Fosi Audio Box X5 Phono Preamp Review

[Roland68]

I agree SMD’s are more difficult to replace than more “traditional” components. One could postulate that if you are making a device that is basically “non-repairable”, then it better be very reliable and long-lived from an environmental perspective. Phono preamps are hardly a “commodity”, so making them “cheap” is just a way to entice limited demand sales. Obviously, the Fosi X5 hasn’t been around long to understand its reliability.

This has nothing to do with Fosi, but is now a worldwide problem in all areas where particularly small and inexpensive SMD components are used.
SMD components like simple resistors and capacitors are very easy to replace, and there are special tools for this.

However, since many of these components are connected to ground, there is no way to isolate the fault if one fails. These problems almost never occur with through-hole components and larger SMD packages.
The savings in production are very small, but increase with the number of components.

 

[Heinrich]

... SMD packages.

The FOSI X5 represents a compromise. This lies in the use of the OPA1612 as the input amplifier. For MC and MI (moving-iron, e.g., Grado), it is an excellent choice that drives these systems to the limits of what is physically possible. For MM (moving-magnet, AT, Ortofon, etc.), the situation is different.
The current noise of the OPA1612 quickly becomes disadvantageous here. A NE5532 would be more suitable. The reason is that the comparatively high current noise of the OPA1612 is converted into voltage noise by the high generator impedance of MM cartridges. The disadvantage amounts to roughly 8dB.
Using the NE5532, a Fosi X5 would be better equipped if only MM were considered. Fosi would have been well advised to make the input op-amp replaceable, but that isn’t feasible due to the different package formats—SMD versus DIP, of the opamps in question.

So much on SMD and its drawbacks.

When analysing the situation with MM, the X5 isn't as superb as it is with MC and MI, but it doesn't fall back by much. Even with MMs the results are very good even by today's standards (-75dB or there abouts). Admitted, the high selectable gain settings allow to raise the volume to levels that would promote some little hiss to the level of audibility. But then dare to drop the needle into the groove, and watch your speakers decompose ...

Below is a comparison of the Fosi X5, once with a Grado Green and once with some AT model. Please note: the difference is only partly an effect of the X5, because MM cartridges inherently produce significantly more noise than MI. MM just generate more Johnson noise to begin with. A fact well known back in the day--by engineers. Caption: red the AT, green the Grado, level is calibrated such that a 1kHz input @5cm/s (standard) would yield -6dB full scale (0dB). Ignore the hum, it's from my setup, really. The green curve is contaminated by my Scarlet ADC--that quiet is the Fosi X5. MC would read quite similar to the MI even if the amplification is adjusted by +20dB.

 

[Roland68]

The FOSI X5 represents a compromise. This lies in the use of the OPA1612 as the input amplifier. For MC and MI (moving-iron, e.g., Grado), it is an excellent choice that drives these systems to the limits of what is physically possible. For MM (moving-magnet, AT, Ortofon, etc.), the situation is different.
The current noise of the OPA1612 quickly becomes disadvantageous here. A NE5532 would be more suitable. The reason is that the comparatively high current noise of the OPA1612 is converted into voltage noise by the high generator impedance of MM cartridges. The disadvantage amounts to roughly 8dB.
Using the NE5532, a Fosi X5 would be better equipped if only MM were considered. Fosi would have been well advised to make the input op-amp replaceable, but that isn’t feasible due to the different package formats—SMD versus DIP, of the opamps in question.

So much on SMD and its drawbacks.

When analysing the situation with MM, the X5 isn't as superb as it is with MC and MI, but it doesn't fall back by much. Even with MMs the results are very good even by today's standards (-75dB or there abouts). Admitted, the high selectable gain settings allow to raise the volume to levels that would promote some little hiss to the level of audibility. But then dare to drop the needle into the groove, and watch your speakers decompose ...

Below is a comparison of the Fosi X5, once with a Grado Green and once with some AT model. Please note: the difference is only partly an effect of the X5, because MM cartridges inherently produce significantly more noise than MI. MM just generate more Johnson noise to begin with. A fact well known back in the day--by engineers. Caption: red the AT, green the Grado, level is calibrated such that a 1kHz input @5cm/s (standard) would yield -6dB full scale (0dB). Ignore the hum, it's from my setup, really. The green curve is contaminated by my Scarlet ADC--that quiet is the Fosi X5. MC would read quite similar to the MI even if the amplification is adjusted by +20dB.

View attachment 492780

I don't know why you're quoting me, when nothing you're writing has anything to do with my post.

 

[Heinrich]

I don't know why you're quoting me, when nothing you're writing has anything to do with my post.

You were discussing SMD devices as inferior, and hardly exchangeable. I exemplified discussing the feasibility of having the input device exchanged for MM or MC/MI use. It wasn't an option for FOSI, because of the packages of competitive opamps. The OPA1612 isn't avalilabe in DIP format. To be made exchangeable, the latter would need an adapter. To make it switchable would have raised the cost by two or so.

 

[Roland68]

You were discussing SMD devices as inferior, and hardly exchangeable. I exemplified discussing the feasibility of having the input device exchanged for MM or MC/MI use. It wasn't an option for FOSI, because of the packages of competitive opamps. The OPA1612 isn't avalilabe in DIP format. To be made exchangeable, the latter would need an adapter. To make it switchable would have raised the cost by two or so.

This shows that you don't know what I wrote about and you apparently haven't even read it.

This has nothing to do with Fosi, but is now a worldwide problem in all areas where particularly small and inexpensive SMD components are used.
SMD components like simple resistors and capacitors are very easy to replace, and there are special tools for this.

However, since many of these components are connected to ground, there is no way to isolate the fault if one fails. These problems almost never occur with through-hole components and larger SMD packages.
The savings in production are very small, but increase with the number of components.

This concerns the cheap SMD resistors and SMD capacitors (not electrolytic capacitors) in small packages that are cut from boards. These have a much higher failure rate and, when they fail, create a short circuit.
This makes measurements and troubleshooting extremely difficult.

This has absolutely nothing to do with op-amps or other SMD components and is a specific problem with these cheap SMD resistors and SMD capacitors.

 

[GuidoK]

The current noise of the OPA1612 quickly becomes disadvantageous here. A NE5532 would be more suitable. The reason is that the comparatively high current noise of the OPA1612 is converted into voltage noise by the high generator impedance of MM cartridges. The disadvantage amounts to roughly 8dB.

If input current noise density is the area to focus on, maybe look further to a more modern opamp than the NE5532.
For instance the OPA1656 (I use those in my chinese hagerman bugle clone), which has 6 fA/√Hz@1khz compared to 0,7 pA/√Hz@1khz (so 700fA/√Hz@1khz), a disadvantage that amounts to roughly 41dB if I calculate that correctly...
Voltage noise density is also still slightly better for the OPA1656 compared to the NE5532 (but here the OPA1612 is considerably better)

Indeed it would be nice if all the opamps in the X5 were dip8 socketed.

 

[Heinrich]

If input current noise density is the area to focus on, …

In some other thread I compared current, voltage, and not the least that peculiar self-noise of the pickup‘s generator with each other. The noise figure of the NE5532 came out to be as low as 2..4(?) dB, differences from other design choices notwithstanding. That‘s good enough for me.

 

[ohnonotagain]

Right now you can pick one of these up on Aliexpress for under $80 with store and site coupons. I just did, so I'll be interested to see how it fares when it arrives in a couple of weeks.

 

[Ian Knight]

Thank Amir for the great review. We hope everyone will enjoy the Box X5.

Here are the links for purchase：
Fosi Audio site: https://fosiaudio.com/products/box-x5-phono-preamp

Can you supply the input impedance and Capacitance please .

 

[KSTR]

Can you supply the input impedance and Capacitance please

Fosi Audio Box X5 Phono Preamp Review

I'm currently running a rather budget TT setup with 2 old B&W speakers (not top line), Marantz slim AVR with phono, Technincs SL1200 and vintage Stanton 881S cartridge but with an after market stylus (not an expensive one). Should I realistically expect a subjective improvement using this phono...

www.audiosciencereview.com

 

[guy48065]

150pF/47k for MM input.
What about MC?

 

[Ian Knight]

Fosi Audio Box X5 Phono Preamp Review

I'm currently running a rather budget TT setup with 2 old B&W speakers (not top line), Marantz slim AVR with phono, Technincs SL1200 and vintage Stanton 881S cartridge but with an after market stylus (not an expensive one). Should I realistically expect a subjective improvement using this phono...

www.audiosciencereview.com

Thanks , that's really helpful to me.

 

[EJ3]

Why is this thread so long?

SO: you just helped make it longer by asking that question? Makes sense in some odd way that I don't understand, I guess.

 

[lini]

(...) For MC and MI (moving-iron, e.g., Grado), it is an excellent choice that drives these systems to the limits of what is physically possible. For MM (moving-magnet, AT, Ortofon, etc.), the situation is different. (...)

Sorry, but I have to contradict there, Because in fact it's not really a question of MM vs. MI, but rather a question of the particular models. I.e., while the Grado MIs indeed have pretty low inductance and impedance compared to typical MMs, there were and are a whole lot of MI models, that don't or hardly differ from typical MMs in terms of inductance and impedance - and there also used to be a few MMs with very low inductance and impedance (e.g. AT22/23/24/25 and Signet TK9/10, just to name one model family).

So in my view in the context of inductance and impedance those Grados are a very questionable choice of an example for a generalised statement regarding MIs - just as the aforementioned ATs and Signets (as well as several Technics models) would also be questionable choices in that regard, due to hardly being represenative for the majority of MMs.

Greetings from Munich!

Manfred / lini

 

[Heinrich]

Sorry, but I have to contradict there, ... MMs with very low inductance and impedance (e.g. AT22/23/24/25 and ...
... Grados are a very questionable choice of an example for a generalised statement ...

Appreciated, don't know to much about the market today. The ATs are a little bit odd, as far as my investigations hold. The inductance comes out to be o/k at well below 100mH, but the output is low likewise at 0,5mV@5,5cm/s. That's in MC territory, but it needs a 47k input**. Other, as I've learned, are deemed to be low impedance--couldn't find actual figures, but then again the output is specified at 1mV@5,5cm/s. Maybe because of the obvious technical compromizes these little guys fell out of favour? Btw, M. Fidler, a maker of acclaimed head-amps discusse dthe topic on ASR--very informative and most of all truely ;-)

You're correct, my generalizations got a step too far. Do you know of current designs with low impedance, matching the X5's elevated current noise? Not that it is a real problem. To really hear noise with an X5/MM combo will need to crank it up like crazy, and then drop the needle to the groove ...

** a perfect match for the X5, right?!

 

[lini]

(...) but the output is low likewise at 0,5mV@5,5cm/s. (...)

That must be a buggy spec, wherever you got it from. Original spec according to catalogues and product manuals for the AT22/23/24/25 and TK9/10 family was 2.2 mV (@ 5 cm/s) - so certainly on the lowish side, but rather high output than low outout MC territory. Audio Technica did offer one really low output (= low output MC territory) MM model, though - namely the AT15SaL, a very rare beast. Other makers of such models used to be Elac and Stanton/Pickering (e.g. EMM 290/390 or 885/981LZS & 785LZE).

Whereas on the current market I'm only aware of really low output MI models by Grado and Soundsmith, but no MM models.

Greetings from Munich!

Manfred / lini

 

[Heinrich]

That must be a buggy spec, wherever you got it from.

You've got me => https://www.hifi-wiki.de/index.php/Audio_Technica_AT_22

It is a pretty nice correct specification, if you can read it: Vs/cm! Have to multiply w/ the standard 5,5cm/s to get where you are coming from, my bad.

The data actually represents quite an achievement considering the explanations from M.Fidler on ASR on the connection between inductance, count of turns in the coil, dc resistance.

Thanks Manfred!

 

[lini]

(...) It is a pretty nice correct specification, if you can read it: Vs/cm! Have to multiply w/ the standard 5,5cm/s to get where you are coming from, my bad. (...)

Well, I'd certainly also prefer the transmission factor stated in a suitable voltage unit per base modulation velocity (e.g. "mV per cm/s"), so it's less confusing.

Mind you, however, that the most usual cartridge output voltage spec style isn't at 5.5, but rather 5 cm/s - most usually in 45°/stereo modulation, although quite a few Japanese manufacturers tend to specify for 5 cm/s in horizontal/mono modulation instead (like Denon, for example), so that one may need to apply a correction factor of 2°0.5 or respectively 1/2^0.5 for comparisons.

Well, and unfortunately Hifi Wiki can't always be trusted 100 % either. For example, in case of the Dual CS 455-1M they claim "Baujahre: 2009 -" and "Neupreis ca. 549 Euro", whereas I've bought my exemplar in 2004 for ~ 240 Euro. And I can prove it - see there: https://web.archive.org/web/20041204205655/http://www.hifi-jaeger.de/dual.htm

Greetings from Munich!

Manfred / lini

 

[FrostedOxen]

I picked up the X5 a couple months ago and I love it. It crushed my Clearaudio Nano and the Schiit Mani 2 so it was the keeper. I'm not sure what the price-point would be to match it ..on sound quality anyway in another brand. Clearaudio was $600 and not even close.

 

[krabapple]

The post above is lacking in detail (not picking on them) just pointing out that fact. It’s a common occurrence on the interweb, even here. Interestingly the Ortofon carts (mm) generally have a recommended loading capacitance much higher than, as an example, Audio Technica. Ortofon DJ’s typically suggest 200-600 pf. AT carts want 100-200 or better yet, lower some perform much better.

Just for comparison--

Ortofon 2M Blue, Red, Black:
Recommended load capacitance 150-300 pF

(Yes, Ortofon makes a bunch of different carts)

This thread is annoying.

Amir's tests indicate the Fosi X5 is a great value. So I bought one, for my meager vinyl-to-digital transfer needs.
I'm just here to research what Fosi gain setting people are using for an Ortofon 2M Blue. (TT is a U-turn Orbit Custom)

To get to that, I wade though (by no means an exhaustive list):
-pages of posts about tariffs.
-a guy calling the Fosi 'horrendous' for stuff no one with human ears will hear.
-people wittering on about swapping out internal components because why?
-endless arcana about circuit design and unit measurement

Are all phono pre threads like this?