What are those circuits, do you know?
There may well be better parts depending on what the function is.
It’s usually not a good idea to do blanket replacement of parts across an entire board without some idea of how it will work, stick to one at a time and evaluate as you go.
Blindly replacing is even worse. Is there a display on the device? Maybe it looks nicer now?
If you can get part numbers from any regulators, there are often clues in the data sheets that tell you what the intended esr is for any output capacitors. Silmics is are good in analog power supplies, but never liked them for coupling, there are better parts for that.
There may well be better parts depending on what the function is.
It’s usually not a good idea to do blanket replacement of parts across an entire board without some idea of how it will work, stick to one at a time and evaluate as you go.
Blindly replacing is even worse. Is there a display on the device? Maybe it looks nicer now?
If you can get part numbers from any regulators, there are often clues in the data sheets that tell you what the intended esr is for any output capacitors. Silmics is are good in analog power supplies, but never liked them for coupling, there are better parts for that.
I did use OsCons where they are adjacent to the DAC (I actually tried a 220 Silmic but can swap out to OScon on the D>A) and providing power to the DAC, Panasonic FR for the OpAmps and for the initial smoothing Mundorf AG, the Silmic and Panasonic FM, I will try to increase the initial 3300 as I have some higher value Mundorfs to try
The top right is supply 5V, 12V and 15V, there are two Op amp board for input and output and there is a logic chip near the PS.
The 10nH or so of lead inductance is no relevant to series coupling in audio circuits, however it is to power supply bypass capacitors.
Dont squeeze the mundorf AG in because they are larger capacity without a reason; its unlikely to fit very well either, they are even larger, with thicker leads. How would you justify a handful of $20 capacitors there? why are you placing oscon where you have? what specific characteristic is going to suit the circuit better than what is installed? what IS the circuit installed? what IS the circuit you have changed to? do you know what they are filtering and have you traced out the GND and flow of power around the circuit? Do you need to add some more GND stitching wires in place of vias to tame/contain the potential HF spray of return current due to changing the way it is coupled by removing it from the PCB? will your change in cap change the filter characteristics due to having a different, perhaps much lower impedance? will increasing to too large a value and too low impedance adversely affect the regulator stability, causing it to oscillate, or potentially creating a 'resonant tank'?. how do your mods lead inductance oriented wrt the coils on the board?
These may seem like we are all picking on you, but these are all just reasonable questions anyone should ask themselves before modifying a mixed signal PCB. If you dont outline your objectives to yourself, how will you know if you have succeeded? Just replacing everything, sitting back and declaring it a win is a good way to spend a bunch of money, but not a great way to engineer an outcome IMO.
These may seem like we are all picking on you, but these are all just reasonable questions anyone should ask themselves before modifying a mixed signal PCB. If you dont outline your objectives to yourself, how will you know if you have succeeded? Just replacing everything, sitting back and declaring it a win is a good way to spend a bunch of money, but not a great way to engineer an outcome IMO.
Yes indeed and thats why it was mentioned by me and now marce. I was under the impression he was blanket replacing parts of the linear and switchmode supply; I dont think any AC coupling/filter caps have been altered as yet, so we arent talking just about audio frequencies. No schematic is posted. I'm not sure any attempt has been made to rough out what the schematic is, before changing it, nor documenting the finished state to aid in trouble-shooting, should such an action be required.
Im no expert as stated but more than happy to experiment, the Mundorf caps I already had they are the same value, low ESR and I have heard good things about them, yes they are larger- physically but the same value. They did provide a sonic upgrade, not life changing but then neither is £25.00
My aspiration is to obtain the best sound from the system as possible, if there was a perfect Processor then I would probably have it, if I thought the £16k Trinnov was it, my journey would be over. But to my ears my old naim AV2 still outperforms a £12k Acurus, which looked very interesting full of stereo DACs, no multichannel DAC which is what is implemented in the competition and sonically it was held in very high regard. Its far from poor but the old naim is better. So to spend £16k and have no guarantee that it will be better isn't palatable. So that leaves me the options with players and I have had them all.... Panasonic, Oppo 205, 203 to be fair I haven't tried the Reavon, but doubt it is worth spending £2k on to find out what I already probably know. So its getting the best out of my Panasonic 9000 and mini DSP
I will upgrade the power supply on the naim AV2 as I have a spare naim Hi-Cap supply I no longer need. but the AV2 is good.
I have had a fruitful journey on my 9000 and want to do the same with the miniDSP. I read lots of upgrades and opinions, and have done my best to understand what I am capable of trying and what might or might not take a difference. Lots and lots of opinions. Maybe I'm lucky but the changes made to the miniDSP have been worthwhile at least so far, only 30 hours in. The last mod to the 9000 was a mixed bag, so I will try another tweak.
I am on a journey here and I have already learnt a lot, I have some people I can take advice form and this forum. I learnt a lot with PEQs and multi subwoofer alignment, and have results that are excellent but it took an awful lot of reading, learning and measurements and listening. I am prepared to do this as its a hobby and passion.
If it helps I have no intention to upgrade my power amps or speakers as they are excellent, but getting truly excellent audio from a multichannel source is not as easy as you might expect. I auditioned the Dataset, Trinnov, Storm, Bryston in very sophisticated set ups and they are not exceptional for music, which for me is an easy way to evaluate a film set up. That and a few well known film tracks
Experimenting with a few pounds on caps and other parts, I am happy to try and see, I had many different coupling caps in the 9000 before settling on the Silmics and that was a big upgrade from the stock caps.
Ideally I'd have a perfect plan but I dont have the knowhow and no solution in real life is that simple anyway is it !?
I am grateful for the feedback, Im here to learn and get better sound in the process
My aspiration is to obtain the best sound from the system as possible, if there was a perfect Processor then I would probably have it, if I thought the £16k Trinnov was it, my journey would be over. But to my ears my old naim AV2 still outperforms a £12k Acurus, which looked very interesting full of stereo DACs, no multichannel DAC which is what is implemented in the competition and sonically it was held in very high regard. Its far from poor but the old naim is better. So to spend £16k and have no guarantee that it will be better isn't palatable. So that leaves me the options with players and I have had them all.... Panasonic, Oppo 205, 203 to be fair I haven't tried the Reavon, but doubt it is worth spending £2k on to find out what I already probably know. So its getting the best out of my Panasonic 9000 and mini DSP
I will upgrade the power supply on the naim AV2 as I have a spare naim Hi-Cap supply I no longer need. but the AV2 is good.
I have had a fruitful journey on my 9000 and want to do the same with the miniDSP. I read lots of upgrades and opinions, and have done my best to understand what I am capable of trying and what might or might not take a difference. Lots and lots of opinions. Maybe I'm lucky but the changes made to the miniDSP have been worthwhile at least so far, only 30 hours in. The last mod to the 9000 was a mixed bag, so I will try another tweak.
I am on a journey here and I have already learnt a lot, I have some people I can take advice form and this forum. I learnt a lot with PEQs and multi subwoofer alignment, and have results that are excellent but it took an awful lot of reading, learning and measurements and listening. I am prepared to do this as its a hobby and passion.
If it helps I have no intention to upgrade my power amps or speakers as they are excellent, but getting truly excellent audio from a multichannel source is not as easy as you might expect. I auditioned the Dataset, Trinnov, Storm, Bryston in very sophisticated set ups and they are not exceptional for music, which for me is an easy way to evaluate a film set up. That and a few well known film tracks
Experimenting with a few pounds on caps and other parts, I am happy to try and see, I had many different coupling caps in the 9000 before settling on the Silmics and that was a big upgrade from the stock caps.
Ideally I'd have a perfect plan but I dont have the knowhow and no solution in real life is that simple anyway is it !?
I am grateful for the feedback, Im here to learn and get better sound in the process
I dont think any AC coupling/filter caps have been altered as yet, so we arent talking just about audio frequencies.
Tks, no small value close surface mounted caps have been swapped out or experimented with yet.
That might not stop me trying some WIMAs though when Im brave enough 🤔
Tks, no small value close surface mounted caps have been swapped out or experimented with yet.
That might not stop me trying some WIMAs though when Im brave enough 🤔
If you truly want that to sound as best as possible, maybe try replacing the lmflyback regulator supply with separate, task appropriate supplies from an r-core transformer.
Might be issues to manage with sequencing on startup though.
Myself, when working on stuff like that, usually just stick to the coupling and small filter caps in the output stage. The rest usually makes such a small difference without major rework, and have made things worse. I didn’t always have the right parts either, but it is how I taught myself the practical end of how inductance can affect things in higher frequency applications.
Those Silmic caps can have a pronounced effect on the midrange, when used in power supplies. I couldn’t believe it until I had heard songs where certain vocals had actually disappeared, and some where it was sporadic, which was really odd. I don’t use them in anything now, except for maybe a subwoofer.
Might be issues to manage with sequencing on startup though.
Myself, when working on stuff like that, usually just stick to the coupling and small filter caps in the output stage. The rest usually makes such a small difference without major rework, and have made things worse. I didn’t always have the right parts either, but it is how I taught myself the practical end of how inductance can affect things in higher frequency applications.
Those Silmic caps can have a pronounced effect on the midrange, when used in power supplies. I couldn’t believe it until I had heard songs where certain vocals had actually disappeared, and some where it was sporadic, which was really odd. I don’t use them in anything now, except for maybe a subwoofer.
I have swapped the 12V supply Linear supply for an eBay (toroidal transformer), I guess you are talking about operate supplies for the 5V,12V and 15V ?
A good point regarding earthing, interesting your experiences with Silmics in power supplies, as this unit is only used for 5-70Hz its very much about bass
A good point regarding earthing, interesting your experiences with Silmics in power supplies, as this unit is only used for 5-70Hz its very much about bass
Wimas are the absolute worst power supply decoupling caps for digital/dsp; please dont do that. you may see quite a few 'audiophile approved' dacs like those from audioGD that surround DSPs etc with a bunch of wima red lollies. Know that they do this out of incompetence and fashion, rather than because its a good idea.
+1 about the Wimas being a poor choice for power supply decoupling caps for digital/dsp. while audio signals are roughly from a few Hz to ~25kHz, the SMPS (100 to 200 kHz SMPS switching frequency means harmonics to several MHz (see fourier series of a square wave)) and the DSP clock and I/O runs much faster than that. at MHz frequencies, parasitic inductances become very critical. a decent 100MHz or faster analog scope or DSO (Digital Sampling Oscilloscope) will be able to display the noise on the power rails.
Decrease the effectiveness of the digital bypassing with increased lead inductance, and all that MHz RF energy is going somewhere-- some of it will find its way into the analog input circuits and the ADC (analog-to-digital-converter), where the DSP will have to deal with it.
if you look at the datasheets for the WIMA caps (and this goes for competing brands of leaded film caps as well), the impedance vs frequency plots with reach a minima somewhere between a few tens of kHz and a few MHz, then start rising again at the self-resonance frequency (at this point the impedance becomes inductive). WIMA (of suitable value depending on other circuit characteristics) are OK for baseband audio signal (2-20kHz) signal coupling caps, but bad as digital and SMPS bypass.
note that ESR is typically specified by the cap maker at a single frequency that makes their part look good on paper.
Decrease the effectiveness of the digital bypassing with increased lead inductance, and all that MHz RF energy is going somewhere-- some of it will find its way into the analog input circuits and the ADC (analog-to-digital-converter), where the DSP will have to deal with it.
if you look at the datasheets for the WIMA caps (and this goes for competing brands of leaded film caps as well), the impedance vs frequency plots with reach a minima somewhere between a few tens of kHz and a few MHz, then start rising again at the self-resonance frequency (at this point the impedance becomes inductive). WIMA (of suitable value depending on other circuit characteristics) are OK for baseband audio signal (2-20kHz) signal coupling caps, but bad as digital and SMPS bypass.
note that ESR is typically specified by the cap maker at a single frequency that makes their part look good on paper.
Indeed, its one thing to take the view that 'everything matters' and that changes in all aspects/parts of digital components are just as audible and/or important as the analogue parts; it is another thing entirely to think that applying the same parts and (often quite sloppy) techniques used in lower frequency, will be tolerated or indeed beneficial in digital/RF/SMPS. You can end up with monstrosities using massive Duelund cast silver capacitors as bypass caps on 45MHz clocks, spruked as being somehow more transparent, or 'creating a more solid and convincing soundstage'.
Tks, I won't try that then !
The challenge is there are a lot of strong opinions on this stuff and they are often diametrically opposed. Much more of a challenge for me as I dont have the Electrical Engineering insight. Many are convinced about things that can by 'theory' make no difference and others have found that it seemingly does. Often for me its just easier to try, I learn I understand and I move forward. Back in the late 90's I took home £5,000 worth of interconnects between Pre and Power, a weekend was lost with blind testing between me a friend and my nearly equally obsessive twin brother. In the end the two favourites were notably better in my system than the 6 or so others. The top was an old BBC microphone cable, made up by a friend who was given a roll during a studio refurbishment. The second best was the second most expensive, the most expensive was near the bottom.....
So what did I learn when my credit card was refunded ? Cables do matter, their is little correlation between price (I didnt say cost !) and performance. That was a valuable weekend. I also found that a lot of silver plated copper cables seem to have a sound that is recognisable, almost a 'tinsel' sound in the higher frequencies.
I am pretty much alone with my AV solution, not moving to overhead channels and using a current HDMI solution. I wish this wasn't the case life would be so much easier ! Having said that my set up is far from poor, I'd just like to get closer to 7.1 competing with some of the better 2 channel set ups and if I can progress with an early Saturday morning with my soldering iron and £60 on caps, that's a very cheap hobby !
The only two AV systems I have heard that were on parr or better were £150k+ and £250k+
The alternative would be to try alternatives to the Acurus ACT4 I sold on when I went back to analogue out. This isn't simple or quick and its certainly not cheap
I'm learning so I'm a happy chap
The challenge is there are a lot of strong opinions on this stuff and they are often diametrically opposed. Much more of a challenge for me as I dont have the Electrical Engineering insight. Many are convinced about things that can by 'theory' make no difference and others have found that it seemingly does. Often for me its just easier to try, I learn I understand and I move forward. Back in the late 90's I took home £5,000 worth of interconnects between Pre and Power, a weekend was lost with blind testing between me a friend and my nearly equally obsessive twin brother. In the end the two favourites were notably better in my system than the 6 or so others. The top was an old BBC microphone cable, made up by a friend who was given a roll during a studio refurbishment. The second best was the second most expensive, the most expensive was near the bottom.....
So what did I learn when my credit card was refunded ? Cables do matter, their is little correlation between price (I didnt say cost !) and performance. That was a valuable weekend. I also found that a lot of silver plated copper cables seem to have a sound that is recognisable, almost a 'tinsel' sound in the higher frequencies.
I am pretty much alone with my AV solution, not moving to overhead channels and using a current HDMI solution. I wish this wasn't the case life would be so much easier ! Having said that my set up is far from poor, I'd just like to get closer to 7.1 competing with some of the better 2 channel set ups and if I can progress with an early Saturday morning with my soldering iron and £60 on caps, that's a very cheap hobby !
The only two AV systems I have heard that were on parr or better were £150k+ and £250k+
The alternative would be to try alternatives to the Acurus ACT4 I sold on when I went back to analogue out. This isn't simple or quick and its certainly not cheap
I'm learning so I'm a happy chap
I completely agree, as far as I can see today, no one cares about the choice of capacitors, it is important that they are audiophile like MUSE or WIMA, and the fact that they are a bad choice around digital circuits does not matter. The ADC/DAC is full of RF interference in the MHz band, not 20kHz or 100kHz and the choice of capacitor is as important as the termination and design of the PCB itself. Poor choice of capacitors at the start already limits the design of PCBs.
On the other hand we have an example where OS-CON is used in analog circuits which is again wrong.
Attachments
I’m learning so appreciate the responses
if I am to experiment or upgrade any suggestions welcomed
if I am to experiment or upgrade any suggestions welcomed
Regarding Wimas they do talk about FKP in digital circuits ?
https://www.wima.de/wp-content/uploads/media/WIMA-Audio.pdf
https://www.wima.de/wp-content/uploads/media/WIMA-Audio.pdf
To really make a difference you need some measurement equipment to check for interaction between the parts, ant-resonance is what the industry terminology is for that.
Higher speed circuits seem to thrive with high Q, low esr supplies, however that can be walking on a knife edge so to speak. I have made improvements, but I doubt are optimal without having done testing. A good ceramic clock decoupling cap came from an old tv box, have no idea what they are, but work amazingly well on the oscillator on my sd card player. Sometimes you get lucky I guess.
The solid polymer caps are great, I guess, when they can be used and not bog down a regulator, or not resonate.
As my collection of digital sources continues to grow, I am thinking my next purchase will be a hand held scope.
Higher speed circuits seem to thrive with high Q, low esr supplies, however that can be walking on a knife edge so to speak. I have made improvements, but I doubt are optimal without having done testing. A good ceramic clock decoupling cap came from an old tv box, have no idea what they are, but work amazingly well on the oscillator on my sd card player. Sometimes you get lucky I guess.
The solid polymer caps are great, I guess, when they can be used and not bog down a regulator, or not resonate.
As my collection of digital sources continues to grow, I am thinking my next purchase will be a hand held scope.
Phase, thanks I have been watching some videos about the ratio between low ESR and high Q, staring to understand a few basics between cap types and their basic performance metrics.