most folks will never take a cover off the instrument pod until one or more lamps fail, or there is some other maintenance reason. Note what i said about the flexible printed board, as even careful folks have damaged the printed circuit board in the pod, CAUSING lamps' problems. When the lamps do fail, they fail at rather inopportune times. I am not necessarily recommending this modification; but I HAVE done it to my own BMW Airhead bikes.
Theory: If the GEN lamp fails, you will USUALLY not have any charging. At key turn-on to the RUN position, the battery positive (+) connects to the ignition switch, then from that switch electricity goes through the GEN lamp, and then through the voltage regulator, to the Df rotor brush. A small current flows, due to the internal resistance of the lamp, and some effective resistance in the voltage regulator, and that small current travels through the Df rotor brush and associated slip ring, into and through the rotor windings, back out the D- slip ring and brush, and then to chassis ground, making a complete circuit and temporarily lightly magnetizing the rotor. That magnetization will completely or almost completely disappear when the key is turned off. Sometimes a bit of residual magnetism remains, which is why the alternator will sometimes produce electricity if the lamp is burned-out, often this takes initialization by rather high rpm. Once the alternator produces electricity, it will continue to produce electricity until the rpm falls to some value well below 2,000, at which time the alternator may continue to produce electricity as the rpm rises, depending on small details and also how long the rpm was, for instance, near idle; and thus the process can be, or might be, repeated.
The rotor must FIRST either have some residual magnetism....or....be slightly magnetized by the lamp and rest of above described circuit, until such rpm is reached that the alternator self-energizes. With an intact lamp and connections, this occurs at some rpm below 2000 rpm; the exact rpm depends on the rotor, etc., in the particular motorcycle. It is true that increasing rpm to, PERHAPS as much as 5000 rpm will possibly allow charging due to some residual magnetism that MIGHT BE in the rotor. You can not depend on this. While you COULD do some temporary jumper wiring, such as jumpering the battery + terminal to the Df terminal (or, better, using a lamp in a series circuit, from battery + to Df), to get the alternator operating in the case of a burned out lamp, it is a hassle, compared to what is proposed below; which is permanent. The modification described below allows the lamp or lamp contacting area at the foil printed material to fail and you still get charging...and you can replace the lamp elsewhere's...at your leisure....rather than by the side of the road during a downpour. The charging with the higher ohms resistor modification, and a burned-out lamp, will not begin at as low an rpm as with an operating lamp, but you do NOT have to spin the engine to quite high rpm either. If the low ohm resistor modification is done, you do not need the lamp at all, as performance will remain just about the same as if the lamp and instrument pod did exist.
NOTE that if you should have a wiring failure in the lamp circuit, this
resistor modification may not help; depending on where and how you install
the resistor. Most wiring failures in the lamp area are at the lamp's
actual contacting via its socket, to the printed circuit flexible material in the instrument
pod....usually due to poor technique in removing replacing lamps. Much more rarely, but has happened, the failure is inside the rubber plug that pushes into the rear of the instrument pod, or, a cracked solder joint at the mating male connection to that plug, inside the pod (fixable by careful cleaning and soldering). If the proposed modification resistor is mounted with good workmanship under the fuel tank, there are no such problems. For the /5 bikes, there is no flexible foil material, so the lamp R/R is not any problem in this regard.
Theory, in more depth: The actual circuit routing is a bit more complex than earlier described. One side of the GEN lamp is fed by the battery (after the ignition switch), and the other side of the lamp not only connects to the voltage regulator input (D+), but that side of the lamp also connects to the alternator positive (+) output of three SMALL diodes on the diode board. When the alternator is not spinning, or spinning slowly, let us say always under 1200 rpm or somewhat higher, the alternator output is very tiny, if any at all. The GEN lamp will be lighted, as the battery current flows through it and then through the regulator and rotor, to engine case ground (battery negative). Once the alternator stator output increases enough, the three SMALL diodes rectify the stator output (rectify means to change A.C. to D.C.) and the voltage on BOTH sides of the lamp is now approximately the same, AND of the SAME polarity (+) ....and the lamp has so little voltage drop across it, that it appears to be, and is, NOT lighted.
The lamp will NEVER supply enough current to come anywhere's close to fully energizing the rotor; when large outputs from the alternator are required..., that function is done by those small diodes, supplying the several amperes needed. As rpm rises, the small diodes pass far more current into the voltage regulator, which supplies the current needed to fully magnetize the rotor, which causes output of the alternator to greatly increase. As the output voltage from the diode board diodes begins to approach the desired amount, the regulator begins to reduce the current flowing into the rotor from the small diodes.
For the NERDY: The voltage (well, current flow) being regulated is that of the small diodes output, and not directly the big diodes output, which goes to the battery and rest of the bike's electrical system. It is the LAMP that separates the main output of the big diodes from the small diodes output, as far as 'sampling' the voltage is concerned. There are some more complex reasons why the small diodes are there, and why this seemingly strange sampling method was done. I will not get into these reasons, they will only complicate things here.
SOME FURTHER INFORMATION: A dim lamp at relatively modest to high rpm usually means some sort of corrosion at connectors, or a bad rotor or overly-worn brushes. A bad battery can also cause the alternator to produce a very large output & the lamp might glow dimly. If you want an explanation of how that works, ask me on the Airheads LIST. A bright lamp at riding rpm USUALLY means an open or shorted rotor or bad regulator, or a bad diode board, and likely there is no charging (no alternator output) at all.
When one or both brushes is worn enough (they tend to both NOT wear evenly), the snail spring that
supplies pressure onto the brush may begin to contact the plastic brush holder,
and that reduces pressure on the brush. This is a very common complaint and shows up as the brush
nears the end of its life as noticeable
GEN lamp lighting (usually dimly) ...but with increase in brightness, as rpm
rises. The reason for this is the SLIGHT wobble (called run-out) of the
rotor...which moves the brushes in and out ever so slightly, and the snail spring
is also slightly bottoming and cannot fully follow the brush
movement. You can think of all this simply as the brushes are not making consistent or high enough pressure in contacting the rotor slip rings. It usually happens on one brush first.
Brushes are 16.5 mm long when new, measured from square end to middle of concave end, and are worthless (but, read on...) once the spring contacts the holder. Brush life depends on riding conditions....dust, dirt....and how much output the alternator is called upon to produce, over time/miles. Generally, 60-80,000 miles is a typical brush life. I have seen the outer (forward-most) brush wear down to unusable, in 20,000 miles, under very dusty conditions. If one or both brushes have worn to the partial contacting point, and you are on a tour, you can put a tiny piece of thick paper between snail spring end and the brush outer end; this will keep the snail spring from contacting the brush holder, and you can ride on, and likely for a very considerable distance, even 5000 miles would not be unheard of with a thick piece of paper. REPLACING the brushes is a bit of a bother, much easier to do it at home, than on a long tour, so, while I do recommend inspecting the brushes maybe once a year for length, the added piece of paper will work well, and is not difficult to do, if a bit fiddly if the inner brush is the problem one.
For those that are interested in the energizing current value into the rotor, with the engine off, ignition switch on, GEN lamp lit, it is on most models ~ 0.2 ampere. The added resistor of The Modification itself passes little, typically around .040 ampere or so, and you can see that the vast majority is passed by the lamp. If the lamp burns out, the much smaller resistor-allowed current will initially adequately energize the rotor, although alternator output becomes first available at a somewhat elevated rpm. The resistor modification does not make the system perform EXACTLY as when the lamp is intact and lamp properly operating in the circuit. Maximum output, and output at any reasonable riding rpm is NOT affected; only the lower rpm area is affected. The reason is that the resistor's purpose is to ALLOW the alternator to produce something, should the lamp fail, and once the alternator IS producing usable electricity, it SELF-ENERGIZES. NOTE again, that the maximum output of the alternator is NOT diminished (and the rpm at which maximum is reached is NOT changed either).
This consists of installing one common and quite INexpensive standard electronic part called a resistor. I will, later here, also describe using an LED in place of the stock lamp. I ran actual tests for the optimum value for the resistor modification for the GEN (alternator, really) lamp circuit.
Any incandescent lamp has a
fairly low resistance when cold (not lit) and that resistance increases a fair amount when the lamp lights up. Whilst a nerdy point, this WAS taken into
account in this article.
The resistor, which substitutes for the lamp if the lamp or lamp connection fails, is connected across the GEN lamp electrically, but not necessarily mechanically at the lamp itself (or its socket), although that is a nice place for it, if you use the higher resistor values in this article.
There are TWO methods for The Modification, that I approve of:
Method #1. Neatly clean the area and solder the resistor across the GEN lamp socket wiring itself. The resistor can be made more secure from vibration by cementing it with a small amount of silicone RTV or similar, although not absolutely needed,...that is because I suggest quite short resistor lead lengths as you do not need leads vibrating nor shorting to anything. The resistor will produce a small amount of heat during the rpm/time the bulb is normally lit, but this is seldom over a minute in duration at a stop sign, and in any event, the heat amount is small...especially with a 470 ohm value. Those installing the resistor inside the instrument pod and across the printed circuit material might well use the 470 ohm value, considering the heat, but I have successfully used a lower resistance value there. I clean the area by starting with a sharp Xacto knife on a very flat angle, scraping away the coating over the copper, very carefully. I may use a small piece of slightly abrasive sandpaper afterwards, perhaps 320 grit. I have also used both pencil and typewriter erasers after the initial knife or sandpaper. For the /5, there is no flexible printed material, and almost any way you want to install the lamp NEATLY in the headlight bucket is OK. For soldering to the flexible material in the pods, you really want the exposed copper to be clean and shiny prior to soldering. Soldering must be done with a small soldering iron, with a reasonably decent sized tip, with rosin-core solder (60/40 common type is best), and quickly.
Method #2. Install the resistor from the terminal of the ignition coil that connects to the battery circuit. This is the terminal that has the green/blue wire. There is often an unused male spade connector available there. Connect the other side of the resistor to either of the blue wires coming out of the voltage regulator plug [this is D+]. You need to do this neatly, with no chance of bare wires, nor vibration breaking them. Insulate the resistor and wiring with shrink tubing, and use proper all-plastic wire-ties, as required. For SOME, this may be a preferred point of attachment, to avoid damaging the flexible circuit at the lamp (if you have a later bike with the instrument pod, not a /5). Mounting the resistor as in this method #2 creates less chance for ham-fisted folks to damage that thin flexible printed circuit board. The drawback of Method #2 is that there are probably more chances for messy workmanship, and vibration problems. This method #2 is the way to go if you are going to use the low ohm resistor, say 50 ohms or so, as described later in this article, as that resistor can get much hotter than the higher ohms resistor that I recommend if soldering the resistor onto the pod flexible material.
This modification can be done to a stock /5, which is somewhat different, no pod, no flexible pod circuit board material, but the same situation applies to the GEN lamp. ...and the modification can be
done at the voltage regulator in the same sort of manner....or in
Resistor discussion: Nearly any type of resistor may be used. It can be the old-fashioned type called 'carbon composition', or any modern type of 'film' or 'deposited carbon' type....or even a 'wire wound' type. The resistor value for installation into the pod or /5 bucket CAN be 470 ohms, but my tests indicate that 330 ohms performs slightly better.
All BMW alternators from all years on our Airheads can use the same value resistor, even though different rotor resistance's were used over the years, because the resistor value is vastly higher than any rotor's resistance.
Resistors NORMALLY come in certain 'standard' values, you could use 270; 300; 330; 360; 390; 470 in the instrument pod...but I prefer the 330 value, as a good compromise between charging characteristics and heat produced. Too much heat MIGHT (not WILL) injure the instrument pod parts...if that is where you mount it, over the long term. Never seen it happen with 330 ohms, ...remember that the resistor is only energized at ANY heat producing level ...at the same time the lamp would normally be brightly lit...that is.....when the alternator is not spinning and the ignition is on (pre-starting)...OR; when the alternator is at a low rpm, idle or somewhat above. Under the very worst conditions I can think of, a 330 ohm resistor would produce about 0.6 watt. If you want to be ultra sure of not damaging the printed material, and you will be mounting the resistor at the printed circuit socket, then use a 470 ohm resistor, which will produce only about 0.4 watt under the worst conditions.
The resistor should be rated at 1 watt or more. However, most modifications have used the smaller physical size of a 1/2 watt resistor, and I have NEVER heard of any failing. Do not use anything smaller than 1/2 watt rated resistor. If you try to use a too large resistor size you may have problems fitting it into your instrument area or neatly installing it under your gas tank. Do NOT mount the ~50 ohm resistor mentioned in the next section, in the pod.
Low ohm resistor information, Cafe bikes, use of LED lamp instead of the incandescent lamp... ETC:
If you are building a Cafe Racer Airhead, and eliminating the stock instrument pod, or want to use an LED anyway....you may be interested in this.
The GEN lamp is rated at 12 volts and 3 watts. Ohms law says the lamp resistance is 48 ohms when energized by 12 volts. NOTE AGAIN that is the resistance when the lamp is fully lighted-up. That is very considerably less resistance than the value of lamp bypass resistance used earlier in this article. Keep in mind that once the alternator produces enough electricity, the lamp is, for practical purposes, not involved...it is extinguished automatically, as charging begins.
It is a property of incandescent lamps
that the LIGHTED resistance is MUCH HIGHER than the UNLIGHTED resistance.
The unlighted resistance of the stock lamp is about 5 ohms. So, the lamp is around 48 ohms lighted
fully, and around 5 ohms not lighted. A lamp is nearly ideal for the purpose of both indication and providing initial rotor magnetization.
If you wanted almost exactly the same charging versus rpm as when the lamp was operational and in the circuit, say, beginning at a typical 1500-2000 rpm... and wanted the same charging curve at the lower rpm area,... then, if NOT having a lamp, you would want to install a standard 47 or 50 ohm resistor, rated at 5 or 10 watts. That resistor SHOULD NOT be installed in the pod...because if you are at idle rpm, the resistor will dissipate 3 watts of heat, and heat up to pretty hot within a minute. If you had no instrument pod, you could install such a resistor elsewhere's. I HAVE installed such a resistor inside a pod, quite successfully, but the resistor was mounted on a small metal plate (for heat dissipation) I made-up, the plate being mounted to the pod housing. Two folks who I chatted with long ago had mounted the low ohms resistor on the back side of the instrument pod or the underside, with wires leading to the inside.
OTHERWISE: The resistor could be mounted/soldered onto brass eyelets you put into a small piece of fiberglass type printed circuit material, and mount the tiny board someplace convenient....perhaps along the frame backbone. You could even use a screw-mountable metal-cased power resistor and mount it on the copper side of a piece of printed circuit board. You do not need a lamp at all with the low resistance modification. For the Cafe Racer conversion, with no pod; a clean and neat mounting of the low ohm (say, 50) resistor someplace convenient and yet protected, is all that is necessary; with some consideration to the heat produced.
Nerdy point: While there ARE special RESISTORS (which may not be their name) available that have characteristics similar to lamps, and some other electronic devices also have such characteristics, these items are NOT part of any modification YOU should be doing....they are not easily found, and usage is possibly more complicated.
HOW-TO-DO-IT, step-by-step, for a 470 or 330 ohm 1/2 or 1 watt rated resistor, inside the /6 and later instrument pod:
1. Go to Radio Shack and get a nice bright LED that is rated at 12 volts. A 12 volt LED already contains internal components and you will avoid having to do a more complex installation. Be sure to get the proper color!
2. Solder the above LED onto the cleaned copper traces. You can do this to traces with the coating on them if you first clean off the coating with a very sharp Xacto knife on quite an angle, scraping the coating off. Make sure the copper is shiny. If not, scrape a bit more, again, quite an angle for the knife....and maybe afterwards use a pencil eraser on the copper too. You can also use 320 or other very fine grit sandpaper.
3. You will need a 25 to 35 watt soldering iron with a reasonable sized tip which has a bit of mass, so you can solder quickly. In general do not use a soldering gun or pistol, they are too difficult for to use.
4. Use 60/40 solder, the most common rosin core type, or the also common 37/63 which may be marked 63/37. Use solder of about 1/16” diameter (.032” is common) which will have plenty of hidden flux in the center of the solder. I suggest NOT using ultra-thin solder, as that has too little flux, and is harder to use for amateurs, takes a bit longer to do the soldering job, and you don't want lengthy heating. You could also coat the bright shiny soldering area with a wee dab of rosin flux. The soldering iron tip should be cleaned less than a minute before the soldering, and the tip well-tinned with solder after the tip cleaning. Apply some solder to the tip, then put the tip on the wire/trace to be soldered, add a bit more solder and finish the job. Hold the wire in place if you have to, to avoid any movement until the solder cools a few seconds. Do the complete soldering in a 2 to 3 seconds.
DO NOT use acid core solder.
5. IF using an LED, they are polarity sensitive, so you must install the LED in the correct direction. If you install it in the wrong direction, it will not light up with the ignition ON, and engine not yet started. LED’s polarities will be on their package, but it is hard for laypeople, sometimes, to determine which side of the pod socket the + and – sides of the LED connect to. Since there are differences between instrument pods and /5 bucket, etc., I have not given specifics here on how to determine which lead of the LED goes where; but a voltmeter will show, ignition on, the polarity. Since 12 volt LED’s are unlikely to be damaged by installing them backwards, if your LED does not light up, reverse its leads.
6. LED's do not draw much current, so you still need a resistor, as in the incandescent lamp situation. You need to solder the resistor wires across the same connections. You MIGHT find it easier to do THAT first. I carefully mechanically wrap the LED bare wires a turn or turn or bit more around the resistor leads, and cut and shape the leads, etc., so when the soldering is finally accomplished, the LED is in correct position.
Values of 330 to 470 ohms are OK.
04/11/2003: add .htm title; edit for clarity, add information on snail spring bottoming, etc.
02/06/2004: clarifications on wording, nothing at all substantial
11/22/2009: more clarifications.
01/10/2011: Add 'A bit of Nerdy Information"
06/03/2011: Clean up a bit
09/29/2012: Add QR code; add language button; update Google ad-sense code; edit article for clarity
and brevity. Later, the troublesome language button script coding was removed.
12/05/2013: Add more details and re-write article for clarity.
02/02/2014: Add some emphasis, clarify details between /5 and later and how charging circuitry works, more on the LED installation, etc.
© Copyright, 2014, R. Fleischer
Return to Technical Articles LIST Page
Return to HomePage