I have a Rotel RSX-1056 surround sound receiver on my bench because the owner is hearing an annoying hiss at even low volume levels. Apparently, this is a fairly common issue that affected the RSX-1056 and some of its siblings from the early 2000s. I've yet to find anything online in the way of an explanation for the hiss, let alone a fix. Glutton for punishment that I am, I'd like to see if this is solvable.
Here's what I've found so far:
• The hiss comes from both channels as soon as the volume is turned up a few dbs from minimum.
• It's audible from a few feet away.
• It's present with or without a line level source connected to the inputs. All inputs are affected.
• Using the Bypass mode has no affect on the level or character of the hiss.
• The hiss is present when music is playing.
• The hiss persists when using the preamp outs of the RSX-1056 to connect it to another amp. This seems to clear the amp board.
• Other owners who have complained of this problem say it does not occur when using the digital inputs. I haven't verified this.
• Just to cross my t's and dot my i 's, I checked the voltages on the main power supply board, and they are spot on with one exception: Instead of 30V going to the front board (that's what Rotel calls it), I'm getting 27V. Not that significant I suppose.
• The DC offset is 40mV left and 4mV right, fwiw.
• The NJM7805FA voltage regulators at IC903 and IC904 seem to be running very hot, though I don't know what's "normal for these.
• It appears that the 7805 at IC903 had been installed and removed at the factory. There's a smear of thermal grease on the heat sink.
• Speaking of the factory, they sure did like the capacitor glue. So far I only see one spot with signs of corrosion.
That's all I got for now. All suggestions welcome.
Here's what I've found so far:
• The hiss comes from both channels as soon as the volume is turned up a few dbs from minimum.
• It's audible from a few feet away.
• It's present with or without a line level source connected to the inputs. All inputs are affected.
• Using the Bypass mode has no affect on the level or character of the hiss.
Analog Bypass: For 2-channel analog in-
puts, there is a special stereo mode that by-
passes ALL of the RSX-1056’s digital process-
ing. The two front speakers receive pure analog
stereo full-range signals with no subwoofer
crossover, no delay, no level adjustments, and
no contour adjustment.
puts, there is a special stereo mode that by-
passes ALL of the RSX-1056’s digital process-
ing. The two front speakers receive pure analog
stereo full-range signals with no subwoofer
crossover, no delay, no level adjustments, and
no contour adjustment.
• The hiss is present when music is playing.
• The hiss persists when using the preamp outs of the RSX-1056 to connect it to another amp. This seems to clear the amp board.
• Other owners who have complained of this problem say it does not occur when using the digital inputs. I haven't verified this.
• Just to cross my t's and dot my i 's, I checked the voltages on the main power supply board, and they are spot on with one exception: Instead of 30V going to the front board (that's what Rotel calls it), I'm getting 27V. Not that significant I suppose.
• The DC offset is 40mV left and 4mV right, fwiw.
• The NJM7805FA voltage regulators at IC903 and IC904 seem to be running very hot, though I don't know what's "normal for these.
• It appears that the 7805 at IC903 had been installed and removed at the factory. There's a smear of thermal grease on the heat sink.
• Speaking of the factory, they sure did like the capacitor glue. So far I only see one spot with signs of corrosion.
That's all I got for now. All suggestions welcome.
Attachments
Check the digital inputs and the multichannel inputs. My Sony STR-DA5ES has the identical issue with the stereo analog inputs but no issue with the multichannel or digital inputs (well, just the standard AVR noise floor, roughly the same hiss as my Denon X2100W). But utterly unusable with the stereo inputs. I haven't troubleshot past trying different inputs, and looking at the service manual now I see it looks like there's a CD/SACD direct mode I haven't tested either. But based on looking at the Sony and Rotel schematics, I think it's the Analog Input Selector IC. The Sony uses what looks like a slightly newer version of the chip in the Rotel (Sony uses Toshiba TC9274N, Rotel uses Toshiba TC9163AF).
A peculiar problem, this.
Where exactly does the hiss set in? Does it remain constant thereafter or does it steadily rise with volume?
In bypass mode in particular, there's not a great deal in the signal path. 1x TC9163 input selector, 1x unity gain buffer, another TC9163 as a multich in / stereo / bypass switch, a TC9482F digital pot (20 kOhms for 0..-7 dB, followed by 50 kOhms in 8 dB steps for 0..-88 dB, specified output noise 1.2 µV but obviously that's going to vary by setting like any pot), and a 14 dB amplifier (later to be followed by 29 dB of power amp gain).
So far, nothing that would suggest wildly excessive output noise.
OK, here's the interesting section.
Looks like every IC and opamp gets an RC filtered +/-18 V with 220 ohms each. That should take a lot of burden off the ground return but this will have to be inspected later.
Hmm. The Gv = 5.5 (+14.8 dB) amplifier with an NJM5532 features 10k and 2k2 gain setting resistors. That's rather higher than necessary for a 5532 at moderate levels, even a 4558 would be happy like that. I might suggest closer to 3k3 and 750R instead, or even 1k8 and 390R... 5532s make half-decent headphone amps, driving 2k2 at ~1Vrms tops isn't going to faze them any time soon, and their voltage noise is (just about) low enough to benefit from these lower values. (That said, an NJM2068 with 3k3/750R would probably still beat both versions.)
Now - when using the pre-out into another power amplifier, does the noise go away at low volume? There is muting between the Gv = 5.5 stage and the power amps which may be involved down there, but which does not apply to the pre-outs. If the noise still goes away there, further testing (shorting signal to ground at strategic points) may be required to pinpoint the problem.
Bandwagon jumpers with other Rotel gear, please be advised that this brand had a history of using rather noisy FET input opamps (like AD712) in the volume gain stage. These can usually be upgraded to lower-noise parts, even bipolar, and feedback network resistors reduced as outlined above. However, the corresponding schematic should be consulted to evaluate your options.
Where exactly does the hiss set in? Does it remain constant thereafter or does it steadily rise with volume?
In bypass mode in particular, there's not a great deal in the signal path. 1x TC9163 input selector, 1x unity gain buffer, another TC9163 as a multich in / stereo / bypass switch, a TC9482F digital pot (20 kOhms for 0..-7 dB, followed by 50 kOhms in 8 dB steps for 0..-88 dB, specified output noise 1.2 µV but obviously that's going to vary by setting like any pot), and a 14 dB amplifier (later to be followed by 29 dB of power amp gain).
So far, nothing that would suggest wildly excessive output noise.
OK, here's the interesting section.
Looks like every IC and opamp gets an RC filtered +/-18 V with 220 ohms each. That should take a lot of burden off the ground return but this will have to be inspected later.
Hmm. The Gv = 5.5 (+14.8 dB) amplifier with an NJM5532 features 10k and 2k2 gain setting resistors. That's rather higher than necessary for a 5532 at moderate levels, even a 4558 would be happy like that. I might suggest closer to 3k3 and 750R instead, or even 1k8 and 390R... 5532s make half-decent headphone amps, driving 2k2 at ~1Vrms tops isn't going to faze them any time soon, and their voltage noise is (just about) low enough to benefit from these lower values. (That said, an NJM2068 with 3k3/750R would probably still beat both versions.)
Now - when using the pre-out into another power amplifier, does the noise go away at low volume? There is muting between the Gv = 5.5 stage and the power amps which may be involved down there, but which does not apply to the pre-outs. If the noise still goes away there, further testing (shorting signal to ground at strategic points) may be required to pinpoint the problem.
Bandwagon jumpers with other Rotel gear, please be advised that this brand had a history of using rather noisy FET input opamps (like AD712) in the volume gain stage. These can usually be upgraded to lower-noise parts, even bipolar, and feedback network resistors reduced as outlined above. However, the corresponding schematic should be consulted to evaluate your options.
Attachments
Last edited:
Regarding the Rotel line of components of that era......
It is my findings that that black adhesive used to anchor components on the PC boards becomes "electrically conductive' over time, and causes multiple issues.
Spill-over onto certain components such as resistor leads, IC chip legs, etc., causes trouble that is difficult to locate.
This results in resistance paths along the glue, causing resistive "shorts", and incorrect voltage levels, or even complete malfunctioning of the unit.
(see D923 in the photo from the OP)
Use a nice sharp set of DVM test leads set to read meg-ohms, and note if the DVM shows resistance readings when the probes are stuck closely together (a fraction of an inch) into that black glue.
Only a high-sensitivity type DVM will show such a reading.
If so...... it must be removed, carefully scratched off and cleaned.
Trust me, I've serviced those type of Rotels, and other brands over the years, and sometimes that nasty glue is a pain in the rear.
It is my findings that that black adhesive used to anchor components on the PC boards becomes "electrically conductive' over time, and causes multiple issues.
Spill-over onto certain components such as resistor leads, IC chip legs, etc., causes trouble that is difficult to locate.
This results in resistance paths along the glue, causing resistive "shorts", and incorrect voltage levels, or even complete malfunctioning of the unit.
(see D923 in the photo from the OP)
Use a nice sharp set of DVM test leads set to read meg-ohms, and note if the DVM shows resistance readings when the probes are stuck closely together (a fraction of an inch) into that black glue.
Only a high-sensitivity type DVM will show such a reading.
If so...... it must be removed, carefully scratched off and cleaned.
Trust me, I've serviced those type of Rotels, and other brands over the years, and sometimes that nasty glue is a pain in the rear.
Last edited:
sgrossklass: Thank you for the analysis. You asked when the hiss sets in. As soon as the volume is turned past "Min" on the display, it's audible. I assume there's a muting taking place at Min because the hiss immediately disappears when the volume is lowered to this point and reappears when it's set to "2," the next highest setting. Now, when the volume is increased further, I can't hear any change in the hiss until the volume is nearly maxed out, at which point the hiss does sound a little louder and is accompanied by a soft, high-pitched tone. But I would expect to hear some hiss at full max volume from any amp.
Everything I just wrote applies to when using the pre-outs into the line inputs of a receiver I use at my bench: The hiss is only present when the volume is set to "2" or higher and does not increase with volume.
What would you suggest next? Try different values for the 5532 gain-setting resistors?
wiseoldtech: Either the glue hasn't turned conductive yet or my DVM isn't sensitive enough. I did as you suggested, but the meter only read open. That said, based on my own more limited experience with similar glues in other equipment, removing it during a recap would be my recommendation to the owner.
Everything I just wrote applies to when using the pre-outs into the line inputs of a receiver I use at my bench: The hiss is only present when the volume is set to "2" or higher and does not increase with volume.
What would you suggest next? Try different values for the 5532 gain-setting resistors?
wiseoldtech: Either the glue hasn't turned conductive yet or my DVM isn't sensitive enough. I did as you suggested, but the meter only read open. That said, based on my own more limited experience with similar glues in other equipment, removing it during a recap would be my recommendation to the owner.
Try finding and temporarily shorting out RFL708 / RFR708 and see what that does.
Noise still there: Might be time for a feedback network mod as outlined. Stlll, even as-is this part shouldn't be contributing more than about 170 µV of output noise, not state of the art but still somewhat better than average and hardly excessive with normal-sensitivity speakers (up to 90 dB/2.83 V).
Noise gone: Probably the digipot. Not sure why though, it's supposed to be essentially a plain pot and output impedance shouldn't be greater than a few hundred ohms near minimum. Maybe injected ground noise? Then I'd take a good look at grounding (i.e. PCB layout) and how / why digipot signal ground might be differing from NJM5532 feedback network ground.
What sensitivity speakers are being used with the unit?
Noise still there: Might be time for a feedback network mod as outlined. Stlll, even as-is this part shouldn't be contributing more than about 170 µV of output noise, not state of the art but still somewhat better than average and hardly excessive with normal-sensitivity speakers (up to 90 dB/2.83 V).
Noise gone: Probably the digipot. Not sure why though, it's supposed to be essentially a plain pot and output impedance shouldn't be greater than a few hundred ohms near minimum. Maybe injected ground noise? Then I'd take a good look at grounding (i.e. PCB layout) and how / why digipot signal ground might be differing from NJM5532 feedback network ground.
What sensitivity speakers are being used with the unit?
Sorry for the delay in responding. A new issue with the Rotel has emerged that's taken precedence over the hiss. I'm now getting intermittent low-level pops in the right channel. Also, when the master power switch is turned off, there's a large rush of DC voltage hitting the speakers. Sigh. Need to think about what's going on here, though the pops at least seem to be another fairly common issue with this model.
Again, I suspect the glue is at least part of the problem.
It causes all sorts of odd issues, and had me cursing up a storm in the shop.
The guy who owned the Rotel unit I was working on realized how frustrating his unit was for me, and appologized even.
And it all boiled down to that nasty glue that turned black and conductive.
Have fun with that unit you have.
It causes all sorts of odd issues, and had me cursing up a storm in the shop.
The guy who owned the Rotel unit I was working on realized how frustrating his unit was for me, and appologized even.
And it all boiled down to that nasty glue that turned black and conductive.
Have fun with that unit you have.
I'm in agreement about the glue. I plan to recap both the power and amp boards, removing all of the glue in the process. That said, I'd like to try and pinpoint the cause of this popping by eliminating some non-glue culprits.
As I mentioned, the low-level pops are only coming out of the right channel and occur even when the amp is I'm Stand-by. Even though all of the voltages on the main power supply board check out, the NJM7805FA voltage regulators at IC903 and IC904 are running very hot. Hitting them with freeze spray didn't seem to make a difference in the popping, though tapping may have -- hard to tell definitively.
Any suggestions for next steps?
I'll give you this one bit of advice.....
Whether you take it or not, is up to you.
Leave the "recapping" out of your head - this unit is not likely to have bad caps.
The "recapping craze" is a sorry, internet-driven bunch of propoganda by people who aren't seasoned old techs like I am, and that isn't the "magic wand" that people are led to believe it is....
Concentrate on the black glue and carefully remove it from around any resistors, IC chips, and board traces.
It must be eliminated completely, as I stated previously, and a sensitive DVM like a Fluke would confirm subtle resistance paths (in the K or meg ohms range in it leading to any odd behavior.
I dislike having to repeat myself.
My rationale for the recap was based on my own past experience with corrosive and/or conductive glue. If it's possibly affected surrounding components, then it may also have affected some of the caps themselves. For example, I've seen cap leads practically corroded through by the stuff. Otherwise, no, I don't follow the "recap cures all" approach.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.