AU-717 Hiss/Cracking noise

[skhounsombat]

Help please!
I read several posts regarding AU-717 noise but I could not find anything that could be related to my problem. I bought the unit with the known hiss/cracking noise. I completed the upgrade/restoration per well documented Leestereo's thread http://audiokarma.org/forums/index....estoration-upgrade-of-a-sansui-au-717.641945/
After completion of the rebuild I still get constant hiss noise at zero low volume and even without any signal inputs into the F2721/F2722. I believe the hiss noise is identical on both L and R channels. I am confused on how could the noise be identical since the each seems to be fed by its owe power supply, including its own transformer.

 

[super98lsc]

Does it do it with the pre-amp switch selected to Separated?
Presume you used De-Oxit on that switch as its been known to cause trouble on some units.
Noisy transistors can cause hiss, if it does it with the switch in the Separated position then we know the issue is likely in the power amp stage. Its possible someone worked on the unit prior to your acquisition of it and accidentally used a non polar cap in place of a polar one etc, this makes the like for like replacement game a pile of fun cross checking against the schematics.

I have 2 AU-717s, I rebuild them both to Leestereo's specs and they are awesome.

 

[Hyperion]

A few questions...

As above - it does this whether rear panel switch is set to 'connected' or 'separated' ?

Does the protection circuitry function normally?
Can you set DC Offset and Bias without any difficulty and to manufacturers spec'?
Any deviations AT ALL from Leestereo's component choices, and source for those components?

Please post some pictures of your work - we like pictures

 

[skhounsombat]

Does it do it with the pre-amp switch selected to Separated?
Presume you used De-Oxit on that switch as its been known to cause trouble on some units.
Noisy transistors can cause hiss, if it does it with the switch in the Separated position then we know the issue is likely in the power amp stage. Its possible someone worked on the unit prior to your acquisition of it and accidentally used a non polar cap in place of a polar one etc, this makes the like for like replacement game a pile of fun cross checking against the schematics.

I have 2 AU-717s, I rebuild them both to Leestereo's specs and they are awesome.

Thanks for your comments.
Yes, there was still noise when switch selected to separated but very fainted one. I de-soldered the signals wires to bypass the switch and there was still noise. I unplugged the signal wires from F2720 and the noise was fainted like pre-amp switch at Separated. I love the sound of the AU-717 with combination of its pre-amp so I don't want to think of using its amp only.
Correct, I used De-Oxit on that switch many times. This is the same as de-solder the input wires to bypass the switch.
I will recheck the caps per your last comments. However, when restoring/replacing caps per Leestereo's specs I followed the polarity per the board markings.
One thing that I have not mentioned. My power supplies caps need replaced as I was able to measure the ripple voltages by oscilloscope last night. The four channels are amongst 80-140mV.

 

[skhounsombat]

A few questions...

As above - it does this whether rear panel switch is set to 'connected' or 'separated' ?
yes, fainted one
Does the protection circuitry function normally?
Protection circuitry worked perfectly and the relay was replaced as Leestereo suggested.
Can you set DC Offset and Bias without any difficulty and to manufacturers spec'?
Initially, DC offsets and Bias were adjusted correctly until I blew TR12 and TR13 on F2722 board and now my bias can only go up to 5-6mV. That beg another question. I used MJE15032 and MJE15033 for the TR12 and TR13. Could that be explained why I could not adjust the bias higher?
Any deviations AT ALL from Leestereo's component choices, and source for those components?
The only few caps I did not replace were some film caps as I could not find them in small capacity and sizes.
Please post some pictures of your work - we like pictures

I will try to post my work tonight when I get home.

 

[Leestereo]

Yes, there was still noise when switch selected to separated but very fainted one. I de-soldered the signals wires to bypass the switch and there was still noise. I unplugged the signal wires from F2720 and the noise was fainted like pre-amp switch at Separated.

If I understand correctly, when the pre-main switch is on "separated", the noise is significantly diminished? This suggests that the problem is on the F-2720 and as was suggested by super98lsc, it is likely noisy transistors. My guess would be the 2SA726 transistors in the differential inputs of each channel; this particular transistor has been known to sometimes (bad batches?) get noisy/leaky as they age. You can replace with hFE matched KSA992 transistor pairs.

 

[LBPete]

You need to determine if the noise is from the amp or pre amp. Separate them and drive another amp with the 717 pre amp and the 717 amp with another pre amp. Odds are it's in the pre amp.

I've been chasing a similar noise in a 717 and it's driving me nuts. I've got it isolated to the tone board. i've replaced all the transistors on that board and it still does it to some degree. Here's a link to the thread. It may give you some ideas.http://www.audiokarma.org/forums/index.php?threads/head-scratcher.685103/

- Pete

 

[skhounsombat]

If I understand correctly, when the pre-main switch is on "separated", the noise is significantly diminished? This suggests that the problem is on the F-2720 and as was suggested by super98lsc, it is likely noisy transistors. My guess would be the 2SA726 transistors in the differential inputs of each channel; this particular transistor has been known to sometimes (bad batches?) get noisy/leaky as they age. You can replace with hFE matched KSA992 transistor pairs.

Thanks, I will try your suggestion.

 

[super98lsc]

When you repaired the nuked F2722 what blew on it?
Tr06 biases TR12/TR13 (John correct me if im wrong its been a while) I know if the bias is low on the pre-drivers TR12/13 you wont have the necessary voltage to properly bias the power output pair.
Did you recheck all the resistors and the diodes D04,D02,D01, and the Zener Diodes as well.
Typically, as in this AU-417 that nuked its outputs (similar circuitry just single power supply) when the outputs go it can burn traces, ruin resistors etc.

I have not yet worked on a unit with a blown FET01 ( the small IC on the board)

 

[skhounsombat]

When you repaired the nuked F2722 what blew on it?
Tr06 biases TR12/TR13 (John correct me if im wrong its been a while) I know if the bias is low on the pre-drivers TR12/13 you wont have the necessary voltage to properly bias the power output pair.
Did you recheck all the resistors and the diodes D04,D02,D01, and the Zener Diodes as well.
Typically, as in this AU-417 that nuked its outputs (similar circuitry just single power supply) when the outputs go it can burn traces, ruin resistors etc.

I have not yet worked on a unit with a blown FET01 ( the small IC on the board)

Oh yes, the big TR702 and TR704 also blew. I will trace back other components that might be blown. So on the bias, it seems that I could only adjust the VR03 for 2 to 4-5mV for a full turn. This was unusual than before. It used to be, or at least from the other side, the adjustment was very sensitive and could go wide range. Anyhow, even with this low bias I could still hear music on the headphone fine with no difference between the L and R channel. Is it necessary to get the bias up to manufacturer's recommendations.
On the FET01, I did try to remove it for noise isolation/troubleshooting with no good result. Lesson learned for me is not troubleshoot noise with parts removed.

 

[skhounsombat]

You need to determine if the noise is from the amp or pre amp. Separate them and drive another amp with the 717 pre amp and the 717 amp with another pre amp. Odds are it's in the pre amp.

I've been chasing a similar noise in a 717 and it's driving me nuts. I've got it isolated to the tone board. i've replaced all the transistors on that board and it still does it to some degree. Here's a link to the thread. It may give you some ideas.

- Pete

Interesting. Mine is very identical on L and R channel. Any thought or success on the volume controller. But when it was turned ways down to zero, there was still noise so now I am not sure. I will keep looking further. I am going to repeat this experiment again as I was not 100% sure. I used Marantz pre out and AU-717 amp and I thought I did not get good results. That was why I kept working on the power amps.

 

[super98lsc]

Does the noise change with the balance knob?

 

[skhounsombat]

Does the noise change with the balance knob?

No change with balance knob or input selection.

 

[super98lsc]

Ok that helps rule out a bad/ungrounded input. Swap out all those crap transistors (2SA726) as John and Leestereo suggested, I would be the issue lies in that board F2720. If there was a broken input wire between boards typically the ground would induce noise but vary as the balance knob is moved from channel to channel, reason is the channel with the bad connection is "finding ground" through the balance knob via the ground on the opposite channel. Case in point my CA-3000 when I first got it. Took me 1.5 hours to find that broken wire inside the jacket of an insulated 2 conductor pair. You got this man it will be an easy fix. Need to address the bias issue on the F2722 after you solve this.

 

[skhounsombat]

Here are the pictures of the outputs. Just to update on the noise description. When the pre-amp switch is selected at connected, the cracking noise is loud. When it is at others, the noise is still there, same cracking noise but not as loud.

 

[skhounsombat]

Update from tonight.
Working on replacing transistors on F2720 as Leestereo suggested. Here are what I found
1) TR01,02 called for 2SA798 but the board was installed with 2SA726 and with correct pin locations, more to come on the pin locations.
2) TR03,04 were installed with 2SA726 but I believe the pin locations were incorrect, see pics.
3) TR5,6 called for 2SC1400 but were installed with 2SA726 and with incorrect pin locations (see attached pics)
4) the rest were installed as called out and with correct pin locations.
Web search on 2SA798 resulted with dual transistor. Could the manufacturer make a mistake on the BOM

 

[stevo1214]

Then you need to replace the TR05 and TR06, They are suppose to be NPN not PNP. KSC1845 should do it. TR01,02 are right according to mine.

 

[Hyperion]

Update from tonight.
Working on replacing transistors on F2720 as Leestereo suggested. Here are what I found
1) TR01,02 called for 2SA798 but the board was installed with 2SA726 and with correct pin locations, more to come on the pin locations.
2) TR03,04 were installed with 2SA726 but I believe the pin locations were incorrect, see pics.
3) TR5,6 called for 2SC1400 but were installed with 2SA726 and with incorrect pin locations (see attached pics)
4) the rest were installed as called out and with correct pin locations.
Web search on 2SA798 resulted with dual transistor. Could the manufacturer make a mistake on the BOM

This is a classic mistake - relying entirely on the parts listing in the SM.

 

[skhounsombat]

This is a classic mistake - relying entirely on the parts listing in the SM.

I see what you meant. How would you do it in order not to trust the last guy who might be working on the unit.

 

[Hyperion]

I see what you meant. How would you do it in order not to trust the last guy who might be working on the unit.

This is quite difficult. With an un-molested unit you can start by assuming that every component in the unit now is correct - as the unit was passed by the factory and worked for years in this condition. The schematic and parts lists must (with Sansui gear) be used only as a guide, and not necessarily 'gospel' when it comes to component types and values. Those with experience have often learned the hard way, this is when electronics understanding and knowledge is priceless in these circumstances as it can reinforce and confirm suspicions about component changes/schematic/parts list differences and what to do about them. Although truth be told, some of the changes are far from obvious, and not detected until an issue rears it's ugly head.

Asking here about such differences very often helps, but sometimes you have to wait a while until the right person chimes in with that golden piece of information, gained by bitter experience, that unlocks a mystery. In short, if you don't have experience, ask here about every discrepancy you do notice, nice clear pictures can often help the wise unravel the mysteries of the ones you don't notice.