Compact AC/DC power supply for IMP-002

I am testing the following circuit…

Is it necessary to put the recommended 100pF capacitors near the VDD pins 18,33,34 or I can avoid them in this configuration ?

Any other consideration ?

Thank You!

Do you just have the Imp002 all by itself, or is your imp mounted on a board with the blink-up circuits, authentication chip, power supply, and other things?

If it’s on a board, you can use a 5VDC supply if you want. The April board up to 16VDC. But you don’t have an April board with Imp002. Find out what the power supply regulator is for your Imp002 board.

I’m certainly not an authority on bypassing, but I’ll typically follow the generally accepted ‘rule of thumb’ to include some sort of high-frequency bypassing directly adjacent high-speed devices, especially for double-sided PCBs. Being right next to the device helps mitigate noise due to power trace inductance (from a typically more distant bulk bypass capacitor), and provides a good low impedance ground path for high-frequency transients.

As there is typically very little space/cost associated with small, local bypass capacitors, I’ve always used them as a ‘best practices’ measure. (I use the same RAC device you indicate, and maintain the recommended local bypassing for the imp-002).

Perhaps a more interesting question related to your circuit might however be ‘what is the minimal value bulk bypass capacitor that would work reliably’? Here I think there may perhaps be a more significant cost and size consideration. Don’t have an answer to this at the moment, but it is on my list of things to look into.


Also just noted that you’ve indicated a RAC03 series (i.e. 900ma) device. Not sure what you’ve got going, but that’s quite a bit of add’l capacity beyond the imp-002 250ma max spec. Dropping down to the RAC02 (600ma) series would save you a couple dollars, and a bit of board space, and still leave with you with quite a bit of excess capacity. Just a thought.

I am using the RAC02 and I have very minimal load beyond the imp. I have also tried another brand that is sort of a drop-in replacement for the RECOM module.

I don’t remember which unit it was; I had a problem in that I had to add extra load to get the output voltage down. There is some small print that says the specifications are only valid for 10% minimum load.

I’m guessing you’re referring to the CUI VSK series? Interestingly, the 'fine print" you mention is also there for the RECOM devices, but perhaps they’re just more well behaved under minimal load conditions.

You should have something with lower ESR close to the imp002, see the recommendations. A 330uF electrolytic is not very low ESR.

The 100pF caps are more for RF compliance tests than functionality, so we recommend them on a commercial product but if this is just a one-off for personal use you’ll be fine without.

RAC02 is not so smaller than RAC03 and I prefer to have less heat generated.
Have I to submit this circuit to any kind of RF test if I sell it ? (in Europe)
The IMP-002 RF certification is not sufficient ? Sorry for the question but it’s the first experience for me :slight_smile:

In either country you need to do approvals on any device that is shipped. In the US/Canada you just do unintentional emitter, and refer to our modular approval (and label your device appropriately).

In the EU it’s self-certification, but generally you need to follow the same process - ie you need to test the rest of your product confirms to EU rules too, which include susceptibility and ESD in Europe. imp002 has had european emitter testing performed which you can reference, but some labs will advise that you need to fully re-test.

I finally tested the RAC03-3.3SC with the IMP-002 in powersave mode --> imp.setpowersave(true);
In addition the IMP controls a bistable relay (coil 100mA @ 3.3v) thru 2n2222A, with an impulse of 0.4 sec (one every 5 minutes)
I see the RAC03 is really oversized and I am thinking to downgrade to RAC02 or RAC01 (thanks to LarryJ)

Do you think the RAC01 (300mA) can also work for me or I need RAC02 (600mA)?

I other discussions I also read that a couple of AA alkaline batteries can live for about 50.000 wakeups. Is it correct ?


300mA should be fine, as long as you don’t have much of any other current draw in the system. When the imp002 wakes, it draws about 250mA for a few microseconds while it turns everything on. If you don’t have at least that much it won’t connect. If the imp is the only thing drawing current, 300mA should be ok, but I imagine you will want to power other components as well, correct? The safe bet might be the RAC02.

Of course, having large enough capacitors in the circuit might take care of that easily. The imp uses much less current under normal conditions.

Circling back around on this thread for a moment for anyone who might have an interest in AC/DC power modules - I had some time today to get a RECOM RAC01-3.3 on a load tester to do some basic characterization … my primary interest being output current capability. (All measurements were made using nominal US line voltage - 120VAC - at 25 deg C). As a comment was also made earlier about heat, I mounted a thermocouple to the case to also have a look at that. (Case temperature was allowed about 5 minutes to stabilize between readings).

Attached is a chart of both Vout and case temperature vs. load. A few observations -

  • an approximate 10ma minimum load was needed to get within the spec’d
    +/- 5% Vout tolerance
  • Iout capability was significantly above the 300ma specification
  • about a 5 degree C rise at 300ma out … an add’l 5 degrees at 500ma out

I’m fairly conservative, so would likely stick with the higher spec’d RAC02 for any production needs above 300ma, but the RAC01 module would appear to have more than enough horsepower to handle many projects. (Given the identical size/footprint of the RAC01 and 02, kinda makes you wonder if there might just be a production selection process at work here).

  • Larry

many thanks! for sharing this. I have been using the RAC01 (and have also used RAC02) in my projects though I have only built a handful of them. I didn’t realize the minimum load requirement until just in the last few weeks. I don’t put the imp into sleep mode but if I did I guess I should expect the supply voltage to be too high.

I have just added a load resistor to my design so that I can be assured the supply will be in spec.

Any further thoughts about bulk and decoupling capacitors to use in conjunction with this supply? I know it is a tough topic as each design is different. I don’t have good enough tools (or skills?) to properly check the output for noise and ripple and such in this context.

Excellent info. Thanks @LarryJ.

You’re welcome @mjkuwp94 and @jwehr … glad to hear it was useful.

Re: minimum loading - yea, this is something I hadn’t paid too careful attention to either … perhaps I’ve just been lucky as I really don’t do a lot of power minimization given the AC source. But it probably is a reasonable idea to make sure it’s loaded somewhere near the 10% spec.

Re: decoupling - I haven’t really worried about this either, so perhaps I’ve been lucky here too. That said, I always include a bulk capacitor not too far from the imp power pins per previous recommendations. (I’ve used 470uF electrolytics … and have found a couple with fairly low ESR … but have noted @Hugo’s comments, so probably should review this). And I’ll include a few smaller bypass caps (typically 0.1uF) adjacent the imp power pins. I think the issue here is perhaps less about ripple and noise, and more about how to handle the high transient current demands of the imp in WiFi/TX modes. So far, so good for me … I haven’t really seen any problems that I believe are power related.

(I haven’t been concerned with EMC compliance to date, so have just typically omitted the 100pF bypass caps at the power pins … see @Hugo’s earlier comment).

  • Larry

Hi all,
Recom released a new version of AC/DC (RAC) with all protections inside (fuse,varistor). This can save space on PCB.

Someone already tried RAC02-3.3SC/277 ? 4$ more than SC

I think this new one requires a minimum load of about 2%.