What do you guys think about using the LT1763 for other power supplies in a DAC? Such as digital supplies, the SPDIF reciever analog supply, or even the DAC analog supply?
I'm looking for something small (8-pin SOIC is even a little big, but it's OK), low-noise and fast so doesn't require huge caps after the reg, and low quiescent current because it's battery-powered.
Is there a negative-supply version of the 1763? EDIT: Found the 1964 is the negative compliment to the 1761 series. 200ma version only though. Would trouble me a little bit for a DAC to have two different kinds of regulators on the pos & neg supplies, but then again I doubt most pos & neg regulators are really 'matched' even if they are similar.
I'm looking for something small (8-pin SOIC is even a little big, but it's OK), low-noise and fast so doesn't require huge caps after the reg, and low quiescent current because it's battery-powered.
Is there a negative-supply version of the 1763? EDIT: Found the 1964 is the negative compliment to the 1761 series. 200ma version only though. Would trouble me a little bit for a DAC to have two different kinds of regulators on the pos & neg supplies, but then again I doubt most pos & neg regulators are really 'matched' even if they are similar.
I used the 1963 in my CS43122 DAC (digital, analog, and clock supplies) and they are fine. Small and quiet. Presumably the 1964 is just the 19634 upside-down, but I haven't read its datasheet.
------------------pedro said:
I used the fixed voltage 5V version and simply shorted out and sense. The load is 3 cm away.
Re: your mods to XO
Not at all, just used the LT1763 as per data, but with a 5cm by 3cm ground plane to screen out 50 Hz.
---------------------pedro said:
Not at all, just used the LT1763 as per data, but with a 5cm by 3cm ground plane to screen out 50 Hz.
Also, regarding a JUNG regulator, will you get less noise than the 8-10uV quoted here for the LT solution?
------------------------------------------------------------
The Jung regulators is massive compared to the LT chip. I have seen 5 uV to 100k quoted for the Jung regulator.
However, from measurements and investigation, I have no confidence in quoted numbers and base my mods on insitu measurements using 1 MHz bandwidth, comparing peak to rms voltages (as low as possible and 1.5 in my case). A lot of 'sales' mesaurements have been made in idealised situations and do not turn out to be the case when connected to a clock or digital circuit.
------------------------------------------------------------
The Jung regulators is massive compared to the LT chip. I have seen 5 uV to 100k quoted for the Jung regulator.
However, from measurements and investigation, I have no confidence in quoted numbers and base my mods on insitu measurements using 1 MHz bandwidth, comparing peak to rms voltages (as low as possible and 1.5 in my case). A lot of 'sales' mesaurements have been made in idealised situations and do not turn out to be the case when connected to a clock or digital circuit.
That's true for every single number, on every single data sheet. They always quote the conditions under which measurement is made - they don't lie!
Their performance in real-world circuits relates to the load placed upon them and their dynamic o/p impedance and bandwidth / stability.
This is where Jung type reg's can be better, the dynamic o/p impedance is usually MUCH lower over a wider bandwidth. You're never going to actively regulate error signals at HF in a sensible reg circuit, for this passive schemes are needed to augment the reg's performance (LC), in order to retain the LF performance of the reg and add the HF performance of a good cap.
The noise level can be a red-herring too - the error amp I use in my clock supply has a high noise corner, relative to the audio band, but sounds better than a much quieter audio type amp in this application. The noise spectra is very important and may not be universally applicable between applications.
As a designer, these are just the details one needs to *listen* to and determine the best solution for a given application, as almost all solutions involve compromises somewhere.
Andy.
Well, it's all relative, they are about 5-6dB quieter than an LM317 (and the figures quoted on the data sheet, which are the same as a 317), but you seem to assume noise is the most important factor in a PSU, which for analogue supplies, it isn't (in the terms referred to here, anyway).
That's where the ears come into play. My analyser produces lovely plots, they tell me nothing whatsoever about how a circuit sounds, unfortunately.
I'm far from proposing a universal solution that works everywhere, but you cannot measure for that either. Any solution, in audio, for a specific application relies on measuremetn of basics then listening, backed up with measurement of details as one discovers them to further the design .
In a PSU that involves details that conventional measurement of noise, impedance and line rejection will tell you absolutely nothing about.
By all means provide examples to prove me wrong, but every major semi's data sheet I've just looked at explicitly quotes the conditions under which the measurements were made, maybe you need to look harder?
I'm not so sure we do actually, I did refer to application-specific solutions. Often designers take a different route to similar goals, but as someone who has done a lot of correspondance-based design, it's been interesting to note the absolute correlation of two pairs of ears, when assessing the same circuit, but in VERY different systems with diamatrically-opposed music.
I find that fascinating.
BTW, if I did make the comment that only comparative measurement was necessary (please direct me to the post) I retract it, unreservedly 😉
Andy.
In a thread about battery preamps I read it was usefull to add a capacitator to raise or lower (can't remember which) impedance.
Does this count for the XO3?
I plan to use a couple of 9v batteries and a RDE 2200uF@16v capacitator.
Does this count for the XO3?
I plan to use a couple of 9v batteries and a RDE 2200uF@16v capacitator.
I think answering that could be libellous 😉
I think it's partly related to being part of a system though, a genuine improvement in a system is likely to be perceived similarly, but the end result may be different.
e.g. I could give you a PSU design that would bring an improvement that I'd guarantee you would like, as would I. The difference is I may not like the resultant sound in your system, and vice versa.
My work has been with people who love music first, HiFi second, which may explain the total convergence of views.
I could of course be talking crap 😉
Andy
Connecting a TL1763 LDO, however, I managed to get noise down to 8-10 uV, with reduced hf components but some 50 Hz evident in the trace. The sound has now become even better, with clear and smooth presentation. There was no need to use inductors at all.
------------------------------------------------
I have now finally had the patience to rig up the 100 mA LT1761-5. Powering a 45.1584MHz XO, I am in the 10nV/sqrtHz noise again, with a 150 uF OSCon at the output .
This opens the possibility of a 2cm x 1cm XO plus low noise ps module for clock replacements.
Like the LT1763, the 1761 needs a 3x2 cm ground plane connected to -ve but under the module to reject a slight hum pickup.
------------------------------------------------
I have now finally had the patience to rig up the 100 mA LT1761-5. Powering a 45.1584MHz XO, I am in the 10nV/sqrtHz noise again, with a 150 uF OSCon at the output .
This opens the possibility of a 2cm x 1cm XO plus low noise ps module for clock replacements.
Like the LT1763, the 1761 needs a 3x2 cm ground plane connected to -ve but under the module to reject a slight hum pickup.
Re: Re: More on Clock Power Supply
Hello all,
Nice thread, to start with. Often, in supply discussions, I note that no difference is made between:
- The noise generated by the supply and / or other circuits, entering the circuit
- The interaction of the circuit (current) with the supply (and eventually other circuits)
Though both aspects have things in common, they again should be treated differently as well.
Furthermore, any common mode to differential mode conversion of "noise" currents, may affect the circuit, depending on layout. This could be considered as a 3rd aspect, whereas concept and implementation of any regulator can be spoiled, quite easilly.
Hum is a fine example of that........
all the best
fmak said:
Hello all,
Nice thread, to start with. Often, in supply discussions, I note that no difference is made between:
- The noise generated by the supply and / or other circuits, entering the circuit
- The interaction of the circuit (current) with the supply (and eventually other circuits)
Though both aspects have things in common, they again should be treated differently as well.
Furthermore, any common mode to differential mode conversion of "noise" currents, may affect the circuit, depending on layout. This could be considered as a 3rd aspect, whereas concept and implementation of any regulator can be spoiled, quite easilly.
Hum is a fine example of that........
all the best
Do I hear hints of a Guido Tent XO*+ PSU project?
Most other clock makers have some form of PSU... GT XO's are next?
Most other clock makers have some form of PSU... GT XO's are next?
- Status
- Not open for further replies.