Bad advice, Raindrop's 1541s are trash. You can't 'tune out' bad grounding and audiophool component footprints.
I posted one of Thorsten's (AMR) critiques here but don't have the patience to look back in this thread.
Here's a collection of T's good advice on the 1541 from another thread.
http://www.diyaudio.com/forums/digi...any-good-tda1541a-dac-kit-55.html#post3505444
Last edited:
'Trash' seems a tad harsh ALD. Eldam did say they're good for playing with and tuning, I'd agree with him based on other boards I've bought from the same designer. The layouts suck that's for sure, but they're a great way to get all the parts together cheaply and for learning about how decent grounding affects the SQ. Incidentally Raindrop Hui is just the seller, not the designer.
Yes this mysterious one !
By the way it is in the AMR777 but the flag ship have still the TDA1541 if I remember. Some like G Tent prefer the lively 1543 and some the difficult biased TDA1545 for the pristine treble... Think the AD1865, AD1862, PCM54, PCM63 are good too if designed around by experts.
just think it depends on the quality of the design around like ECDESIGNS (John) show and teach us with patience. And the quality of the setup like TL do too (caps, wirewound Rohpoint resistor... et tout ça...).
@ Z600,
fuller advice :
A 100 dollars Raindrop Hui is not too expensive to listen music rapidly then beginn tunning, swap parts... it's not smd. I learned a lot with mine. And I didn't write the AD1865 is better tha TDA1541, just think it's different and another school !
that's why i talk to you about raindrop Hui: you talk about old and big regulators (can be wonderfull if fine tuned) and this thread is dedicated to ECDESIGNS development... here it's smd and micro LDO regulators...
But if your skilness is ok for smd and pcb design or beginn from scratch, Audiolapdance advise is not bad too... I should say it's a good advice !
if the goal is just listening music : you have too here on DIYAUDIO a bare pcb like the Red Baron for TDA1541 based on the authorised ECDESIGNS findings (but smd parts...). Maybe cheaper with the parts but don't know exactly. Never hear it.
oh, thank you Max, Best whishes for all
Last edited:
@Eldam
thank
happy new year
i search: http://www.diyaudio.com/forums/group-buys/167414-reference-tda1541a-dac-i2s-bus-architecture.html
thank
happy new year
i search: http://www.diyaudio.com/forums/group-buys/167414-reference-tda1541a-dac-i2s-bus-architecture.html
Last edited:
Hey Braxi, ya 'trash' is a bit harsh.
Here's T's critique:
Okay, lets do some analysis of the kit from pictures, shall we?
First, one quick look suffices to note that this "kit" is mostly a "cut'n'paste" job from a number of sources, not something that I would call a "design".
Let's start at the beginning. The WM8805 circuitry is "bog standard". But the decoupling, instead of following best practice, is based on audiophoolish notions of using through hole film cap's with arrangements that produce long loops on chips that run at frequencies of 10's of MHz all the way to around 100MHz (WM8805 internal operation). The powersupply are generic 3-Pin regulators, including for the critical reference clock for the DPLL in the WM8805.
The "Os-Con's" may look similar to the best series from Sanyo but are the cheaper Nichicon alternative. I have tested these myself and they perform anywhwere between three times (resonance frequency) and twenty times (minimum impedance) worse than real Os-Cons.
So the WM8805 section will work after a fashion, but there is no way it is ever going to offer anything like the phase-noise / jitter levels the WM8805 is capable of. In fact, I think despite all it's shortcomings, a well applied Cirrus Logic CS-8412/14 will have lower jitter.
Okay, the SAA7220 is a MASSIVE noise generator. It's decoupling matches that of the rest of the digital section, that is it has for all intents grassmudhorse all decoupling. The single ferrite bead between the SAA7220 PSU will help, but likely not remotely enough, especially as the "far side" of it is not decoupled...
Just poking a 'scope around the PSU pins on this PCB will be feast for the eyes, I'd predict noise levels in the 100's of mV around the filter and probably no better for the rest.
Okay, now our heavily jittered signal goes to our TDA1541.
This has the same kind of decoupling arrangements on the supplies as the rest of digital section, that is grassmudhorse incompetent and the DEM decoupling is about as bad as can be.
By the looks of it the regulators for the TDA1541's are some discrete ones, does not like shunts though. No matter how good or bad they are, they are too far from the IC's to do much good no matter their nature.
Overall the level of "engineering" around the TDA1541's seems an attempt to recreate the 1980's Philips/Marantz approache and getting an epic FAIL at even that task (I would recommend these players primarily as a lesson in "how not to do it").
Okay on to the analogue stages. These seem a mixture out of one of the common base I/V's we have seen here (there where numerous variation on these all with noisy LEDS's as references), I'd probably rather have a CEN or SEN, but they are probably the part of the whole circuit, thanks to the Chinese Kopy-Mao that calls itself "Engineer" having used a rather decent source to Kopy.
Sadly the Kompetence of the Kopy-Mao did not extend to kopying a discrete diamond buffer to place after the I/V so we find an Op-Amp.
I think almost anyone can build a better TDA1541 DAC on a piece of veroboard and using some copper foil, using what can be learned in the forums here.
Certain the "Reference" DIYA Groupbuy project would be a better choice, wretched as it is...
That's not just a trainwreck, that's an epic, high speed, Wenzhou level trainwreck, complete with completely clueless Wang Yongping style apologists saying "This is all very good and proper, whether or not you believe, I believe it."
CAVEAT EMPTOR MONITUS ES
Ciao T
---
From:
http://www.diyaudio.com/forums/digi...-any-good-tda1541a-dac-kit-4.html#post2853067
By chance my 1865 AD kit from Raindrop'hui has a CS8414 input circuit and I have to say that i didn't hear her TDA 1541 kits. I know the link you give.
I learn a lot by myself, step by step with try and error some advises and some reading like TL's post you show.
Still thinking Raindrop hui isn't to expensive because a good enough starting point (as Abrax understand it in a good way: not a final buy). As TL wrote this is because parts are not good enough, the design not too good, some parts sometimes fakes that you can learn a lot. The most important is to see, hear, understand, experiment what swaps, setups, new parts, differents PS feeding work to change the sound and final result improve the cheap PCB. 100 dollars to learn and have "music" from shelve before work on it is not too risky. I don't say it's nothing, 100 dollars is not nothing... i think it was the sense of your "bad advice". But the sound is not too bad and it's less expensive than commercial on shelf products for an already populated pcb.
Not by chance, my DIY kit Subbu maid by good designer (We can think it is with a Mark III !) has a WM88005 input circuit ! S...t ! But, saved... it depends on the PS before... and think the designer are good enough. Hé a man which doesn't like to write like to use its iron pen (private joke) and it's the sign of a good design. John use both but John is John... A Linux man, a sharing man.
The Subbu kit is the same price because of the free time spent by designer and the poor cost of bare chip. here you learn too because populate the pcb with half of the skill & knowledge needed can be... problematic. Altthough with an error of polarity printing of the pcb and with the after work of "diyer customer" most of them suceed with assembly and I think with a better result than a raindrop kit.
here we don't know what the poster really want but sure, the TDA 1541 Red Baron can explose you if you are not sure to be abble to populate the smd parts and understand where the trouble is if you break something.
All these 3 options (free given design 'Subbu), commercial (raindrop Hui) and half-free (Red Baron) are nearly the same price... with different philosophies an final results. This is not my goal to advertise Raindrop'hui kits and thinks that at more than 100 euros they are too bad. but at no more than 100 euros my AD1865 kit is good... enough... and a future second step after tweaking (with no regrets in relation to the price).
I read twice all the ECDESIGNS threads and more with some posts. To apply you must havr skill enough, do to smd and have some aditional knowledge to ber able to design a multilayer pcb from the bare advices.
No surprise here, the more expensive is the best... still DIY. Here after many years of tweaking at a low level, I can advise to buy a commercial product like the one of John (ECDESIGNS). It's not expensive, because tweaked month by month by an audiophil designer. I think because John share with focused adverstising (because you share you advertise) without spent many dollars in printed adds in print reviews : you can have a good player at near 1000 euros but not the double ! But maybe I'm naive. Myself I think John wanted half to maid a gift and make his advertising... which is normal, we are not in a communist world after all. John gave us a lot. Like others before : Thorsten, Tent, and many others.
John didn't decide of the best commercial way with a product which is not a disc player (it's a SD card player) because he choose the best sounding results; he said too, the DAC option alone is not the best in relation to the feeding with SD cards. Hé ! That's what we want : THE BEST QUALITY WITH SOUND ! Bravo.
At the end the Diyer decides. It's really rare if you have money "and" time at the same time. Most of us live with "or". But the pleasure of DIY is to learn too. One hand in the pocket, one hand on the iron pen.
i have the Subbu and the raindrop Hui and learn with both. If i have money to ECdesigns final product I will buy too !
Sorry John for the pollution on this long thread. I hop you continue to share. thanks for that. Interested to read... with my weak level in electronic but great needing in quality.
Eldam.
I learn a lot by myself, step by step with try and error some advises and some reading like TL's post you show.
Still thinking Raindrop hui isn't to expensive because a good enough starting point (as Abrax understand it in a good way: not a final buy). As TL wrote this is because parts are not good enough, the design not too good, some parts sometimes fakes that you can learn a lot. The most important is to see, hear, understand, experiment what swaps, setups, new parts, differents PS feeding work to change the sound and final result improve the cheap PCB. 100 dollars to learn and have "music" from shelve before work on it is not too risky. I don't say it's nothing, 100 dollars is not nothing... i think it was the sense of your "bad advice". But the sound is not too bad and it's less expensive than commercial on shelf products for an already populated pcb.
Not by chance, my DIY kit Subbu maid by good designer (We can think it is with a Mark III !) has a WM88005 input circuit ! S...t ! But, saved... it depends on the PS before... and think the designer are good enough. Hé a man which doesn't like to write like to use its iron pen
(private joke) and it's the sign of a good design. John use both but John is John... A Linux man, a sharing man.The Subbu kit is the same price because of the free time spent by designer and the poor cost of bare chip. here you learn too because populate the pcb with half of the skill & knowledge needed can be... problematic. Altthough with an error of polarity printing of the pcb and with the after work of "diyer customer" most of them suceed with assembly and I think with a better result than a raindrop kit.
here we don't know what the poster really want but sure, the TDA 1541 Red Baron can explose you if you are not sure to be abble to populate the smd parts and understand where the trouble is if you break something.
All these 3 options (free given design 'Subbu), commercial (raindrop Hui) and half-free (Red Baron) are nearly the same price... with different philosophies an final results. This is not my goal to advertise Raindrop'hui kits and thinks that at more than 100 euros they are too bad. but at no more than 100 euros my AD1865 kit is good... enough... and a future second step after tweaking (with no regrets in relation to the price).
I read twice all the ECDESIGNS threads and more with some posts. To apply you must havr skill enough, do to smd and have some aditional knowledge to ber able to design a multilayer pcb from the bare advices.
No surprise here, the more expensive is the best... still DIY. Here after many years of tweaking at a low level, I can advise to buy a commercial product like the one of John (ECDESIGNS). It's not expensive, because tweaked month by month by an audiophil designer. I think because John share with focused adverstising (because you share you advertise) without spent many dollars in printed adds in print reviews : you can have a good player at near 1000 euros but not the double ! But maybe I'm naive. Myself I think John wanted half to maid a gift and make his advertising... which is normal, we are not in a communist world after all. John gave us a lot. Like others before : Thorsten, Tent, and many others.
John didn't decide of the best commercial way with a product which is not a disc player (it's a SD card player) because he choose the best sounding results; he said too, the DAC option alone is not the best in relation to the feeding with SD cards. Hé ! That's what we want : THE BEST QUALITY WITH SOUND ! Bravo.
At the end the Diyer decides. It's really rare if you have money "and" time at the same time. Most of us live with "or". But the pleasure of DIY is to learn too. One hand in the pocket, one hand on the iron pen.
i have the Subbu and the raindrop Hui and learn with both. If i have money to ECdesigns final product I will buy too !
Sorry John for the pollution on this long thread. I hop you continue to share. thanks for that. Interested to read... with my weak level in electronic but great needing in quality
.Eldam.
Back to the power supply, I believe the -15v supply is the most important for the best sound. It feeds current sources in the 1541. As far as the regulator for this -15v, I was told that a current source should not feed a current source (1541). It is better to have a voltage source (series reg) feed a current source (1541). Not a shunt reg (current source) feeding the current sources in the 1541 (-15v).
Anyone have any thoughts on this?
Anyone have any thoughts on this?
I would not call a shunt regulator current source. This is confusing.
For supply voltage -15 V can be used as a shunt regulator and serial stabilizer.
Personally, I prefer a shunt regulator. He has a constant current consumption.
Hi Rick,
Yes, -15vdc is not only important, but hypercritical. I've used a lot of kind of regulators. Finally the one I'm using is this:
A separate power transformer, hyperfast diodes, a discrete regulator, 4x22.000uf/25v capacitors and closest as possible to the -15vdc pin, a 1uf capacitor from Illinois capacitor, the smd 1210 size ones for high frequency.
The regulator has to be kept as simple as possible, using only one active element to regulate the voltage, do not use op amps for error amplifier. I have been using a lot of transistors, finally the BD244C and BD243C for the positive has been chosen. These ones have the best HF response.
Of course for the +5&-5 you must use the same principle, but, if you want, you can use instead of 2x22.000uf/25v, only 2 units........
Kindest regards, and I hope this helps a little.
Yes, -15vdc is not only important, but hypercritical. I've used a lot of kind of regulators. Finally the one I'm using is this:
A separate power transformer, hyperfast diodes, a discrete regulator, 4x22.000uf/25v capacitors and closest as possible to the -15vdc pin, a 1uf capacitor from Illinois capacitor, the smd 1210 size ones for high frequency.
The regulator has to be kept as simple as possible, using only one active element to regulate the voltage, do not use op amps for error amplifier. I have been using a lot of transistors, finally the BD244C and BD243C for the positive has been chosen. These ones have the best HF response.
Of course for the +5&-5 you must use the same principle, but, if you want, you can use instead of 2x22.000uf/25v, only 2 units........
Kindest regards, and I hope this helps a little.
These two "current sources" are different beasts, as far as I see, the internal one tries not to maintain a stable voltage like the first one.
Thanks for all the replies, I have found that the -15v supply to the 1541 has more affect on the sound than any on the other supplies. That is why I am trying to make it as good as possible. I see Galeb has also found out how important the -15v supply is to good sound.
I wonder how many people are using series regs and how many are using shunt regs? I am using a series right now.
I wonder how many people are using series regs and how many are using shunt regs? I am using a series right now.
Hi Rick,
You damned right,
How the -15V are important is a known fact since the first tweaks in the end of the 80s'.
Look for old e-papers on the web learn you a lot too about it. ECDesigns explain a lot how the ground design and routing are important too like others :T Loesch and some famous diyers also.
I would like to have a benchmark of the most efficient and simple ones. A very stable - 15 v with a triple Pi filter (coil no resistor) can be a beginning... I'm curious to know the results of the smd low noise reg. ?!
Tell us if you find interesting results
You damned right,
How the -15V are important is a known fact since the first tweaks in the end of the 80s'.
Look for old e-papers on the web learn you a lot too about it. ECDesigns explain a lot how the ground design and routing are important too like others :T Loesch and some famous diyers also.
I would like to have a benchmark of the most efficient and simple ones. A very stable - 15 v with a triple Pi filter (coil no resistor) can be a beginning... I'm curious to know the results of the smd low noise reg. ?!
Tell us if you find interesting results
I still follow the thread when I have time.
Anyway, the last TDA1541A module I tested had local, discrete shunt regulators driven by external capacitance multiplier & current limiting circuit.
In the new projects like the SD1-player and the UD1-streamer (attached picture shows early prototype) I use “quatrode” series regulators exclusively. These consist of a filtered ultra low noise precision voltage reference that drives a unity gain super buffer. These voltage regulators are more efficient compared to shunt regulators and allow the use of small mains transformers with very low stray capacitance.
TDA1541A -15V is most sensitive to power supply ripple (check datasheets) and it powers sensitive on-chip analogue circuits. So clean -15V power supply is advisable.
Attachments
Hi Rick Miller,
Sorry for the late reply, too much to do, too little time.
The quatrode regulator basically consists of a reference voltage and a buffer circuit. I use a LM329 (approx. 7V). It’s output voltage (and its noise) are attenuated by a resistive divider. The resistive divider output is filtered in order to get a clean reference voltage. In the SD1-player I need multiple output voltages (4 in total) that are all derived from a single LM329.
The quatrode circuit consists of MOSFETs and Darlingtons, it is an universal building block with 4 connections, in, out, minus, and plus. It has very high input impedance, very low output impedance, large bandwidth, low noise, excellent step response and line / load regulation.
It is also possible to use the quatrode circuit as shunt regulator. For the quatrode shunt regulator the reference voltage equals Vout - 0.6V, for quatrode series regulator the reference voltage equals Vout +0.6V.
Quatrode series regulators are useful when load current changes significantly, digital circuit (processor) for example. Shunt regs have a specific bias current so peak current is limited by it, these shunt regs are suitable for powering loads with (almost) constant load current.
I use 3 quatrode shunt regulators in the new TDA1541A project (The TDA1541A DAC project has been finally picked up again).
“small and nice” as we call it over in Holland. This USB DAC prototype is only 12 millimeters high and fits in the palm of your hand. It is based on SD1-player circuits like local, discrete masterclock, slaved source (asynchonous USB), quatrode regulators and so on.
Sorry for the late reply, too much to do, too little time.
The quatrode regulator basically consists of a reference voltage and a buffer circuit. I use a LM329 (approx. 7V). It’s output voltage (and its noise) are attenuated by a resistive divider. The resistive divider output is filtered in order to get a clean reference voltage. In the SD1-player I need multiple output voltages (4 in total) that are all derived from a single LM329.
The quatrode circuit consists of MOSFETs and Darlingtons, it is an universal building block with 4 connections, in, out, minus, and plus. It has very high input impedance, very low output impedance, large bandwidth, low noise, excellent step response and line / load regulation.
It is also possible to use the quatrode circuit as shunt regulator. For the quatrode shunt regulator the reference voltage equals Vout - 0.6V, for quatrode series regulator the reference voltage equals Vout +0.6V.
Quatrode series regulators are useful when load current changes significantly, digital circuit (processor) for example. Shunt regs have a specific bias current so peak current is limited by it, these shunt regs are suitable for powering loads with (almost) constant load current.
I use 3 quatrode shunt regulators in the new TDA1541A project (The TDA1541A DAC project has been finally picked up again).
“small and nice” as we call it over in Holland. This USB DAC prototype is only 12 millimeters high and fits in the palm of your hand. It is based on SD1-player circuits like local, discrete masterclock, slaved source (asynchonous USB), quatrode regulators and so on.