iFi Audio Model "micro iPhono2": Creating Schematic - help needed

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
There are a lot of shortcomings on iFi's model iPhono2 (©2015-V1.3), from which I have create the schematic.

A) Printing error concerning the device inscription at SW3 for RIAA EQ Graph on bottom of this device

1) eRIAA (DIP 4/5) means in real life IEC (International Electrotechnical Commission) - user selectable IEC-RIAA option, i. e. additional 6db subsonic- resp. rumble filter. eRIAA "On" means actually subsonic "off", because the "ON" position create a bypass across C17/37/40 for the right channel and across C54/53/18 for the left channel. The correct labeling there should actually be "Subsonic on" in the "off" position of DIP4/5 and not "eRIAA".

2) what here IEC means in real live (DIP3/6) I don't know (maybe "eRIAA ? - go to
File:ERIAA.svg - Wikipedia).
In this case one could simply say that the terms "eRIAA" and "IEC" have been reversed, but I don't think, this is right.
DIP3/6 controls the CMOS IC's U4/U5 (4066). In the "ON" position PIN8 and 9 of U4/U5 and thus R74/53 in the RIAA network is shorted.
I doubt if the ordinary RIAA function works in the right matter, because I miss a resistor between the output stage (emitter Q14/9) and the input of the RIAA network, that is in the NFB path (C68/46, R69/87). Therefore, the eRIAA function is also doubtful.

B) Shortcomings affecting reliability and lifetime

1) To ensure a constant current consumption and a higher power consumption than with the iphono2 circuit itself (for additional 100mA), an additional resistor of 150 ohms was introduced between the inputs of the positive and negative rail.
This have folllow consequences:
- With the included, slightly too weak in-wall power supply from AMR Audio model "LN-1530" sometimes rises up an undesirable 50Hz hum (by use of an enhanced power supply, where are better parts in use, this issue disappears).
- Despite the fact that 3 pieces 51 ohm resistors were connected in serial mode to distribute the thermal load, some parts become seriously hot, particularly the electrolytic capacitor C4 and the others arround it.

C) Shortcomings affecting sonic quality

1) internal SMPS technology to create internal a neg. rail vs. GND - go to
https://www.onsemi.com/pub/Collateral/MAX828-D.PDF
This means one SPS supply in the pos. rail and two SPS supplies in the neg. rail. Not a really good approach - at least from my view. But a very cost effective solution, because the in wall power supply is an already exist part in a great amount for various other applications.

2) CMOS switches; for the EQ modes "Columbia", "DECCA" and "eRIAA" (what ever it is in real live) there are CMOS switches (4066) in use. A very great benefit at first look is here the fact, that the PCB wires to the actually switch are very short. OTOH - there is an important disadvantage: - see also the paper under
http://www.ti.com/lit/an/slyt612/slyt612.pdf
For example, the resistance of a CMOS switch in the closed position, referred to as the on-resistance or R-on, changes depending on the input voltage. This behavior is usually undesirable and can significantly distort the signal in some applications. This disadvantage prevails the advantage of short leads between the parts - so I think.

3) low quality capacitors both for voltage regulator and RIAA network
except the elcaps from ELNA, series 1644 (for C13/23) there are only no-name, low-cost smd elcaps arround the Non NFB voltage regulator in use - unfortunately with only 16VDC and 6V3DC. Consequence is bad sound already after a short period of use. In general - 50VDC until 63 VDC provides the lowest ESR at 50Hz and 10KHz according the datasheets and thus in this application with low voltage operation very long live time.
Even with the SMD RIAA caps one get the benefit of short wires and small space requirements, but also here prevails the disadvantage: Unwanted microphonic effect. Even in SMD versions are top quality caps available - go to
https://www.ibselectronics.com/pdf/wima_smd_capacitors.pdf
Unfortunately no small design is to realize by use of top quality parts.
OTOH there is no need to design such a RIAA preamp part as extremely small than this model.

4) Mute control unit: There are a jFET with drain at the output and source at GND. Most designers assume, that there are no disadvantages if there are only switches from Signal part to GND, but no switches in serial to the signal path. That's just right for mechanical versions. Unfortunately not with CMOS and like here jFET's.

The next step is the development of possibilities for improvement.
Where is the difference between "eRIAA" and "RIAA" - I see only the difference between shorted (black line) and not shorted subsonic filter capacitors (blue line) ? - go to the second image under "LAB REPORT". There should actually be a difference at the raised area above 18 KHz.
Strange is also the raised area below 100 Hz without subsonic capacitors at the outpout (bass boost). What is therefore the reason (in the schematic there is no such function to discover)?

Additional questions rises up:

1) what benefit is provided upon the use of R81 (4K7, for schematic go to second image post #12, connection betwen the L+R outputs in front of the 100R resistors R62+R27) ?

2) why subsonic function by C53/37, C18/17 and C54/40 in parallel mode located at the output instead at the input and why only 6db ?

3) where is the patented and
so called TubeState ClassA Tri-Brid Amplification circuit ? Maybe the corrected sketch on first attached image under post #18/21 ?

4) by selecting MC mode (MC = ●) the MM stage operates in the inverted mode (MM = -). Whithout consequences for the RIAA-EQ network ?

Unfortunately there are no answers under
Mono and Stereo High-End Audio Magazine: Interview with Thorsten Loesch of iFi - AMR about iFi iPhono

Thank you for advices.
In my PM box I discover the associated statement from the developer (for me no longer of interest, because the iPhono2 was replaced in the meantime by an own created version based on the PEARL-circuit)

I see you dissecting my little phone stage, but you seem to lack critical understandings why certain things are done the way the are.

I'll readily agree that fundamental circuitry (MC stage, EQ/Gain and buffer are very trivial, the good stuff happens in many of the details of the implementation, especially much of the signal circuit switching, which is transparent under operation.

As the MC Pre-Pre inverts polarity, it needs to be set up inverting and by doing so we can use shunt switches to mute the unwanted signal-path, instead of switching lo level signal.

Correctly applied electronic switches can be extremely transparent, more so than most mechanical solutions at the same price, which is how they were used. There are also reasons why some switches are mechanical (e.g. LF rolloff) as in those positions electronic switches were not appropriate.

The labelling of some switches is misleading if you attempt to interpret them in a traditional sense (as in switch X does this). If you set them as covered in the manual (yes, RTFM) the results are as intended.

By not reading the manual and drawing incorrect conclusions about what a switch is supposed to do instead of setting as the manual indicates - well the results are wrong.

The LF rolloff is a very clever trick. If you think about it it will become clear. It is actually very effective (more than you think) despite only being 6dB/8tve because of the way record warp (the main problem with a good turntable) interacts with the cartridge.

The power supply is not really a problem in the units being sold now. At the time the manufacturing of our power supplies (it is our design) was shifted from one subcontractor to another and the subcontractor tried to cut corners, which resulted in problems with this batch of power supplies, which normal production testing failed to pick up and which affected some but not all shipped units.

That contractor did not get more jobs and we worked with another, but it took time and by the time we could offer free replacements for the early batch power supplies (which we did to customers who had problems and went through the correct channels) a number of reviews had been published and the myth "the power supply is a problem" had been perpetuated.

All in all, if you would rather than trying to put down what was done in the iPhono (and II) you tried to understand why it was done, it may prove interesting and enlightening. It measures exactly as intended.

Both objectively and subjectively it performs very well, too good actually, few dealers want to stock it as they prefer to sell inferior sounding units at 2K Euro Phonos they make more money on.

Final one, you are still not very good decoding SMD's, there are still misidentified parts. We have very few parts that are not generic, though parts are generally of high quality.

So, try harder and see if you can fully decode what is going on. Start thinking why the CMOS Switch variable resistance with signal voltage matters jack in the circuit as it's done and what could have been done wrong and created major problems.

Thor
 
Last edited:
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.