Need some advice on Sony TA-N7B

Harolda

Well-Known Member
I am looking at a defect TA-N7B. Before starting to work on it i first want to check the status of the V-fets. I have worked on these units before, but they where in full working order when i received them.

I tried to find a procedure to check these devices. I found some on AK, mainly for 2sk18 and 2sk60 for other Sony amps like TA-4650, -5650 etc. But i guess the procedure is the same.

So what i did
1) Resistance measurement between S - D (Source - Drain)
all measure between 1.5 - 2.1 Ohms

2) Diode test (+/- and -/+):
S - G all measure around 0.5V
D - F all measure around 0.5V

I attached some pictures of how i used to Multimeter to perform these measurements. Is this correct, and if so can i conclude that the devices are OK?

I also did some first voltage checks on the amp (without the V-fets). I have a service manual, but it is a scan and it is difficult to read the exact voltage requirements. What i found thus far is that at some point where the manual indicates 0V, i get ~45V. I will post more data on the voltage measurements, but it would be helpful if someone has a better scan of the service manual.

thanks in advance for your advise.
 

Attachments

  • P1050020.jpg
    P1050020.jpg
    75.3 KB · Views: 49
  • P1050019.jpg
    P1050019.jpg
    70.9 KB · Views: 33
  • P1050018.jpg
    P1050018.jpg
    76.6 KB · Views: 29
It sounds like the VFETs are fine. The following document corroborates your findings. https://solidstateamps.files.wordpress.com/2014/02/sony_service_bulleting_62.jpg

Still I strongly recommend that you pack up the amplifier and send it to me for an extended evaluation! ;) It could take a decade, though, possibly longer. :)

I do have an original service manual but no scanner handy but I can send you close-up photos if you tell me the page/figure numbers are having trouble with.
 
It sounds like the VFETs are fine. The following document corroborates your findings. https://solidstateamps.files.wordpress.com/2014/02/sony_service_bulleting_62.jpg

Still I strongly recommend that you pack up the amplifier and send it to me for an extended evaluation! ;) It could take a decade, though, possibly longer. :)

I do have an original service manual but no scanner handy but I can send you close-up photos if you tell me the page/figure numbers are having trouble with.

I would appreciate it very much if you could help with a close-up photos of the
- Mounting diagram L-CH Power amp board, page 6
- Mounting diagram R-CH Power amp board, page 7
- Power supply board, page 8
- socket and connection boards, page 11 & 12
 
I performed some checks on all the fusibles and caps i replaced. Nothing strange there

the tantalum (C109/C209) has been replaced with a film type?? is this possible? I will put it back to the original because this is suspect.

I also replaced the MV21N's (D101,D102,D201,D202) by two 1n4148 in series. I also did this in the other amp i restored. Now i wonder if this is correct and if this does influence the bias settings.


regards,
 
the tantalum (C109/C209) has been replaced with a film type?? is this possible? I will put it back to the original because this is suspect.

I'm no expert on this amp but I'm thinking 2.2uF films are fine there.

I also replaced the MV21N's (D101,D102,D201,D202) by two 1n4148 in series. I also did this in the other amp i restored. Now i wonder if this is correct and if this does influence the bias settings.

If it were mine I'd put the MV12N's back ... unless you think there's something wrong with them.:thmbsp:
 
Where did you get the idea of checking voltages without the VFETs installed????
The amp will not operate correctly if the VFETs are not installed. Neither will it operate correctly if not all of them are installed. Also, there has been a LOT said on these forums about this amp and rebuilding it (search!) where one of the most common reasons for not being able to set the bias properly is the failure of the output cascode BJTs. They will probably test fine on a regular DVM with a transistor test fixture and also when tested as diodes, but not work correctly because they develop leakage over time. It is HIGHLY advisable to replace these with D44H11 and D45H11 even if they appear OK and even if you can get a hold of NOS replacements for the originals.
 
Close-ups of pages 6-8, not sure if there is enough resolution to help or not.
 

Attachments

  • IMG_2840.jpg
    IMG_2840.jpg
    121.6 KB · Views: 34
  • IMG_2841.jpg
    IMG_2841.jpg
    119.7 KB · Views: 27
  • IMG_2842.jpg
    IMG_2842.jpg
    124.8 KB · Views: 20
  • IMG_2843.jpg
    IMG_2843.jpg
    121.9 KB · Views: 21
  • IMG_2844.jpg
    IMG_2844.jpg
    121.3 KB · Views: 23
Where did you get the idea of checking voltages without the VFETs installed????
The amp will not operate correctly if the VFETs are not installed. Neither will it operate correctly if not all of them are installed. Also, there has been a LOT said on these forums about this amp and rebuilding it (search!) where one of the most common reasons for not being able to set the bias properly is the failure of the output cascode BJTs. They will probably test fine on a regular DVM with a transistor test fixture and also when tested as diodes, but not work correctly because they develop leakage over time. It is HIGHLY advisable to replace these with D44H11 and D45H11 even if they appear OK and even if you can get a hold of NOS replacements for the originals.

Yeah i know, i got this idea, because i did not want to put the working Vfet's back in without knowing what is wrong with the unit. Ofcourse it will not operate correctly and i think i almost read every thread on this amp multiple times.

Is there no way to check everything before putting the Vets in?

I already replaced the 2SC1173 and 2SA473 with the same types i bought some years ago when i worked on the other amp. I will try with the D44H11 and D45H11. Thanks for the tip.
 
Yes there is a way. The problem with pulling out VFETs is that the output stage then has no power, and the output transistors degenerate to B-E diodes, which means that the current from the drivers increases a lot unless you first reduce the idling current to zero (turn the bias pot all the way down). The reason why I recomend replacing the originals is that they have a very low breakdown voltage (30V if memory serves) and therefore you can's simply strap power to the output stage to get the amp working provisionally because they will be dead in an instant. Unfortunately, in the properly working amp they will also die, although it will take a long time - there are circumstances when Vce is going to rise over 30V (like clipping) even with everything working. There is a sort of inherent mechanism that protects them through the action of the cascoded VFETs but exposing them to such conditions WILL eventually kill them. D44/45H11 are 80V transistors so this is far less of an issue. They also can withstand almost double the current and have some other very desirable characteristics like almost constant beta over a huge range of currents - which is EXACTLY what you need in this amp.
Now, the way to test this unit, AFTER you have gone through the usual DMM testing of components known to fail such as fusible resistors, and you have tested the main power disconnected from the amp, is to remove all the VFETs and all but one pair of the BJTs. There must be no load connected for any further test. Reduce bias to zero (careful with the direction you need to turn the bias trimmer!). If you can disconnect the power from one channel completely and do it one channel at a time, even better (as in disconnect the mains side of the power transformers for that channel).
Powering up the unit should be fairly safe if you have ascertained that there are no obviously damaged components. Then first check the power supplies. Everything else goes after that.
At some point when you are ready to test the fixed unit, if you want some insurance that it does not turn all bad if you have missed something, you can power it up with one pair of output transistors installed (and I mean one pair of BJTs and VFETs). This is sufficient to check voltages and even idle current if you chose the one pair that connects to the emitter resistors used to check bias on the test points. My advice would be to just check if you can get some idle current going at a sensible setting of the bias trimmer. You can even check for output signal with a scope. Even a load is possible but because you are running on only one pair, the load impedance should be 3x rated load, so 24 ohms and up. If it all checks out and assuming the rest of the output pairs are OK, reduce bias back to zero, install the rest and finalize the amp.
 
still in the process of checking all components. Capacitors are all new and replaced correctly. All fusibles are new and checked several times, no faults there.
I just received the d44h11 and d45h11, so i will put them in this week.
I checked the 5w metal plate resistors (mpc5), although they should be 0.47ohm, all 12 measure 0.3ohm.
I took out and performed diode tests on the 2sc1173 and 2sa473, but everything looks ok, just wanted to check their function.

ilimzn, i am going to disconnect one channel from the power supply and only work on one side. If i put in only one pair of D44H11 and D45H11, will i be able to check all voltages and the correct functioning of that channel?

regards
 
fusibles R127/R227

Yesterday i found two fusibles, which i did not notice before (also in my other amps).
These are R127/R127, they are connecting the centre of the large power supply caps to ground. These resistors are not easy to spot, because they are mounted in the cable.
I replaced these as well and put a heatshrink over them. Can anyony tell me the function of these resistors?

regards,

Harold
 
You should be able to get the amp going in a basic sense (no load!) with only one pair, and starting with lowest bias. This essentially just connects the driver stage to the output through the B-E junctions of the output transistors. This closes the NFB loop and should normally set the output to near zero V and you should even be able to adjust the offset. Of course, there is no sense in trying to adjust bias as there is no power to the output stage so no bias current will flow. To do this, one pair of VFETs has to be inserted. Be sure to check the fusible resistors that go to the gates of the VFETs!

The resistors you ask about are ground isolators, they connect the chasis to the separate ground of the two channels without actually making a hard connection which could result in a ground loop. In many cases these are also bypassed by two antiparallel diodes.
 
bourns pots

Finally i found the issue with the amp and it was my very own stupid mistake.
I am really ashamed about this but when i put in new trim pots i put the 100 Ohm version at the location of the bias adjustment and the 22k Ohm version at the DC offset location. :thumbsdn:

It was difficult the recognize because of the small letters on the side of the trimpots but after rechecking everything a few times i noticed that i had 101 on the bias pot and 223 on the dc offset pot.

if will unsolder them and put them in the correct position this week and will keep you updated.

regards
 
Finally everything is ok. I have a healthy 70mv for bias on both channels and approx 2mv on DC. I will Now finalise the amp and recheck bias and DC when the unit is warm.

Thx for all the support. I will post Some pictures when it is finished

Regards
 
What a sound. This is better then the other 2 N7's i have. I think the d44/d45 make the difference, especially in the bias setting. This amp is now running at 75mv. With the other amps i could not set them higher than 50mv. It is also running much hotter now.
DC offset is stable at 5mv on both channels.

What a combination ta-e88 --> ta-n7 --> Yamaha NS2000


Regards
 
Finally everything is ok. I have a healthy 70mv for bias on both channels and approx 2mv on DC. I will Now finalise the amp and recheck bias and DC when the unit is warm.

Good to hear you have it up & running nicely :)

Personally, I prefer to bias around 35 to 40 mV for a more natural sound; beyond 45~50 mV it becomes (IMO) a bit overly 'creamy' in the lows. Class AB sounds better to me than driving it into Class A.

Also, at 70 mV the VFETs run very warm, and semiconductors like these suffer from that over long term; keeping the VFETs at around 35 degrees is better.
 
Last edited:
Back
Top Bottom