Since I didn’t use the capacitor in the first place and still didn’t experience any RF interference, there could be two possible reasons:
- There’s little or no RF in my area, so the capacitor simply wasn’t necessary. If the amp were placed near a strong RF source, like a radio tower, it might start picking up interference.
- There is RF present, but the designed circuits and PCB layout are good enough to make the amp immune to RF interference.😉
Could also be that this "RF issue" is something one person concluded based on insufficient evidence and reported as the truth from there on. Confirmation bias then led them to strengthen their belief, after all, all capacitors and inductors near an LM3886 must indicate trouble with RF, right?! I asked them to provide evidence for their claims and they responded with "prove me wrong". So there's that...
Between my customers and I I'm sure we're deep into the thousands of LM3886es used at this point. I know I've gone through a few thousand of them through the past 11 years. Yet, it is in this thread that I encounter this "RF issue" for the first time.
Of course it is possible that there is an issue that I missed and that my customers never noticed. After all, I do use EMI/RFI filters on all my amps. But I doubt it. There are countless LM3886 circuits out there and nobody is clamouring about RF issues that I can tell. If you come across a thread where people are having issues with RFI in an LM3886 that can't be attributed to poor PCB layout or poor wiring please let me know.
Tom
Between my customers and I I'm sure we're deep into the thousands of LM3886es used at this point. I know I've gone through a few thousand of them through the past 11 years. Yet, it is in this thread that I encounter this "RF issue" for the first time.
Of course it is possible that there is an issue that I missed and that my customers never noticed. After all, I do use EMI/RFI filters on all my amps. But I doubt it. There are countless LM3886 circuits out there and nobody is clamouring about RF issues that I can tell. If you come across a thread where people are having issues with RFI in an LM3886 that can't be attributed to poor PCB layout or poor wiring please let me know.
Tom
Don’t have anything to do with commercial LM3886 designs but as indicated I have experience with the “RF issue” in devices and SMPS and like to solve such issues.
It is not a one person “RF issue”. It depends on local factors and varies over the day. It became worse when 5G was introduced. Quite a few opamps are sensitive to it too. If it does not happen to you and your customers that is fine. It does not exclude the issue elsewhere on the globe and the local situations there. Don’t downplay anything you have not witnessed. It is not a complaint against you and your products.
It is not a one person “RF issue”. It depends on local factors and varies over the day. It became worse when 5G was introduced. Quite a few opamps are sensitive to it too. If it does not happen to you and your customers that is fine. It does not exclude the issue elsewhere on the globe and the local situations there. Don’t downplay anything you have not witnessed. It is not a complaint against you and your products.
Last edited:
I'm not downplaying anything and I'm not perceiving what's posted here as an attack on me or my products. I made that clear in Post #33.
All I'm asking for is evidence of the issue before an issue is claimed to exist. So far no evidence has been provided.
Sure.
Tom
There is absolutely no need for the OP to show "evidence" to you of a phenomenon he experiences with his DIY design. The phenomenon is described, a schematic has been posted and details are given and he asks for help. That is way more information than usual. Michael Chua experienced similar stuff. It could be a PCB layout imperfection combined with local circumstances. It could even be the LM3886 depending on batch/date of production, who knows at this stage? Why doubt the OPs experiences? Are you and your stuff the measure of things?
Sure:
https://laas.hal.science/hal-02319469v1/document
https://www.st.com/resource/en/appl...etic-interferences-emi-stmicroelectronics.pdf
https://www.ti.com/lit/an/sboa128a/sboa128a.pdf?ts=1748376997969
https://www.renesas.com/en/document...tMCChVNiUrHVgKhUAJRWSq7FSXHAnxgmj1hyUqBC-tfqn
Sure:
https://laas.hal.science/hal-02319469v1/document
https://www.st.com/resource/en/appl...etic-interferences-emi-stmicroelectronics.pdf
https://www.ti.com/lit/an/sboa128a/sboa128a.pdf?ts=1748376997969
https://www.renesas.com/en/document...tMCChVNiUrHVgKhUAJRWSq7FSXHAnxgmj1hyUqBC-tfqn
Last edited:
Jean-paul, I think it would behove you to read the entire thread and view my comments in their proper context. I've never suggested that OP should provide evidence for anything. I asked Michael Chua to provide evidence for his claims that the LM3886 is particularly susceptible to RFI and he provided nothing substantive to back his claims.
I asked because if the LM3886 is particularly susceptible to RFI I would like to know about it so I can deal with it. But without anything of substance I have nothing to go by.
I'm not suggesting that opamps cannot be susceptible to RFI. All I'm saying is that in my experience I haven't found the LM3886 any more susceptible to RFI than other power opamps. I, therefore, don't see any need to churn the rumour mill.
You can turn just about any opamp into an RF receiver (or oscillator) with a crappy PCB layout. Is doing so the fault of the opamp or the fault of the PCB layout engineer?
Tom
I asked because if the LM3886 is particularly susceptible to RFI I would like to know about it so I can deal with it. But without anything of substance I have nothing to go by.
I'm not suggesting that opamps cannot be susceptible to RFI. All I'm saying is that in my experience I haven't found the LM3886 any more susceptible to RFI than other power opamps. I, therefore, don't see any need to churn the rumour mill.
You can turn just about any opamp into an RF receiver (or oscillator) with a crappy PCB layout. Is doing so the fault of the opamp or the fault of the PCB layout engineer?
Tom
The RF signal can be sneaking into your circuit through any long trace that is a part of your audio input circuit. Wires or circuit board traces act as antennas for RF. If you place a resistor like 1Kohm in series with a wire then you can kill the antenna effect significantly. You could try adding a series resistor in the longest wire or circuit trace in the input section of your circuit to see if it has an effect. If your layout was perfectly symmetric with respect to pin 9 and pin 10 then the input circuit would be immune to RF because the RF would be common in amplitude to both inputs and therefore would not get amplified. This cancelling effect would likely be lost if the RF signal is larger than the common mode range of the amplifier inputs.
After three pages, I think it's time to verify it is actually the LM3886 who's to blame. Get any random other analog amplifier and hook it up in the same way at the same location as a test. Preferably try a few different ones.
When some radio amateur next door is broadcasting high power with poor quality uncertified equipment, any amplifier will likely pick it up.
When some radio amateur next door is broadcasting high power with poor quality uncertified equipment, any amplifier will likely pick it up.
Not the question of the OP. Please see post #1.
He wants his amplifier (that happens to have LM3886) to be quiet not other amplifiers.
He wants his amplifier (that happens to have LM3886) to be quiet not other amplifiers.
I know. But some levels of emi just can not be shielded or filtered. And if that is the case here, adressing the root cause might be a better approach.
That could certainly be amateur radio or CB radio. If you could find out which neighbour transmits the signal, maybe you could ask what frequency bands they are using. Then at least you know what frequencies need better suppression. Maybe you could also ask for a test transmission when you want to test an immunization attempt.
Actually... If a licensed HAM radio operator is interfering with your electronics you can require that they remedy the situation at their expense. 'Tis the law. This probably only works if you're using commercially available equipment. I doubt you'd have a leg to stand on if you're having issues with your DIY build, though the HAM operator may be interested in helping you anyway.
CB on the other hand relies on certification of the equipment rather than of the user. CB equipment should not be interfering with commercially available equipment. "Should"...
Either way, if you pump enough RF into an audio circuit you're bound to get some odd behaviour. I add EMI/RFI filters on the inputs of my amps because that keeps switch arcs, cellphones, and such out of the audio. Without the filter, you'll often hear a pop in the speakers when a heavy load such as a furnace or fridge kicks in. It gets annoying after a while. Relays are good arc transmitters it turns out. 🙂 But in that case the RF gets in through the input cabling. Not the fault of the amp.
Add an LC low-pass filter with Q ≤ 1 and a cutoff frequency around 1-2 MHz on the input.
Tom
I'm not familiar with Canadian law. As far as I know, over here, radio amateurs just need to keep the field strength at their neighbour's property below some specific value. If the neighbour's equipment then still suffers from interference, that's the neighbour's problem. Most radio amateurs would be happy to help solve the issue anyway, though.
Dears, Im reading this thread with interest. Finally I have to comment : go balanced input and get rid of all interference by 80db or more CMRejection, eliminate it for good. No more worries about neighbors with CB station or bad air-conditioning systems, or data servers.
Just actual is this thread where OP is handling this with LM3886-s, there are more options of course, but input to amp must be balanced.
https://www.diyaudio.com/community/threads/a-new-ec-composite-lm3886-amp.419558/post-7836205
Just actual is this thread where OP is handling this with LM3886-s, there are more options of course, but input to amp must be balanced.
https://www.diyaudio.com/community/threads/a-new-ec-composite-lm3886-amp.419558/post-7836205
Really?
I was a licensed HAM operator in Denmark many moons ago (1989, I think). The law there was that the HAM operator was to remedy any interference caused at their own expense. But, of course, that could have changed many times since then.
With CE and FCC testing these days there should be a lower probability of EMC trouble ... assuming the CE/FCC marked equipment actually meets the standards.
I'm definitely in favour of using a balanced input, but mostly because it moves the ground connection out of the signal path. I doubt a typical balanced audio input will have appreciable CMRR at RF. I suppose it could if you used something like an LM7171 as the receiver, but then you get different issues. 🙂
Tom
Hi, Correct, CMR drops with frequency. But I think that RF is not such issue as we don't hear it.
It is its modulation that reaches audio range (that I heard many times, AM radio for example, also big inverter's noise(A/C) or whatever around, mainly just mains power frequency) . This modulation that enters audio range is nevertheless well rejected by balanced input. CMR should be effective up to what we hear, IMO.
Hmm... Let me try to unpack that.
You are correct that we don't hear RF. But RF coupled into an audio circuit can result in DC offset shifts. That's what caused the BRR-BRR-BRRRTT noise when a 2G/3G GSM cell phone received a call or text message and was placed close to, say, a transistor radio. Specifically, it was the 200 Hz frame rate of GSM that caused those noises.
National Semiconductor (now TI) has a good white paper on the topic: https://www.ti.com/lit/wp/snoa817/snoa817.pdf. It's part of the LMV851 collateral.
The RF can also be demodulated. You can demodulate AM with a diode and a lowpass filter and there are plenty of those in a typical audio circuit. I find it doubtful that an audio circuit would be able to demodulate FM or PM.
Tom
On an evening just over two decades ago, I was in the studio of the local FM radio station here when a lady who lived right across the street called the studio because she heard our audio coming out of her active computer loudspeaker. She hadn't heard that before, so she wondered what had changed and what could be done about it. As it so happened, we were temporarily transmitting straight from the studio building because of a malfunction at the normal transmitter site, so the field in her house must have been stronger than normal. When I explained that it was very temporary, she was OK with it.
AM is certainly more likely to cause problems than FM, but there's always slope detection and resonating transmission lines (and many other things) that can act as filters.
AM is certainly more likely to cause problems than FM, but there's always slope detection and resonating transmission lines (and many other things) that can act as filters.