I think KNNA is also building a small sphere or has built one already. Your sphere needs to be calibrated regularly. For that purpose we have official calibration lamps, there are several ways to do it though.
Yes, we have spectrometers too. That's for example how we measured the plasma lamp spectrum. For the UV range you would preferably have special sensors, I think the OceanOptics shows a lot of noise in the low bandwidth region.
The problem with the lamp tests was that the lamp/ballast combination is also important: on some ballasts some lamps do not reach optimal 100% power conditions, for example the only get to 960W instead of 10000. But we found a solution for that too: we modified an electronic ballast so we were able to tune the ballast to every lamps optimal output power, and optimal overrun/underrun values. So when we measure at 1000W we indeed measure 1000W going to the lamp, according to the lamp specs. With the power analyzer do not only measure the ballast input, but also the ballast output current, voltage, frequency, THD, harmonics etc.
I just had a look at the preliminary list after one series of measurements, it includes the Ushio, Hortilux, Venture, Gavita, Philips USA, I believe even Digilux. We have quite a collection of lamps here already. I have lumens and micromole measurements. All new lamps after minimal burn-in period in a calibrated ulbricht sphere with optimal lamp power (electronically tuned with a high frequency ballast (>90 kHz). I really hope that other manufacturers will now also publish their ppf values for their lamps. And of course everyone who has a sphere at their disposal can repeat every test, where it is important to measure at the same conditions, that is ballast adjusted to optimal output to the lamp. Some lamps are at 100% already relatively high in their efficiency curve, performing less efficient when you overrun them. Other lamps still have enough headroom and even show improved efficiency overrunning. The actual lamp voltage is a good indication for that. Over time the lamp voltage rises, until the lamp runs out of specs. That's when a good electronic ballast shuts down. That also can happens for example when there is a manufacturing problem with a lamp and the lamp voltage rises too fast or when the lamp already has a high lamp voltage and then is overrun for a while. It will still start at 1000W but shut down by the ballast when it it boosted.
Interesting stuff
This is the sphere in which we test lamps:
More pictures of the equipment can be found here Also a good example that some chinese ballasts do not come to 100% output. They do not use a lot of power and they do not get warm but you do miss up to 10% of your light. You can also see that with these particular lamps the efficiency increased when overrunning. You can also see in the first table that, at optimal input voltage, the magnetic ballast performs as good as the lumatek running at superlumens (+5%), obviously not as efficient of course.
Yes, we have spectrometers too. That's for example how we measured the plasma lamp spectrum. For the UV range you would preferably have special sensors, I think the OceanOptics shows a lot of noise in the low bandwidth region.
The problem with the lamp tests was that the lamp/ballast combination is also important: on some ballasts some lamps do not reach optimal 100% power conditions, for example the only get to 960W instead of 10000. But we found a solution for that too: we modified an electronic ballast so we were able to tune the ballast to every lamps optimal output power, and optimal overrun/underrun values. So when we measure at 1000W we indeed measure 1000W going to the lamp, according to the lamp specs. With the power analyzer do not only measure the ballast input, but also the ballast output current, voltage, frequency, THD, harmonics etc.
I just had a look at the preliminary list after one series of measurements, it includes the Ushio, Hortilux, Venture, Gavita, Philips USA, I believe even Digilux. We have quite a collection of lamps here already. I have lumens and micromole measurements. All new lamps after minimal burn-in period in a calibrated ulbricht sphere with optimal lamp power (electronically tuned with a high frequency ballast (>90 kHz). I really hope that other manufacturers will now also publish their ppf values for their lamps. And of course everyone who has a sphere at their disposal can repeat every test, where it is important to measure at the same conditions, that is ballast adjusted to optimal output to the lamp. Some lamps are at 100% already relatively high in their efficiency curve, performing less efficient when you overrun them. Other lamps still have enough headroom and even show improved efficiency overrunning. The actual lamp voltage is a good indication for that. Over time the lamp voltage rises, until the lamp runs out of specs. That's when a good electronic ballast shuts down. That also can happens for example when there is a manufacturing problem with a lamp and the lamp voltage rises too fast or when the lamp already has a high lamp voltage and then is overrun for a while. It will still start at 1000W but shut down by the ballast when it it boosted.
Interesting stuff
This is the sphere in which we test lamps:
More pictures of the equipment can be found here Also a good example that some chinese ballasts do not come to 100% output. They do not use a lot of power and they do not get warm but you do miss up to 10% of your light. You can also see that with these particular lamps the efficiency increased when overrunning. You can also see in the first table that, at optimal input voltage, the magnetic ballast performs as good as the lumatek running at superlumens (+5%), obviously not as efficient of course.
