Silencing the noise floor of an older amplifier

Jump to Latest
#22
Non necessarily...even famous B1 buffer from Nelson Pass has capacitors in signal path, and to paraphrase the master himself, yes, there are capacitors in signal path...get over it!

I got quite a number of preamps, (being a preamp junkie) and guess what? The best sounding to me is sanyo plus c55, with back to back capacitors in signal path, followed by nikko preamps with similar circuitry.
The onkyo 306 i have is dc coupled and does not sound so musical, and nikko dc coupled full fet i already sold...not to mention a bunch of tube preamps i have, all sport caps in signal path, for obvious reasons.

Are you all familiar with 'infamous' howland preamp sporting tantalum cap for a special flavor? There is really interesting thread about it.

And try to do passive crossover without caps...
 
#32
The only way you are really going to know if you are making any improvement is to do one mod at a time and measure the noise floor after each mod. Could be very time consuming. Is the old girl really worth all the effort? You can't make a silk purse out of a sow's ear.
 
#33
Willi Studer said:
I wouldn't call it an issue, it is rather a higher level of inherent background hiss of this amplifier in general, when compared to more modern equipment. And it is present on all inputs. Of course, as expected, much more on phono input because of high gain, and the hiss there is more of brown noise flavour.
Click to expand...
I've simmed the phonopre in LTspice, and it looks like it's got a bit of a power supply rejection problem, +PSRR in particular. 10 dB at 100 Hz, 23 dB at 1 kHz, 37 dB at 10 kHz, all referred to output - that's knot grate.

Splitting the 220k resistor R5 into 22k + 220k and adding a 22µ filter cap to ground gets that up to 56 dB at 100 Hz, 72 dB at 1 kHz and 86 dB at 10 kHz. I hope 22k is high enough in impedance that inserting noise into the ground is not too much of a problem - otherwise a dedicated ground connection back to the power supply may be in order.

At this point we're limited by -PSRR, which is better but still only average (-29 dB @ 100 Hz, -40 dB @ 1 kHz, -37 dB @ 10 kHz), and unfortunately can be tamed only by brute force - making the negative supply as quiet as possible.

For kicks, you could try out what happens when you replace either C11 or C33 by something much smaller, maybe 10n instead of 10µ. And I do mean only one of them at a time. Of course, you can also play with bigger values.

Speaking of trying stuff, see what kind of difference the tone defeat makes, how noisy the preamp and power amp are individually when separated, how the balance control affects noise (balances changes gain in the balance amp circuit). I wouldn't be surprised if it were all down to noisy supplies, PSRR issues and grounding practices, as it was in the Sansui amp.

BTW, the "electrolytic across diodes" tweak basically applies to all the other balance and tone amplifier circuits as well. They're all variations on the same theme (SRPP output). It may be more beneficial there, given that phonopre output levels are quite low and distortion appears to be negligible as-is - going from -111 to -132 dB is not particularly interesting. (I haven't checked RIAA response. The stock cartridge loading options of 25k, 50k and 100k also seem too coarse, 36k for ATs and 68k for Shures may be more interesting.)
 

Attachments

  • b750-phono-sim.png
    b750-phono-sim.png
    42.4 KB · Views: 208
Last edited:
  • Like
Reactions: Freedom666
#34
Willi Studer said:
Don't pay attention to this imposter, we had a lot of problems with him on our local forum until he finally got banned.
Click to expand...

vulejov said:
I remember local forum.. in that place you claim that audio signal is DC.. I think that someone here still keeps screenshots..
Click to expand...

:cop: arguments brought from other forums are not welcome here and will not be tolerated. :cop:
 
#37
sgrossklass said:
While transistors can become noisy and intermittent in old age, it is the exception, not the rule. Something appearing this consistently suggests a problem that lies deeper. Clearly, noisy old voltage regulators could play a role, but it takes subpar inherent PSRR of the circuitry or subpar grounding for that to become a problem.

Looking at the schematic, I can already see that the preamp section has the same kind of pass-me-on grounding that the Sansui AU-5900 from the same time had. Not necessarily a problem as long as you don't try to bypass any power rails to it. (See the last thread on said amp for how this can backfire.) The input board also shows some ground loops, but that could just be the schematic ("ground is ground") and would have to be verified by looking at the actual board.

The input circuitry is mostly followers, these should have decent PSRR.

Power setup is complicated by them stacking two sets of +/-20 V rails to obtain +/-40 and +/-20 V (plus +/- 10 V using some unregulated voltages). Capacitance on the regulator outputs is kept to a minimum (a few µF at best), maybe so that they could skimp on protection diodes. The input board uses a pair of cap multipliers with BD139/140 to generate a set of filtered supplies. Unfortunately both sets of supplies are called +/-20 V, so you can't tell which is which - morons. I assume all the input amp circuitry (including phonopre) uses the filtered supplies though.

I would suggest:
1. Obtain some nice, low-noise 78M20/79M20s - a lot of them are, but not all of them (ON Semi or NJR are good, for example). Unfortunately 20 V is a bit of an oddball value. Skip this step if you can't find any.
2. Up output capacitance for all 4 regs significantly - to like 220-470 µF if you can fit those.
3. Add the customary protection diode from output to input (and reversed for the 79xx) - 1N4002 or something. This is just so that the cap from step 2 can discharge upon poweroff without damaging the reg.
4. 78GU1C does not strike me as anything that would still be in production, but you can try the old LM3x7 low-noise trick: Add a small electrolytic (3.3-10 µF) across (R1||R5) and (R3||R6), respectively - observing polarity. A small protection diode (oriented like D3 or D6) should also be added in parallel, 1N4148/1N914 is plenty.
5. On the input board, up the cap multiplier filter capacitors from 10µ to maybe 100-220 µF, space permitting. (MkI: C15/C26, otherwise: C11/C33.) If space is tight, consider going from single RC (2k2/xxµ) to RCRC filtering. I would also experiment with separating filter capacitor return from GL/GR and using the unused 3A terminal (present in the 1st edition) to attach a new wire that runs alongside +/-20 V back to ground on the power supply board. Maybe they had planned to do this originally but the other solution proved "good enough".

I would start with step 5, then 2+3, 1, and 4. And look at grounding on the input board, as mentioned.

Bonus tweak: A few 10 µF of electrolytic across both of the bias / level shifting diodes D1+D2 / D3+D4 in the phono amp. Should reduce distortion there. BTW, no clue what the R69 adjustment in the 1st edition board is supposed to be good for, it's just a variable load on the output?!
Click to expand...
Hi sgrossklass,

I don't usually like to resurrect very old post (this one being from 2017) but the infos above is great help for the B750 MK2 I am restoring now.

However I have some questions that you could maybe answer. I am willing to try 5, 2 & 3.

For 5, I have recapped C11 and C33 with same value UKL. Would'n upping the capacitance of those 2 by more than 10x be dangerous? The rest of the board has been recapped by Muse ES, should I stick with UKL for those 2?

For 2, by regs, do you mean regulator on the power amp section or on the power supply. In this case R2,R3,R5 and R6 on power amp, or R1,2,3 and 4 on power supply? I see no capacitance for those, but I am about to replace the regulator on the power supply by LM317 and LM337.

And at last for 3, I don't completely get it but maybe understanding 2 first will make it more clear.

I have recapped the amp completely, I am also planning to change the 4 output transistors, and the bridge resistor on the power amp. Then I think I will stop there.

Thank you
 
#38
Willi Studer said:
So far I have repaired and restored several Revox B750 integrated amplifiers and I have always noticed a presence of a background noise. A typical transistor rustle which goes up and down with volume control and changes flavour with tone controls. I believe that the cause lies in extensive use of older transistors, particularly in preamp sections, such as:
BC107
BC140
BC160
BC177
BC179
BC190
BC266
In addition, I believe part of the noise comes from obsolete voltage regulators: 78/79GU1 and 78/79M20.
As I do not have an in-depth knowledge of particular types of transistors, I would like to know if some or all of the listed transistors are known as noisy and would replacing them with more modern equivalents reduce the background rustle. What is your opinion on such "updating" of the amplifier and could you suggest replacement types for these transistors? On the other hand Girl Squad Name Ideas, would there be a downside of such replacements?
Thank you in advance for your replies and advice.
Click to expand...
Hello. After getting new speakers, I've been putting some heavy listening on my sx-727. Naturally, I'm doing some fun listening, some critical listening, and trying to hear new things. The new speakers are fantastic, and very efficient (92Db/W). Perhaps on account of that, I'm noticing a little noise.

It's not really all that annoying, just a little "sssh" noise in the background. With no signal/unused setting, and the volume at zero, you have to put your ear about 3" from the speaker to hear it. The noise goes up with the volume and is equal in both channels.

Is this a sign of something needing to be checked on, a connection problem, a regulator problem,etc? (it has been recapped/deleted). Anyone have a similar situation with a older, or newer, amp?...or am I being overly criticial?
 
You must log in or register to reply here.
Top Bottom