Project files: IRM Switching PSUs

What is it?
Since I first discovered the IRM-series of compact switching supplies from Mean Well I’ve grown quite fond of them. They are compact, cheap and very easy to implement so they are perfect for everywhere an “aux-voltage” is required to power non-critical circuitry. Through the different applications I’ve found for these I have managed to build up a full series of boards suitable for the IRMs.

While some of the boards can be (and are intended to be) used for “serious” stuff (to be shown later on), a very obvious application for most of these boards are as AUX-supplies for powering relays, displays, logic circuitry etc. where a bit more or a bit less ripple and noise are of no consequence, but where the compact size and low standby consumption is a real plus.

There are four board versions, suitable for the IRM modules in all versions from 3-30W output power (the 30W board is missing from the pictures as I couldn’t find the prototype when they were taken – sorry! 😀 ).

How big are the boards?

  • The 3W board measures 1.8” x 1.5” (app. 46 x 38 mm.)
  • The 5/10W board measures 1.2” x 2.65” (app. 31 x 67 mm.)
  • The 15/20W board measures 1.25” x 2.95” (app. 32 x 75 mm.)
  • The 30W board measures 1.6” x 3.6” (app. 41 x 92 mm.)

What is the status of the boards?
All of the board files are version 1.0 or higher. Some tweaks have been done after the initial protoypes for a few of them, mostly because of errors/issues with the IRM module footprints.

Does it use any special/expensive/hard-to-find parts?
No, none. Several places to get the IRM-modules them selves (Mouser, Reichelt, TME etc.) and everything else on the boards is more or less optional 😀

Anything else I need to know?

  • The modules have worse specs for ripple and noise than most linear regulators, but obviously the switching frequency is quite high (66-100 kHz depending on model), which means that passive filtering like an LC or a CRC (“pi”) filter would be an ideal way of reducing the output noise. I have a couple of examples for that which I might show later.
  • I haven’t been able to find a spec for how much capacitance the modules will tolerate on the output, but it probably should not be overdone.
  • Remember that obviously one side of the board carries mains voltage, so take the necessary precautions when working with them.

Download design files here

Related information:
Note: Always read the “intro post” for additional important information about my designs.

Evolution of a concept…

I’ve been playing with the (somewhat) unusual combination of a class A amp and a switching PSU before (here). The previous version worked quite well, but even as I was building it I though that there might be scope for optimisation and for making it smaller – which to be honest was my goal all along.

The amplifier section is now my own version of the JLH 1969 with MJL21196 output transistors (because I had them available). The heat sinks are the same surplus items I’ve used before – because they are cheap and exactly the right size for what I wanted to do 🙂

As I observed in the previous post, the chosen PSU was more than a little overspec’ed for the job, but as luck would have it a much better choice came up recently – the Mean Well EPP-150-27. These are “next generation” supplies (i.e. more compact at 2” x 4”), there’s a convenient 12V aux supply for a fan etc. and the 27V output voltage is spot on for the JLH. There’s a smaller 100W version that could be used as well, but since I could get these as surplus items off ebay there was no incentive to do that 😀

I’ve also been looking at getting some more quiet fans, but I’ve now realised that the most annoying noise from fan-cooling doesn’t really come from the fan itself but from the sound of air rushing through the various holes in the enclosure. Once I’ve done the final design I’ll have to test what can be done to keep the amps as quiet as possible. The noise level on the first iteration was OK, but definitely with room for improvement.

To marry the parts together I came up with a “central core” design where the amp and the PSU are bolted to the heat sink tunnel and that’s about what I have done so far. The rest of the mechanics are in the works and there are a few (as yet undisclosed) additions that I will add as well in due time 😀

More JLHs…

Yeah I know, I should probably stop making these at some point 🙂

We’ll do this one quickly then: Standard JLH1969 Ebay-board with upgraded components and better transistors (MJ15003). Fan-cooled heat sink with temperature control (we’ll see how well that works…). Industrial-grade 10A switching PSU per channel. Monoblock configuration in Modushop GX288 chassis.

The PSUs (Artesyn NLP250) are overkill for this application but they were cheap (surplus items). And besides, more is better – right? Not necessarily here though, because some of these industrial-grade PSUs have a reputation for being extremely noisy at low power output. As a class A amp, the max. current consumption of the JLH should be twice its quiescent current (so app. 2.6A) and each PSU will deliver four times that before the limiter kicks in. Whether this is a real problem here or not I don’t know yet, but there are no audible artifacts at all so I am not overly concerned right now.

The heat sinks were also surplus items and I am not sure exactly what their rating is. At full speed the fans are a bit too noisy for my liking (the heat sinks add some flow noise as well) but if the fan speed is lowered a little I don’t think it would be really noticeable once the amps are seated in a rack. For the time being I have installed some small temperature speed control boards for the fans (hence the somewhat messy wiring) but depending on how well that works and how hot the amplifiers get, I may go back to fixed resistors. The PSU already has a dedicated 12V fan output so the resistor doesn’t need to drop a lot of power.

The heat sinks had a cutout in the side that was just too narrow for the angle brackets to fit into. The best solution would have been to mill the ends of the brackets to fit the cutout so there is only one contact surface. Unfortunately I don’t have access to a mill anymore so I had to find another solution: A copper “heat slug” to fill the gap. Just a piece of copper bar in the right thickness cut to size and with thermal grease on both sides and that should provide the best possible thermal transfer under the circumstances.

So, will I stop building these amps? Erm, no! 😀 I am out of the ebay-boards but I do however still have my own version of the 1969 JLH that isn’t cased up yet. This build has given me some inspiration for how I can build that into small monoblocks so while it might take a while to do at least I know what parts to keep my eye out for 🙂