Anytime! Taking you questions one at a time:
1. The output frequency in Hertz, or kilohertz, on a ballast impacts the cost of the design more than anything else. Higher frequency means smaller inductors, and the resonant inductor (used to strike the bulb in these designs) is the single most expensive item on the board. Higher frequency means cheaper to make, which is why high-performance DC-DC converter chips are screaming along at 4 MHz, 40 times higher than 100 kHz ballasts. Higher frequency at the output means more work for the design engineer to keep the EMI low, true enough, but that doesn't mean "high frequency = bad". Low freq square wave is like comparing an gas car to a Diesel car -- they can both run very clean and efficiently, but the engineers have to work a lot harder with the Diesel. A low frequency output at a hundred Hertz is 1000 times further away from the AM radio band and telephone DS1 and DOCSIS basebands than a 100 kHz ballast. Some are really quite good, but a low frequency design doesn't have this problem to solve in the first place which is why a lot of the majors (Hortilux, GE, me, Philips) went that way, and technology is generally going that direction. Long term, most HID ballasts will probably be square wave.
On THD:
Harmonic distortion is meaningless in this context. If we're building audio amps, yes, it matters, but we're not. Heating plasma can be done with a square wave, sine wave, or DC. Harmonic distortion measures how close to a pure sine wave the signal is. But it doesn't matter! DC is perfection, and is 100% THD. A square wave works beautifully (which is why GE, Hortilux, Revolution, etc. use them!) but it is 40%+ THD. The lamp doesn't care or notice how close to a sine wave the current is. So bah and pooh on THAT. It's just something for marketing people to talk about "Ooh twin turbo, low THD". It's crap. On the INPUT side, it does matter. The PFC stage is drawing power from the mains and they ARE sine waves. So low THD there is important to not distort the power line for everyone else like your other ballasts, stereo, TV, etc.
For a quick read on THD in PFC stages, Lambda has a good article here:
https://www.digikey.com/Web Export/Supplier Content/Lambda_285/PDF/TDKLambda_pfc_switchmode_powersupplies.pdf?redirected=1
Power Factor Correction (PFC) stages are pretty much required to meet European standards, and these standards include THD computed to the 39th harmonic. THD in the output stage is pure marketing spew. The lamp doesn't notice or care what the waveform is.
2. The Phantom guys are not lying, I think you mixed up two separate things: The frequency of the ballast output is how fast the light switches from positive to negative, and that frequency can radiate on longer lamp cords. The output is 100 kHz. The processor speed, the 32 MHz clock speed of the CPU, is how fast the internal microcontroller is running and has nothing to do with output frequency. That's like the clock speed on a computer, just s-l-o-w compared to a laptop since it's just an MCU, and not a full-blown fire-breathing Intel chipset. My Deva MCU runs at 20 MHz, clock speed, so anything in that range is pretty common. Below 10 MHz is getting too slow to be precise, anything above 16 MHz is enough, 20 MHz is overkill, 32 MHz is a shit-ton of fast. Hope that helps! BTW, I build a lot of gear for Hydrofarm, and if they are going to print something that isn't technically right I say something about it, and they've always fixed it when I do. I think they want to get it right, or maybe I'm just that big a pain in the ass, but either way they're pretty good about specs.
3. Give me enough Ferrites, and I shall move the earth.... Uh, not really.
Ferrite is a last-ditch, desperate, ball thrown from mid court while the buzzer sounds. The right way to suppress conducted EMI is with a two or three stage fillter, which costs money. That's why some cheap-asses don't do it, and very few did it well a few years back. Here's what these input stages look like:
The problem is Leakage Current. See the capacitors marked Cy1 and Cy2? These provide a path to earth for high frequency common-mode noise. That's good. They're also made by mortals and leak current at 60 Hz from the powerline. That puts a tickle of leakage current on the case of your ballast. A little is fine, is safe, and passes UL. Adding another filter stage adds more Cy caps, and is NOT safe. Leakage currents can be dangerous (as in ZAP YER DEAD dangerous) which is why we take the Last Chance For Gas approach with ferrites. Adding another common mode core like Lcm in the pic doesn't help without Cy's to tune it, so if your filter stage is Sheisse, it's redesign the filter, throw ferrites at it, or buy a ballast designed by adults and hasn't been value-engineered to death.
So: Better to start with a good product than try to fix it later with likely mixed results. If they cheaped out on the EMI filter stage, you KNOW they cheaped out on the filter capacitors and output transistors, and the ballast isn't going to live long or maintain brightness over the years.
4. Ferrite is safe, filters with additional Cy capacitors may not be. Ferrites will help, but may not be enough depending on how bad it was to start with. The ballast curve in my previous post won't be legal (forget good, just shooting for LEGAL) without redesigning the front end. No amount of ferrite will get you there. Best answer is start with a good ballast. There are good ones out there, and some very good ones too.
5. Yeah, those are my design. I built a bunch of them to fix some noisy ballasts causing problems. They work well, and I put them out for about cost. I think that's why Nanolux dropped theirs, since mine was just so cheap, and also technically very good.
6. Our lights are an all-new design and we've got them going to people for test and review. We've got a few guys on this board going to test them too.
Fans suck. My background is avionics, and you don't want to ground a plane because a cheap-ass fan took a dive. The heatsink design on the Deva is pretty slick -- you can put your hand anywhere on it while it's running. The fins are just MONSTERS, and she runs warm but not hot. Ceiling clearance is, maybe, 10 inches for air movement? If you have fans moving air in the room so it won't get blazing hot near the ceiling you can go less. Glad you like -- they're very, very different than any ballast you've seen. No resonant ignition, low frequency output, very, very low EMI, and intelligent adaptive control means constant brightness at any setting. Most ballasts are +/- 15% in brightness from ballast to ballast. We designed for 1% and guarantee 1.5%. What you set is what you get. There's more, but you asked about EMI not my new light...
Hope that helps!
Greg
