Why You Need the BladeBallast™
Comments from a friend and Corporate Strategist at a major electrical/electronics manufacturer “I am sure that you have done your homework on why this solution and packaging are the most suitable for the application. And it looks brilliant.
What really intrigues me is the modularity: you now can offer the customer an option to try and then upgrade the system as needed. Typically customers value the flexibility (i.e. they are willing to leave a larger margin with you). Think Sony memory players compared to the iPod. Sony was superior, it looked as good or better than the iPod but it lacked accessories (other companies that acted as competitors, cooperative competitors in the accessory market).
Customers loved it: “I can get this for Christmas, then add that with the help of Grandpa Al, then add that and that with the help of Aunt Jane…”
So, stretch your thinking: how can you find a competitors , or better a complementary product (think IBM + Microsoft in the early days) that enhances your quest for the answer to customer’s questions?
Here is the ‘consequence’ of my customers’ questions, what I am calling BladeBallast™ product line (see attachment). I believe that I now have the answer to the water treatment market for both potable and waste water.
Cost Analysis of Various Ballast Architectures
I have one a cost analysis of the PFC, Buck, PP, and SR architectures, the circuit topologies available to the ballast designer and what I come up with is my PFC/Buck/PRPP ballast looks pretty good, even against series resonant. My ballast will operate either 2 1800 ma lamps or 4 800 ma (900 ma) lamps, with uC monitoring, control and communications. The more dominant SR ballast would have trouble driving the wide variety of lamps from a single product.
The New BladeBallast™ Concept
While coming up with this concept, I looked to the server industry. Even though I am not in that business, I have seen pictures of early server rooms and they ended up with banks of PC computers sitting on racks. So someone came up with the ‘Blade’ concept, where while the basic computing technology didn’t change a whit, the packaging changed radically. Operators can now form very dense server farms where hundreds of servers reside cheek by jowl and are fan force cooled, saving a lot of money by consuming very small footprints in server rooms where that space is very expensive because of the air conditioning and humidity control.
In visiting Artesyn, Intepro’s largest customer, I saw that their big business now is blade power supplies. Someone evidently figured that what is good for the servers is also good for the power supplies. They too take up a very small footprint that allows for every high power density as they are all racked and cooled in mechanical enclosures that cost of fraction of all the old boxes.
My reasons for using the same type of packaging are that for water treatment, customers may run between tens, hundreds and even thousands of lamps. When you do that, you find that the potted or unspotted can ballast, a unit optimized for mounting in a light fixture and driving a maximum of 4 lamps, becomes very unwieldy. I applied the same packaging technologies to my parallel resonant ballast line. This ballast technology, like the servers and power supplies
that are now blade based, is old technology. The innovation is in the packaging. I have used a disadvantage of my designs, the fact that I need large transformers but can only buy multiples of small ones, to my advantage. This allows me a ballast that is very flat; I only need 1.60” (40 mm in a eurocard rack) per card and can therefore get 10 of them in a 19” retma card cage.
Benefits of the BladeBallast™
My ballast is flexible anyway. I can now drive either four 800 ma (driven at 900 ma) lamps, two 1800 ma lamps, and the ballast probably has the ability to drive one 4.4 amp lamp, a lamp that is not currently in production at the moment, but is coming. If you figure that I am driving just shy of 500 watts per ballast, a card cage full of them is driving 5 KW of lamps! The user can of course buy multiple card cages.
Cooling is easy too. The black potted cans are cooled fairly effectively when the are used singly in a commercial sign or in a light fixture, but you put ten or twenty of them together in close proximity and you have some real cooling problems. The BladeBallast™, however, is unspotted, has no enclosure other than the card cage, and is easily cooled by a 1U fan rack mounted under one or multiple card cages. Because of the parallel resonant, bipolar technology, I need no heatsinks on the board and the transformers, the hottest part of the board, are very happy with this very small airflow. Transistors are barely 15 degrees above ambient with airflow.
You are right on the mix and match approach. If I had to sell ballasts tomorrow, they will be a parallel resonant (PR) half bridge (HB) design with single phase input and no PFC. If you look at my matrix, however, you will see that the real sweet spot for my parallel resonant design is a PFC front end that can accommodate either 120, 208 or 240 single phase, a buck chopper intermediate stage, and a PR push pull (PP) output stage. With this design, I get low THD, lamp current regulation and tuning to 50%. For the longer term, the series resonant (SR) looks pretty good, but it will fit on the same form factor and I am working towards using the architecture to drive one, two or four lamps also. Even my first model will offer three phase and that appears to be the lowest cost option at all.
The three phase solution will take some selling. Right now, all ballast designers seem to be ‘jelly bean’ guys. They design a ballast that is used one at a time. My ballasts are designed to be part of an integrated system and when you are talking 5 KW per card cage, this would be the equivalent of a 5 hp motor, and you seldom see single phase motors over about 1.5 hp, or about 1.5 KW. To most effectively utilize facility power, we will offer three phase. Not only does it automatically offer balanced phases, high power factor, and low line current, it lowers the cost of the ballast by eliminating a lot of the DC bus capacitance, which are quite expensive, from the design.
With a visible lamp system, you know when your lamps are burned out; your desk is dark. UVC can cause temporary and permanent blindness and second degree skin burns. The user needs remote sensing for notification of defective lamps and on large systems, this could cost almost as much or more as the black can ballasts themselves. The BladeBallast™, in it’s premium form, will have an integrated uController and RS485 communication. A card cage back plane will carry power to each individual ballast and also handle communications. Each of the two output connectors will be couple directly to the lamp wiring harness and those connectors are mechanically attached to a metal section of the backplane.
The user has the ability to select which model is best for his application and then at a later date they can change to a different model as all the BladeBallasts™ have the same form factor, communications hardware and software protocol and electrical connections.
BladeBallast™ for UVC White Paper Rev. B