In preferred mono mode, where the DACs are parallel they have high current for high S/N.
Laserscrape, I dont think SLA is in the same league as modern LiFePo4 battery, the internal resistance is incredible small on a high end LiFePo4 battery. If you want absolute lowest esr go for a big Lipo racing battery for quadcopters, small batteries can deliver hundreds of amps! Extreme short current combined with an unstable lithium ion/polymer chemistry makes them extremely dangerous to short out!
I assume SLA is wins, but the increase in ESR means that you need more of them to have small ripple. Earlier there was a user on this forum that was going to try a small psu combined with a lithium polymer battery and I think it would work quite well 
Going lithium instead of SLA would take up a heck of a lot less space also. I see block LifePo4 batteries on ebay all the time on ebay for cheap btw. Another benefit with large polymer batteries is that the cells are soldered internally so wire resistance is extremely low and the wires becomes a lot shorter since the packs doesn't take up much space. 22.2V lipo packs are very common. You could even buy an ebike battery 48V which comes with a BMS. There is a lot of alternatives to choose from at the moment
Going lithium instead of SLA would take up a heck of a lot less space also. I see block LifePo4 batteries on ebay all the time on ebay for cheap btw. Another benefit with large polymer batteries is that the cells are soldered internally so wire resistance is extremely low and the wires becomes a lot shorter since the packs doesn't take up much space. 22.2V lipo packs are very common. You could even buy an ebike battery 48V which comes with a BMS. There is a lot of alternatives to choose from at the moment Yes that's me. And I'm sad that i was born and i grew up on that way becaming that knd of guy who simply try only one thing and states that is good. Sorry for being in life.
Speaking of RF, have you ever tried using inductors or ferrite beads in series with the dac outputs to suppress RF coming from the dac?
No. Afraid it would distort the audio. Although I believe Benchmark DAC-3 has ferrites on the single-ended outputs (not sure if its to keep RF from getting out or from getting in, maybe both I guess).
Also, I do know of other proprietary solutions others have been used, but rarely see them used in most dacs.
Last edited:
Szabolcs,
Please accept my apology for framing my words in a way you found offensive, it was not my wish to be offensive to you. Sorry.
Mark
I made the stupid assumption that high voltage SLA and lifepo used higher voltage cells, not just more cells in series. 3.6V LifePo cells are easy to find at at similar price to SLA, and probably end up being cheaper than SLA due to shipping cost.I assume SLA is wins, but the increase in ESR means that you need more of them to have small ripple. Earlier there was a user on this forum that was going to try a small psu combined with a lithium polymer battery and I think it would work quite well
That surprised me though, so the higher voltage batteries will naturally end up with higher ESR?
All very interesting questions which I cant answer, but I've noticed that high current Lipos are short and fat vs long and thin for lower C rated batteries. So each layer have shorter pathlength to the wire tabs. So high current batteries probably do better in high frequencies than very long thin lipos. Batteries that are rolled probably have worse performance if only one end of the roll is connected to a tab inside.
Many people who tried batteries really like the sound it gives but its not easy to understand why it would sound different than an LDO. Perhaps the LDO have varying impedance at some point, they obviously have some overshoot.. Perhaps batteries doesn't have this characteristic?
Many people who tried batteries really like the sound it gives but its not easy to understand why it would sound different than an LDO. Perhaps the LDO have varying impedance at some point, they obviously have some overshoot.. Perhaps batteries doesn't have this characteristic?
eziitis,
After thinking about it privately for some time, I recently came across a document translated from Russian essentially saying the same as my thought: Sometimes for audio it isn't just low output impedance or precise voltage regulation that we care about.
Voltage regulator linearity may be the missing factor, some people are already convinced it is the key. It comes down to linearity of the error amplifier and other internal circuitry, or linearity of particular discrete regulator designs.
If one thinks about it, such a consideration may make a lot of sense from the perspective of some audio circuits that have poor or nonexistent PSRR. One can hear the sound of the associated voltage regulator come through in the audio. Do you like distortion from that source better or worse than from any other source?
Doesn't have to be just nonlinear distortion either, some types of linear distortion can be pretty objectionable if one realizes what is happening.
After thinking about it privately for some time, I recently came across a document translated from Russian essentially saying the same as my thought: Sometimes for audio it isn't just low output impedance or precise voltage regulation that we care about.
Voltage regulator linearity may be the missing factor, some people are already convinced it is the key. It comes down to linearity of the error amplifier and other internal circuitry, or linearity of particular discrete regulator designs.
If one thinks about it, such a consideration may make a lot of sense from the perspective of some audio circuits that have poor or nonexistent PSRR. One can hear the sound of the associated voltage regulator come through in the audio. Do you like distortion from that source better or worse than from any other source?
Doesn't have to be just nonlinear distortion either, some types of linear distortion can be pretty objectionable if one realizes what is happening.
Last edited:
@Markw, a few questions:
- What do you mean by linearity, if the relationship between voltage drop and current draw is proprtional?
- What was the best "sounding" regulator you had in use for a DAC like the ESS with no PSRR?
@all:
- A while ago I found a comparison of noise spectrum, output impedance against frequency etc. for different regulators somewhere, it included the AD797 in buffer config amongst others. Can't seem to find it anymore... Anybody?
- What do you mean by linearity, if the relationship between voltage drop and current draw is proprtional?
- What was the best "sounding" regulator you had in use for a DAC like the ESS with no PSRR?
@all:
- A while ago I found a comparison of noise spectrum, output impedance against frequency etc. for different regulators somewhere, it included the AD797 in buffer config amongst others. Can't seem to find it anymore... Anybody?
Since it has never been carefully studied that I know of, I would have to give an opinion. Maybe later.
Hopefully opinions are okay to post for this:
Depends on what part of the dac circuitry we are talking about. I like good LDO regulators for clocks, DVCC, VCCA, and that type of thing. Digital and RF.
For AVCC, a dual AD797 buffer with a low noise reference running from 15v rails is one of the very best I have found. When used for ES9028PRO I ended up adding some extra electrolytic filter caps not to far from the AD797 opamps to allow for good bass response. Have to keep an eye on caps though since it is known local bypass caps for AD797 can cause stability issues, but they still need to be used of course. Some series resistance for larger caps can serve to provide damping. See AD797 data sheet applications section for some specific advice.
For ES9038Q2M, used OPA1612 for AVCC buffering. They were chosen because some film caps were on the +-15v rails for the output stage (and too close to the AVCC buffers for no-fuss AD797 stability). Worked fine with OPA1612.
For the output stage I would say that conventional good linear regulators should be used. NJM7815/7915 are probably better than other parts with that number and better than LM317/337. Pre-regulators or RC filter stages may be needed before 15v regulators to fully suppress line harmonics from rectification. With the same considerations about pre-filtering or pre-regulation a Jung regulator can be a good choice too. diyaudio store sells a PCB to build one.
Good to use R-core transformers if possible since they don't conduct much HF/RF AC line noise into dac circuitry.
I like to provide an intermediate voltage bus of 7v or 8v to power any LDO regulators. 5v is a little too low for best PSRR from some LDOs so my preference to go a bit higher.
Last thing to add: If possible, a used Monster HTPS 7000 MkII from ebay can be a very good thing to add to a stereo system or to an audio test bench. Trust me
Last edited:
OK, looked at the datasheet yesterday because we were discussing about ESS AVCC regulation then. Would you care to elaborate what value the electrolytic caps were? What range for the resistors, I assume in the low single ohms? Did you also use ceramic caps next to the AVCC pins, I would assume yes.
If so won't the ceramic cap much more endanger the stability of the AD797 because of their much lower esr at high frequencies?
Ah, and one more question: Did you mean to put the resistor in series with the electrolytic cap only, or did you mean to put in series with the output and take feedback back to the negative input from after the resistor? Hope you understand what I mean
The AD797 data sheet shows to use a small resistor in series with an electrolytic bypass if larger than some specified value (would have to check data sheet again for the value). The X7R ceramic bypass caps should not have added resistors.
For larger electrolytic filter caps, presumably it depends on how far away they are from the AD797 since distance will add series R and L. For big film caps with very low resistance at high frequencies, you would have to figure it out.
The big electrolytic caps I added were only inches away but no problems. Although specified for very low ESR, the ESR didn't stay low very long as frequency went up. Apparently it wasn't at issue at frequencies where AD797 might have power pin stability considerations.
Since its hard to know all about caps from limited data sheet info and hard to know all about AD797 from published info, one is left to run some tests to see what works in one's application. I would only suggest to make sure the dac chip is otherwise fully powered up before making a test connection to an AVCC supply thought to be stable but untested for stability when connected to the dac chip. Particularly so if the AVCC buffer is powered from 15v rails. Don't know how much abuse the dac chips can take at the AVCC pins, is all.
For larger electrolytic filter caps, presumably it depends on how far away they are from the AD797 since distance will add series R and L. For big film caps with very low resistance at high frequencies, you would have to figure it out.
The big electrolytic caps I added were only inches away but no problems. Although specified for very low ESR, the ESR didn't stay low very long as frequency went up. Apparently it wasn't at issue at frequencies where AD797 might have power pin stability considerations.
Since its hard to know all about caps from limited data sheet info and hard to know all about AD797 from published info, one is left to run some tests to see what works in one's application. I would only suggest to make sure the dac chip is otherwise fully powered up before making a test connection to an AVCC supply thought to be stable but untested for stability when connected to the dac chip. Particularly so if the AVCC buffer is powered from 15v rails. Don't know how much abuse the dac chips can take at the AVCC pins, is all.
Last edited:
Hello! assembled i / v output according to your circuit on opa1612. the sound wheezes in both channels. if you pull out the left opa1612, then the sound in the right channel becomes clear, but in the left there is still sound and it wheezes. and vice versa, if you pull out the right opa1612, then the sound in the left channel becomes clear, but in the right there is a sound and it wheezes.