Can the LD1117V33 handle short peaks of up to 900mA

designpower supplyvoltage-regulator

Got a project with the typical 5V and 3.3V power requirements. My initial prototype I used the LM3940 (datasheet). This worked pretty well but these regulators are damned expensive here in Australia ranging from $5.00 to $7.00 for one.

I have seen people use the LD1117V33 regulator for this as well. This one is considerably cheaper at +/- $2.00. This got me wondering a bit if I am really going down the right road with the design.

From my understanding the only differences are:

  1. LM3940 only takes 5v in and deliver 3.3v out. While the LD1117V33
    can take 15v and smaller and outputs 3.3v.

  2. LM3940 is rated at 1A while LD1117V33 is rated for 800mA.

  3. The LM3940 has limited packaging available (TO-220,TO-263,SOT-223) compared to the LD1117V33(DPAK,SO-8,SOT-223,TO-220).

My design is pretty modular so users can slot new modules/sensors into the board. Thus the current needed might vary from user to user depending on the configuration of the board/modules.

I feel that the LM3940 gives me extra current compared to the LD1117V33 but 20% more current at nearly 4 times the price. I doubt the board will ever get close to 1A of power but I am pretty sure they can peak at about 900mA when some of these sensors/modules gets power hungry.

I noticed that the datasheet for the LD1117V33 says that the output current can range from 800mA to 1300mA. If my device generally stays under 800mA usage with occasional spikes up to 1A will the LD1117V33 not be a much cheaper and viable alternative?

Just to clarify I am not asking for a recommendation on the parts.

My concerns are as follows:

Can the LD1117V33 handle short peaks of up to 900mA? I see it can output in that range on the official datasheet? However I would rather go for stability than shaving off some cash.

Best Answer

No it's a pretty crap regulator and cannot be counted on to source even 800mA. The datasheet says this.

enter image description here

800mA is only guaranteed with Vin-Vo = 5V and Tj = 25°C, and you have Vin-Vo = 1.7V (nominal) and hopefully you want it to work at other than 25°C. They do guarantee it will be no more than 1.3A and typically 950mA but again that's under conditions that are unrealistic.

As I said in the comment, choose a better regulator if you have a solid spec you are trying to meet. The one you mention is probably good for 500mA with proper engineering tolerances (I didn't bother to try to figure it out). AZ1117EH-3.3TRG1 (fixed 3.3V) made by BCD in Shanghai and now available through Diodes Inc. will deliver 1A minimum and is about 0.09 USD in 1K - 7 cents if you buy a reel. A reel of 4K pcs is the same cost as 40 $7 regulators.

enter image description here

Of course you may also be limited by the power dissipation but that's a whole 'nother issue.

I do disagree with the comment from my elegantly moustachioed friend about the wisdom of dealing with Chinese suppliers for small quantity parts through eBay and Aliexpress- at least for non-hobby situations. You're better off to deal with a franchised distributor that supplies clean current date-code parts that actually are what is marked on them, unlike the cesspool of counterfeits that you may find elsewhere. For example, you can easily find 'National Semiconductor' (now gone) marked regulators with a current date code (so obviously fake) and a part number that NS (or TI) has never made. Downside is you might have higher shipping costs or a minimum order. Digikey ships to Australia for free for a fairly modest (200 AUD) order.