I haven't seen any test data on how losses change when extracting extra power, but...

A handful of years ago, I rigged up a switchable resistance to load my 1st gen SON while riding. I attached a multi-meter to the handlebars, and went out to take data at different speeds and different loads.
This was mostly to get accurate data, since I'd seen data from someone else and pretty much knew what to expect. My results were largely what I expected, and allowed me to put together a suitable electrical model for my SON.

One of the lessons from this is that you can only get more power from a SON by adjusting the load.
Well, you can also add some capacitance in series with the load, and that will help draw more current from the SON, but it's much easier to adjust the load resistance by means of a small switching regulator that you need to have anyway. Up to a point, increasing the load resistance will let you get more voltage and power from the SON while slightly reducing the current.

In regards to drag, some is due to the resistive losses in the SON's coil and the wires between the dynamo and light. If you draw more power and less current, then these resistive losses will go down.
There is also energy loss due to the changing direction of the magnetic field in the steel core of the dynamo. I don't think this changes much due to changes in the dynamo current, but would be happy to learn about data that sheds some light on this.
Based on this, I'm not sure that drawing more power from the dynamo will increase losses in the dynamo. It will certainly increase the power required to rotate the wheel, but only by the amount equal to the extra electrical power extracted (... I think...).

The one thing that confuses me is the mention of 12 watts. I'm not sure that Schmidt is saying that it can get 12 watts out of the dynamo. Based on my data with my SON, at 20 mph, the most I can get out of it (with a resistive load) is 7.7 watts. Maybe Schmidt is just saying that 12 watts will occur at some higher speed, such as 30mph?
Beats me, but I'm curious to see what develops.

Steve in Peoria

I don't keep up on the different technologies or chemistries, but I keep hoping that a good low temperature rechargeable battery shows up.
I've made some standlight circuits for my bikes, and they have used AA nicads. These don't mind (much) being recharged in very cold temperatures, and the 1.3V works well with my circuits.
Last year(?), I had one that died and I was looking for a good replacement. It turned out that I did have some untabbed AA nicads stashed, but I would like to find a good replacement.

I did check some of the lithium chemistries, and didn't find anything that I thought performed as good well as a nicad. .. oh... it seems that nicads are gone from the marketplace, so the next best option for that form factor is probably NiMH.

The best cold temp performance for a lithium was the fairly new lithium titanate. Nichicon was marketing something called the SLB12400L151. It could be charged down to -30C, according to the preliminary datasheet that I've got. Or maybe that's just the discharge temperature? It's a very brief datasheet. Unfortunately, the capacity is very poor. 150mAhr for a cylindrical cell that is 40mm x 12.5mm.

The lithium iron phosphate batteries are rated for discharge down to -20C, but only 0C for charging, which isn't sufficient for my standlight use.
Is there something suited for charging at -15C (around 5F)?
Just wondering if I need to convert my standlights to supercaps. I've got one headlight using a 100F supercap, and it's not bad. It's a rather bulky item to retrofit where a AA used to be, though!

Steve in Peoria

2 i: I would like to know more about how to set up the Bike Energy Harvester. Please tell me more about your setup and some pictures would be nice. I do some winter camping and like the cold weather batteries. I am also looking at a cell rugged phones for navigation that can be charged at some lower temps such as - 20c or -30c

Rick, do you have any data on how many watts your phone will draw while navigating? I am sure that backlight (on or off, brightness setting) plays a big part of that. I am not asking about instantaneous data, more a question of how much power you would need in watt hours for a typical day of touring. And I am sure that the way you do that (bluetooth or other communication in use or not, on-line data or not, etc.) would also play a role. And yes, what you are doing that day would play a role too, are you keeping the screen visible for constant navigation or following a rail trail for hours that requires almost no screen time. I think an estimate of watt hours you would need each day on average would be an important part of figuring out what you need for charging capability.

And, if you are using a high power taillight like some people do, that could be another consideration. If you are charging up taillight(s) every day too, ... you get the idea.

Yesterday I did a 200k randonneuring event, one person used a phone for navigation, within a few hours after the start she had to plug into a powerbank to keep her phone running. I do not know if she started the day with a full charge or not, but most people that use their phone daily all day long try to charge up their phones overnight.

I suspect that phone navigation consumes a lot more power than dedicated GPS units. If so, that could dramatically increase your power needs for self sufficient touring.

My general recreation (non-cycling) GPS consumes about 7 to 8 watt hours in about 12 to 14 hours of use with backlight off. Full backlight, cut that time in half. It does not have any wireless capability, so that is an electric load I do not have to consider.

Tourist in MSN: I am looking at a rugged phone with a large battery that can still be charged at temps near freezing. I intend to use the Bike Energy Harvester or the ladelux headlight from Schmidt to keep the phone that I use for navigation charged. I believe having Bluetooth and Wi-Fi off and just having the phone off when not needing the navigation on will be manageable. The other large user of power will be my tail flasher. In addition to the rugged phone I am going to purchase the Dinotte D6 The slow pulse setting lists operating time at 40 hours. The D6 has a USB C charging port on it. I will also carry a fast charge 18 watt cell phone charger. I was looking at the Voltaic solar panel setups. I am thinking a 10 watt Panel would be enough.

One other big power savings with a phone is downloading the maps. Google lets you do that and so does RWGPS. It takes a few steps with either app. RWGPS suggests turning the phone to airplane or offline mode.

I have ridden with someone that uses rwgps for navigation, even on long rides. He has a power bank that's bigger than his phone. Not sure that's fully necessary, but the longest I have used rwgps for navigation was about 60 miles.

unterhausen: I intend to use some downloaded maps but didn't know that the GPS would work in airplane mode.

GPS only receives data from satellites. No need to transmit. So, with GPS on, data, wifi, cell and bluetooth can all be off.

My free standing GPS unit (not a cycling version) receives only, does not have any capability to transmit, but that means I have to load stuff into it using a USB cable.

A few months ago, I got a reconditioned iphone 12. It's in such good condition that I can't tell it's been used, and even the battery is strong. Anyway, it gives me the impression that the GPS receiver is not a power hog. I haven't done any measurements, though.

I am familiar with Android phones, not Apple. I have an app on my Android phone called AccuBattery Pro. It calculates all kinds of statistics on my battery usage. I have not tried to figure out what would happen if I left the phone on but no apps running and airplane mode for a day, and then repeat with GPS on and a GPS navigation app on. But if I was interested in trying to figure out my power needs, I would probably do something like that.

Or skip the power usage app, just let phone sit turned on, airplane mode and all apps off for a day and look at percent change in battery level. Repeat for a day with GPS and navigation app on for a day and then check percent battery usage change. You lose a bit of precision looking only at difference in percent battery usage, but that probably is more than enough for making estimates for bike touring needs. If you are relying on dynohub for your power needs, you get a lot of variation from day to day riding depending on how hilly the terrain is, you can't really say that you expect to get X number of watt hours of power per day of charging by dynohub. That is why I want a pass through cache battery that has a capacity of several days of usage.

My GPS (a Garmin 64 general recreation GPS, not a cycling specific one) uses roughly a half watt when on and backlight off. I figured this out from letting it sit turned on with a set of freshly charged NiMH AA rechargeable batteries until it shut off. The track it maintained as it sat there gave me the time of operation until it auto shut off. With backlight on full, the same kind of test told me that I can cut that run time in half. With screen off 99 percent of the time appeared to have almost no difference than screen on and backlight off, which surprised me. Apparently the screen on with backlight turned off uses almost no power difference than with screen off. Thus, I now just leave the screen on full time, but I am obsessive about keeping my backlight at only the minimum that I need which is usually none. This GPS has no connectivity for wheel sensors, HRM sensors, no wifi or bluetooth, simply just doing the GPS thing without devoting any power to other uses.

I routinely use my phone to find details of the terrain over which the plane is flying. I have maps for the whole world downloaded, and the GPS is likely to work when I sit by the window.

You can charge LiFePO4 below 0C, provided the charging current is at 0.05-0.1C. This is about what you get from a hub dynamo anyway w/o special provisions. I have used 26650.

Sorry for the slow response. I am traveling, fortunately, with a Harvester on my Brompton. I put it into a pocket of this bottle pouch, together with USB cables. The pouch is lousy, but good enough for the purpose. The cable from the dynohub comes in through a hole in the bottom of the pouch. I melted it out with a soldering iron, I think. I modified the zipper to open from two sides. Usually, the pouch is closed, and I open it slightly when charging the phone or something else. Through the opening, I can see the control LEDs.

Compare the charging output to the same drain rate on a e-bike battery; The torque/thrust from the e-bike equals the drag of the generator. If you have the ability to charge the same size battery as the e-bike, in the same time it takes to deplete the e-bike, you're probably talking a lot of drag.

Were this not the case, I would plan on charging my CPAP battery on the bike. But no.

just wondering... does the manufacturer say it can be charged below 0C, or are you saying that you are operating the battery outside of what the manufacturer specifies?

Steve in Peoria

Most manufacturers state that you should not charge below 0C, but most will assume that most consumers cannot handle the charging rate. RELION who is an important player in the low-temperature battery market, particularly in Canada, says about generic batteries: "Important tips to keep in mind: When charging lithium iron phosphate batteries below 0°C (32°F), the charge current must be reduced to 0.1C and below -10°C (14°F) it must be reduced to 0.05C. Failure to reduce the current below freezing temperatures can cause irreversible damage to your battery." There exist chargers in the market for charging lithium batteries at low temperatures, and I expect that they adapt the charging rate to the temperature. I put my LifePO4 batteries on my main bike in 2017 and have commuted mostly on that bike every workday, every winter. I usually return when it is dark and can tell that the battery capacity has not markedly dropped from operating a horn parallel to my lights: The horn takes up to 2A, and the lights do not dim. The conditions under which I activate the horn are usually when I am slow or stopped. My batteries are exclusively charged from the dynohub. The bike is parked outside when I am at work and it freezes in the garage at home, too.

As an EE I suspect the overall architecture is that of a rectifier, which inherently converts a high power siinewave into a series of high power half-sine waves. In instantaneous light sensor would register the light as a lit of flickering. It does not do that because the response of the eye is not instantaneous. The flickering frequency is related to the wheel speed so flicker becomes apparent at lower wheel speed, and smoothes out as the wheel speed increases above a threshold frequency.

By assisting the system filtering improving on the natural human eye filtering frequency, nearly all of the flickering can be reduced. There is always a price for more complsx, more effective circuit design to provide more system performance, resulting in better flicker performance, but probably costing more money..

The circuit designer has many techniques avaailabl to use in a design or design improvement, especially one with knowledge of advanced power electronics. There were notes online for the design of such things, but I haven’t reviewed them for quite a while. Plus, today there should be newer options for the design improvements.

I like putting the stand light on a small super cap, except for the voltage regulation. How much current does the small light draw, and how much runtime do you need? Is there a limit on recharges in the below zero temperatures?

A slightly silly solution: sealed lead-acid could be used below 0C.

The voltage regulation for a supercap isn't too bad. My design just uses a voltage reference followed by a transistor set up as an emitter follower.
The matter of design requirements is very much driven by personal requirements and how it will be used.
I like a standlight that drives a white LED with maybe 50mA? It's been quite a while since I've thought about it, honestly.
With the 100F supercap powering a small boost converter, it runs for 10 minutes or so.
The Cooper Bussman "PowerStor" HV series supercap is rated for operation down to -40C, so there's no issue with recharging (for me, at least).

Lead-acid might be solution, at least in theory. Is there a model that would be suitably small for this sort of application?
The nicad AA really did work well for me. The capacity worked well and it wasn't very large. The lifetime was probably fine for most commercially built lights, and I'd get 7 or 8 years of use in my standlight designs. The 100F supercap has less capacity, is considerably larger, and is quite a bit more expensive, but the longer lifetime is certainly appealing.
Retro-fitting a 100F cap (18mm diameter and 60mm long) into a light designed around a AA nicad (14mm dia x 50mm) is a heck of a challenge!

Steve in Peoria

So the Soma Fab shop seems to show the IQ-XL 300 lux headlight in stock for $329. I put it in my cart and it didn't complain. It's also on the Merry Sales website (same company). If they have it, I wonder why Peter White doesn't. Merry Sales says if it's not in stock, you can't add it to your cart. It let me add it to my cart.

Anyway, I liked this bit from Peter White:

. .
So, high beam has less light than low beam? Hmmmm, I wonder why nobody thought of that before?

It's aimed higher. I wonder if that means it's harder to see potholes.

unterhausen: I believe you are looking at the ebike version.The dyno and the ebike version are both called the IQ-XL. The ebike version has been available for quite some time. The dyno version isn't available for another month or two.Starbike will allow you to put it in the cart but there is no delivery date and there is a very high demand for it. Of the eventual three high beam dyno lights, B&M is probably going to win out for my purchase.

It has to do with the way LUX is measured. LUX is not about how many photons are emitted from the LED. That would be LUMENS. LUMENS is a rating of what the LED is emitting. LUX is a rating of how brightly illuminated is a surface on the ground a specific distance from the headlight. So, LUX takes into account the optics of the headlight, and the horizontal surface (what we usually call a road) - in front of the bicycle. So consider a Sinewave headlight. It produces a round beam if projected on a vertical surface, very different from any Busch & Müller or Schmidt beam. If a particular Busch & Müller headlight's LED emitted exactly the same number of photons as the Sinewave Beacon, at different locations in front of your bike, the two headlights would be illuminating the road surface at wildly different amounts. The Sinewave puts out an even amount of light within the circle of its beam, while the Busch & Müller headlight, when aimed correctly, puts a relatively even amount of light on the road, which of course is a horizontal surface.

Now, what about the high beam in the Ladelux and IQ-XL? Since the hub's output is limited, when the second LED is switched on for the high beam, there is less power available for the primary LED, so the LUX rating, which is a measurement of the light projected on the road surface, must be reduced.

The Ladelux is using the same reflector as the Edelux II, which uses the Cyo Premium reflector that they buy from Busch & Müller. So no, they are not waiting on Busch & Müller for the reflectors. The earlier Edelux used the reflector from the original CYO headlight. I have a prototype Ladelux in transit from Germany right now. I should have it on Thursday. I'll do a bit of testing here, and then I'm handing it off to an experienced brevet cyclist and she will give it a more thorough test.

The Busch & Müller IQ-XL is not shipping yet, so my friends at Merry Sales/Somafab don't have any. We are expecting to have them late this year, and they will have them about the same time I do.

We have had the Supernova M99 DY Pro for about a year and a half. I haven't ridden with it, but our customers have been very happy with it.

Thank you for a good explaination.

I put the prototype SON Ladelux on my bike yesterday and took it out for a spin last night. I only rode about 7 miles so it was not a test of the USB charging but just a test of the headlight itself. The low beam is the same as with the Edelux II I have been using for several years. There's a 30MPH speed limit sign about a ¼ mile from my house. With my Edelux II it's invisible at night, and it was also invisible with the Ladelux in low beam. But when I switched to high beam after a car passed me the speed limit sign was brightly illuminated, as you would expect from a high beam. Also, all of the signs for intersecting roads, which are normally invisible, were well illuminated by the high beam setting. I had no occasion to flash a high beam at an oncoming car as with the Ladelux in low beam people were dimming their high beams. People around here are very polite.

Switching between high and low beam is done by pressing briefly on the top of the high/low switch box which mounts to the handlebar. The switch doesn't actually move, at least not on the outside. It just senses the pressure of your finger to change modes. There is a small light on the switch that changes color to let you know the mode you're in. It's a sort of yellowy green in low beam, and blue in high beam. To turn the headlight on or off completely, you press the button for a bit over a second. Where I live there is no overhead street lighting, so for me it was obvious from the beam whether I was in high or low beam. But in a city with all of the ambient light you might not be able to tell just by the projected beam if you're in high or low beam, so the indicator LED would be quite useful. Not that I have any wish to move back to Boston!

My headlight beam page is long overdue for an update, so in the next week or so I'll round up all of the newer headlights and make more images. I'll post here once the updated page is ready.

I forgot to mention that I was wearing full fingered leather gloves. So the on/off switch does not rely on skin contact to be activated.