Ballast for low pressure mercury lamp

[nascheme]

Sorry for the offtopic post but I hope someone here has some advice. My parents live on a farm and use a UV sterilizer for their water. The ballast on the low pressure mecury UV lamp keeps blowing caps and fuses. I'm wondering if it's due to bad power, bad components or just bad design.

A little searching reveals the following patent: US 6274988 (. The circuits in the patent appear to match the board I have. The half-bridge driver is a International Rectifier IR2151. The two MOSETs are IRF830. The big electrolytic caps (47 uF, C1 & C2 on page 3 of the patent) don't seem to have a name, CE EB (stylized 3 or M), 105 degC, 250 V. The board has the year 2002 on it.

The board around the MOSETs is darker, I suppose due to overheating. Is this thing worth trying to repair? How would I go about verifying the operation of the circuit? I have a 10 MHz DS oscilloscope.

[Scenic]

Re: Ballast for low pressure mercury lamp

The big electrolytic caps (47 uF, C1 & C2 on page 3 of the patent) don't seem to have a name, CE EB (stylized 3 or M), 105 degC, 250 V.

?
that would be Panasonic a.k.a. Matsushita

[nascheme]

Re: Ballast for low pressure mercury lamp

[...] Panasonic a.k.a. Matsushita

Yup, that's it. Thanks.

[Toasty]

Re: Ballast for low pressure mercury lamp

If it's blowing C1 & C2 & fuses, then I would say that at least one of the rectifiers D1,2,3,4 is bad (shorted). That would allow AC from the mains into the caps and take them and the fuse out.

Check also C4 for shorting and R1 for value.

Toast

[nascheme]

Re: Ballast for low pressure mercury lamp

If it's blowing C1 & C2 & fuses, then I would say that at least one of the rectifiers D1,2,3,4 is bad (shorted).

Good call, thanks for your help. All four of the bridge diodes have 0 Ohms of resistance both ways. I have another ballast that blew a fuse but the diodes appear to be okay on that one.

Check also C4 for shorting and R1 for value.

I couldn't find these after searching for a while. I guess the circuit in the patent is not exactly the same.

A few more things I found:

• The circuit appears to be some type of flyback converter. The circuit seems very simple compared to other real world switchers.
• There is no in-rush limiting for the two 47 uF caps connected to the recifier. Based on Horowitz & Hill, I understand normally there would be some protection (e.g. a thermistor).
• As far as I can see, there is not much for protection circuity (e.g. over voltage crowbar, etc).
• There is a variable resistor that is supposed to tune the circuit. I'm wondering if I replace the caps and diodes, if I have to re-tune it.

If it's useful, here is a picture of the board:


The black and white wires at the top are line power. The four wires below go to the low pressure mercury bulb. I'm wondering now if I should:

• Fix this thing by replacing components.
• Replace some components and also make some improvements (no sure what those would be).
• Buy a new ballast from the same company (hopefully the new one is more reliable).
• Trash the whole system and buy a new UV sanitizer.

Again, sorry for the offtopic post. I was thinking that this circuit would be similar to what drives the lamp on some displays. BTW, I'm guessing it would not be wise to start poking around with my oscilloscope. It's rated for 300 V, I think and this thing probably produces more than that.

[crtfool]

Re: Ballast for low pressure mercury lamp

The circuit board and transformer look very toasted - your best option would be to buy a new ballast, in my opinion.

[Toasty]

Re: Ballast for low pressure mercury lamp

Replace the diodes and the caps. You need to test the 2 MOSFETs also. If they shorted, that would overload the rectifiers D1-4, which when shorted could allow AC into the caps blowing them. It's a cascade failure and who knows when the fuse finally made up its mind to open.

I am also concerned about the small ceramic cap just below and right of the transformer. The top looks cracked. If so, that needs to be replaced also.

If the ballast transformer is shorted, then it's time for a new unit. The newer units seemed to have been revised as 1 unit for all voltages. Previously there were 2 - One for 120v and another for 240v.

R1 is probably one of the 3 large resistors. A backside shot of the board would make it simple to determine. Since none of the 3 large resistors appear burnt, cooked, or overloaded, at this time I don't believe that the control circuit has failed. R1 is inline and serves as a voltage dropping resistor and current limiter to the chip.

If you need the diodes, e-caps, ceramic cap, and MOSFETs, it should run less than $20 incl s/h. I think I saw a price yesterday for the whole unit of about $75.

Toast

[Toasty]

Re: Ballast for low pressure mercury lamp

Uploaded pic to forums so it is not lost to future users.