Marcturus is right, and what he says is really important, so please forgive my shouting: Lumens are the wrong thing to be talking about. What matters to the safety performance of a headlamp, tail light, brake light, marker light, etc is intensity, measured in candela at a specified location -- could be straight out from the axis of the light, could be 15 degrees left and 2 degrees up relative to the axis, could be anywhere within the range of about 60° left and right, 20° up and down. A lamp with a light source that produces more lumens is not necessarily better than a lamp with a light source that produces fewer lumens. It's all about how efficiently and effectively the light is focused and distributed.
Hello again to CPF, after a long absence... The proper way to look at any light that is directly to be seen, is using the light density in cd/m^2 but you also need a certain luminous flux (lumen, for the non-experts this is sort of the 'amount of light', though light is a flow so therefore amount is not really the right expression), or by specifying the emitting area, these are of course equivalent in that you need just 2 of these and the other follows.
As to why cd/m^2 is important, it is because light gets sent out by an emitter of a certain area and received by the eye which uses a lens, onto an area in the eye. I've explained this in my 2015 proposed standard, in which I didn't specify exact details of light intensity etc. but using the ones from StVZO is mostly ok so the essence of my proposal was to show what needs to be in a proper standard both from a technical perspective and a perspective of making it understandable to anyone (within reason), so reasons for all requirements are an integral part of my standard. See:
https://swhs.home.xs4all.nl/fiets/tests/verlichting_analyse/verkeersregels/whs-2015/index_en.html
and for the explanation of why size matters, see section 1.0 and 1.3.
Note that a lens is not a proper distributing optic! It changes nothing about the emitting area, though it reduces how much the eye will see of the total light sent out, if it's a wide angle lens. Also about the original question, 40 lumen is way more than any dynamo lamp and none of those are 'too dim'... As someone else said "If they don't see you with 40 lumen they won't see you with 160 lumen'. I can lower that: "If they don't see you with a few lumen as in say the Lumiring, they won't see you with a 160 lumen lamp".
I could discuss more details such as about DRL, which similarly needs to be treated as a to-be-directly-seen light, as this is similar. So I will tell you a bit about this topic and the discussions I had with a researcher working on the new TA that should be about ready, finally, a long time after the changes in StVZO that got into effect mid 2017 (end of 2016 they were already approved), and since which the KBA and labs doing approvals have been using a mish-mash of standards. The 30 cd signal function has not been in effect I think for a quite a while, there was the option it seems to use that at the same time as the new brake light requirements which are brighter, but the KBA makes final decisions and they choose which rules to follow and they can deviate where needed from the rules. E.g they do not use the 3W rules of DIN 33958, only those of TA, i.e. a company can not choose itself which rules are to be followed for an approval.
I tried to get the person working on the new TA to deal with point sources and other issues, but either he is not interested or constrained in what he can do, in any case he told me that there will likely not be a size requirement for bicycle DRL in the new TA (and in the intermediate rules being used since mid 2017 there isn't a requirement for emitting area either, i.e. they are using ECE's intensity but not the light emitting area, although the latter is also regulated in the ECE rules (for cars)!). I don't like DRL as explained in that standard and I think it was here on cpf that long ago I made the comment about visibility that you can question where it ends, i.e. there is a trade-off everywhere between safety and convenience. My comment was that you might as well install a water canon on all cars to spray cyclists and pedestrians, then they will certainly notice you! ;-)
I gave this comment too, to the researcher in question, after he told me that his predecessor said "It's better to be annoyed a bit, than to be dead". I find such non-arguments not funny, so in return I told him my water cannon suggestion and the suggestion to lower the speed limit for cars to 20 km/h, then safety is assured! This is sort of a reductio ad absurdum to show that in everything there is a trade-off, and I don't see any evidence that DRL is useful, I do see that it distracts. DRL or pseudo-DRL (not separately approved DRL) in bicycle lamps is even worse, I told him that e.g. the Luxos should never have been approved. In any case, it seems he wasn't amused by my response, and what he then wrote showed me that he is constrained in what he can do. I then made the following page about inconsistencies and problems in StVZO/TA:
https://swhs.home.xs4all.nl/fiets/t.../verkeersregels/de_stvzo/inconsistencies.html
To end: I've not done much testing last few years, caused by a few things. One of those is that I've been very disappointed by the fact that there has been just about no progress in dynamo lamps since the Saferide 60, in beam shape and light colour with the still predominant and poor choice of cool white light in most lamps. I've been meaning to do some more testing of pedelec lamps and perhaps I will test the front wheel Spinup F12W-pro if there is enough interest (eg. enough people donating to buy buy one, I could then draw lots for one of those people to get it after testing). In any case, if you (the readers of cpf) want to see more no-nonsense and rigorous real-world testing of bike lamps and dynamos such as the Spinup F12W (instead of e.g. the things that GCN does on youtube, one video that I saw recently was as usual more about entertainment than having any substance), and possibly of more visits to bike lighting makers, subscribe to my youtube channel (
https://www.youtube.com/channel/UC8glnbTXbm7YfVqdX3VjQsg ), comment there, and if you are serious about seeing more reviews in the style that I've done before, consider supporting me/my site, see for that my website where I give a few options:
https://swhs.home.xs4all.nl/fiets/tests/index_en.html
I may start a new thread to get some input or support on how to continue, if there is not enough interest I will likely do only reviews of stuff I want to and already did buy for myself.