I’ve just had a nice conversation with my energy advisors, Select Solar, about some of the issues I’ve seen with the tests I’ve done so far.  The main problem being that Li-Ion batteries will not trickle charge. They require a certain level of stable voltage/current before they start to charge and this wastes the possibility of charging a battery at lower sun power. In fact, on a hazy day or with light cloud cover, the sun power might not actually go above 70% and you have no chance to charge anything. This could be  major problem and I’ve decided to try and address it.

A solution to this problem, as I mentioned before, is lead-acid batteries. These are the type you will find in your car and are a tried, tested, reliable and relatively cheap solution. The problem is that they are also extremely heavy – at least 400% of the weight for like-for-like power storage. They also operate at 12v which means voltage conversions (and more electronics and power-wastage) in order to operate the UMPC.

The decision I’ve made is to buy a 70W/hr 12v lead-acid battery (2200g) and to try and use this as a charge buffer. Here’s a little sketch of the planned set-up.

The idea would be that I connect devices to the lead-acid battery as the solar power increases during the day. For example, I might add a couple of AA batteries or my mobile phone in the morning. At mid morning I might replace it with the Li-On battery and during the peak hours I could re-attach the AA battery charger with 2 or 4 batteries depending on sun power.

Two problems I can see here:

• How do you know the charge on the lead-acid battery?
• Will the lead-acid battery be able to feed up to 1.5 Amps?

There’s only one way to find out. I’ve just put an order in for a 70W/hr lead acid battery and charge controller along with some fresh AA batteries, a powerbank tip adaptor for the Samsung Q1, a 12v charger for the mobile phone and something I’ve always wanted, a flexible USB LED-lamp!!!

Later today I will probably be ordering the solar panel. I won’t be using the P3 Panel as the only advantage with that was that it could drive 19V into the PowerBank. Now that I’m going for a 12V source solution I’m going to be looking at the 25W version of this Sunlinq panel which is the same price as the 15W version of the P3 panel. There’s a risk that its going to be too big for the bike at 1m long but even if I fold 25% of it away, its still going to give more power than the 15W panel and during mid-day pause and work sessions I hope to get the full 25W out of it.

I’m quite lucky to have been able to test a lot of UMPCs out over the last year. Most have been good in their own target markets but only some have been good for me. One of my favorites, the Kohjinsha SA1, an AMD based device is also one of the most-power efficient too and I took it to CeBIT this year where I used it exclusively for 5 days. It worked well and it demonstrated to me how much I might use a PC in one day. During the daytimes I was able to get by with a single charge – 4 hours. In the evening I was using it for another 2 hours, minimum! OK. CeBIT was a worst-case scenario. There was breaking news all over the place and there was one point where I was walking to a meeting with Intel and trying to edit and post a news entry at the same time. I don’t expect to be doing that on the bike! I do expect to be stopping 2-3 times per day for 30-60 minutes at a time though. Primarily to update my location, photos and geotags and to make quick reports here and also to work on any news that’s breaking. In the evening I expect to do a small amount of video editing, uploading and maybe a longer post, chat, IM etc etc. 2 hours is what i’ll give myself in the evening depending on the forecast for the next day.

At 9W drain (hopefully less in evening with low backlight and careful control of HSDPA and WiFi) that’s about 36W per day.

In addition to the UMPC I will have:

Mobile Phone. This is a Nokia 6380 with the 3.3v 900mah battery. 3watts. I will use this as my email and quick blog news checker using the gmail java app and the opera mini browser with bloglines mobile. These are combinations of applications that I’ve found very efficient and there’s really no need to turn on the UMPC to read email and RSS text. I can even do short emails on the phone too. If something interesting comes up then i’ll take the UMPC out of hibernate and do further reading, research or even email or blog. If I throw an hour or so of MP3 or FM radio in, I estimate that one charge per day on the 6280 will see me through. That will take about 5W of power from the power bank or I might be able to charge it in the early/late part of the day when the suns power falls below 70%.

Camera. This takes 4 x AA batteries. I use 2300mah 1.2v rechargeable batteries that should (if I use the camera carefully) last up to two days. That’s again, 5W per day. Charging will have to come from the solar panel during the day. Hopefully this can be done for an hour earlier or later in the day. Testing and research into different types of chargers needed. Smaller capacity batteries could also be useful.

GPS logger. This takes 2 x AA batteries every 15 hours. I hope that 2 batteries lasts me 2 days so that’s 2.5W per day. I will take a Bluetooth GPS device in case the logger fails. I think it lasts 10 hours on a 3W battery and can be charged from a USB port so its much the same in terms of power.

Lamp. This will be a very efficient LED lamp. Its possible to get ones that are solar powered or hand-cranked. I expect this to be a negligible drain. I will even take a few candles just in case. Bear in mind that it will be light until 2200 and will get light again at 0500. I don’t expect to be awake much during these hours!

Finally, something I haven’t mentioned yet, a backup PC. Well its more of a web tablet really. Its the Nokia N800 and it will run for 3-5 hours on a 5W battery. That’s very efficient and could serve as an email tool, IM tool and, at a pinch, blogging tool (although i’ll have to look into a mini keyboard solution for it.) This will only be used if the worst happens. I.e. I run out of power or the main UMPC breaks. Otherwise I won’t use it and it will stay packed away.

Total daily usage:

UMPC: 36W

Mobile Phone: 5W

AA batteries for camera: 5W

AA batteries for GPS logger: 2.5W

Total daily power usage: 48.5W

In the previous post I estimated I could average 48W solar power collection and storage per day. Bingo! Now, did I forget anything?

The other luxury I will give myself is a fully charged set of batteries before I go. I don’t think its unreasonable as anyone in their right mind would ensure that their batteries were charged before they went. Here’s the battery power that I will take with me:

• Powerbank: 56W
• Extended battery: 60W
• 2nd extended battery. Empty 
• 8xAA batteries for rotating through camera. 20W
• 4xAA batteries for rotating through GPS. 10W
• Phone charged. 3.5W
• Emergency UMPC charged. 5W

Total starting power: about 150W.

In theory, that lot will give me a 1 day buffer which means at some point in the tour, I can have a day of bad weather and continue as normal.

I will talk to select solar next week to see what they think of my estimates and if they give it the thumbs-up Ill start real testing – assuming the sun comes out!

The weather is still bad here in Germany and there’s no chance of getting reasonable testing done on a solar panel at the moment. The forecast is also bad so I’ve decided to send the test P3 panel back. In the very short time I was able to test it I was able to learn a lot though and its given me some thoughts about how I will move forward. This post goes over some of the stats, figures and calculations so its not exactly the most interesting post for the average reader. For those that are thinking about doing a similar thing though, this could be a very interesting post for you!

A very important lesson was learnt on the first day – the charging system doesn’t even start working until the sun reaches 70% of full power. This is due to the Li-ion charging system on the powerbank. It wont trickle charge so only kicks into life when there’s enough power to drive it. This renders about 8 hours sunlight per day as unuseable with this set-up.

The second thing learnt is that the charging system doesn’t work faster if there’s more power available. It seems that the power bank needs 12W to charge and will not take any more power if, say, 15W is available.  Thinking back to my basic electronics, I guess this makes sense! This means more waste though and requires balancing the charging requirements with the power available if i’m to use this set-up.

One very nice feature of the P3 panel was that it was able to charge the TabletKiosk powerbank directly, without voltage conversion. That helps a lot because the other solar panels i’ve looked at only drive 12V. That would mean converting voltage up to 19V and losing power in the process. Taking a few of the figures learnt, there’s a simple equation gives me the theoretical power possible per day.

Power available (W/hr per day) = Powerhrs x ChR

Powerhrs is the number of hours per day when the sun is over 70% power. (taken from sun power records. June. Bonn, Germany) This is about 7 hours in May, June and July.

ChR is the charge rate of the power-bank (must be about 0.7 of solar panel rating for best efficiency.) = 12W

This assumes a sunny day. No cloud cover. No shadows and the panel facing the sun. The changes of that happening are slim. I have the advantage that I’ll only leave the house if the 6-day forecast is good so I can reduce this risk to a minimal amount but I think I’ll lose another 20% through shadows and positioning while I’m moving. If cycling around becomes a problem with shadows and sun positioning I might have to carefuly plan my stops in order to maximise solar panel positioning. Maybe a stop at 10-11 and again at 3-4 would help get the best out of the early and late light.

So lets plug the figures into the equation and see what we get with the P3 solar panel and the TabletKiosk MP3400 power bank.

7  x 12 = 84W/hr energy per day assuming 15W from panel at max sun power.

Now for the second part of the equation:

1. There’s no such thing as a perfectly sunny day. Especially when you’re cycling. I’ve been advised to use a figure of 4 hours per day for sun. In my tests, on two sunny periods, I managed to load a 56W/hr battery in about 8 hours so the equation seems to be roughly OK. At a 12W load rate, that’s about 4 full hours of sun so I agree with the 4 hours rule.  
2. You can not charge and use the bank at the same time. This means that a second power source is needed for the daytime. In the evening its fine because I run from the powerbank. I could use the battery on the device UMPC in the daytime of course but what happens when its empty? Can I charge it from the solar panel (that’s stupid because I would need a second UMPC just to charge the battery – the solar panel will not be able to run AND charge a UMPC. Also, if there’s not enough voltage or current protection on the UMPC I would risk killing the UMPC) There’s is the possibility of topping up the UMPC overnight from the power-bank but If I do that, I have to consider the following…
3. To charge other devices from the bank, there is something like 20% power loss through the charging circuitry.

There’s one other point too. The powerbank can only store 56W/hr of power. Taking point 1)  into account brings me down to 48W/hrs on a sunny day which I guess solves the problem of limited powerbank capacity!

I think the best solution for point 2) is to use a second power bank. This also has the advantage of giving me a spare powerbank should one fail, and, being able to fill a second device should I find myself with sun and nothing to charge! It also reduces the risk of killing the UMPC and avoids the wastage of 3)

So we’re now at 48W/hrs per day. Its incredibly poor considering the power of the sun. In theory, the suns power is equivalent to 1Kw of energy for every m2 and that makes solar energy capture and storage horribly horribly inefficient. Consider that in 4 hours, about 2KW/hr of energy will be hitting the solar cells on the panel I tested. Yes 2000W/hrs. Enough energy to power a small UMPC for 1 month at 8 hours per day! The efficiency is a crazy 48/2000. Just over 2%.

How can this be improved? One thing I could do is to find a lower-power charging solution. That is, start storing energy before the sun reaches 70%. If you can start storing energy when the sun is at 50%, you have a few more hours per day and if you can start gathering energy at 30% power, you can even get some when its cloudy!

The second variable is the power-loss on charging. This is generally because the Li-ion batteries need a fairly strict charging voltage and current and there’s a fair bit of circuitry there to protect them. Lead-acid batteries could be better but, wow, the weight! To store just 70W/hr of energy you’re looking at 2.5KG. The 56W Li-Ion power-bank I have is just 500grams. 

The third option is to get a panel with more efficient solar cells.

Having said all that though, this is an off-the-shelf system that appears to work. I didn’t need to make any modifications at all and I like the idea that this project could go ahead with off-the-shelf equipment. It makes it easier for others to replicate it.

The big question now is, will 48W per sunny day be enough? If so, I don’t have to worry about the terrible efficiency and high cost. It will be enough to provide me with power where I can’t get it. I’ll be analyzing the power requirements in the next post to see if 48W per sunny day is enough. If it isn’t, i’ll have to go back and look at other options.

Its decision time on the UMPC and I think I’ve just about come to my conclusion – The Samsung Q1b HSDPA. Its not perfect but it covers the important requirement of efficiency very well. Maybe in a few more months there will be better options but right now I need to focus on one device and start testing and getting used to it.

Let me take you through the reasoning though. You might be thinking about the same thing yourself so I’ll list my requirements and then go through the options.

Requirements.

• Power efficient. Obviously. I’m looking for sub 10W average power drain for a total power budget of under 60W per day. (More on the power-budget in another post.)
• Lightweight. All UMPCs are lightweight. This shouldn’t be a problem.
• Bright screen. Some are better than others and when you’re outside, this can make a huge difference. Put it this way, I’d rather be sitting on the terrace of a cafe rather than inside on my own!
• Keyboard. I’m going to be doing a lot of typing so this is going to be an important requirement. Nearly all built-in keyboards are nothing more than thumbboards. I will have to consider external keyboards
• Protective case. A slip case isn’t going to be enough. I need a very good protective case. Preferably rugged.
• Ruggedness. 8 days on the road is going to be tough on a UMPC. I’m concerned about dust and the hard drive.
• Bluetooth connectivity (for data through mobile phone). There are other connectivity options including connecting a mobile phone via USB and getting a UMPC with built in 3G data modem.
• SD card slot for digital camera images.
• Extended battery. Generally the batteries with UMPCs are around 20-30W/hr types. an extended battery is going to make charging easier.
• Reliable. A proven track record is important.
• Cost effective. Upper limit is around 1500 Euro for all UMPC equipment.

I have already chosen VIA as the platform I will use. Intel is powerful but can start to sap power very quickly is not used carefully. One stray process and you can eat 15W in an hour. I don’t want that to happen. AMD is another option and there are some nice low power solutions around. The only problem is that the processing power of the AMD solutions is limited. I have edited a couple of movies on an AMD platform and it was a long process. I can not afford to leave the PC for one hour crunching videos and sapping energy.

VIA have been on the UMPC scene since the Origami launch in March 2006 and the core of their solution is the C7-M ULV CPU. Its available in 1, 1.2 and 1.5Ghz versions but until recently, only the 1Ghz versions have appeared in UMPCs. Amtek (OEM), Samsung and Uren are really the only manufacturers that already have VIA-based UMPCs on the market and although there are three more available this month, they are new, untested and not available for immediate delivery.

The T700 was the first VIA-based UMPC. Its a basic device compared to some of the UMPCs we’re seeing today with a simple style, no SD and no web cam. It has proven to be a solid device though. It has an extended battery and there are bump cases available for it. Its also good value. You can pick up  branded versions for under $800. The big problem with the Amtek T700 though is its average battery life. 11-12W drain is normal. It is also lacking the SD card slot required.

The second option is the Q1b from Samsung. This is the VIA version of the Q1, another UMPCs released as part of the Origami project from Microsoft. Its a slightly more attractive design and comes with a 280nits screen. Brighter than the T700. Again there’s no mouse and no SD card slot but bump cases and even a specially made Samsung organiser pack with keyboard is available. In fact, I’m using one  to type this right now and its a very very nice keyboard. The other major advantage of the Q1b is its battery life. It can average well below 10W and is even reported to run in a battery saving mode with low-backlight for 4 hours. That’s about 8W average drain. Again, its a good value buy although its not available in its basic form in Europe because Samsung introduced it here in Germany with a built in HSDPA modem, bigger disk and more memory. Potentially a great advantage but the price is quite a bit more than the T700 is available for. It would stretch the budget when taking into consideration the keyboard and protective cover.

Finally, there’s the Uren V1. Unfortunately this one is very quickly out of the running because it’s not available at outside Asia. Its also a car-focused UMPC. The most difficult thing about the Uren though, is that no-one has really written any reports about it in English. Its certainly not a UMPC that I’ve been able to get hold of and test so at this stage it goes straight to the bottom of the list.

So there you have three UMPCs. One looks like good value and capable. The other looks like its perfect for the job but expensive and the other is out-of-reach. I mentioned a budget of 1500 Euro for the UMPC. This includes the cost of protective cover, keyboard and extended battery. When I add up the costs of the Q1B HSDPA, the organiser pack (with the wonderful keyboard) and an extended battery I come to just under 1500 Euro. Phew! I’ve even got some spare change for a USB-SD adaptor. The Q1B HSDPA it is then and as you might have seen on UMPCportal.com, I have already taken delivery of one. Unfortunately it had a faulty USB port so its gone back but the short time I spent with it was really nice. The HSDPA modem is going to be a great advantage and I’ve already transferred the data portion of my Vodafone tariff to a new SIM card. No messing with Bluetooth or cables and no worrying about keeping the mobile phone charged. If the phone dies there’s always skype!

Thanks to Select Solar UK I received a 15W P3 Solar panel this morning. As I write, it’s out in the garden with 660W/m2 of sun power beaming down on it! Temp in the shade is about 22 degrees.

P3 Folding Solar Panel

Its smaller and lighter than I expected too.

The first test was to measure the open circuit voltage and at about 450W/m2 sun power (I’m using this web site to give me my sun power reading – its very local to me.) it was over 20V. At 650w/m2, its open circuit voltage is about 28v which is higher than I expected.

I connected a AA battery charger with 2 AA batteries in it and it was charging straight away. With 4 batteries I had to wait until about 10:30. I also tried connecting the TabletKiosk PowerBank but it wouldn’t charge until the sun power reached about 650w/m2. I haven’t connected the Q1 yet because I’m missing a connector for a DC-DC converter I will use for protection.

Look at the graph below. Its the sun power from today and I’ve added some notes to it.

Click to enlarge.
A – 2 AA batteries. B – 4 AA Batteries. C – PowerBank

You can see by the wobbly graph that there’s a few clouds around today. Yesterday was perfect and using that part of the graph I can work out how much time there is for a given level of sun power. The sun is at or above the ‘C’ level for 6 hours a day for example. The area under the curve gives us the power too. I’m not going to get my integration maths book out right now but its fair to say that if the sun peaks at 800w/m2 (80% of Panel power) the average is something like 0.7 (RMS?) of (600+(800-600) for 6 hours. Tap tap tap. That’s 75% of 15W panel power for 6 hours….a possible 67W/hrs of power. (average 11w for 6 hours.)

67W/hrs of power is a full power bank or full Samsung Q1 extended battery which is what I think I need every day to keep me going. I might be able to get that down to 40W by being careful with the UMPC but if there are 2 days of clouds, I’ll be out of power.

This is just a preliminary test. I need to do more double checking on this. And don’t forget, its sunny today. The barometer dropped sharply this morning so tomorrow is going to be very different. It will be interesting to see if I can power anything under cloud cover.The other thing to consider is that I’ll be moving on the bike. There will be shady times! The big question is, do I take a risk and challenge and go with the 15W panel or do I take the next step up and go for a 24 or even 30W panel. Like I said, more testing needed!

I’ve put up a gallery of the P3 panel.

More testing going on today and tomorrow.

Thoughts. Should I take a simple sun meter? How much power will I lose through using a DC-DC converter? Should I buy a small multimeter to take with me?