A while ago I posted about my problems with a MAX1771-based PSU that I was designing and building, asking for advice. I did get some advice, which was right -- but only partly so. Read on if you're interested in the details.
My board was based on Nick de Smith's design. What I wanted, though, is higher current capability: I wanted to be able to draw 80mA and possibly more. I also wanted to have a second HV output through a voltage doubler (to get ~400V for dekatrons) and some additional stabilizers for digital circuitry. I followed the advice given by Maxim in the datasheet and by Nick on his site -- reduced RSENSE to 0R025 (25mOhm) and used a large inductor (rated for 5A). The inductor is heavier than all the rest of the board combined :-) However, while the voltmeter happily indicated 180V at the output, the circuit was not stable: the MAX1771 was going into shutdown and there was a huge (30V) low-frequency ripple on the output. Ugly. The advice I got from this group was that this was likely caused by noise on the feedback pin, and that my FB trace routing was no good. Several board prototypes later (thank heavens for toner transfer, you can't beat the 30 minute turnaround time)... As it turns out, the advice about the feedback trace was right, but only partly so. My problems were due to noise on the FB pin allright, but no amount of careful routing would have solved them. One way to get the circuit to behave in a stable way was to increase the current through the voltage divider. Even with 0805 resistors there is room to spare and going from 1.5MOhm total to about 300kOhm total meant that at least this part of the circuit was less susceptible to interference. However, I couldn't find a way to make the thing work with a (vertical) potentiometer and I don't think there is a way. A 0R025 RSENSE resistor means that we're ramping the current up to 4A before releasing the energy stored in the inductor. 4A is a lot of current. The magnetic field produced by the inductor is so strong that it affects everything around it, even very carefully routed traces. A scope probe held in the air near the pot picks up almost .6V of interference! This only goes up as one approaches the inductor. Depending on how the inductor is mounted (which side of the board) you can get voltage increases or voltage drops on your FB pin. I did some experimenting and: -- moving the inductor away is a solution, but it needs to be *far* away, 8cm or so, -- shielding the inductor, while improving the situation, increases the current consumption 2x, produces audible noise and heats up the shield considerably, so it is not a solution. All in all, I don't think a circuit with an RSENSE of 0R025 is feasible, at least not with a 5cm x 5cm board. I'll be going back to 0R050 and 0R100 and smaller inductors. I can post some scope traces if people are interested. I have learned a lot in the process, and as a nice side effect I have a pretty clean PCB layout that is realizable with a single-sided board (requires one jumper) -- this might be useful for people who make their own PCBs. Once I get it verified and tested I plan to post it for people to use. --J. -- You received this message because you are subscribed to the Google Groups "neonixie-l" group. To post to this group, send an email to [email protected]. To unsubscribe from this group, send email to [email protected]. For more options, visit this group at http://groups.google.com/group/neonixie-l?hl=en-GB.
