Circuits to avoid low speed front LED FLASHING with hub generators?

[SA_SA_SA_SA]

I don't like  the annoying fast  flashing of front hub dynamo led lamps at low speed: also what is the point of, e.g. the UK government, limiting UK flashing cycle lamps to 1-4hz for the sake of epileptics if any old fast flashing is accepted from a hub dynamos front led when ridden at low speed or being wheeled?

perhap one day German epileptics will persuade the Stzvo to tighten up their requirements on front hub dyno LED lamps but until then:....

Has anyone built themselves a circuit to fit betwixt hub and front lamp* to mitigate this?
It would be a nice if a battery could be avoided (cos otherwise one would just be dusting off ones copies of dynamo backup circuits... )

Presumably a relay to switch the extra antiflicker  circuitry out at moderate speed is the most efficient if clunky way to retain efficiency


*(I am presuming all rear lamps will be connected via front lamps  terminals for that purpose)

[Brucey]

I don't think such a system can easily be made as a 'bolt-on' because it would likely require conversion to DC and many (most) dynamo lights are not rated for use with  DC, presumably because the rectifier parts are likely to be overloaded (all the current passes though half the parts if the rectifier is powered via DC).

The simplest way of getting some benefit as you desire would be to install a much bigger smoothing capacitor between the rectifier and the rest of the circuitry in a headlight, presupposing that a normal rectifier is used.

The downside of this is that at a walking pace the light output of the system might be steadier but it may (net) also be reduced somewhat.  There may be other problems too.

 BTW if you fit a relay to switch over so that there is briefly a 'no load' state, it might switch at speed in which case the open circuit voltage of the generator might get to >50V or so; this may be too much for a relay unless correctly specified.

The fact that dynamo lights flicker slightly at low speeds is not significant IMHO, not in practical terms.

cheers

[grams]

One of the reasons headlights flicker is because the standlight brightness is lower than the normal brightness, so you're seeing the normal brightness circuit kick in every time a pulse comes from the dynamo. A simple hack would be to tweak the circuitry so that it's all the same brightness, at the expense of standlight life.

I think the reason lights make no attempt to solve this is that it's a moderately hard problem. I think you'd have a hard time doing it without a microcontroller and actively monitoring the pulses from the wheel, then making a decision when to switch in, which adds a lot of components to your circuit. The decision point vary significantly between dynamos and lights, so it'd be hard to do as a general purpose external box.

[SA_SA_SA_SA]

I don't think such a system can easily be made as a 'bolt-on' because it would likely require conversion to DC and many (most) dynamo lights are not rated for use with  DC, presumably because the rectifier parts are likely to be overloaded (all the current passes though half the parts if the rectifier is powered via DC).

I know the owners of B&M Cyos were advised DC must be below 7V5 (I had presumed due to 'inverse-series-zener' style voltage protection, which I thought would be best assumed to apply to other lamps, but the rectifier rating didn't occur to me: as 4 alkaline cell (4.5--6V) battery backups used to be available I presumed manufacturers should at least make sure their lamps withstand such voltage. The lightspin bottle dynamo actually produced DC from a switch mode PSU I think.
Also, setting lamp aim is easier if DC can be used (and B&M sell a battery device to do that): I had to make an battery to AC generator for my DC hating Herrmanns One S(V2) aiming....

If the rectifier rating is a problem (surely it implies cutting things a bit fine...from AC the diodes may see higher peak currents than from DC if feeding a smoothing cap) then I suppose would could add a DC-AC oscillator but this is getting rather complicated: so if using batteries I would probably just rely on the lower voltage of the DC backup.

But the battery-free solution I half have in my head involves effectively chopping up the slow sine wave hub output to give a faster non sinusoidal output: this is probably both complicated and incurrs an efficiency loss hence my asking for any preexisting circuits and mentioning relay bypass
It would be much easier, if on behalf of epileptics, the Stzvo required the manufacturers to do this flicker reduction work inside their front lamps

...I think the reason lights make no attempt to solve this is that it's a moderately hard problem....

Which is why I would like the Stzvo to force them..

I dislike the fact that B&M E-bike lamps are  separate from the generator ones: surely increases stock variants (cost) etc as well as reducing flexibility. A dynamo lamp with a DC input would suit me (enough dynamo voltage to avoid flicker would take precedence over battery power).

....The simplest way of getting some benefit as you desire would be to install a much bigger smoothing capacitor between the rectifier and the rest of the circuitry in a headlight, presupposing that a normal rectifier is used....

From bottle/FER spoke dyno experiments with reflectalite-style but-DC-only-LED bulbs, I am not sure about that:
Adding a large cap to a simple bridge rectifier and said LED bulb worsened flickering*: cos presumably  cap charged, so LED PSU saw extra voltage, went to higher power and drained the cap before the next current peak recharged the cap, hence a gap==flicker....

*Without cap, at very low speed it just 'warbled' which seemed bearable and much better than the flickering a cap induced.

Also, modifying a STtzo approved lamp would nobble its approvedness I expect .

....BTW if you fit a relay to switch over so that there is briefly a 'no load' state, it might switch at speed in which case the open circuit voltage of the generator might get to >50V or so; this may be too much for a relay unless correctly specified....

Interesting, I have never seen that mentioned in magazine dynamo battery backup articles. In filament bulb days there were chattering relay battery backups like the Pifco (halfwave rectifier diode and big cap driving relay coil), but my proposed way to avoid relay chatter at switch point would be to use a DPDT relay and switch the generator into a 12R-ish 3Watt resistor rather than leave it open circuit... Plus, perhaps some inverse-series zeners if needed to stay within relay rating.

....The fact that dynamo lights flicker slightly at low speeds is not significant IMHO, not in practical terms....

While I merely find it annoying, presumably front bright LED flicker is as significant to epileptics as out of allowed 1-4Hz range flashing battery lamps (even in range ones I think caused some mumbling to CJ).

[Brucey]

the time constant of any 'clever circuit' circuit that powers the LED ought to be such that it is long enough that the circuit won't cause flickering at low speeds.

cheers

[Cudzoziemiec]

First thing would be to find out whether it is actually a problem for epileptics. The flickering of dynamo lights at walking speed looks to me to be slower than that of most flashing bike lights, and is certainly less intense (because flickering from bright to dim, not bright to off).

[Kim]

First thing would be to find out whether it is actually a problem for epileptics. The flickering of dynamo lights at walking speed looks to me to be slower than that of most flashing bike lights, and is certainly less intense (because flickering from bright to dim, not bright to off).

A friend of mine has had migraines (of the go and sit in a dark room rather than whatever you were supposed to be doing kind) as a result of their Brompton's dynamo light flickering, so it's certainly an issue for some.  Of course, if you're that sensitive, vehicle headlights moving around are a hazard anyway.


It doesn't bother me, as it's only a transient state when starting and stopping, and I think a flicker is possibly more useful than a dim light (which would seem to be the choices, unless you get some extra energy from somewhere) for seeing where you're going.

If you do go down the route of getting some extra energy from somewhere, 4DToplight Multi approach would seem appropriate: Light runs permanently at full brightness, using dynamo power preferentially.  Otherwise it's powered by a couple of AAs.  Electronically simple, and battery life is extremely long, as it's only doing standlight duty.  You lose the purity of dynamo power, but it might not be a bad approach if you're a slow climber.  (My approach for when slow climbing is likely to be an issue is to have an independent battery light as a backup.)

[SA_SA_SA_SA]

...It doesn't bother me, as it's only a transient state when starting and stopping, and I think a flicker is possibly more useful than a dim light (which would seem to be the choices, unless you get some extra energy from somewhere) for seeing where you're going.
....  (My approach for when slow climbing is likely to be an issue is to have an independent battery light as a backup.)

I was thinking the annoyance of me /botheration of epileptics is more when a front LED lit hub dynamo-d cyclist is coming towards me/other at walking speed resulting in annoying bright fast flashes which would be illegal in a battery LED lamp cause out of 1 ..4Hz range.

From a bottle dynamo, I have never found LED dyno lamps a problem when slowly going up hill, in fact it seemed to be the biggest advantage of even ancient LED front lamps such as the B&M DTopal (17lux) over filament halogens (slow and road disappears making unlit roads cycling like the film Speed ). I just don't want the thing to flash at walking speed. Plus I want switchover to be automatic .

I had been thinking of a relay based 4AA  backup along your mentioned 4D light(albeit I don't mind reduced brightness):
at low speed a Greinacher rectifier circuit is switched in supplemented by the battery till switched out and hub drives lamps directly. The hub is always driving the relay coil.
Or more simply just a plain relay backup but that switches hub into a resistor with a DPDT relay to reduce relay chatter.

But it would be nice if there was a pre-existing roadtested  battery-free solution

I suppose another approach would be the 1994/5 Wireless World circuit of Greinacher rectifier circuit followed by step down regulator (LM2575)  updated by driving DC Ebike lamps with a (rechargable)  backup battery...but that involves many more new lamps. Of course if manufacturers had to make the same lamp cope with either E-bike DC or  a hub dynamo (even if different inputs).....

[Jakob W]

Is this really a widespread problem? Surely 99.9% of cyclists using hub dynamos (in the UK at least) are going to be doing more than the crawling pace at which flashing might be a problem. Flashing 800 lumen DEATH BEAMS are far more prevalent.

[Kim]

Is this really a widespread problem? Surely 99.9% of cyclists using hub dynamos (in the UK at least) are going to be doing more than the crawling pace at which flashing might be a problem. Flashing 800 lumen DEATH BEAMS are far more prevalent.

Quite.  Other than long slow climbs in the dark, it's mostly an issue when walking the bike, and you can always turn the lights off for that.  (Indeed, I sometimes turn the lights *on* for that, specifically because a light means people are more likely to notice and not walk into the front of my bike when I push it through a crowd).

One of the things I like about dynamo lighting on other people's bikes is that it mostly comes from countries where front lights aren't allowed to flash.  A bit of flicker on start-up is an acceptable trade-off.  I don't see many dynamo users riding at walking pace, unless there are cars in the way.

[SA_SA_SA_SA]

With the introduction of hire bikes with hub dynamos, I have noticed fast flashing front LED dyno  lamps a fair bit,  plus on the less common non-hire  bikes(perhaps wheeled, cycled slowly less often?). At night I would rather leave my lamps on when wheeling in pedestrian areas/pavements (in case forget to turn back on). I would prefer my lamps to either be steady of remain in the legal 1..4Hz (otherwise why bother having such laws?). In would of course be best if manufacturers were required to do the work.

Another approach would be the geared Renak hub dynamos but these are rather expensive if you want two wheels, one spiked tyres , one not.

[Brucey]

....I would prefer my lamps to either be steady of remain in the legal 1..4Hz (otherwise why bother having such laws?). ....

the law is there to prevent folks from using lights that flash like that VERY BRIGHTLY ALL THE TIME (which NB is liable to be bad for riders of the bike in question!) rather than fairly dimly, a tiny fraction of the time when it doesn't really matter.

As others have pointed out there are far bigger problems to solve in the world of bicycle lighting, with riders habitually using dangerous or antisocial lighting seemingly without giving a thought to how it affects others.   

cheers

[Jakob W]

TfL bikes have a be-seen flashing front light by design (and those silly green cycle symbol projectors), but I don't recall ever seeing a dyno flicker from them. TBH unless the Stzvo regs change I can't see this ever being fixed.

[SA_SA_SA_SA]

The LED Dyno front lamp ones I see flashing/flickering are a lot brighter than the little AAA style flashers that begot the  1..4Hz legislation  was introduced (NB at the time they considered an upper brightness limit and rejected it: there's some farsighted government for you ).  So a horribly bright 1..4Hz flasher is only possibly illegal by the vague causing unnecessary dazzle rule: but is probably unpleasant even if below that vague level.


I still think its worth fixing: and that thats a better use of lamp design electronic engineers' time/electronics cost than adding pointless and  feeble(compared to 1200cd car ones) DRL modes to cycle lamps.  But, yes, I suppose it requires the Stzvo to flex its muscles on behalf of epileptics/ consistency.


(I suspect some B&M lamps already have microcontrollers in them, so perhaps most of the effort would actually be in writing code, with few/no  extra components cost....?)

[grams]

The Luxos U has a microcontroller and doesn't flicker at low speed unless (a) the bike hasn't been ridden for weeks and the internal lithium battery is dead or (b) it's succumbed to water ingress/age and isn't working properly.

The Toplight Brake Light rear light has a microcontroller for the brake light function, which probably also controls the off button and the standlight timeout. These also switch back on smoothly if the standlight capacitor isn't discharged (and hasn't, as happens frequently, snapped off the circuit board).

The others I've taken apart (various brands) have simple analog circuits to regulate the LED voltage/current.

[McWheels]

Fortunately for everyone, a clever chap over at CandlePowerForums has worked it all out:

https://www.flickr.com/photos/kurtsj00/5420960597/

Now you just have to build it, test it on the bench, and install to your bike. Good luck, let me know how you get on, MoSFET H-bridges and all.

I actually run a much simpler circuit, https://www.flickr.com/photos/jdp298/11212392525/ which does flicker at (very) low speed for the rear lights. The front is now a B&M Cyo Senso summat thingymabob https://www.rosebikes.com/article/b--m-lumotec-iq-cyo-premium-t--senso-plus-front-headlamp/aid:709237, but it doesn't really flicker except at walking, and it's only the DRLs then anyway.

[Diver300]

I have a some comments.

Firstly, it's hard to see how the rectifiers in an AC lamp could be so close to their limit that you couldn't run on DC. The peak current would be more with AC, and rectifiers are cheap and would usually be generously rated. I've run a few mains powered devices on DC of a suitable voltage without problem.

Secondly the supercapacitors may have quite a high ESR (Equivalent Series Resistance) so putting largeish electrolytic or tantalum capacitors in parallel may help.

Lastly, the current through an LED, and therefore it's brightness, varies a lot with a small change in voltage. McWheels's circuit gets over that somewhat by regulating the supercapacitor voltage, so the current doesn't vary as the supercapacitor voltage varies. A suitable resistor in series with the LEDs could also reduce the flicker.

(McWheels's circuit has two potential weaknesses. The output voltages of the regulators might be different, meaning the highest one takes the most current.  Also, the LED voltage might be too high or too low for the regulators. Slightly higher voltage regulators and a current limiting resistor for each regulator would get over both those issues. It could also be that the regulators are already running as constant current devices, in which case the exact type is very important.)

[Brucey]

Firstly, it's hard to see how the rectifiers in an AC lamp could be so close to their limit that you couldn't run on DC. The peak current would be more with AC, and rectifiers are cheap and would usually be generously rated. I've run a few mains powered devices on DC of a suitable voltage without problem.

I think you perhaps underestimate the penny-pinching that goes into designing circuits that are mass-produced.  The problem is not the peak current (after all by their very nature rectifier diodes are intended to accept AC...)  it is more that the thermal loads  become excessive.

Basically instead of (say) having ~0.25W in each of four diodes, if you run the same power using DC through a rectifier you will have ~0.5W in each of two diodes. The other two diodes are doing no work at all and don't see any thermal load. Needless to say the two working diodes may fail immediately (few components are rated for twice the usual service load, especially if they are to be connected to a source with an intrinsically limited current) and even if they don't do that they may have a greatly reduced life. 

BTW there is no need to guess; decent quality lamps will say whether they can be run on DC or not, and with most B&M lamps for use with hub dynamos,  the answer is 'not'.

cheers
 

[grams]

I think you perhaps underestimate the penny-pinching that goes into designing circuits that are mass-produced.

That penny-pinching will also push them to off-the-shelf components, and rectifiers in particular skew towards much higher rated applications (i.e. mains).

I have in front of me a B&M Lyt BN Plus circuit board (a very cheap light) which has a BT05S rectifier which is rated for 0.8 amps DC, which is well above anything it would need to do in service.

(also, for a packaged bridge rectifier, if the limiting factor for a continuous rating is the heat extraction ability of the package then it'll be the same on AC or DC)

BTW there is no need to guess; decent quality lamps will say whether they can be run on DC or not, and with most B&M lamps for use with hub dynamos,  the answer is 'not'.

Manufacturers advise against a *lot* of things that are perfectly ok. Life would be very boring if we stuck to them.

[Brucey]

fair points.  However when connected to a DC source  the current drawn by a lamp is not limited in the same way as when connected to a typical hub generator. Most hub generators won't give more than ~700mA even into a dead short.

If the DC source is not well regulated and the lamp uses certain means of controlling the current in the LED (eg a 'crowbar' circuit for example) then the rectifier  or other parts might bite the dust.

cheers

[aidan.f]

Plenty of dynamo lamps use a crowbar. it shorts the generator for part of the a.c. cycle at high speed to prevent excessive voltage damaging the electronics. With a constant current source this is an efficent method.
I did, in the early days of high power LEDs design and build a few regulators. (with very low flicker) It's just not worth messing about these days given what is available commercially.

[SA_SA_SA_SA]

Plenty of dynamo lamps use a crowbar. it shorts the generator for part of the a.c. cycle at high speed to prevent excessive voltage damaging the electronics....

(or in olden days stopping bulbs blowing downhill .... by crowbar I presume you mean something like two 7V5 zeners in inverse series(ie back-to-back) (which is what protected my filament bulbs in the past) although  I would describe that as a voltage 'clamp' (it chops the top of the sine wave off above the clamp voltage) and is presumably the reason for B&M stating a 7V5 DC maximum for a Cyo.
Commercial Halogen 2.4W lamps would use something like a P6KE7 SCA transient voltage suppressor which is a more convenient single component (rather than 2 zeners)....

A crowbar circuit as I understand it would not work as a voltage regulator in that way
 because it would lower the voltage to the bulbs too much. https://en.wikipedia.org/wiki/Crowbar_(circuit)

[Kim]

No, a proper thyristor crowbar, upstream of the rectifier.  Consider the operation of a trailing-edge^[1] incandescent dimmer and you'll see how it can be used to efficiently limit voltage from an AC current source: Voltage gets too high, crowbar triggers - effectively shorting the dynamo, the regulator (if present) operates on stored energy in the input capacitor (which at this point is fully charged) until the next zero-crossing when the crowbar resets.  If the load is sufficient to keep the voltage down, the crowbar never triggers.  Unlike a zener clamp, the power dissipation is negligible, due to the low forward voltage of the thyristor/TRIAC, and some hubs may exhibit less drag in this condition than when open circuit, due to reduced eddy currents in the windings.

Such a circuit would readily release its magic smoke if connected to a DC voltage source, unless the voltage was kept below the point where the crowbar operates.


_{[1] Although the crowbar circuit has more in common with a leading-edge dimmer, by shunting the supply we're effectively chopping the trailing edge of the waveform.}

[SA_SA_SA_SA]

No, a proper thyristor crowbar, upstream of the rectifier.  ....you'll see how it can be used to efficiently limit voltage from an AC current source: Voltage gets too high, crowbar triggers - effectively shorting the dynamo, the regulator (if present) operates on stored energy in the input capacitor (which at this point is fully charged) until the next zero-crossing when the crowbar resets.  If the load is sufficient to keep the voltage down, the crowbar never triggers.  Unlike a zener clamp, the power dissipation is negligible, due to the low forward voltage of the thyristor/TRIAC, and some hubs may exhibit less drag in this condition than when open circuit, due to reduced eddy currents in the windings.

Such a circuit would readily release its magic smoke if connected to a DC voltage source, unless the voltage was kept below the point where the crowbar operates....[/sub]

Ah, very interesting, I have been stuck in the past thinking the same zener-style clamps circuit as was used with old fashioned halogen bulb lamps would be used for lamps with smoothing capacitors and electronics . 
The reduced power dissipation when triggered is  an advantage  for component cost/thermal design I presume.

It never occurred to me that it might reduce drag at the hub, as whether zener clamp or shorted triac the hub is still supplying  current compared to being open circuit.

Given that the Stvzo accepted the DC lightspin dynamo I wonder why they don't require lamps to cope with either DC or AC at least up to a certain value of DC (eg 7V)) at least for the more expensive high  lux lamps (marked according).

[SA_SA_SA_SA]

I have corrected the title and OP in this thread for clarity:
It is really the fast on/off flashing at low speed I complaining about, I can live with warbling.

(I sometimes get the feeling YACF 'corrects' what I type to the wrong meaning, although not in this case: a straightforward mistake here ).