I am planning on making a simple electronic device which I might eventually try to market if it works out O.K. Before taking the plunge and trying to mass produce it, I would first try to sell a few on the internet to see if anyone even wants it.

It turns out, however, that to sell anything in the United States you need to pass FCC testing, or meet the criteria that makes it exempt. From what I hear FCC testing costs upward of $10,000, which I am not willing to put up right now. After much searching online for the official FCC documents (which were near impossible to find) it seems like one of the conditions is that that makes you exempt from FCC testing is if there is no oscillator or frequency in your circuit which is over 1.705 MHz (please let me know if this is wrong, I asked on here before finding the documents and everyone said it was 9 kHz and closed the thread).

Here is the link to the regulation:

Title 47: Telecommunication PART 15—RADIO FREQUENCY DEVICES Subpart B—Unintentional Radiators § 15.103 Exempted devices.

The following devices are subject only to the general conditions of operation in §§15.5 and 15.29 and are exempt from the specific technical standards and other requirements contained in this part. The operator of the exempted device shall be required to stop operating the device upon a finding by the Commission or its representative that the device is causing harmful interference. Operation shall not resume until the condition causing the harmful interference has been corrected. Although not mandatory, it is strongly recommended that the manufacturer of an exempted device endeavor to have the device meet the specific technical standards in this part.


(h) Digital devices in which both the highest frequency generated and the highest frequency used are less than 1.705 MHz and which do not operate from the AC power lines or contain provisions for operation while connected to the AC power lines. Digital devices that include, or make provision for the use of, battery eliminators, AC adaptors or battery chargers which permit operation while charging or that connect to the AC power lines indirectly, obtaining their power through another device which is connected to the AC power lines, do not fall under this exemption.

Does anyone know of a micro that has a clock speed and all oscillators lower than 1.705 MHz? I found some micros that have clock speeds of 1 MHz, but oscillators are 4 MHz. A clock speed over 500 kHz could probably work, but 1 MHz would be best!

  • 1
    This page ecfr.gpoaccess.gov/cgi/t/text/… says "(u) Radio frequency (RF) energy. Electromagnetic energy at any frequency in the radio spectrum between 9 kHz and 3,000,000 MHz." and has many other places of mentioning 9 KHz. But it does look like your page is pretty clear that you don't have to worry until 1.705 MHz – Kellenjb Jun 3 '11 at 3:34
  • And as far as your comment about your prior question being closed, Kevin Vermeer gave you 2 days to explain how or if your question was different from the others linked to, and since there was no response it was closed. – Kellenjb Jun 3 '11 at 3:51
  • That provision doesn't say anything about the oscillator, it says "Digital devices in which both the highest frequency generated and the highest frequency used are less than 1.705 MHz". Even a 100khz square wave generates and uses frequency content well above 1.705 MHz. – Mark Jun 3 '11 at 3:56
  • @Kellenjb I'm certainly not an expert on this but the clause pertaining to the 9 khz lower limit is much more specific, the wording is "An unintentional radiator (device or system) that generates and uses timing signals or pulses at a rate in excess of 9,000 pulses (cycles) per second and uses digital techniques;" – Mark Jun 3 '11 at 4:07
  • 1
    Thanks guys, I must have misunderstood the statement, since any type of square wave is going to have an infinite number of harmonics, so I don't understand why they even put that there then? – w1res Jun 3 '11 at 21:13

Many microcontrollers are fully static, i.e. that the clock can be completely stopped, or that you can run it at frequencies like 0.1 Hz, for instance, for 1 instruction per 10 seconds (might be useful for debugging). Some components on the die may require a minimum clock frequency to work, however, like an ADC: the sampling capacitor will discharge if you don't complete a conversion within a certain time.

That being said, the controller's clock will cause EMI over a band much wider than just the clock's frequency. The shorter the rise/fall of a digital signal the more energy there will be in the harmonics. To decrease EMI some microcontrollers, like Freescale's MC9S08, have (switchable) slew-rate controlled I/Os.

Well, where do I start listing...

Many microcontrollers have low clock modes. For example Atmel's AVR line can be used with 32.768 kHz crystals. From what I've read, you can use them with even lower frequencies, like a 555 timer running at couple of kHz.

Another microcontroller I've used is Parallax Propeller which has internal 32 kHz clock source, but it isn't as accurate as a crystal.

I've read that PICs too can work with low frequency clock sources, but I have no experience with them.

  • PICs are much like you describe Atmel's. – kenny Jun 3 '11 at 12:21
  • indeed, AVRs can go very low :) – alkar Jun 15 '11 at 1:20

The document you linked to says:

"Digital devices in which both the highest frequency generated and the highest frequency used are less than 1.705 MHz."

The problem that you'll have with this is the definition of "highest frequency generated". Just about any digital circuit will generate frequencies higher than 1.7 MHz, even if the highest clock frequency is only 9 KHz. The reason for this is the edge rates (a.k.a. Slew Rate) on the digital signals have lots of harmonics on them.

Here's a cool web page that shows how a square wave is just the sum of a bunch of sine waves.

The best way to do what you propose is to simply design and build the circuit correctly, paying close attention to EMI and RF, so that when you do go into FCC/CE testing you will pass on your first try. Sure, you have to pay the money. But you can minimize the money by making your device as easy to test as possible, and eliminate any re-testing required.

  • Thanks for the answer. I actually understand signal processing quite well, however I must have misunderstood the clause in the document. I don't understand why they would word it that way however since any type of square wave is going to have infinite harmonics (like a sinc function). So I guess there really is no way to get around to this other than to do the testing. – w1res Jun 3 '11 at 21:11
  • 2
    I'm curious what the FCC means by "highest frequency generated", given that the only way one could determine the highest frequency at which harmonic content is significant would be to test it, and the purpose of the spec seems to be to indicate which devices are exempt from testing. The 9KHz limit seems to say that it doesn't matter too much what harmonics the device generates if they're all confined to brief events separated by at least 111us; I wonder what the effect would be of the FCC pushing that rule up to 35Hz? – supercat Jun 4 '11 at 21:38
  • 1
    Whether or not anyone here agrees with it, the FCC (like other regulatory bodies) has chosen exemptions based on system frequencies (those directly used or generated, not indirectly generated from harmonics and such - which can't be known without testing in the first place and so wouldn't make sense to be used as an exemption). So if you are running digital clock speeds under 9kHz, no radiated emissions testing needed (same goes for Europe). The FCC definition of a "digital device" includes this frequency requirement, so there is no question that the exemption excludes harmonic effects. – bt2 Jul 23 '16 at 4:33

I think maybe you're missing the point of the FCC regulation. What they are primarily concerned with is spurious emissions being kept to below some power level within some spectrum. It has little to do with what speed your processor is running at. It has everything to do with what radiation is emitted from your device in aggregate - i.e. all your radio front-end in conjunction with all the switching noise of your computing elements.

The way this stuff is typically dealt with is with properly matched RF circuits, EMI filters (e.g. on power supplies), properly dsigned/employed external interface connectors, and emission reducing enclosures; not by using slower microprocessors...

You can run a huge number of microprocessors, microcontrollers or DSPs from clock sources in the 1MHz range. I remember when we got the high speed 6502 processor that doubled the clock speed to 2MHz. Most processors that have low power states that allow the clock source to be disabled without losing their execution state will allow the processor clock to be set to any speed between DC and the processor maximum.

The MSP430 processor has an internal DCO clock source that can be tuned to a frequency of your choice and calibrated against a 32KHz clock source. Getting this to run at 1MHz is quite straightforward.

Many processors have an internal PLL that multiplies the crystal frequency from the external MHz range up to a much higher frequency. You would obviously have to disable this sort of feature.

The only processors that you would have trouble with are those that have a minimum clock frequency for the external oscillator. These can lose their internal state (muck like unrefreshed DRAM) if the processor clock is too slow.

Your Answer


By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

Not the answer you're looking for? Browse other questions tagged or ask your own question.