I have this Sansui AU 2900 which I had all the capacitors replaced.
The problem is that the left channel continues to produce crackling. I have already taken it back to the audio store technician for a second time and after 3 months of waiting they told me that they tested it for a long time and in the end it was just a fuse making a false contact. The technician also told me that he checked the transistors and put plastic and new thermal paste.
The whole operation to date has cost me 700E and in the gallery you can see all the capacitors replaced. I took pictures of other details of the board and the transistors. From the gallery you can see the first 4 photos (before) of the original state while all the others are post operation (after).
Unfortunately, I lost confidence in the technician who did the repairs, and could probably have spent 700E on a new, modern amp.
I am very despondent and would like someone to help me understand and maybe try to repair it myself. I'm not an electronics expert, but I do have good dexterity.
If you need more details, I am willing to send more pictures.
Thanks to anyone who can help me find a solution.
Edit 1: I've seen in other thread and forum that this is a common issue. I've downloaded the service manual and first thing I learned is about bias. Comparing the pictures before/after it looks like that was not touched. Is it possible that after recapping the full unit there is no need to adjust the bias? In the service manual they suggest to check every 3 months. I have a multimeter and I think that this is the first thing I'll do tomorrow morning, but not sure it will fix the issue.
Edit 2: the cables I'm using for connecting the speakers are custom; I made them by myself using an high quality 4mm^2 section cable . I'm not sure if I'm doing the right thing but I'm connecting them this way with spades at the very end. I realized only at the end that they are a bit bigger but they sound great.
Edit 3: I've recorded 2 videos:
- After the recap
- Today, when I got i back home and the technician told me that was only a fuse...
As you can see on the left channel the sound is completely distorted with lot of crackling.
Edit 4: I used my multimeter to adjust the bias. I was able to adjust the right channel to 13mV but I have alway 0mV on the left one. I measured several times, ampli was running for 30min or more, speakers connected, volume at minimum and I can measure 12.9/13/13.1 mV always on the right channel and always 0mV to the left.
Edit 5: I let the ampli cooling down and I'm able to measure the on both channel ~13mV and there is no distortion. I'll redo the measure in 30min.
Edit 6: I listened at low volume (3 on the knob) for half an hour. I adjusted the bias to 13mV in both channels. No distortion. I continued listening at low volume and then increased the volume to 5. As soon as it started to distort, I turned the volume down and measured 13mV on the right channel and 0mV on the left channel. @anatech Is this still related to TR03 TR04 and IC01 and IC02?
The problem is that the left channel continues to produce crackling. I have already taken it back to the audio store technician for a second time and after 3 months of waiting they told me that they tested it for a long time and in the end it was just a fuse making a false contact. The technician also told me that he checked the transistors and put plastic and new thermal paste.
The whole operation to date has cost me 700E and in the gallery you can see all the capacitors replaced. I took pictures of other details of the board and the transistors. From the gallery you can see the first 4 photos (before) of the original state while all the others are post operation (after).
Unfortunately, I lost confidence in the technician who did the repairs, and could probably have spent 700E on a new, modern amp.
I am very despondent and would like someone to help me understand and maybe try to repair it myself. I'm not an electronics expert, but I do have good dexterity.
If you need more details, I am willing to send more pictures.
Thanks to anyone who can help me find a solution.
Edit 1: I've seen in other thread and forum that this is a common issue. I've downloaded the service manual and first thing I learned is about bias. Comparing the pictures before/after it looks like that was not touched. Is it possible that after recapping the full unit there is no need to adjust the bias? In the service manual they suggest to check every 3 months. I have a multimeter and I think that this is the first thing I'll do tomorrow morning, but not sure it will fix the issue.
Edit 2: the cables I'm using for connecting the speakers are custom; I made them by myself using an high quality 4mm^2 section cable . I'm not sure if I'm doing the right thing but I'm connecting them this way with spades at the very end. I realized only at the end that they are a bit bigger but they sound great.
Edit 3: I've recorded 2 videos:
- After the recap
- Today, when I got i back home and the technician told me that was only a fuse...
As you can see on the left channel the sound is completely distorted with lot of crackling.
Edit 4: I used my multimeter to adjust the bias. I was able to adjust the right channel to 13mV but I have alway 0mV on the left one. I measured several times, ampli was running for 30min or more, speakers connected, volume at minimum and I can measure 12.9/13/13.1 mV always on the right channel and always 0mV to the left.
Edit 5: I let the ampli cooling down and I'm able to measure the on both channel ~13mV and there is no distortion. I'll redo the measure in 30min.
Edit 6: I listened at low volume (3 on the knob) for half an hour. I adjusted the bias to 13mV in both channels. No distortion. I continued listening at low volume and then increased the volume to 5. As soon as it started to distort, I turned the volume down and measured 13mV on the right channel and 0mV on the left channel. @anatech Is this still related to TR03 TR04 and IC01 and IC02?
Last edited:
If it's a steady crackling (unrelated to signal levels or vibration), that could be a noisy transistor. I had one of those in my Hafler preamp; I tracked it down by signal tracing, threw in something I had on hand that the Radio Shack transistor substitution guide said was reasonably close, and that made it quiet again.
Hi boredcollie,
Let's do some basic checks before zeroing in on anything. Try not to make assumptions, and bias is not going to be your problem unless it is a very weird problem. To adjust bias you need a decent voltmeter, something over $100 because you are measuring millivolts. Cheap meters are not accurate enough. One you set bias, it should be stable, but it will vary depending on temperature and mains voltage. It isn't that exact. So, for now ignore bias current unless the heat sinks are getting very warm or hot.
A "recap" should never be done just because for starters. You were given bad advice and wasted a lot of money.
With the volume control all the way down, do you hear the noise?
If so, the two most likely parts causing your problem are IC01 or IC02, and TR03 or TR04. I didn't check to see which channel is which. Now, the IC numbers area dual transistor. It can be replaced by two singles, but it is critical that they be matched closely. Most technicians can't do that, I doubt you can. The beta test feature on a meter will not work for this because the transistors must be at the exact same temperature when they are measured. I mean exact. I designed a jig to do this and gave the design to our members, but it doesn't make sense for you to build one for this and your level of knowledge. Please believe me, I am not insulting you at all. I am trying to help you.
There are other parts that can cause this as well. You really need to find and put your trust into a good technician.
-Chris
Let's do some basic checks before zeroing in on anything. Try not to make assumptions, and bias is not going to be your problem unless it is a very weird problem. To adjust bias you need a decent voltmeter, something over $100 because you are measuring millivolts. Cheap meters are not accurate enough. One you set bias, it should be stable, but it will vary depending on temperature and mains voltage. It isn't that exact. So, for now ignore bias current unless the heat sinks are getting very warm or hot.
A "recap" should never be done just because for starters. You were given bad advice and wasted a lot of money.
With the volume control all the way down, do you hear the noise?
If so, the two most likely parts causing your problem are IC01 or IC02, and TR03 or TR04. I didn't check to see which channel is which. Now, the IC numbers area dual transistor. It can be replaced by two singles, but it is critical that they be matched closely. Most technicians can't do that, I doubt you can. The beta test feature on a meter will not work for this because the transistors must be at the exact same temperature when they are measured. I mean exact. I designed a jig to do this and gave the design to our members, but it doesn't make sense for you to build one for this and your level of knowledge. Please believe me, I am not insulting you at all. I am trying to help you.
There are other parts that can cause this as well. You really need to find and put your trust into a good technician.
-Chris
I think I remember that the preamp on the AU 2900 is the same as the 3900 and 4900.
And the very common problem is that the TR 2SA798 (differential torque, you won't get it, you have to make it with two conventional TRs) gets noisy.
There is a thread here:
Sansui AU 4900 problem
And the very common problem is that the TR 2SA798 (differential torque, you won't get it, you have to make it with two conventional TRs) gets noisy.
There is a thread here:
Sansui AU 4900 problem
I have this one:
But didn't cost more than 100$ so I doubt it can measure 13mV 🙁
There is a 200mV on the left side, but I'm not sure if it will be accurate to measure the 13mV as stated in the service manua
I can't sleep because I'm so disappointed, can you believe it? This evening I opened it up and saw that the whole preamp part is untouched, it has all the original capacitors. I don't have a turntable but I'm sure that doesn't work well either.
No, it disappear.
I have identified IC01 and IC 02 as well as TR03 and TR04. They are both in an easily accessible location. I hear you when you say that you are not insulting me, but it makes me even more discouraged to know that it is difficult even for a technician because it will be probably impossible to find someone with these skills where I live....or at least someone who is willing to work hard at it.
Can you share the jig, at least to understand if it something I can do.
Another question I have: does it make sense replace only TR03 and TR04 and leave the TR01 and TR02 original? I've see on another forum that they should be replaced with others similar but modern.
TR01 and TR02 are 2SA726 and if it is just matter of unsoldering/soldering it's something I can do.
I can also replace the insulator with these, or it is not needed?
If you know someone in French Riviera that can help me, it'd be great 🙂
Thanks, I'm reading that thread trying to understand better. I've added 2 videos of the problem btw.
I've just uploaded 2 videos of the problem. I'm using the term crackling but not sure if it is correct. Sound is distorted and really noisy compared to the other channel.
Hi boredcollie,
If the sound disappears with the volume control down, the fault is before the power amplifier section (great news!).
That meter can't really be trusted for those low measurements. Don't touch the bias please. I'm sorry for the news, but most of those meters don't even meet their own accuracy specs new out of the box. I used to calibrate and certify test instruments in a lab, so I do have experience there. I hate to say it, but the most common brands you can really rely on will be Fluke and Keysight. Get something just up from the bottom of the line and watch the accuracy specs for what ranges you need. The other thing to worry about is if the meter will stay in tolerance. Most do not, but Fluke and Keysight will for decades. There is the odd other brand that is good as well, and they will also cost more. Don't buy Extech, they look really good, but they aren't really.
Sorry to hear you didn't even get the service you paid for, but maybe they did you a favour as the less they touched, the less they did wrong.
At this point, to troubleshoot what should happen is that a sine wave should be fed into the unit and the signal traced with an oscilloscope, also without a signal. To try and do this without equipment is really all chance and luck. The previous service person did not do what they should have done, and they may not even have that equipment to start with. Maybe a friend or member close to you can give you a hand.
The jig I mentioned was in "another Adcom GFA-565 thread" (I think) It was a long time ago, however it made accurately matching transistors easy. You did need a meter that measured down to 0.0 mV at least, and a bipolar power supply. A pair of 9V batteries would work.
No, it doesn't make sense to work in the power amplifier area since you proved the fault to be before the amplifier section. What you might want to try is to operate the various controls to see what make a difference in the sound. That might help narrow down the section where the problem is. That is what I would do on the bench. After that you are pretty much down to experience and signal chasing.
It's a good amplifier and well worth fixing. At this point you're going to have to forget what you have paid this far, it isn't the amplifier's fault. Start fresh with either a well respected audio technician, or maybe a member here that is close to you that can help further. I did not watch the Youtube videos (sorry, I don't do that). But what you are describing tells me what I have to know. I would need to see the scope waveform of the signals to tell you anything more intelligent at this point.
Using a good respected audio tech is almost always less expensive than someone "cheap". They tend to have good morals and will treat you fairly. Many other TV and computer techs have much less training, less equipment and rely more on the internet to extract as much money as they can from you. Those good techs will also warn you if they see anything more serious before you sink money into something. Your story is a pretty common one I see everywhere. It makes me both sad and angry, so I am really sorry this has happened to you. Hopefully someone can give you a hand.
-Chris
If the sound disappears with the volume control down, the fault is before the power amplifier section (great news!).
That meter can't really be trusted for those low measurements. Don't touch the bias please. I'm sorry for the news, but most of those meters don't even meet their own accuracy specs new out of the box. I used to calibrate and certify test instruments in a lab, so I do have experience there. I hate to say it, but the most common brands you can really rely on will be Fluke and Keysight. Get something just up from the bottom of the line and watch the accuracy specs for what ranges you need. The other thing to worry about is if the meter will stay in tolerance. Most do not, but Fluke and Keysight will for decades. There is the odd other brand that is good as well, and they will also cost more. Don't buy Extech, they look really good, but they aren't really.
Sorry to hear you didn't even get the service you paid for, but maybe they did you a favour as the less they touched, the less they did wrong.
At this point, to troubleshoot what should happen is that a sine wave should be fed into the unit and the signal traced with an oscilloscope, also without a signal. To try and do this without equipment is really all chance and luck. The previous service person did not do what they should have done, and they may not even have that equipment to start with. Maybe a friend or member close to you can give you a hand.
The jig I mentioned was in "another Adcom GFA-565 thread" (I think) It was a long time ago, however it made accurately matching transistors easy. You did need a meter that measured down to 0.0 mV at least, and a bipolar power supply. A pair of 9V batteries would work.
No, it doesn't make sense to work in the power amplifier area since you proved the fault to be before the amplifier section. What you might want to try is to operate the various controls to see what make a difference in the sound. That might help narrow down the section where the problem is. That is what I would do on the bench. After that you are pretty much down to experience and signal chasing.
It's a good amplifier and well worth fixing. At this point you're going to have to forget what you have paid this far, it isn't the amplifier's fault. Start fresh with either a well respected audio technician, or maybe a member here that is close to you that can help further. I did not watch the Youtube videos (sorry, I don't do that). But what you are describing tells me what I have to know. I would need to see the scope waveform of the signals to tell you anything more intelligent at this point.
Using a good respected audio tech is almost always less expensive than someone "cheap". They tend to have good morals and will treat you fairly. Many other TV and computer techs have much less training, less equipment and rely more on the internet to extract as much money as they can from you. Those good techs will also warn you if they see anything more serious before you sink money into something. Your story is a pretty common one I see everywhere. It makes me both sad and angry, so I am really sorry this has happened to you. Hopefully someone can give you a hand.
-Chris
Noisy transistors are a common fault on this Sansui series (2900, 3900, 4900). In addition, the 2900 hasn't got an output relay, and could burn the woofers.
By the way,I live in Milano. Please feel free to PM me if you wish.
By the way,I live in Milano. Please feel free to PM me if you wish.
Hi Orit,
While noisy transistors are not unknown, they are still not extremely common. The problem may well be a noisy transistor, but keep an open mind and trace the fault to the problem. Don't assume what the problem is before you actually locate it.
Just keep an open mind.
-Chris
While noisy transistors are not unknown, they are still not extremely common. The problem may well be a noisy transistor, but keep an open mind and trace the fault to the problem. Don't assume what the problem is before you actually locate it.
Just keep an open mind.
-Chris
Yes, as I said, it was a very common failure in the 2SA798 series, there are hundreds of cases and many well documented on the web, including mine, but it is not a job for everyone.
The 2SA798 didn't go noisy as much as the transistors drifted way out of match. Compared to the number that were used, a high percentage have gone noisy but the most common reason for replacement is an out of match condition.
I think the 2SA798 noise issues are on par with the singles that were produced using the same die and similar transistors. All I am really saying is that you shouldn't jump to the conclusion that the 2SA798 is at fault right off the bat. I was never a fan of that part even when I serviced products in the 70's, but don't condemn it just because many have gone noisy.
I think the 2SA798 noise issues are on par with the singles that were produced using the same die and similar transistors. All I am really saying is that you shouldn't jump to the conclusion that the 2SA798 is at fault right off the bat. I was never a fan of that part even when I serviced products in the 70's, but don't condemn it just because many have gone noisy.
I think you have not had to deal with one of those TRs that get loud over time and in a totally random way ... they are terrible! You think you have everything solved and within days of having it working perfectly, (or only a few hours, or a few minutes and then for several days everything OK), the noise returns.
There are no instruments that can detect the origin of the failure previously.
The solution is simple, because I have never heard that two differential TRs (takes one per input from each preamp) failed at the same time, so my advice to the OP, from my own experience, is to take them out and change the channel.
There are no instruments that can detect the origin of the failure previously.
The solution is simple, because I have never heard that two differential TRs (takes one per input from each preamp) failed at the same time, so my advice to the OP, from my own experience, is to take them out and change the channel.
As a general question related to amplifier, does it make sense to upgrade things like the speakers connectors to a more modern speaker jack like this one?
In terms of restoration which is the approach? To keep everything as much as possibile original or to improve the unit?
In terms of restoration which is the approach? To keep everything as much as possibile original or to improve the unit?
The term restore means:
"Put a thing in the state or estimate that it had before"
Swapping the original pressure terminals for "modern" binding post will make absolutely no improvements to the sound.
"Put a thing in the state or estimate that it had before"
Swapping the original pressure terminals for "modern" binding post will make absolutely no improvements to the sound.
How many TR's of "differential pair" encapsulated audio amplifiers do you know that have made these kinds of faults? Those random clicks, being located in the first stage of the amplification chain, becomes ++++ DB in the speakers.
I do not know of other cases other than the "infamous" 2SA798, it would be nice to know, thank you!
Hi academia50,
The 2SA798 was pretty much the only dual transistor used, and it was super cheap. However, I have seen many singles also go noisy as well.
The dual transistors you get from other suppliers are a lot more expensive and much higher quality. But, they didn't use them in normal consumer equipment. So you don't really have another option to look at.
The 2SA798 was pretty much the only dual transistor used, and it was super cheap. However, I have seen many singles also go noisy as well.
The dual transistors you get from other suppliers are a lot more expensive and much higher quality. But, they didn't use them in normal consumer equipment. So you don't really have another option to look at.