Tom's Hardware > News > Special > Miscellaneous > Solar PC Project Ends After Uptime of 245 Hours

Solar PC Project Ends After Uptime of 245 Hours

12:10 - Tuesday 9 October 2007 by Daniel Schuhmann - source: Tom's Hardware UK

Tuesday, October 9

For three weeks our Solar-PC based on AMD-hardware was powered exclusively by our solar panels – completely independent of the power grid. As such, our system represents a so-called isolated system. Our readers were able to follow the status of our photo-voltaic power plant “live”, thanks to our automated cameras that monitored the setup from various vantage points. Our carefully configured AMD-system consisted of an Athlon X2 BE-2350 (2 x 2.10 GHz), a Gigabyte GA-MA69GM-S2H motherboard with an AMD-690G chipset, 2 GB of memory and an 18” monitor. We decided to use AMD components because a similarly configure Intel system would have needed more power, resulting in shorter uptimes. At any rate, our PC-system is easily sufficient for everyday office tasks – few home or office users are likely to require more processing power. This does not apply to the performance hungry gamer, as the solar panels would not be able to supply enough power to a gaming machine unless the area of the polycrystalline Solar Cells were doubled or the buffer battery’s capacity increased. However, gaming viability was not on our agenda for this test

Solar PC Livetest

Now let us move on to the highlights of the past 22 days. Our two solar panels with a nominal capacity of 130 Watts each and an area of about two square meters reached their peak power output of 277.2 Watts on September 27. However, the best daily yield of 1375 Wh was achieved on September 23. Had we fed the electricity produced that day into the power grid here in Munich, we would have received a remuneration of €8.23 (€0.49/kWh). Over the course of a year, that would mean €142 – not bad at all ! On September 22, the Solar-PC set its record for the longest uptime, running for 18 hours and 43 minutes without interruption. The average sun-time during our Solar-PC Live Project (September 18 through October 9) was 8.5 hours per day, while the Solar-PC ran for an average of 11 hours per day.

This illustrates once more that a normal office PC can easily be run on solar power self-sufficiently. That’s not to say that outages don’t happen, as was the case on September 26 when the battery was drained after a series of rainy days and was unable to recharge.

Over the course of our project, the system was up and running for a just under 245 hours. Considering that we only got 189 hours of sun-time during the same period, we think that 245 hours of operation constitute a very respectable result.

If you’d like to go back and start reading about our solar-powered PC project from the beginning, have a look at the articles leading up to our live test :

Thursday, 4 October 2007

Our Solar-PC has been running for over two weeks now, so it’s time to take stock and look at the results so far. Overall, our system has generated 13.4 kWh of power - very respectable considering the time of year and the short time frame of only 17 days. After all, the weather hasn’t always been on our side, which meant the Solar-PC has remained switched off for longer than we would have liked on some days.

In theory, we can be happy from a financial perspective as well. Our solar cells with an area of just under 2 m² would have netted us a remuneration of €6.70 from the Munich public services. On Tuesday, the PC ran for almost 10 hours and 45 minutes, completing a long day at the office.

The weather was also quite good on Wednesday, 3 October, when our system generated 1109 Wh of power. This allowed the Solar-PC to run on solar energy for over 11 hours ! The battery capacity was finally drained at 2:10 at night, forcing the PC to shut down. Nonetheless, so far we have been unable to repeat or even break our current record of 18 hours and 27 minutes, which we achieved on 24 September when our solar panels generated almost 1300 Wh. We recorded the highest peak energy production of 277 Watts on 27 September, a day that was otherwise characterized by very cloudy skies.

As a result of the dreary weather on Thursday morning, the Solar-PC only start up around 13:00. Before then, the battery charge was too low. The weather forecast isn’t looking to promising either, so that we are not expecting a lot of uptime today.

Monday, 1 October 2007

On Saturday, 29 September, our network experienced a problem. Because of this, the monitoring system was no longer able to determine the status of the Solar-PC. The result was that for the rest of the day and on Sunday, the website displayed incorrect values for the battery’s charge level. Nonetheless, the system itself continued to run without any problems, since it is independent of the web presentation. Thus, the Solar-PC ran on Sunday as well, just as planned.

Another small issue affected the collection of some of the live data for the PC on the weekend. These values for the CPU-Temperature and the CPU fan speed were not gathered and were therefore not included in the graphs. We tracked the problem down and found its cause in a malfunction in one of our data capturing tools. We have now corrected the problem, and all data is now up to date.

The display for the battery charge was restored Monday, 1 October, at 9:00. Using the data from the performance log, we were able to calculate the correct current charge level for the battery and set the live monitor to this new value at 10:30. The system is now running smoothly and error free again.

Fortunately, the weather has improved greatly since last Friday, when it was absolutely miserable. While Friday’s power yield was only 99.1 watt hours, energy production sky-rocketed to 1212.37 Wh on Saturday. Despite being cloudy, Sunday was still good for 803.11 Wh. These yields are reflected in the Solar-PC’s uptimes on the weekend - 16:08h on Saturday and 11:06h on Sunday. That’s pretty decent, in our opinion.

Friday, 28 September 2007

Actually, we really didn’t have very high expectations for Thursday, what with the continuing rain. However, the cloud cover suddenly tore open for a short while around noon, letting the sun shine through. This led to a curious situation : Despite the bad weather overall, this sunny interlude let our solar panels set a new record of 277.2 watts for (short term) peak power generation. The previous record of 211 watts had been set the previous Sunday, 23 September 2007. The only explanation we can come up with is that the air must have been much clearer after the rain, which must have literally "washed away" the usual air pollution from the large city of Munich.

Sadly, this effect did not last long. Shortly after our moment in the sun, the clouds were back. At least the day gave us an acceptable energy collection time of 8 hours and 42 minutes, giving the Solar-PC enough power to run continuously for 4 hours and 15 minutes.

So far, Friday is once more looking rather bleak. We’ve had nothing but rain all day since this morning until now. With a little luck the sun may peek through the cloud cover this evening and send some solar energy our way.

Thursday, 27 September 2007

It seems we don’t have the blessings of the weather deities for our Solar-PC project. Wednesday’s weather here in southern Germany was as miserable as predicted, so that for the first time, the Solar-PC remained shut off for an entire day. Unsurprisingly, our energy yield also fell to a new all-time low, from 183.6 Wh on Tuesday to 90.82 Wh on Wednesday - and the PC doesn’t even start under 100 Wh.

Currently, the Solar-PC is running, since the battery stored 60 Wh of power yesterday. Coupled with the fact that it’s not raining here at, this may allow the PC to run for 2 to 3 hours today. Well, and there’s always that shred of hope left that maybe the cloud cover will suddenly open, revealing the sun...

Wednesday, 26 September 2007

All good things must come to an end, and the same held true for our run of sunny days. Tuesday marked that end, with a full day of rain. Of course, this also impacted the uptime of our Solar PC. 3 hours and 5 minutes show that at least minimal operation is still possible under these less than ideal circumstances.

Our yield fell from 1292.6 watt-hours on Monday to only 183.6 on Tuesday. The Solar-PC only booted up at 12:38 - far later than on the previous days, when it got an early start well before 10:00. It seems we were spoiled by the good weather. Since the panels weren’t able to generate enough power, the PC already shut itself off at 15:38, when the battery charge once again fell below the 100 Wh mark.

Things aren’t looking any brighter on Wednesday. The weather is so dreary that our lab techs don’t even feel like going to the Oktoberfest in the evening, which is currently on in Munich. After all, who wants to go and stand in the rain...

Tuesday, 25 September 2007

Our results for Monday, 24 September 2007 were just as expected. While your yield of 1292.6 Wh fell just shy of the 1300 Wh-mark, the results are still very satisfying. This time, the Solar-PC ran for 18 hours and 27 minutes, shutting itself off at 4:13 in the morning when the battery’s charge dropped below 100Wh.

Today was just as rainy as predicted - it’s raining cats and dogs. As a result, our energy yield will likely be very low, unless the weather clears up in the course of the afternoon.

Monday, 24 September 2007

On the weekend, our solar panels were able to show off their full potential. Sunday was indeed a sunny day here in Munich, resulting in a very respectable energy yield of 1375 watt-hours. Producing 1335 watt-hours, Saturday was only slightly less "productive" and even shone with a sun-time of 11 hours and 9 minutes - 10 minutes longer than Sunday’s (10 hours 59 minutes). The panels got an early start both days, beginning generation at 7:30, so that the PC was only ever offline for 5 hours, give or take. Friday, Saturday and Sunday, the Solar-PC’s battery lasted until around 4 am - providing late-night entertainment for the insomniacs among our readership. Once again, the system set a new uptime-record on Saturday, 22 September - 18 hours and 43 minutes of uninterrupted operation is our new best mark.

So far, Monday’s weather has been very kind to us. However, we are expecting to experience a massive drop in energy production tomorrow, Tuesday, 25 September - if the weather forecast is correct.

Friday, 21 September 2007 - Noon Update

This time, the battery held on until 03:28 in the early morning. Remember, two days earlier, it had already run out of juice at 23:30. The weather forecast looks very promising at the moment. We shall see whether the Solar-PC can beat its own record of the previous day. Let’s hope for clear skies.

Thursday, 20 September 2007

Conditions on Thursday were even better. You could almost say we had ideal weather, a perfect late summer’s day. This is reflected in the solar panels’ energy yield : Between 7:30 and 18:47 they generated a total of 1304.52 Wh - a full 43 percent more than on the previous day (909.94 Wh). The surplus energy not needed to run the Solar-PC was used to charge the battery. Thus, the system was able to continue running until long after sunset. In fact, it passed the midnight-mark for the first time for a total uptime of 17 3/4 hours (previous day : 13 1/2 hours).

Wednesday, 19 September 2007

Heavy fog lay on the THG-lab on Wednesday morning. Luckily, these bad conditions changed for the better very quickly.

The best September weather we could have wished for allowed the solar-panels to generate so much power that our PC was able to run for 13 hours - without a single interruption. The total yield today was 1000 Wh, which our local power company - the Munich public services - would remunerate with 52 Euro-cents, if we were to feed it into the power grid.

The energy bonus the system had accumulated by the early evening of the previous day enabled the Solar-PC to start at 10:04. Furthermore, the battery had also stored enough energy for the system to continue running until far into the night, only shutting down at 23:30 !

Many readers commented that they were unable to tell from the still images of the Solar-PCs monitor whether the system was actually running. To solve this problem, the display will show a slideshow of various images so there will be a visible change from one snapshot to the next.

Tuesday, 18 September 2007

After weeks of preparation, we begin our live project. We conducted numerous checks and tests on the weekend and on Monday – dozens of test runs to ensure that nothing goes wrong. After the sunshine of the last few days, we were filled with confidence. Then came Tuesday and nothing but rain all day long.

Although the battery kept charging bit by bit, it took until the middle of the day – 12:59, to be exact – until the solar panels had generated enough power for the battery to reach 100 Wh, allowing the PC to start up. The uptime was just as disappointing. The PC was only able to hold on for two and a half hours and had to shut down at 15:24, when the battery charge fell below 100 Wh again. Even though the battery continued to charge for the rest of the afternoon, there was not sufficient energy to start the PC a second time.

Let’s hope that Wednesday blesses us with a little more sunshine. Our preliminary test showed us one thing – we can only achieve acceptable results with a slightly cloudy sky and occasional sunshine.