DIN 1.5W gen std: seems to absolve gen of regulation

[SA_SA_SA]

From the website of the velogical rim dynamo** which has gained German Stvzo via DIN approval as a 1.5W generator for LED only, it would seem the generator is allowed to assume voltage regulation will be performed externally by lamps, it just has to produce enough power at certain speeds.

** http://www.velogical-engineering.com/rim-dynamo-en-1

I wonder how much meeting the standard Stzvo voltage/current knee curve costs in efficiency: could all hub dynamos gain easy gains in efficiency from this?
Would it allow the generator coil/magnets to be shrunk enough that more poles could be added easily to reduce flicker*?
Filament bulbs are on the way out even for cheap lamps....

* I don't see the point of forbidding fast flashing battery lamps due to epilepsy concerns if you allow front hub dynamo lamps to do it when cycling slowly

[bobc]

Just my 1/2p worth - any half decent LED light will have some sort of switching converter in it, so seems daft to put another one in the genny....
i.e. it sounds sensible to me....

[axel_knutt]

any half decent LED light will have some sort of switching converter in it

The lamps they’re using have no switching regulator, so I assume they haven’t found any that do. They’re all shunt regulation, so they’ve had to put the series impedance back in the circuit in the form of an external PTC thermistor, which largely defeats the point of improving the dynamo. (Their “thermistor” is actually a 12V 20W halogen bulb.)

So yes, the dynamo is more efficient, but the efficiency of the system as a whole is still poor, because of the shunt regulation. The publicity claims figures as high as 60-75%, but their test results are showing an average of about 54% at the dynamo output, and 41% at the lamp input. Taking into account the shunt regulation in the lamps though, the efficiency at the LED varies from about 35% at 5mph to less than 5% above 25mph.

It’s a good starting point for someone who’s inclined to replace the shunt regulation with a switcher.

[SA_SA_SA]

The lamps they’re (velogical, I presume) using have no switching regulator, so I assume they haven’t found any that do. They’re all shunt regulation, so they’ve had to put the series impedance back in the circuit in the form of an external PTC thermistor, which largely defeats the point of improving the dynamo. (Their “thermistor” is actually a 12V 20W halogen bulb.)...

How do you know that? A switching regulator seems the best way to get power into an led from the varying output of cycle generator while avoiding low speed flicker etc. I bet the more expensive/more powerful front led lamps will all have some sort of switching regulator in them.

But the lamp circuitry (switching regulator) will probably have some sort of shunt regulator/protection to ensure the components voltage ratings are not exceeded, eg lights off with a hub dynamo: some ICs wont stand more than 40Volts hence would need a clamp to ensure that is not exceeded.

Also, I think the PTC is optional, needed for protection when going fast so at lower speeds is presumably like a piece of wire....

As thermistor we use a 12 volt 20 watt halogen bulb with
soldered cables. The thermistor takes care of the optimized
cooperation between dynamo and LED frontlight at higher
riding speeds, especially with LED daylight running lights. The
thermistor balances the overvoltage protection of the LED
frontlight and optimizes the extremely low drag of the dynamo over wide ranges of riding speeds.

Their efficiency figures seem OK....

But anyway, other similar generators may not need the PTC thermistor, I was just making the point that abandoning the very restrictive Stzvo 6V knee/curve might make it easier to make more efficient/lighter/simpler generators for LED lamps only....

[axel_knutt]

How do you know that?

By looking at the Velogical test data in your own link:

All six lamps they’ve published data for are consuming more current as the voltage rises, not less. A switched mode regulator consumes a constant power as the supply changes, so the input characteristic should be a hyperbola and have a negative slope resistance, but the lamps are nothing like that. You can also see the knee in their characteristics where the zeners start to clamp. The dotted black line shows a constant power locus, the sort of curve you should expect to see at the input of an SMPS.

If you’re happy with the efficiency then you’re happy with it, but it’s not really any better than any other shunt regulated system that you can buy cheaper:

The dotted line shows the overall efficiency you might expect from a 75% efficient SMPS. A plot of the power makes it even more obvious, a switcher doesn’t need 15W to drive an LED:

I think manufacturers might be wary of selling low impedance dynamos because of the risk of getting a bad reputation if people connect them to unregulated lamps.

[SA_SA_SA]

Yes but how could they have the usual switching regulator ICs if they was no shunt regulator to protect them from higher voltages above eg 40V (as opposed to the 9Volt-ish zener clamp in a plain halogen lamp)?

Perhaps derivatives of mains rated switcher ICs could be used if such generators became more available, these could then need less shunt protection.

It doesn't prove the absence of a switching regulator, just that any they have fitted needs some high voltage clamp protection. So you lose some efficiency at high speed but the efficiency is still good, and the generator lightweight.

But my point is that the Stzvo curve comes from a claw pole generator, not requiring it allows different solutions that could be cheaper/lighter/more efficient...

[axel_knutt]

Firstly, if you want to protect a switching regulator against voltages above 40V you don’t clamp it at 6V and defeat the whole point of the switcher. Secondly, if you clamp the input of a switcher, the negative input resistance will still be evident below the clamp threshold. Thirdly a switched mode regulator is liable to go unstable if run from a high impedance source. And finally, their test data doesn’t just show you what isn’t in the lamp, but also what is in the lamp:

Led Cct.JPG (4.68 KiB) Viewed 551 times

So, yes, it does prove that there’s no switching regulator in any of the lamps Velogical tested. I don’t know quite what your definition of good efficiency is, when the overall system is no better than an ordinary dynamo with a high leakage inductance.

[SA_SA_SA]

Firstly, if you want to protect a switching regulator against voltages above 40V you don’t clamp it at 6V and defeat the whole point of the switcher.

I never suggested that! I meant use a zener rating of just under 40V (by a safe margin) to protect a 40V IC.

However, I suspect rear lamps do have a 6volt-specific clamp, so I suppose it depends whether the front lamp simply passes on the raw dynamo feed or converts/processes the rear lamp voltage itself. I notice in your? 'Lamp VI characteristics by brand' graph, the supernova E3 and RL2 is clamped higher than the others: this has a dumb rear LED lamp controlled by the front lamp circuitry, which could support the idea that the rear lamp is doing the aggressive clamping in other systems. Oh and I misread the scale; before I zoomed it I read 10V as 30V . So I agree using a 6Vstyle clamp would seem to cause unnecessary inefficiency.

... Thirdly a switched mode regulator is liable to go unstable if run from a high impedance source.

I have seen a switched mode based battery backup by PW Fry in ELectronics world Dec 1994 / Oct 1995: an LM 2575 from voltage doubling rectifier from high impedance bottle (union 6701: a very draggy dynamo; internal resistance 6ohms I think!) and protected by 27Volts of zener diode clamp...
Seeing actual lamp circuits is the only definitive answer for me

And finally, their test data doesn’t just show you what isn’t in the lamp, but also what is in the lamp:
Led Cct.JPG

I can't find this diagram in the velogical data pdf but it looks like a simple test circuit rather than one I would expect in expensive high power dynamo lamps: it won't be able to extract as much power as theoretically possible. I'd be disappointed if an expensive lamp contained that! Also I remember B&M commenting on the difficulty of getting the circuit for their first front LED, the DTopal right which would suggest a circuit more complicated than the above.


But anyway, leaving arguing about velogical's design aside, other similar generators may not need the PTC thermistor, I was just making the point that abandoning the very restrictive Stzvo 6V knee/curve might make it easier to make more efficient/lighter/simpler generators for LED lamps only....

[axel_knutt]

Firstly, if you want to protect a switching regulator against voltages above 40V you don’t clamp it at 6V and defeat the whole point of the switcher.

I never suggested that! I meant use a zener rating of just under 40V (by a safe margin) to protect a 40V IC.

You said:

how could they have the usual switching regulator ICs if they was no shunt regulator to protect them

Which implies that you were arguing that the 6V clamping in the lamps was to protect a switching regulator.

However, I suspect rear lamps do have a 6volt-specific clamp, so I suppose it depends whether the front lamp simply passes on the raw dynamo feed or converts/processes the rear lamp voltage itself.

It’s irrelevant how it’s done, the system is grossly inefficient. If there's a switcher in there it's doing absolutely nothing at all.

... Thirdly a switched mode regulator is liable to go unstable if run from a high impedance source.

I have seen a switched mode based battery backup by PW Fry in ELectronics world Dec 1994 / Oct 1995: an LM 2575 from voltage doubling rectifier from high impedance bottle (union 6701: a very draggy dynamo; internal resistance 6ohms I think!) and protected by 27Volts of zener diode clamp...
Seeing actual lamp circuits is the only definitive answer for me

I haven’t said that it will go unstable, I said it’s liable to because any circuit with negative resistance is potentially unstable. If you insist there’s a SMPS in the lamps, it leaves you with the job of finding a sensible explanation as to why it’s no more efficient than a handful of resistors can achieve for a few pence, why they’ve clamped it at about 6V, and why the characteristic has no negative slope.

I can't find this diagram in the velogical data

The information is in the test results:

The slope of a VI characteristic is the resistance of the circuit. Between the conduction threshold of the LED at 2 or 3V and the onset of clamping the resistance is about 10.7ohm, and above the clamp threshold the resistance is 3ohm.

But anyway, leaving arguing about velogical's design aside, other similar generators may not need the PTC thermistor, I was just making the point that abandoning the very restrictive Stzvo 6V knee/curve might make it easier to make more efficient/lighter/simpler generators for LED lamps only....

Any dynamo will need some method of regulating the output in order to prevent damage to the lamps, and if it's not switched mode it will be inefficient. As I said at the top, I agree that the dynamo in isolation is an improvement, but you can't use it in isolation so you have to look at the efficiency of the system as a whole.

[Brucey]

call me simple minded if you like but I find it easiest to specify an LED in the front lamp that will happily cope with all the current that the generator can possibly dish out, i.e. I'm relying on the internal resistance of the generator (and its current output being limited for other reasons) to limit the total output to some extent. This has the advantage that this also limits the voltage across both lamps. The system is isolated using a mechanical switch so the only parts that need a high voltage rating are those in the generator and wiring before the switch.

With nearly all the hub generators I have seen the open circuit voltage can rise almost linearly with speed (as you might expect) but the current generating capacity does not increase pro-rata in the same way. The reasons for this are severalfold but one important one is simply the AC frequency at which the generator runs; this becomes such that the soft iron laminations no longer saturate fully, and this effectively limits the total value of flux linkage change per unit time and therefore the generators power output.

Since the eddy currents (which limit the flux penetration into the soft iron) are generated in the soft iron similarly regardless of whether the lights are on or off, there are always going to be losses in the generator anyway. This leaves somewhat less scope for altering the efficiency of the generator by altering the characteristics of the load/regulation. Only once the speed is high enough that switching regulation can supply enough current is there any real gain, and at that speed the odd 1-2W saved will probably be small beer anyway.

cheers

[SA_SA_SA]

Firstly, if you want to protect a switching regulator against voltages above 40V you don’t clamp it at 6V and defeat the whole point of the switcher.

I never suggested that! I meant use a zener rating of just under 40V (by a safe margin) to protect a 40V IC.

You(SA) said: 'how could they have the usual switching regulator ICs if they was no shunt regulator to protect them'
Which implies that you were arguing that the 6V clamping in the lamps was to protect a switching regulator.

Why would I use a 6V-suitable shunt to protect something that can take 40V when that would cost efficiency? So I don't see how that is implied, I (mistakenly ) thought a just under 40V shunt was being implied .

However, I suspect ((most)) rear lamps do have a 6volt-specific clamp, so I suppose it depends whether the front lamp simply passes on the raw dynamo feed or converts/processes the rear lamp voltage itself.

It’s irrelevant how it’s done, the system is grossly inefficient. If there's a switcher in there it's doing absolutely nothing at all.

I meant, suppose the front lamp has a clamp that is just under the voltage needed to protect its hypothetical switcher, from lets say 40V, so lets say a 35V clamp, and it uses some circuitry to make regulated 6V for a tail lamp terminal (ie the tail lamp is NOT wired directly across the generator, and thus the tail lamp clamp should not be activated be cause its getting regulated voltage from front lamp ) then the only clamp seen by the generator will be the higher Voltage one (my hypothetical 35V one) and efficiency would be gained over wiring both front and rear lamps across generator in parallel (which would cause the usual 6V-suitable taillamp clamp to activate as in the test data). The test data doesn't seem to state whether the tail lamps were wired direct to generator or via front lamp terminals (eg switched on senso/hub lamps).
NB The Supernova pair has a higher clamped voltage and its special tail lamp is powered by the front lamp circuitry.

Also, as for 'doing nothing at all' an LED with a series current limiting resistor is wasting power in the resistor: surely a switching based 'current driver' would be able to use some of that wasted power to light the LED more brightly....

said 'I can't find this diagram in the velogical data' .

The information is in the test results: ...
The slope of a VI characteristic is the resistance of the circuit. Between the conduction threshold of the LED at 2 or 3V and the onset of clamping the resistance is about 10.7ohm, and above the clamp threshold the resistance is 3ohm.

Ahh, it is what you deduce to be the 'equivalent circuit'; where is the rectification part of the circuit? So I would still be curious about the actual circuits. Perhaps, someone has taken one of the mentioned lamps apart?
Anyway I know there was a (DT/Bisy I think) dynamo rear lamp that had a LT1073 switcher IC in it (I dismantled it), driving a chain of 6 LEDs. So expecting a switcher in an expensive front dynamo lamp wishing to be as bright as possible over varying speeds seems reasonable to me.

Any dynamo will need some method of regulating the output in order to prevent damage to the lamps, and if it's not switched mode it will be inefficient. As I said at the top, I agree that the dynamo in isolation is an improvement, but you can't use it in isolation so you have to look at the efficiency of the system as a whole.

OK. But I still think there is a possibility that some expensive front lamps have a higher voltage clamp and they supply the tail lamp voltage from their own circuitry, the generator thus will not 'see' the low voltage tail lamp clamp and it should not be activated by the regulated output if the front lamp tail terminals.

A pity they did no tests with front lamps only. Or mentioned how the tail lamp is wired. I suppose for a bottle, there is no need for a switched/auto front lamp so they may well have simply wired front and tail lamps in parallel (apart from the Supernova S3 with its higher clamped voltage ). I suppose I could ask velogical... .

Perhaps, other than the Supernova, no such lamps currently exist ....

[edocaster]

call me simple minded if you like but I find it easiest to specify an LED in the front lamp that will happily cope with all the current that the generator can possibly dish out,

Hear, hear. I can't say I understood all the discussion, but I don't know of any dynamo which outputs over an amp, and there are many LEDs which can withstand 3 times that.

The only reason they burn up in battery lights is if the system allows them to draw more amps. That's not going to happen with a bike dynamo.

The 'Lamp VI Characteristics by Brand' chart looks a bit suspicious to me, and looks like the results of lights being driven on a bench power supply rather than a real bike dynamo, assuming the figures along the X-axis are current in mA.

[SA_SA_SA]

....I find it easiest to specify an LED in the front lamp that will happily cope with all the current that the generator can possibly dish out,

But then you are either wasting any voltage over and above the Vf of your single white LED (eg 3.6-4.0 volt-ish) across a series resistor/linear current source or if none fitted then lowering the voltage seen by the rear lamp to Vf + any rectifier drop: a switching current driver could avoid that inefficiency = brighter front lamp. Eg Vf of 4Volts multtipled by 0.4Amp is only 1.6W out of an 'official' available 2.4W and fancy-ish electronics should be able to extract more at higher speeds.

Yes, nice and simple for home brews, if bright enough for purpose, but I would expect more efficient circuitry in an expensive bought front lamp than from home brews....

[Brucey]

yes, I daresay I could have a more efficient arrangement, but when I'm rattling along I'm getting more like 0.6A which seems to be enough. With most generators/lamps complex regulation only gives you any real benefit between about 10mph and about 15mph in terms of light output. Above that the LED is probably going flat out and below that there is no 'voltage overhead' to waste anyway.

Arguably complex regulation could give you a real benefit in light output at higher speeds too, but only if the LEDs can make more light and you need it. If you use more power (in more light or inefficient regulation) at speed you will pay for it in terms of extra drag but

a) this becomes a smaller fraction of the whole (you must be a more powerful rider to start with) and
b) the effect of a small change in efficiency becomes smaller and smaller in terms of speed, because of the effect of aerodynamics.

Efficient regulation might net you an efficiency gain worth 1-2W at normal speeds; worth having, but not if it comes with a penalty of some kind.

As a general rule I don't think having complex regulation (in return for a small benefit of some kind) is a very good idea; it tends to just make the whole thing less reliable, and if the water gets in, it is far more likely to cause problems in the lamp.

'Simple' is often a better idea, if the outcome is 'good enough'.

cheers

[edocaster]

But then you are either wasting any voltage over and above the Vf of your single white LED (eg 3.6-4.0 volt-ish)

Then add a second LED in series, or use an LED with a double die. Fortunately, you can fit two LEDs under 6V - the only reason that's an issue anyway is playing along with standard rear lights.