The CD-7 still is a beast!!! 😳
I was trying to get the schematic from AMR home page. But I was directed to Ifi .. if you think the products are bad.. just go to there’s support site 😆
I was trying to get the schematic from AMR home page. But I was directed to Ifi .. if you think the products are bad.. just go to there’s support site 😆
Yes. iFi became the tail that wagged the AMR Dog and is now a Chinese company in Dongguan.
The products are not bad.
But like any small scale product it is possible to encounter long term reliability issues caused by either supply chain or things insufficiently accounted for in the design process.
There it matters that the manufacturer stands behind the product and gives owners support under FRAND, which implies for example fair pricing for spare parts not available on the open market.
Thor
I hope you understood I didn’t meant the AMR products? … I was referring to Ifi products… never have been impressed by them
I hope I will get a answer from them
I hope I will get a answer from them
Oh, ok.
Thank you. I designed most of these too. But much more constrained in cost.
What iFi products did you hear?
I suspect an Answer as such will be forthcoming. If it is the answer you hope for is another story.
Thor
Well it was some time ago.. think Zen dac? Or something. Got it from a friend cheap. Sharp cold sound 😬 and also the psu Supply products . Maybe good stuff but didn’t hear any difference
The Ifi silent switchers were also made to be thrown away it seems. Good performance but for a short time.
Ifi stuff is below the SMSL/Topping level with regards to longevity.
Ifi stuff is below the SMSL/Topping level with regards to longevity.
Thor. If you would modify an old player with the saa7220. Would you use a Regulator or a separate psu?
Zen DAC was meant to be a 99 USD Product.
Well, compared to most cheap gear out there, no. But I agree. Next to anything AMR, cold fish. It carters to a different market.
All the various iFi "problem fixers" are there to fix probems.
Like I had a guy who compared a 50 USD SMPS against a really expensive LPS and was disappointed it didn't go better that that. I mean what the fudge?
I have some I had since around '17, most run 24/7 and all are still going strong.
The little iFi SMPS were designed for 25 USD retail (nobody wanted to carry them at that price) and are basically "fully tricked out" commodity SMPS with an additional little module that kills noise.
Lifespan will inverse to current drawn. Usual design for these is ~ 5,000 Hours at rated current. Not ifi, any "commodity" supply. So if you see ANY switching power supply rated 5V/2.5A it is safe to assume that running it 24/7 at 2.5A will kill it in a year or so.
I know quite a few were used by peeps to non-stop power Raspberry Pi's that basically drew steady current beyond the rated limit. And yes, I know S&M oversold these on all sorts of things, metrics etc.
Again, i have many iFi products here, most from the very first production batch, so far non have gone T.I.T.S.U.P. (Totally Incapable To Support Usual Performance) on me.
I'm sure your experiences are genuine. From my own experience and being in the final line of tech support as well as on the board where we got regular reports on problems while I was there, my experience and impression is a little different.
I am not sure I understand the question?
If we are talking about cleaning up the noise from the notoriously noise SAA7220 I would use a chunky choke (TM) with a resonance beyond 2 X BCK out. Probably ~ 1mH or so.
And a mix of capacitors, including something like a 1,000uF/6.3V Os-Con and as many 1uF 1206 X7R capacitors as I can fit under the IC footprint with adhesive Kapton tape and adhesive copper tape forming ground and power planes, with the power plane partially overlapping the near full ground plane. You have ~ 32mm, so that allows for at least 10pcs 1uF 1206 capacitors easily, if you pack them 20 will fit. Here the impedance for one of these:
So 10pcs parallel will give < 1 mOhm impedance around the BCK Frequency (~5.6MHz) and < 100mOhm from 150kHz - 130MHz. With 20pcs we are looking at < 100mOhm 75kHz-300MHz.
The distance between Pin 12 & 24 will cause a bit extra inductance, but with a 32mm wide ground and power plane a lot of this probably disappears.
Thor
That was more than I understood 😅
But thank you anyway. Meant more of a simple solution of a separate power supply. Or use Like LM 7805 regulator
But thank you anyway. Meant more of a simple solution of a separate power supply. Or use Like LM 7805 regulator
Well, LM7805 will do jack.
A separate power supply seems overkill. Especially given that it's probably desirable to switch the DAC to Non-OS and only use the SAA7220 to generate 176k WCK used to synchronise the DEM Clock.
Thor
PS, add 2.7k series resistors on BCK, WCK and DATA before the TDA1541 pin's, after cutting the traces.
These should be SMD type 1206 size (so you can still see them) and one side should be directly at the pin, the other side linked to the trace with the solder resist scraped away.
Yup.
These should be SMD type 1206 size (so you can still see them) and one side should be directly at the pin, the other side linked to the trace with the solder resist scraped away.
Yup.
PPS, ignoring the how for a moment, on a top level here is what the CD-77 did (differently) from your usual TDA1541 Player.
1) No SAA7220, SRC4190 was used SYNCRONOUS as 2X/4X oversampling filter.
2) Separate reference clock with separate supply.
3) Shunt regulated power supply with CCS feed, completely floating (isolated from all else, grounds link at TDA1541)
4) ECL Buffer based signal conditioning after 74AC based reclocking with limited voltage swing and slew rate, to minimise logic signal feedthrough into the output.
5) Syncronise the DEM Oscillator to the audio clocks. I found that this step especially dropped measured HD on all chip's and wiped out most differences between generic and crown chip's (Taiwan marked chips tested only).
6) SMD Film Capacitors for DEM decoupling, plus large value OS-CON on MCB(s) with star ground to AGND.
7) Special "XY" feedthrough decoupling capacitors on all power pin's of all the IC's dealing with audio/clocks.
Most of this cannot easily be retrofitted to generic TDA1541 designs,
I have here a Marantz PM-75 integrated Amplifier with a TDA1541 based DAC inside that finally needs a refurb (random banging noises in one channel, very frequent now).
I will document my refurb plus any modifications and post in a separate thread on DIYA.
Thor
1) No SAA7220, SRC4190 was used SYNCRONOUS as 2X/4X oversampling filter.
2) Separate reference clock with separate supply.
3) Shunt regulated power supply with CCS feed, completely floating (isolated from all else, grounds link at TDA1541)
4) ECL Buffer based signal conditioning after 74AC based reclocking with limited voltage swing and slew rate, to minimise logic signal feedthrough into the output.
5) Syncronise the DEM Oscillator to the audio clocks. I found that this step especially dropped measured HD on all chip's and wiped out most differences between generic and crown chip's (Taiwan marked chips tested only).
6) SMD Film Capacitors for DEM decoupling, plus large value OS-CON on MCB(s) with star ground to AGND.
7) Special "XY" feedthrough decoupling capacitors on all power pin's of all the IC's dealing with audio/clocks.
Most of this cannot easily be retrofitted to generic TDA1541 designs,
I have here a Marantz PM-75 integrated Amplifier with a TDA1541 based DAC inside that finally needs a refurb (random banging noises in one channel, very frequent now).
I will document my refurb plus any modifications and post in a separate thread on DIYA.
Thor