D-Noizator: a magic active noise canceller to retrofit & upgrade any 317-based V.Reg.

Jump to Latest
K

kumori

Member
Joined 2011
I have a question about the output voltage stability.
Is it „normal“ that the output voltage increases after installing the denoiser?
Also, I get load depended output voltage variation.

My setup was first tested only with the lm317, set to 51 V with R1,R2 = 383r and 15k.
Protection Zeners are in place and it works as it should.
After adding the denoiser with ZTX851 and R4, R5 = 8,2k and 300k, -> Vce= 5V-ish, the output voltage increases to around 54V. Also, the output is quite load dependend. It increases a couple of 100 mV when going from no load to 10mA to 20mA load current. At 20mA, the output is above 55V.
Somewhat strange for a regulator, usually one expects it to go the other way with increasing load...

Obviously something is wrong, and I would be glad about any pointers...
The Lm317 stage runs a little bit low on current, the 383Ohm give around 3,2 mA, but this should be fine. The current through the ZTX851 is around 5-6mA, this should probably also be fine.
Might this regardless be a problem of DC stability caused by the lower currents?

Thanks,
Florian

Update:
Further measurements show:
The voltage across R1 rises to 1,35 V, which explains the voltage increase.
After removing R4 and R5 the output voltage is stable again with 51V.
 
Last edited:
T

tombo56

Member
Joined 2019
Paid Member
I’ve done my backlog of catching with 3 months worth of posts in this thread. Spotted Elvee’s correction of LT implementation schema. It doesn’t change much in the simulation but real circuit probably could not work well.

@carlmart
I see now why did you ask if there is better LT1963 model as you have surely done simulations and evaluation of this regulator. As for the PSSR, basic results are worse then with LM317. I believe that this is right. LM317 starts with 10 – 30 dB better PSSR compared to the LT1963.
However, PSSR and low noise is IMO not everything we need for audio. LT1963/LT3015 are optimized for fast transient response. That is why I suspect that they could be better in total over LM solutions.

Do you consider denoiser regulators for use with DAC? I do, among other usages, and that is why I’m poking around this design. Cdsgames didn’t come back with any result and he is the most qualified to make judgment on their usage for DACs. This could mean three things:
- He didn’t have time to evaluate
- Results are bad
- Results are so good that he doesn’t want to educate his competition
Let’s hope for the third outcome
 
C

carlmart

Member
Joined 2001
preciseaudio.com
Tombo56

I'm quite aware of the limitations on whatever simulations you can perform, be it with LTSpice or any other program.

The fun or practicality in using the Denoiser or Dienoiser with 3X7 regulators is that you may already have some or you can get them quite cheap from eBay, with little chances of the regulator chip being fake, I believe. Not the case with some LTs, particularly those with low noise LDO types, like LT3042 or 3045. I'm not sure about the LT1963 or LT3015, but they are not cheap on Digi-Key.

You can take (or should prefer) a non assembled 317/337 regulator kit such as this, add two Denoiser/Dienoiser piggbacks, and end up with a top regulator, comparable to the best around, including Jung's superregulator

LM317 LM337 AC/DC Adjustable Filtering Power Supply Voltage Regulator PSU DIY | eBay

My application would be DACs, RIAA preamps (discrete and IC) and several others,

I wonder who is Cdsgames and where I can find the tests he performs.
 
T

tombo56

Member
Joined 2019
Paid Member
“cdsgames” is user on this forum and from company Allo, producers of Boss, Katana and Revolution DACs (IDK if owner or/and chief designer). They also produce dedicated LPS for their DACs. He said that he will test denoiser designs with LDO regulators and report back.
 
Elvee

Elvee

Member
Joined 2006
kumori said:
I have a question about the output voltage stability.
Is it „normal“ that the output voltage increases after installing the denoiser?
Also, I get load depended output voltage variation.

My setup was first tested only with the lm317, set to 51 V with R1,R2 = 383r and 15k.
Protection Zeners are in place and it works as it should.
After adding the denoiser with ZTX851 and R4, R5 = 8,2k and 300k, -> Vce= 5V-ish, the output voltage increases to around 54V. Also, the output is quite load dependend. It increases a couple of 100 mV when going from no load to 10mA to 20mA load current. At 20mA, the output is above 55V.
Somewhat strange for a regulator, usually one expects it to go the other way with increasing load...

Obviously something is wrong, and I would be glad about any pointers...
The Lm317 stage runs a little bit low on current, the 383Ohm give around 3,2 mA, but this should be fine. The current through the ZTX851 is around 5-6mA, this should probably also be fine.
Might this regardless be a problem of DC stability caused by the lower currents?

Thanks,
Florian

Update:
Further measurements show:
The voltage across R1 rises to 1,35 V, which explains the voltage increase.
After removing R4 and R5 the output voltage is stable again with 51V.
Click to expand...
The symptoms point to an oscillation problem.
I cannot say exactly what the primary cause is, but you should inspect the bypassing scheme and the way the original regulator and the denoiser are wired.
If you post some pics of your setup, it would probably help.

Note that the denoiser is supposed to be used in the "normal" range of the regular 317: up to ~30V.
Yes, in theory only the differential voltage matters, but in practice things are somewhat different, and using a 317 with voltages higher than the abs max delta rating is a disaster in waiting.
I have seen many such disasters in decades (and had to find a fix).

Anyway, a 317HV should work more safely, and I provided guidelines about using a TL783, which should be applicable here too.

kimschips said:
Can I also ask what about lm317 /337 output voltage about 5 and 8V DC can the denoiser work with that low voltages?
Kim
Click to expand...
It can, but with very slightly inferior performances.
If you scale the resistors value by a factor of ~2, the performances will remain the same.

kumori said:
Input/Output differential is 10V maximum.

Anyway... how does it happen that there are 1,35V across R1 when installing the denoiser?
Click to expand...
Probe the collector of the transistor with an O-scope, you will probably see an oscillation.
 
D

dotneck335

Member
Joined 2012
Elvee said:
Yes, in theory only the differential voltage matters, but in practice things are somewhat different, and using a 317 with voltages higher than the abs max delta rating is a disaster in waiting.

I have seen many such disasters in decades (and had to find a fix).
Anyway, a 317HV should work more safely, and I provided guidelines about using a TL783, which should be applicable here too.
Click to expand...
What do you recommend for the negative side? The LM337HV devices appear to be no longer available.
 
C

carlmart

Member
Joined 2001
preciseaudio.com
From my experience with higher voltage regulators, where I opened a specific thread for and where Elvee provided generous help, discrete regulators seem to be the answer.

On sims they do no look as good, but on the application I was using it for (low current power amps regulators) I think it was more than enough.

To go better I think using the TL783 with Denoiser seems to be a good option. The only problem that chip has is the rather high input/out voltage ratio you will need.
 
terranigma

terranigma

Member
Joined 2007
Elvee said:
kimschips said:
Can I also ask what about lm317 /337 output voltage about 5 and 8V DC can the denoiser work with that low voltages?
Kim
Click to expand...
It can, but with very slightly inferior performances.
If you scale the resistors value by a factor of ~2, the performances will remain the same.
Click to expand...
You mean all resistors in circuit? I think most of people here planing to use single rail de/dienoiser on their DAC projects like myself. It seems we need confirmed values (or complete circuit diagram) for 5v and 3.3v versions.

@tombo56 can you generously provide 5v and 3.3v BOM too? Are you working on 15V analog version specifically?
 
Last edited:
K

kumori

Member
Joined 2011
Elvee said:
The symptoms point to an oscillation problem.
I cannot say exactly what the primary cause is, but you should inspect the bypassing scheme and the way the original regulator and the denoiser are wired.
If you post some pics of your setup, it would probably help.

Note that the denoiser is supposed to be used in the "normal" range of the regular 317: up to ~30V.
Yes, in theory only the differential voltage matters, but in practice things are somewhat different, and using a 317 with voltages higher than the abs max delta rating is a disaster in waiting.
I have seen many such disasters in decades (and had to find a fix).

Anyway, a 317HV should work more safely, and I provided guidelines about using a TL783, which should be applicable here too.


It can, but with very slightly inferior performances.
If you scale the resistors value by a factor of ~2, the performances will remain the same.


Probe the collector of the transistor with an O-scope, you will probably see an oscillation.
Click to expand...

So, I scaled down the voltage to around 32V and still have the same problem, the voltage rises under load. The problem should not immediately be connected to the higher output then. I need a negative version, and alternative regulators and the HV type are not applicable to the negative rail. Or is there anyone I don‘t know?

Indeed there is oscillation at the collector of the transistor, around 400mV p2p at 12kHz. Thanks for the tip. It was not really visible at the output, there it was only around 4mV. (I used 1:10 probes so it was only barely visible 400uV at the output)

For the layout, I used the pcbs that I showed earlier in this thread:456
I will check layout and solder work for anything suspicious.
Thanks for the help. :)
 
You must log in or register to reply here.
Top Bottom