What purpose the flash is, might be almost for a cycle lamp. Even if other purposes, it comes into question to be visibly. For pratical use it's better to keep high visibly and to hold down electric power consumption. In principle, that's possible due to make one flash short and keep current strength for one flash.

For an example, case of same current rating and period 0.3 seconds, it's not so different visibly between flash 0.1 seconds and flash 0.05 seconds though, its electric power consumption ratio would be 2 vs 1.
Flash 0.1 seconds with period 0.3 seconds is better visibly and holds down lower consumption than flash 0.2 seconds with period 0.5 seconds.

Admitting that, it has limitation to make flash time short. Maybe around 0.05 seconds is limitation, which depends on the circuit and its program, voltage, current.

Relaxation Oscillator is none other than the discrete circuit which could control such requisites. As a discrete, it would be only one. Of course it's kinda difficult to control, e.g. it comes to be difficult to keep current strength due to flash too short period. However relaxation oscillator is better by far to be made visibly and lower consumption than vibrator and ring circuits. It would be more accurate to say that vibrator and ring circuits are no good for efficiency.
If it needs to control period and flash time, current and so on, it has no choice but to use AVR or PIC, IC programmabled or to divert IC of ready made products.



A strong and momentary LED lighting is so beautiful.
Physical presence of luminary has come to be not sensated and only light leaves an impression. That's going into rhapsody too much though, a momentary lighting kept to be illuminative is beautiful.
It's tradeoff to be momentary and be illuminative.

That possibility depends on the circuit and the program.



A rechargeble NiMH battery is desirable to be long life. Most important approach is making no degradation due to over discharge though, it's easy to over discharge due to use for light.
Recently there are much lights with a Li-ion battery. A Li-ion with protection circuit would stop to discharge at proper final voltage, i.e. it would not be used at unexpected timing. It's not that no need protection circuit which has a risk to fire and explode due to over discharge.
It's passable and valid to use a NiMH battery for light in guessing, cause elctric density of a NiMH has come to close to a Li-ion's. Besides for hand-build light such a risk should be fended off in guessing.

As for a NiMH there is almost no risk though, its life come to be very short due to over discharge.
As discharge features of a NiMH, voltage drop is very small, when drop of intensity of illumination could be noticed it might be over discharged. Voltage drop of gadget with dcdc could not noticed especially.

In shortpoint it's concern to know the timing for exchange.
It's best to have a battery indicator. It's better to have a monitor lamp turned on in case of closing to the final voltage. Without such items, if it could be exchanged earlier it doesn't make to over discharge. And case of earlier exchange, it's suggested to use a discharger cause continuation charging makes battery memory effect. A discharger will stop at proper final voltage and help to keep battery good condition.
By the way proper final voltage is certain voltage should be stopped with certain load. Proper voltage might be exchange due to impedance of load.
Please refer to this page named "discharger" in depth.

A small front lamp could be made due to make in common system of front and tail lamp. In case of that battery and circuit would be mounted to the tail lamp. Front, around handle, is desired to be neat and clean especially for a roadbike or a sport typed bicycle.

There are many merits due to use the battery and the circuit in common. I had wished to use a security alarm in common though I couldn't find a good shell for it inmounted to a saddle.

As for a left picture there are parts of a head lamp. It would be most small class one in guessing. The battery and the cicuit would be installed to the unit inmounted to a saddle.
Of course bombshell LEDs could be inmounted to an atatchment or a holder though there is not reason to make so small.

Demerit of that is the cable which should be brought out and be splined in top tube.

Bombshell LEDs could be inmounted to any parts even if it's small part, e.g. a mounter of cycle computer, any belt, a brackets, and so on.
It's might be good to affix a plate mounted high performance LEDs to around head, to make an atatchment for bombshells, and so on.

At the end of consideration, --
I made an atatchment to be mounted to bottom of the belt of a cycle computer. It would be possible to take off and be in the shade. Around the head would be clean too. As for the atatchment, it was already made for another mono-head lamp.



There is a right IC to be diverted which is made in China, and has seven patterns of flash, is used for a cheap tail lamp and toys.
There are two types as old and new. New one has four control lines which is convenience to divide to front and tail. Old one has three control lines in which case it's good to use one line in common of front and tail.

Old and new chips have same count of pins. Old chip has five pins for three control lines, new one has four pins for four control lines and has one NC pin. New one come to stand higher voltage than old one.
Its basement come to be used better material.

Cut its basement with enough margin as a right picture. It could be cut more small for installation.

As for the IC, I wrote a chaptor named "divert IC flashing by Open Collector" in depth




Divides control lines to front and rear.
It's aim to charge pump due to oscillation without dcdc circuit, which could turn on lighting white and blue LEDs by two NiMH cells.
Focal points as follows:



Cause one flash time is very short by the IC, resistors as current limiter are no need. Its efficiency would be good in guessing.

It might be better to use resistors for blue LEDs. It's a matter of taste though a blue LED has high visibly respect to illumination.



As usual, tested on a breadboard. At first I used a dcdc though, which could not flash as expected. It could not turn off completely. I did investigation of cause, but I couldn't find it.
And I decided to adopt the charge pump method.

As a matter of fact, at the outset I had an intention to use a dcdc cause that's easy and simply. And I changed the plan to design the charge pump used oscillation. Thus I couldn't test the circuit on a breadboard well.


On a basement checked.





Cut the basement kinda larger than battery box and mounted some parts on it.


It needs a cable with three wires and put a connector on it. Three wires are for Vdd and two collector control lines.












It could be used as mono-tail lamp without cable to the front lamp.


Heaped LEDs of a front lamp. It's interesting that the flash lighting is moving ups and downs.





Set the front lamp as a right picture.





The light set to an atatchment mounted on bottom of a holder of a cycle computer. That's small and in shade.


As for an acrylic dome, it has visibly for side angle lower illumination though.





A tail lamp inmounted to a saddle.


Flashing due to momentary lighting is so pretty.




this way. Installed resistors for blue LEDs of a tail lamp.


A blue LED is easy to secure visibly respect to its illumination.

Resistors as current limiter of a tail lamp would make electric power margin for a head lamp, cause charge pump due to oscillation. That's efficiency in guessing.



This circuit drew after adding a Power LED for front and adding red bombshell LEDs to rear before exchanging count of cells to three.


Blue LEDs of a tail lamp had come to turn off during the Power LED of head lamp was lighting constant. That's voltage droped down. Added four red LEDs to tail lamp which could be exchanged by switch, which didn't turn off while a Power LED was lighting constant cause red LED's Vf was lower.

As for a right picture, three white bombshell LEDs were inmounted to the atatchment before a part of a Power LED mounted on it.

Increased count of cells to three which made blue LEDs not turn off while a Power LED was lighting constant.




This is a part of a tail lamp inmounted to a saddle.

Exchanged the battery box for two AAA cells to one for three AAA cells.
It would be located over side of LEDs.



Blue LEDs didn't turn off while a Power LED was lighting constant.






I wondered when the battery dropped down it's not so wrong cause constant lighting of a Power LED is used rarely.



One flash lighting had come to be long due to change the voltage of battery. Illumination is enough but it had been impossible to flash quickly and that made runtime short.
That could be adjusted due to exchange a condenser or resistors though which had been left.



Expected to be high efficiency, high illuminative and momentary one flash due to three cells 3.6V.


There is no constant current circuit nor resistor as a current limiter.
Puts the output of a dcdc into a Power LED directly. HT7750A boosts 3.6V of battery to 5.0V. The current is limitted by the indactance. That might be a wrong method though it had worked well. Maybe good efficiency?

Old part for bombshell LEDs is used an atatchment of an adaptor as same as before. There are four LEDs diameter 3mm. A recent high performance LED is very luminous.

I didn't draw on the diagram though, added a switch with 2 circuits 2 positions which could shift a Power LED and bombshell LEDs due to one motion.
It's possible to turn on both of them due to shift to bombshell side and turn on the own switch of a Power LED.




It's ideal to tun on the warning lamp though that would be need many parts. Simply I added a small red monitor lamp which would be turned off to reached proper final voltage of NiMH battery.
Voltage dropped is depends on value of load, adjusted to be standardized constant lighting of a Power LED.

Concretely, connected a zenner diode and a resistor, a red LED in series among Vdd and GND. Several mA current is enough for that.

This cheap tail lamp made in China has three output control lines and drives five red LEDs. It has seven patterns of lighting. Very good IC.
IC molded on a basement, a tact switch due to electric rubber, a hard shell and five lenses. The price was 100yen.
( never mistake a tail lamp with one control line and three red LEDs only )

As a mode of constant lighting which is one of seven patterns is drived due to oscillation which duty ratio is 0.25. That's why there is no resistor as a current limiter.

This IC has been used to various products e.g. for OEM and a lighting stick, toys, electric small objects, and so on.
As for OEM, it seems to be supplied to manufactures providing bicycle accessaries, and the same tail lamp is saled twenty fold in price. Its shell is colored red conformed to the traffic regulations though.

As products made in China flooding.

However superior aspect is only functional design of IC. Maximum allowance of voltage is too low and its solder is too frail. It might be broken due to put 5.0V into as an example, i.e. It could not be divert to use white or blue LED.
And it's not useful to exchange seven patterns by tact switch. That's bother to push repeatedly for every using. Case of momentary battery choking off on impact, that makes IC to reset and to turn off lighting.

Tried to educe positive aspects from the IC.



Aim as follows:



As for blue LEDs, it's enough three cells 3.6V to illuminate but to use a dcdc with one cell. A dcdc is diverted from a charger for celular.
It's not needed to high illuminate for a tail lamp and it could be lighted due to two cells 3.0V. However in case of two cells drop down it might not be lighted even if remaining battery capacity. It would be possible to use up the battery and to save weight due to a dcdc.

It has a risk to break the IC due to put the voltage boosted into IC directly. Simply I tried to divide the voltage by resistors and found that the IC was driven till around 0.7V.
However, lower voltage made period flashing slower especially around 1.0V though it would be no problem to select quick pattern.

Its load is very small and battery doesn't drops down to drive the circuit with a dcdc. As for a test, case of using battery droped down and used up by a discharger, the battery rebounded to 1.1V during driving it.
i.e. the load is very small, and one flash time might be very short in guessing.

Thus, decided to put the IC from a cell directly. The period of flash would not be slowly due to use battery normally. If the period is noticed to be slowly the battery would be already over discharge.

There is no main switch on the original circuit of the product. Cause electric power consumption of the IC is uA level, the IC itself is kept to be supplied power. The battery capacity would not decrease due to the IC only whether LEDs turned on or off.
It's no problem to connect the cell and the IC directly as same method of the original product.

It would be possible to fix a specific pattern of flash due to prepare another switch with only two positions which could turn LEDs on or off. i.e. that's main switch.
The tact switch which exchange patterns of flash is not need to use routinely.

In that case it's a problem to choke battery off which would make the IC to reset and turn off LEDs. Cause the tail lamp is used for bicycle which is oscillated and is on impact. Simply a condenser could prevent the problem.
A condenser would be enabled among main switch turned on cause it has few self-discharge even if slightly.

That's about it skimmed the cream.

By the way, Used two control lines of three ones. Cause I would like to make the flash pattern symmetrical and so selected three LEDs and two control lines which could be simply made a symmmetrical flash pattern. It might be a symmetrical flash pattern due to select three control lines and five LEDs though which are too much LEDs.

For an example, 6th flash pattern is repeat to turn on no.1, no.2, no.3 in order, which would be assigned with center LED, bilateral LEDs and none, that could make two patterns as follows:


Impression would be convergence or diffusion.



Test on a breadboard.





Mounted some parts on the basement as a right picture, and checked to light.





Cause it would be used in the open air it should be filled resist painting.


The basement of the IC was cut too small. Its soldering would be peeled easily and I fixed it by resin.










After testing, it should be proofed by something e.g. resin or hotbond, epoxide resin.







Set to be inmounted to a saddle.



I noticed as follows due to use this tail lamp.
Compared to alternately flashing by a multi vibrator and a LMC555, the lighting was lovely and its runtime is so long as mentioned already. Besides the momentary lighting visibly was no less than, contrary to preconception. The momentary flash lighting visibly would be preferable due to flash quickly.

in brief,


as functions,


The tail lamp had a good reputation. I had made few ones for my friends, which could be made easily.

Due to multi control lines, in this case three control lines, it would repeated to turn on and off in order.
I tried to control time of one line flash lighting only which could control period time of one rotation.

The load of a ring circuit is large comparatively, which could not be boot by one AAA cell. After booting at once it could be driven easily by one cell though. And for its booting, I built in a CR2032 battery as trigger which could be a reserve battery.


A dcdc was diverted from a charger for celular.



Decide values of condensers, resistors and so on. Its scheme is same as the multi vibrator.

Test, try and cut, on a breadboard.

As for R3, R4 --
Those are factors to decide time of two lines flash lighting of blue LEDs which would be desired to be under 0.1 seconds.

A volume resistor is ready for control time of a line flash lighting of white LEDs which could control time of a rotation around 0.3 ~ 0.7 seconds.




A shell was a battery box for two AAA cells. The circuit was installed to space of one cell as same scheme as the multi vibrator.
Same design too.

A time of flash lighting of white LEDs could be controlled by the VR. I had an intention to mount two white LEDs though, mounted only one small white LED.

A CR2032 battery was ready for trigger for booting.



On back side of the shell, LEDs and wires, switches were exposured which should be filled with something for proof e.g. solder resist and epoxied resit.









LEDs were protected by a acrylic plate.

Compared to the tail lamp due to the multi vibrator, its illumination could be contained though which was still too illuminative for a tail lamp.

Due to getting decrease the time of flash lighting of a white LED, a white LED was turned off though, there was still short interval which was all light out. Curiously, it could be found the timing to exchange battery due to set a white LED lighting very short time which would be turned out completely when the battery droped down.






I don't know why though it would be all lighting due to turn VR toward the end of short time side. As precious opportunity? I took a picture as right one inmounted to a saddle.

There were puzzles about the ring circuit.



It's a shame to need an asist battery for boot, and I tried to exploit charge pump by a condenser. If that had been succeeded its procedure would be kinda bother.


That was failed as a result.
Certainly that could make voltage 3.0V though it seemed to could make too low current due to a condenser.

It's a circuit of standard IC named LMC555 withought artifice. That would flash lighting in alternate shifts and be exchanged period of flash lighting due to a VR. A right side of the circuit diagram is a dcdc which was diverted from a charger for celular.
A trigger battery CR2032 as same as the ring circuit which was not drawn though shold be needed. And the switch had three positions as off/on/trigger.

In case with a dcdc, two resistors as current limiters had a problem that required attention. That might be not unique to the dcdc in guessing.
When voltage of LMC555 OUT dropped down, the load would came to be small and the voltage of the dcdc output would be increased. A value of R4 should be half of a value of R5 normally though, in case of this circuit, R4 should be twice value for current balanced.
If resistors had been located to each LED particularly, two resistors might be ready as quadruple value of R5.



Shaped parts.

A period of flash lighting could be exchange due to a hemi fixed resistor which could be adjusted by +driver.




Tried to operation check after some parts mounted on a basement.



It would be used out of doors and the parts exposed should be draped over something e.g. regist, hotbond or epoxied regin.
With those processes, finished.



Once again as same as the ring circuit, a trigger battery CR2032 would be atatched.


That said, the trigger battery was no need for boot to lighting with full charged main battery, but need to work out the main battery.





After full test, it should be proofed by epoxie resin or something.



A switch had three positions, which was not drawn on the circuit diagram, off/on/trigger as same as the ring circuit.



"can perform the perfunctory tasks with the 555"
"the 555 extends sanctuary from discrete"
Those wordages are commonly used, and the 555 is actually so much.

However, Those are almost experiments for beginning electric craftwork. As for an item of practical use, 555 has an aspect which might not be useful.

Its runtime was only seven hours due to one AAA NiMH battery 800mAh fullcharged.
There was no time turning off all lights, one flash lighting was so long. It's not suitable for practical use compared to the relaxation oscillator and the capacitor charge pump those could be running hundred hours.

A ring circuit and multi vibrator are same as about that.

This is a typical circuit as an astable type. Its period of flash lighting could be adjusted due to a VR. It's a point of the circuit that one side is always lighting in alternate shifts. A dcdc was diverted from a charger for celular which was drawn on upper half of the circuit diagram. An astable multi vibrator on lower half.
I drew 1.5V for battery though, An AAA NiMH had been used for which is practically 1.2V.




As for the program or the application, in plain words that's about condensers, resistors and values of them, due to cut and try.
On a breadboard.


Cut and try is a good method for study about functions of each part and for no design mistakle. It might provide excellent performance unexpected.

As for resistors, there are high performance LEDs recently and it's an open question that the rated current is needed for a tail lamp. Its runtime could be longer due to under drive. However under drive would make quick flash difficult.
i.e. that would be trade-off among illumination, period of flash lighting and runtime, due to one resistor.

That trade-off would be responded to transistors and velues of condensers. For an example, transistors for strobe would make flash lighting to be illuminative easy, but to be quick flash difficult. A large capacity of condenser would cause the same affairs.




Inmounted the circuit to a cell space half of a battery box.
Of course a battery to other half.

The battery box was inmounted to a saddle to be in shade. There is a possible saddle and an impossible one.
Only LEDs were shown up under shade.

There were two VRs which could control time of one flash lighting both center LED and two sides LEDs.



LEDs and wires, switches became exposed which should be proofed due to regist and epoxie regin or something.









LEDs were protected by an acrylic plate.



Flashing in alternate shifts.

It's too illuminative and quick flash lighting had come to be difficult. Its brightness could be visibly around road surface. That's an error in the evaluation of the test phase.

Flashing three LEDs just all together. That's one of dcdc due to condensers discharged and charged.
Oscillation could make both flashing and switching to discharge and charge at the same instant. As for this circuit, it could boost to triple voltage of the battery that could illuminate white and blue LEDs. It's impossible to be so illuminative though its runtime would be so long and its momentary lighting would be lovely. That's no problem for a tail lamp in guessing.


Blue LEDs could be turned on without particular dcdc.



At first, decided an adjustable range of period of flash lighting. That's appropriate and effective around 0.4 ~ 0.7 seconds cause it's impossible to adjust short period due to be needed time for charge. A longer period had an advatage as the circuit though it's desired to be short as a tail lamp.
With that, decided values both VR and C1. It would be discharged in a moment and almost of time of period spent for charge. There were many trade-off.
Small C1 could not be illuminative but could flash quickly. The focus was the time of period. Decided values of R4, R5, R6 and R7 for charge enough C2 and C3.

It's no need to limit the current though a resistor several ohm was installed for protection. No, it's really no pleblem without the resistor.




Its circuit was inmounted to one cell space of an AAA battery box as same as the multi vibrator and the ring circuit.
Its scheme is same as those.

A toggle switch was kinda clunky...



By the way, It's not made in accordance with the circuit diagram.











A momentary flash lighting of ble LED is so lovely.
That's impossible to make someone realize due to pictures though.



The runtime could be handred hours easilly, possibly over a thousand hours.



It's illuminative modestly like this. It's a sign of low illumination to get a view outline form of a LED by camera though, it's not so wrong as a tail lamp.