http://www.manson.com.hk/products/detail/42Some time ago I picked up a couple of these very cheap when Dick Smith decided to get out of the business. However, as noted on this forum, they're pretty horrible things. Internally there's very little to them, a transformer, tap changing relay, two MJ15003 series pass transistors, an op-amp and a small handful of discrete components.
A thread on here the other day was talking about using a power amp IC as a linear supply, so that got me thinking since I have half a dozen LM3886's doing nothing. Could make a nice conversion to a programmable PSU, and since there's got to be thousands of these PSUs out there, it's probably make a nice open-hardware project.
With one of the power supplies torn down, it's evident how easily most of the main parts can be re-used. The tap changing and rectifier is on a seperate PCB, the heat-sink with all it's mounting etc will be perfect. It's also got a plethora of secondaries off the main transformer, so having separate rails for the op-amps and digital circuitry is a snap. No need to touch any of the main wiring.
So I set up an LM3886 in the standard application circuit but with all the DC blocking components removed, and a closed loop gain of 20 (the LM3886 is unstable at Gcl<10). I've done a fair bit of mucking around with this setup to familiarise myself with the behaviour of the IC. Overall it's fairly well behaved, however as can be expected with an audio amp IC, it has quite a wide bandwidth and conditional stability, which make it reasonably easy to turn into an oscillator. Some further compensation will be required. However, I've tried driving quite large capacitive loads and there's no hint of that unsettling it. So at least it appears to be a winner, to make best use of the heatsink two of them would be paralleled.
So the question now is what to do about it being able to source and sink power? The transformer has only one high current winding (with a CT), and the LM3886 still requires a significant negative bias supply (~>10V) otherwise it goes tits up. So the negative rail of the IC is currently run off a low current -15V rail. This works fine for sourcing current, but the moment it tries to sink any real current, it's going to pull everything out of whack. My thoughts at the moment are to set-up the current limiting circuit to be a window arrangement, with adjustable current sourcing (0-2.5A) and fixed current sink of just 100mA, which will stop it from killing the -15V rail.
Thoughts?
Home
Help
Search
Login
Register
