|
Post by brucew268 on Aug 24, 2021 19:34:06 GMT
I thought I'd put here what I've tried lately in DIY or or tweaks and what I messed up, or found, instead of cluttering up other parts of the forum.
|
|
|
Post by brucew268 on Aug 24, 2021 19:36:44 GMT
This week I got a nice window into the sound of a filter resistor (PI) in the power supply of my preamplifier. I had a couple smoothing capacitors feeding the voltage regulator and some months back decided to change it from CC to CRC without really paying any attention to what that was doing or how, so it was implemented in a way that really didn’t affect anything positive or negative. Last week I corrected that to a proper CRC but with the resistor attenuating a fairly large voltage (4VDC). It worked well and I was having significantly better sound than I’ve heard out of my preamplifier in the last couple years.
I did notice that the metal film attenuating resistors were getting a bit warm and that they were using close to half their rated power. Not an emergency, but I thought to change over from a 0.6W resistor to a 2W to give a lot of spare capacity and it would run cooler.
To my complete surprise, it did not sound the same or better but worse! All the extra depth, dimension, and nuance and results of a low noise floor moved the wrong way felt a bit fuzzy and indistinct.
In an attempt to understand why, I decided to have a closer look at the datasheets (doh!). Well the 0.6W Tyco has a temperature coefficient of a nice low 50ppm/C, has a 1% tolerance, and is described a “very low noise”. The 2W Vishay metal film has a mediocre coefficient of 350ppm/C, a 5% tolerance, and is not described in regards to noise. Obviously I had gotten a rather better 0.6W than a 2W and that was the only change I had made.
So, I changed back to the 0.6W to see if the sound quality difference was repeatable or my imagination. Presto! Back to very good depth, dimension, nuance, and a sense of a lower noise floor, that despite the lower current headroom.
All that SQ change just from the choice of resistor that attenuated the voltage in the power supply CRC!
Next, I’ll try a wirewound to see if I can hear any difference.
|
|
|
Post by firebottle on Aug 25, 2021 9:07:50 GMT
Good info there. You could just double up the 0.6W resistor with two of twice the value in parallel, still low noise.
|
|
|
Post by brucew268 on Sept 5, 2021 14:40:10 GMT
A week ago I tried a wirewound as the filter resistor, a compact Vishay Sfernice RWM resistor which happens to be favoured by Joachim Gerhard in speaker crossovers. This is a very different application in a power supply filter, but the only noise should be from heat and at 3W on a 70mA circuit it should be 3x what is recommended. Interestingly, this resistor had to burn in for a bit. The bass was uncontrolled and boomy, while the HF details were nice and all there but just de-emphasised compared to the bass range, almost cartoonish. Found that the sound quality bounced back and forth for a 3 days then seemed to settle. After a week the bass was better controlled but still a bit over-emphasised and the treble details were a little soft. This could be that the RWM is not a good choice or that shaving off 4 volts with one resistor has too much effect on the SQ, though the metal film cap handled it reasonably (low noise, low temp coefficient). I've gotten a little sloppy in unsoldering several times on a stripboard so the trace is lifting in this spot next to a trace cut. No more experiments in that spot on the board! This weekend I've redone the power supply filter from CRC to be a CRCRC and will see how that does... used different traces than above. Since it has all new filter caps I won't feel comfortable that the sound is reliable until it has some hours on it, and Sunday afternoon HiFi always sounds like rubbish anyway.
|
|
Bigman80
Grandmaster
AA Founding Member & Bigbottle Audio Creator
Posts: 16,142
|
Post by Bigman80 on Sept 5, 2021 14:55:24 GMT
A week ago I tried a wirewound as the filter resistor, a compact Vishay Sfernice RWM resistor which happens to be favoured by Joachim Gerhard in speaker crossovers. This is a very different application in a power supply filter, but the only noise should be from heat and at 3W on a 70mA circuit it should be 3x what is recommended. Interestingly, this resistor had to burn in for a bit. The bass was uncontrolled and boomy, while the HF details were nice and all there but just de-emphasised compared to the bass range, almost cartoonish. Found that the sound quality bounced back and forth for a 3 days then seemed to settle. After a week the bass was better controlled but still a bit over-emphasised and the treble details were a little soft. This could be that the RWM is not a good choice or that shaving off 4 volts with one resistor has too much effect on the SQ, though the metal film cap handled it reasonably (low noise, low temp coefficient). I've gotten a little sloppy in unsoldering several times on a stripboard so the trace is lifting in this spot next to a trace cut. No more experiments in that spot on the board! This weekend I've redone the power supply filter from CRC to be a CRCRC and will see how that does... used different traces than above. Since it has all new filter caps I won't feel comfortable that the sound is reliable until it has some hours on it, and Sunday afternoon HiFi always sounds like rubbish anyway.
View AttachmentThe measurements from Joachim's speakers are generally poor in comparison to others in the price bracket, yet they get good press over in Germany, saying that, so did the Naim speakers. In truth, they'd have served better purpose as a bee hive. Anyway, interesting idea that.
|
|
Deleted
Deleted Member
Posts: 0
|
Post by Deleted on Sept 6, 2021 0:16:32 GMT
I don't know what the buggery bollocks you're talking about but regardless of that, what's you're favourite Fall album Bruce ? As this is now the only official Fall forum you do have to answer. Any gear stroking wire freezing screw head aligning ever giving Pepsi "hi fi" skullduggery can be attended to later. Thanks in advance for your cooperation. Yours ...The moderator.
.
|
|
Bigman80
Grandmaster
AA Founding Member & Bigbottle Audio Creator
Posts: 16,142
|
Post by Bigman80 on Sept 6, 2021 5:58:46 GMT
I don't know what the buggery bollocks you're talking about but regardless of that, what's you're favourite Fall album Bruce ? As this is now the only official Fall forum you do have to answer. Any gear stroking wire freezing screw head aligning ever giving Pepsi "hi fi" skullduggery can be attended to later. Thanks in advance for your cooperation. Yours ...The moderator. . 🤣🤣
|
|
|
Post by brucew268 on Sept 6, 2021 8:08:13 GMT
I don't know what the buggery bollocks you're talking about but regardless of that, what's you're favourite Fall album Bruce ? As this is now the only official Fall forum you do have to answer. Any gear stroking wire freezing screw head aligning ever giving Pepsi "hi fi" skullduggery can be attended to later. Thanks in advance for your cooperation. Yours ...The moderator. . I've been enjoying Sharon Robinson's "Everybody Knows". She was a collaborator of Lenonard Cohen's for many years and co-wrote some of his music. I really like her stuff and the different flavour she brings to some of those songs as woman lead singer.
|
|
|
Post by misterc on Sept 6, 2021 8:53:54 GMT
the Naim speakers. In truth, they'd have served better purpose as a bee hive. Oli I believe you meant expensive firewood
|
|
Bigman80
Grandmaster
AA Founding Member & Bigbottle Audio Creator
Posts: 16,142
|
Post by Bigman80 on Sept 6, 2021 8:56:14 GMT
the Naim speakers. In truth, they'd have served better purpose as a bee hive. Oli I believe you meant expensive firewood Absolutely
|
|
|
Post by brucew268 on Sept 6, 2021 17:00:00 GMT
Today’s lesson: if your power supply uses an adjustable voltage regulator (mine: LT1085) and the datasheet says to connect the voltage-adjust resistor directly to the case, DO IT. I had it that way until I was trying to fit heat sinks to the voltage regulators and decided I could deftly connect the resistor instead to the pin instead of the case. It looked good, but among a few other changes when I was fiddling with the power supply, I just wasn’t getting the sound performance I was used to, slightly lacklustre clarity, and low end definition just wasn’t what it used to be… and if anything my power supply changes should have made the sound better. Weeks later when it finally occurred to me that my issues might be related to this resistor, a lunch time task fixed it. Wow, that was a lot better, like I just bought a better class of preamplifier! BTW: the CRCRC caps from yesterday's post probably have more burn in hours to go in producing an open and transparent sound with only 24 hours on them, but it sounds pretty good. Does the CRCRC filter make the SQ better? I can’t really tell with any certainty. It may take a longer eval, but I know this config objectively produces a cleaner signal to the voltage regulator. See an example on Wayne Colburn’s scope below:
|
|
|
Post by misterc on Sept 6, 2021 17:22:19 GMT
Bruce consider using specific voltage units they can have a much improved impact o the sound plus a much improved PSRR
|
|
|
Post by brucew268 on Sept 7, 2021 6:55:06 GMT
Bruce consider using specific voltage units they can have a much improved impact o the sound plus a much improved PSRR That would be simpler, provided I can use 12v. But am I misreading the datasheet in the Fixed Voltage LT1083/LT1085? It looks to me like it says the adjustable regulators have much improved ripple rejection, and the charts on the two datasheets show 10-12db better on the adjustable, right?
"In applications that require improved ripple rejection the LT1083 series adjustable regulators should be used. With LT1083 series adjustable regulators the addition of a bypass capacitor from the adjust pin to ground will reduce output ripple by the ratio of VOUT/1.25V. See LT1083 series adjustable regulator data sheet."
|
|
|
Post by misterc on Sept 7, 2021 9:31:08 GMT
Bruce, personally I would be using different regulators other than stock monlithic or use these as pre regs and depending on current you could construct some LT3045 (can use in parallel) to further improve things. However if you have to use them, then I would suggest these for the +ve rail
Or there are plenty of Jung style regs on ebay which will outperform the monoithics easily even dual rail version, Oli uses one in his dac.
|
|
|
Post by brucew268 on Sept 7, 2021 12:21:19 GMT
Bruce, personally I would be using different regulators other than stock monlithic or use these as pre regs and depending on current you could construct some LT3045 (can use in parallel) to further improve things. However if you have to use them, then I would suggest these for the +ve rail
Or there are plenty of Jung style regs on ebay which will outperform the monoithics easily even dual rail version, Oli uses one in his dac.
I'm sure there are newer, quieter regulators than the old 3-pin, however, this was a part of some experiments by John Bau in 2009 to explore the under-appreciated role of impedance phase linearity rather than than ripple and noise as the primary focus. So, this power supply is based on his results as applied to the LT1085. He eventually moved on from current sense resistors to inductance experiments and ended up patenting a resulting design.
I did also read in passing someone recommending Walt Jung Shunt regulation for digital.
|
|
|
Post by misterc on Sept 7, 2021 12:48:33 GMT
Bruce,
Havig spent a fair few years using full TDR capability to board level/sripline and layer demarkation with more power suppiles for many industries than I care to remember, personally I would be using other methods of line voltage linearity.
Impedance pathways are very important far more crictal than most given crdence for, but you need to look at the TOTAL transmission line including the board, layouts, layer workings, copper pours, veer placement,numberand type of interface connectors, number and type of components etc. The list is long and deep lol and calculations quite complex. However as my father said cheat fair, I use curver tracers, 100 Mhz LCR's and full TDR devices these allow me to look really deep into all the aspects of elecrical pathways. I can even measure the impedance from an amplifier output, along upto 40 meters of cabling throught the x/over and back to amplifier. The unit can even dembed the individual componenets in the x/over. Very specialist equipment but essetnial for circuit board design work in high speed digital design and power distribution networks. You can use a redimentry VNA to achieve pretty much all that is need in audio design for a modest outlay. You can easily knock up some basic dual rails jung style and compare them if you wished
Its good to see people really looking into alternatives and thanks for sharing your expliots with the forum.
|
|
|
Post by brucew268 on Sept 7, 2021 13:35:29 GMT
Bruce, Havig spent a fair few years using full TDR capability to board level/sripline and layer demarkation with more power suppiles for many industries than I care to remember, personally I would be using other methods of line voltage linearity. Impedance pathways are very important far more crictal than most given crdence for, but you need to look at the TOTAL transmission line including the board, layouts, layer workings, copper pours, veer placement,numberand type of interface connectors, number and type of components etc. The list is long and deep lol and calculations quite complex. However as my father said cheat fair, I use curver tracers, 100 Mhz LCR's and full TDR devices these allow me to look really deep into all the aspects of elecrical pathways. I can even measure the impedance from an amplifier output, along upto 40 meters of cabling throught the x/over and back to amplifier. The unit can even dembed the individual componenets in the x/over. Very specialist equipment but essetnial for circuit board design work in high speed digital design and power distribution networks. You can use a redimentry VNA to achieve pretty much all that is need in audio design for a modest outlay. You can easily knock up some basic dual rails jung style and compare them if you wished Its good to see people really looking into alternatives and thanks for sharing your expliots with the forum. LOL, 'exploits' implies accomplishments and merit, to my mind, which is far from what I will ever achieve! I'm just bumbling along in a state of semi-ignorance in terms of circuit design and amd trying to learn just enough to recognise someone else's good design and assemble it well. There is the sense of fulfillment in manually assembling and soldering, not getting blue smoke,and having better sound as a result.
I recognise that your experience and expertise is worthy and way beyond my world. I'm cursed with itchy fingers, limited funds, and high sonic expectations.
|
|
|
Post by antonio on Sept 7, 2021 23:35:16 GMT
I like your last line there Bruce, but I think you're doing yourself an injustice with the previous sentence, and also having 'bumbling' in your thread title is more than harsh. Just my opinion
|
|
|
Post by brucew268 on Sept 8, 2021 9:26:03 GMT
I like your last line there Bruce, but I think you're doing yourself an injustice with the previous sentence, and also having 'bumbling' in your thread title is more than harsh. Just my opinion Thank you for your kind words. There are designers and there are assemblers and I'm rather toward the assembler end of the spectrum.
|
|
|
Post by antonio on Sept 8, 2021 11:33:52 GMT
What's wrong with being an assembler? When I look at the diy section on forums, most folk seem to be following a design that they found on the web, and let's be honest, how many of these designers are coming up with something new, they are usually tweaking previous designs that they have made, to which they got idea off the net in the first place
|
|
|
Post by electronumpty on Sept 8, 2021 11:36:40 GMT
Nothing wrong with assembling! I get a lot of satisfaction out of it. Sometimes I wish I know more about the electronic side of things and the theory etc. But that's a rabbit hole I don't want to go down! With this I think a little knowledge is a dangerous thing so will rely on the advice of those with more experience and better qualifications than me. 🙂
|
|
|
Post by brucew268 on Sept 8, 2021 12:26:28 GMT
What's wrong with being an assembler? When I look at the diy section on forums, most folk seem to be following a design that they found on the web, and let's be honest, how many of these designers are coming up with something new, they are usually tweaking previous designs that they have made, to which they got idea off the net in the first place Nothing wrong with assembling! I get a lot of satisfaction out of it. Sometimes I wish I know more about the electronic side of things and the theory etc. But that's a rabbit hole I don't want to go down! With this I think a little knowledge is a dangerous thing so will rely on the advice of those with more experience and better qualifications than me. 🙂 Exactly, happy to be an assembler! Also props to Hynes, Jacobs, Tony who really know their power supply stuff and design or fettle cutting edge kit.
|
|
|
Post by brucew268 on Sept 27, 2021 17:24:41 GMT
Today's note to self: stop changing more than one thing at a time, or even on successive days!! Supporting memo: more things matter than you tend to believe.
I rewired an RCA shielded IC to be a balanced (XLR) IC and placed it between the preamp and amplifier. Then I decided it was a good idea to move a couple small components which required re-dressing the cables in back, of course. This was 1-2 days after I soldered some changes to the preamplifier power supply.
About a day later, I decided my noise floor was not as low as it used to be, and also lost some clarity and definition both in HF and LF. Sometimes I tell myself that I got more ease and naturalness, but it was actually a loss of definition. A couple days of listening usually corrects that rationalisation.
Then I had to leave town for two weeks without any chance to investigate and correct... how annoying!
So last night I checked my soldering job (power supply) and fixed one possible issue. And noticed the DC offset had wandered in one channel, corrected now. Maybe it sounded a little better today?, but certainly was not back to 'before'.
Let's see... a newly terminated balanced XLR interconnect. one small component moved to a different shelf and it's DC cable rerouted. two small components moved across the shelf and stacked. had replaced the foam rubber feet on them with super soft silicone, as they are lightweight. cables, mostly digital and DC, re-dressed
That's a few things to check today. It turned out that both the soft silicone feet and the new location of the components were the cause of the poorer quality sound. Back to the 'before' configuration and my sound presentation returned!
Stop. Making. Multiple. Changes.
|
|
Bigman80
Grandmaster
AA Founding Member & Bigbottle Audio Creator
Posts: 16,142
|
Post by Bigman80 on Sept 27, 2021 18:43:53 GMT
Today's note to self: stop changing more than one thing at a time, or even on successive days!! Supporting memo: more things matter than you tend to believe. I rewired an RCA shielded IC to be a balanced (XLR) IC and placed it between the preamp and amplifier. Then I decided it was a good idea to move a couple small components which required re-dressing the cables in back, of course. This was 1-2 days after I soldered some changes to the preamplifier power supply. About a day later, I decided my noise floor was not as low as it used to be, and also lost some clarity and definition both in HF and LF. Sometimes I tell myself that I got more ease and naturalness, but it was actually a loss of definition. A couple days of listening usually corrects that rationalisation. Then I had to leave town for two weeks without any chance to investigate and correct... how annoying! So last night I checked my soldering job (power supply) and fixed one possible issue. And noticed the DC offset had wandered in one channel, corrected now. Maybe it sounded a little better today?, but certainly was not back to 'before'. Let's see... a newly terminated balanced XLR interconnect. one small component moved to a different shelf and it's DC cable rerouted. two small components moved across the shelf and stacked. had replaced the foam rubber feet on them with super soft silicone, as they are lightweight. cables, mostly digital and DC, re-dressed That's a few things to check today. It turned out that both the soft silicone feet and the new location of the components were the cause of the poorer quality sound. Back to the 'before' configuration and my sound presentation returned! Stop. Making. Multiple. Changes. That's a very a wise idea....one I too seem to ignore.
|
|
|
Post by firebottle on Sept 28, 2021 6:48:21 GMT
'Supporting memo: more things matter than you tend to believe.'
Oli and I learn't a while back that everything matters. No matter the components and transformers etc, positioning, case material, orientations, conductor and insulation make-up, the list goes on.
|
|
|
Post by brucew268 on Sept 28, 2021 10:37:24 GMT
'Supporting memo: more things matter than you tend to believe.' Oli and I learn't a while back that everything matters. No matter the components and transformers etc, positioning, case material, orientations, conductor and insulation make-up, the list goes on. My experience agrees with that. But since my knowledge and experience and not extensive, I was qualifying it to temper skeptical responses. Of course if I list the differences I've specifically experienced, I'm sure there are those who would balk. So it is what it is.
|
|
|
Post by misterc on Sept 28, 2021 10:48:20 GMT
Remember EVERYTHING that is attrached, sat on or connected to has an effect no matter how small on your system sound, sometimes just routing cables, correcly and making sure all the equipment is leveled up correctly will pay dividens imho
|
|
|
Post by brucew268 on Oct 4, 2021 17:54:53 GMT
Aargh! Yes everything matters. I dropped a different switch in just before the mRendu endpoint to see if it would make a difference to the Zyxel ES105a... well yes it did! Lot's more detail and propulsion, but I noticed that the sound was a bit like performing in a room that was too live acoustically. The switch came with no feet and a mfr warning that said the bottom panel could get a bit hot so my first listen was with generic flat silicone bumpers into the depressions made for such.
"What if I try rather softer bumpers?" but put them on the bottom panel just next to the existing ones. Since they weren't in the formed depression, they sat proud of the flat bummpers. That change took all of 60 seconds and did take away the "too live" sound and put slightly more body on the midrange, but it also created a slight bit of hardness.
OK, no bumpers, just the bottom panel on the bamboo shelf. Right, that's different again. Not hard and maybe more balanced frequency-wise, but it now sounds a bit busy and unseparated.
What if it was the placement of the bumpers rather than the softness? OK, softer bumpers in the depressions. well that's close to the flat bumpers with the "acoustically live room" sound, maybe slightly less than the flat ones.
It's a royal pain when a digital switch makes a noticeable difference with each change of footers, and I haven't found the right combination yet! It just shouldn't be this hard.
|
|
|
Post by brucew268 on Oct 5, 2021 19:09:17 GMT
We're following a theme here whether I like it or not! "It all matters." It turns out that when I changed the switch I also inadvertantly changed which was the last ethernet cable. In trying various things to tame the GF imbalance, I changed the last meter of LAN cable back to what was previous and that seems to have solved it. In theory the 23AWG solid core FTP is a better cable, but in the last meter it didn't suit. That's my story today.
Curiously, when I had the 23AWG LAN cable as the last, I had the sense of a lot of more detail but the soundstage had flattened and had no 3D sense to presentation, kind of similar to the effect of uncontrolled reflections in an acoustically live room. Odd effect from an ethernet cable/switch combo. Before I had router > long cable* > switch > 1M cable* > switch > 1M patch cable > microRendu > cardas USB adapter > DAC *23 AWG solid core FTP CAT6A (with the Cisco I no longer daisy-chain two switches.)
|
|
|
Post by brucew268 on Oct 6, 2021 17:34:25 GMT
A judicious 10mm disc of Silent Coat bitumen sheet is still helpful in one corner of the chassis -- the top, not the bottom. The bottom panel is broken up by fixings whereas the top panel is one uninterrupted expanse of some level of resonance.
|
|