Technics RS-B965

[Alex SE]

i am not that expert on this, but why changing resistors at output?
can't you do something directly at the meters, where your problem actually is?

and, why don't you check levels at dolby chips, where it's more important that it's the right one?

nice solution for those heatsinks, as those transistors don't have a hole for the screw.

Hi Vince,
Dolby levels are as they should be. Adjusted with ANT test cassette, 245mV (varies 240-250mV) .
As we were talking before, 965 peaks little more that it should and I'm just irritated of that. I have solved that part by replacing R41 and R42 with 11K instead of 10K (as ANT told me to do so). There is no another way to deal with peak meters.
Next problem is that while signal is at 0db, output is 365mV instead of 400mV which can be adjusted by replacing R47 and R48 (100K) with something else. As I wrote earlier, it can be something around 110K.
Making at the moment a little pcb with resistors 100K + 4.7K + trimpot 10K which will give me variable resistor 104.7 - 114-7K. That way I will determine what value of that resistor I need. That's because I don't know how to calculate (formula).

Observe that whole solution can impact accuracy of calibration, but I have to test an see if this solution works.

 

[vince666]

ah, great... didn't get it before.

so meters are OK on ANT's advice and you simply need to boost output a bit.

the trimmer method is the easiest way to find the right value, since we aren't able to calculate it directly.
i'd say you can go on like this.

 

[WOPL Sniffer]

BTW, friend of mine told me that Nakamichi use to put in huge caps. One of mine 965 have four 1000uF in power supply. The other one have two 2200uF. I was thinking, If I replace those four 1000uF with four 2200uF, would it give some effect?

No

 

[Alex SE]

ah, great... didn't get it before.

so meters are OK on ANT's advice and you simply need to boost output a bit.

the trimmer method is the easiest way to find the right value, since we aren't able to calculate it directly.
i'd say you can go on like this.

I told to ANT long time ago that peaks are not as they should be and he told me that R41 and R42 can be replaced by something else. I have tested same way as I'm testing R47 and R48 now and found that 11K is doing a job. That way peaks are OK.

Regarding output - unbelievable! Tested with 98K and 115K and there is no difference, output with both values is exactly the same. How that can be? 16K should give at least something, but not even 0.1mV. Should I try with 150K?

 

[vince666]

unfortunately, I am not able to answer your doubts on this.
you know, I am basicly another "try it and then see what happens" guy just like you.

but it might be something easy to answer for anyone with a bit of proper electronics knowledge, i guess, including if that's just the right place to change values to get what you need without influencing anything elsewhere.

 

[Alex SE]

... including if that's just the right place to change values to get what you need without influencing anything elsewhere.

Only logical resistor is that one. You can play even with R43/R44 but that would affect even headphones output by 10% but how wise it is.

 

[vince666]

ah yes... gave a broader look at that part of the schematic and, indeed, before the output there is a place where signal goes also to the meters and headphones.
of course, you don't need to make headphones output higher.

but, i wonder... changing R41/42 to lower the meter's levels did perhaps also lower headphones together with the line outputs?
if so, boosting headphones would put them back as they were before... but it's also true that headphones also have their own volume control, unlike the line outputs.

that said, i am seeing that R41/42 are connected to ground on one end and to the negative (inverting, that's the proper name i guess) inputs of an op-amp on the other end... also, the headphones volume control is connected in a similar way (ground and the negative inputs of an op-amp) while these ones at line out aren't directly connected to any op-amps.
who knows if this might be a reason why you aren't experiencing any differences while changing values of R47/48.

again, someone with some proper knowledge in electronics might explain us hows and whys related to this matter.

 

[Alex SE]

ah ... changing R41/42 to lower the meter's levels did perhaps also lower headphones together with the line outputs?
if so, boosting headphones would put them back as they were before...

I wasn't thinking about that and that sounds correct. Now I have to disassemble a whole deck again

 

[vince666]

it might be useful to find some way to attach the resistors "on the fly" from the top side until you find the right values... and after you find the right solution you apply it the proper way.
doing this way, you will be disassembling the deck just a couple times in total and not a lot of times.

 

[Alex SE]

Deck is now completed and there is nothing that can be done with peaks since that playing with R41 and R42 affects even calibration. I'm afraid that everything have to be as is. There are probably some resistors that can be replaced in a calibration (osc) area but that would be something that only ANT can solve since that he is a doctor for 965.

Bad news, I had to remove that extra PCB from the pictures because it causes disturbances (probably too close to rec amp pcb. As soon as I removed it, everything works normal. Mounted back all of those 5 transistors as they were, added heat sinks on them.

 

[vince666]

ah, nice to know.
but those DIY heatsinks can be applied also with transistors at their original place?
they fit nice?

 

[Alex SE]

ah, nice to know.
but those DIY heatsinks can be applied also with transistors at their original place?
they fit nice?

Not exactly but with little modification they fit perfectly. Bigger problem are clips as they are too wide. Will order these.

 

[WOPL Sniffer]

Change everything....... What's a little smoke and real loud squealing between friends. You guys must be smarter than the engineers who built that thing.


HUH?

 

[Alex SE]

Change everything....... What's a little smoke and real loud squealing between friends. You guys must be smarter than the engineers who built that thing.


HUH?

I'll take this as retorical question and will not try to answer. Everybody should have opinions and I have respect for all of them nomatterif I agree or not.

 

[WOPL Sniffer]

Changing out parts willy-nilly for different sizes just to see what it will do sounds like an excellent idea..... NOT

 

[Alex SE]

Changing out parts willy-nilly for different sizes just to see what it will do sounds like an excellent idea..... NOT

That is a question of interpretation. Respect.

 

[Alex SE]

@vince666, I did a test using your CD and white noise. Here is what Audio tester says:


Peaks and output is in original state, without any mods. Blue is source loop thru a deck, yellow is tape (some cheap Type I maxell S).

 

[vince666]

That's nice!

Do consider that, while using white noise, FR plots will never be as beautiful to see as you can get with a sweep test.
That said, it still gives you a good idea about how things are going on.

PS: i would lower the bias by a hair on that ferric tape, to just avoid the response starts dropping around 8Khz or so.
Anyways, do consider that the internal calibration tones of the B965 are 400Hz and 8Khz then the inbuilt calibration meters would easily tell you it's flat but if you tune the bias in order to leave a hair more at 8Khz then, on most tapes, you won't get a (small) drop at 10Khz and beyond.

 

[Alex SE]

Sweep is still weird but I'm about to test one other solution: to add 600Ω resistor between + and - on the RCA cable input to a deck.

 

[vince666]

sometimes, i read on SM of decks to just add resistors between + and - at outputs when connecting them to a voltmeter.

on my "test tones" CD, did you try to test with the multitones test... it's sort of a sweep with many different individual frequency tones.

there is a good explaination of it (but I guess you just know it), included with the downloadable files, here at my friend Walkman Archive's blog.

http://www.walkman-archive.com/wa/2015/01/bias-calibration-cd-by-vince666/