I have a Rotel RB-1072 (ICE) amp. Soon after getting it, I noticed that any time it is on it emits a lot of RF. It destroys AM radios & my Shortwave radio with interference. I have a Sony HD tuner that is completely un-usable on AM. I know some of you are going to want to comment on the AM part of this. That is not the issue. Also, I have done enough investigation to know that it is the RB-1072 that is the problem.
I'd like to know if anyone else has this problem with their ICE amp or if you've been able to resolve it. - Thanx
I'd like to know if anyone else has this problem with their ICE amp or if you've been able to resolve it. - Thanx
If your amp is transmitting RF then it must be oscilating somewhere at RF.
I would be looking at the input filter to make sure it filters out RF.
I would also look at the capacitor on the VAS to ensure your not allowing RF through just AF.
The Zobel network should be in your design too.
In my experience of amps oscilating then tend to over heat and destroy themselves in time.
I would be looking at the input filter to make sure it filters out RF.
I would also look at the capacitor on the VAS to ensure your not allowing RF through just AF.
The Zobel network should be in your design too.
In my experience of amps oscilating then tend to over heat and destroy themselves in time.
Poor design and sign of times to come
One thing easy to try is get a surplus toroid and run your speaker cable about a 6-12 turns close to the amp. You may have to do the same thing with all the inputs as well incl power (3 cores) . I pretty sure this will relieve your problems (RF design guy). Try using a big powered iron or a smaller ferrite core.
I'm surprised this is a problem on a major brand. No safety ground is not only a bad sign for safety ( means no shields) as well as noise conduction and re radiated noise. I would take it back to the dealer and complain, let them fix it!
One thing easy to try is get a surplus toroid and run your speaker cable about a 6-12 turns close to the amp. You may have to do the same thing with all the inputs as well incl power (3 cores) . I pretty sure this will relieve your problems (RF design guy). Try using a big powered iron or a smaller ferrite core.
I'm surprised this is a problem on a major brand. No safety ground is not only a bad sign for safety ( means no shields) as well as noise conduction and re radiated noise. I would take it back to the dealer and complain, let them fix it!
nigelwright7557 said:
It's a Class D amp. It oscillates by design (at the PWM frequency) using a square-ish waveform that has lots of harmonics. Although, these days, most modern Class D amp designs vastly reduce RFI by spreading the noise spectrum out so there's relatively little power at any one frequency. I'm not sure, however, if Rotel is using the "spread spectrum" technique.
I second the ferrite idea and applying general RFI supression techniques. Commercial gear like the Rotel should be designed for and certified (tested) for FCC and other RFI compliance. But RFI testing is not exactly black and white. I've moved a cable a few inches and made a failing piece of equipment pass at the certification lab. And the Class B FCC level does allow some "possible interference" for home devices.
You might also contact Rotel and see if they have any suggestions? If it's as bad as you say, I suspect they've been down the road before with other customers. They may even send you some free ferrites, etc.
The Apostate brings up an interesting point about subwoofers. I've seen Class D amps used in subs that are not fully enclosed (shielded) internally. As the subwoofer itself is usually made out of MDF (wood) they can be quite nasty for RFI with essentially zero shielding. And some of those are not certified at all.
Follow up
Thanx for the replies.
I've tried some ferrite beads on the in&out points of the cable. They aren't big enough to affect anything unfortunately.
I have called Rotel & am waiting a call back.
The 1072 runs cool as a cucumber. No heat anywhere during normal operation & just a bit warm when it's really working hard. So I don't think it's gonna blow up due to this spurious RF noise.
I like the amp, but not sure if I will keep it unless I can fix this. Some EE friends I've talked with say all class d's will do this to some degree.
Thanx for the replies.
I've tried some ferrite beads on the in&out points of the cable. They aren't big enough to affect anything unfortunately.
I have called Rotel & am waiting a call back.
The 1072 runs cool as a cucumber. No heat anywhere during normal operation & just a bit warm when it's really working hard. So I don't think it's gonna blow up due to this spurious RF noise.
I like the amp, but not sure if I will keep it unless I can fix this. Some EE friends I've talked with say all class d's will do this to some degree.
Re: Follow up
I think nigelwright7557 missed that it's a Class D amp so you can safely ignore his comments about it overheating, etc. If it was a class A/B amp and causing radio interference then you'd have something to worry about.
And yes, Class D amps tend to create some level of RFI regardless. It's especially difficult to keep it out of the speaker cables.
One other thought if nothing else works is to try braided style, or possibly even shielded, speaker cables. And make sure you're using good well shielded line level interconnects to the amp and keep all cables as short as possible. Tightly coil up any excess cabling including the power cord and keep the excess cable close to the amp chassis if possible while as far from whatever is experiencing the interference. You want the cables to be as lousy of an antenna as possible.
Ograv said:
I think nigelwright7557 missed that it's a Class D amp so you can safely ignore his comments about it overheating, etc. If it was a class A/B amp and causing radio interference then you'd have something to worry about.
And yes, Class D amps tend to create some level of RFI regardless. It's especially difficult to keep it out of the speaker cables.
One other thought if nothing else works is to try braided style, or possibly even shielded, speaker cables. And make sure you're using good well shielded line level interconnects to the amp and keep all cables as short as possible. Tightly coil up any excess cabling including the power cord and keep the excess cable close to the amp chassis if possible while as far from whatever is experiencing the interference. You want the cables to be as lousy of an antenna as possible.
Class D can be designed for low EMI/RFI with not much more effort just some forthought and knowledge. But who knows if it will get better or worse since lowest cost is king in consumer audio these days. Less and less attention is paid to analog transmission reception in general as most of these are being migrated over to digital. Untill everything is digital expect more problems and cross compatibity issues with older technologies. Also beware if you going to mix use of class D seperates with analog preamps and SE analog connections too. We need to let manufactures know that spewing more RFI into the public spectrum is unacceptable and is a form of pollution more or less.
FIY ferrite beads are not equivalent to common mode toroid chokes that I described above.
FIY ferrite beads are not equivalent to common mode toroid chokes that I described above.
Re: Re: Re: Follow up
Good point. I was assuming if there was still a problem after trying to establish a quiet ground and following the other suggestions, then the braided cables might be worth a try (i.e. Kimber). It all depends on where the main leak is. If the Rotel engineers did a lousy job of filtering the speaker outputs and a signficiant amount of the PWM carrier is present at the outputs, then braided cable may well help. But that could be verified with a decent scope connected to the speaker outputs (be careful about grounding issues making the amp unhappy!).
It really depends on the source of the RF leak(s). That's why I suggested Rotel might be your best resource. If the RF can be tamed, they probably know the best way.
Otherwise, troubleshooting these problems is unfortunately a matter of trial and error--especially without a decent RF spectrum analyzer and an appropriate "sniffer" probe/antenna for it.
And as for digital radio and internet broadcast, for better or worse, it is slowly replacing analog RF radio for many applications. Now if we could just get the broadcasters to use higher bitrates...
infinia said:
Good point. I was assuming if there was still a problem after trying to establish a quiet ground and following the other suggestions, then the braided cables might be worth a try (i.e. Kimber). It all depends on where the main leak is. If the Rotel engineers did a lousy job of filtering the speaker outputs and a signficiant amount of the PWM carrier is present at the outputs, then braided cable may well help. But that could be verified with a decent scope connected to the speaker outputs (be careful about grounding issues making the amp unhappy!).
It really depends on the source of the RF leak(s). That's why I suggested Rotel might be your best resource. If the RF can be tamed, they probably know the best way.
Otherwise, troubleshooting these problems is unfortunately a matter of trial and error--especially without a decent RF spectrum analyzer and an appropriate "sniffer" probe/antenna for it.
And as for digital radio and internet broadcast, for better or worse, it is slowly replacing analog RF radio for many applications. Now if we could just get the broadcasters to use higher bitrates...
Denial is not a river in Eygpt
Somehow I don't think Rotel will even admit there's a problem. I was thinking of this issue today ie common mode hash needs a good reference to verify, so that's probably why they left off the earth on the AC mains plug. It'd probably pass certifications sitting on a bench by itself, but introduce longish speaker cables and a preamp with an earth connection then it's a system issue
Surprisingly a small portable radio is a good tool for these jobs. AM ferrite bar antenna's are pretty good. FM antenna terminal with a homemade sniffer works good too.
Somehow I don't think Rotel will even admit there's a problem. I was thinking of this issue today ie common mode hash needs a good reference to verify, so that's probably why they left off the earth on the AC mains plug. It'd probably pass certifications sitting on a bench by itself, but introduce longish speaker cables and a preamp with an earth connection then it's a system issue
Surprisingly a small portable radio is a good tool for these jobs. AM ferrite bar antenna's are pretty good. FM antenna terminal with a homemade sniffer works good too.
Thanx again for all the discussion.
I have called & emailed Rotel Tech Support & have yet to receive a reply. When I called on Monday, the tech guy still had his out of office reply for last week. Not a good sign.
I have really appreciated Rotel products for years (RT2100, RX503, RT990x, RSX1067, RA1062). But the lack of response from a premier audio distributor is disappointing.
FYI- my SW receiver (ICOM R75) is a good +40' from the Rotel 1072. My R75 sits atop my desktop computer. It doesn't get any interference from the computer! I have a little hand held portable SW receiver & used it as a "signal strength" meter. The RFI is most extreme right near the amp, but the speaker cables are also radiating quite a bit. The RFI is also volume dependent. More volume, more RFI. Hence, disconnecting the wires eliminates all RFI. I also tried a true earth ground to the chassis of the 1072 & after awhile it went into protection and, BTW, it didn't help.
The issue is one of inconvenience & economics.
I'll let you know if Rotel does or does not respond.
Bob
I have called & emailed Rotel Tech Support & have yet to receive a reply. When I called on Monday, the tech guy still had his out of office reply for last week. Not a good sign.
I have really appreciated Rotel products for years (RT2100, RX503, RT990x, RSX1067, RA1062). But the lack of response from a premier audio distributor is disappointing.
FYI- my SW receiver (ICOM R75) is a good +40' from the Rotel 1072. My R75 sits atop my desktop computer. It doesn't get any interference from the computer! I have a little hand held portable SW receiver & used it as a "signal strength" meter. The RFI is most extreme right near the amp, but the speaker cables are also radiating quite a bit. The RFI is also volume dependent. More volume, more RFI. Hence, disconnecting the wires eliminates all RFI. I also tried a true earth ground to the chassis of the 1072 & after awhile it went into protection and, BTW, it didn't help.
The issue is one of inconvenience & economics.
I'll let you know if Rotel does or does not respond.
Bob
Ograv said:
It sounds to me like the amp has no output filter at all, relying on the speaker itself to filter the PWM signal. Intheory this works fine - except for the EMI. If this is what is going on, the speaker wire itself becomes the radiating element. The more output current produced, the more EMI.
Without a filter, the output voltage will be a rail-to-rail high frequency square wave at 50% duty cycle with no signal. If you see this, then the amp had no business being sold. An iron powder or permalloy toroid choke (not common mode choke, but true inductor) and a polypropylene cap will fix this. And I can see manufacturers being cheap enough to omit them if they think they can.
jaycee said:
wg_ski said:
I'm sure the amp has some sort of output filtering. It just might not be enough. Rotel actually seems to care what their gear sounds like. So they may have taken something of a minimalist approach to the output filtering trying to avoid phase shift in the audio band, etc. The downside to such an approach is more RFI. I've been very impressed with Rotel's conventional Class A/B amps and I know the new Class D amps have received some good reviews.
If the amp went into protection with the chassis grounded it was likely the ground helped form a positive feedback loop with a source component connected to it. It probably isn't fair to blame the amp.
Amps oscillating when grounded in ways the designer never intended is an all-too-common problem. There are posts here on diyAudio, in fact, of amps going up in smoke by just connecting a (grounded) oscilloscope because they start oscillating well above the audio range. It's easy to get phase shift at high frequencies across the inductance of cables, chassis, etc. When you combine that with even small ground loop signals, you may get oscillation that can rapidly destroy amps--especially those with bipolar output transistors--due to output pairs being in simultaneous conduction. A Class D amp is likely somewhat more immune to being damaged by such things (they are, after all, designed to "oscillate" at similarly high frequencies) but Rotel is still wise to protect against such things.
Anyway, I don't think your amp is a "bad design". I'm guessing Rotel made some design trade offs related to optimizing the sound while keeping the price down. Perhaps having to choose between raising the price, lowering the sound quality, or compromising RFI performance, they may have chosen the last option.
Every Rotel product I have or had, has impressed me. The 1072 is the first device that has been a bit frustrating. I love the way it does it's job, tucked away on the bottom "slot" of my equipment shelf. Very efficient while providing effortless, detailed sound. It is extremely well constructed (wrap around cabinet, speaker connectors, electronics, etc). But I think they missed the boat on RFI suppression.
At this point, I am a little more disappointed in Rotel's non-response. I will continue to pursue their help.
At this point, I am a little more disappointed in Rotel's non-response. I will continue to pursue their help.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.