Ferrum OOR and HYPSOS measurements

This unit was sent to me by HEM.
Content and analysis made possible thanks to support from https://headphones.com, and https://patreon.com/goldensound supporters.


Ferrum is a brand of audio products created by HEM in poland. You might not have heard of them, but you definitely will have heard of some of the products they created, as HEM are actually the ones behind the designs for several well known products such as Mytek.


However Ferrum is their first ‘in-house’ brand, and so far they have two products available.
The first, released last year, is the HYPSOS, a premium hybrid power supply with variable output voltage between 5v and 30v.
The HYPSOS was well received and its other aspects like 12v trigger I/O, high power capability, 4 terminal sensing and spread spectrum operation make it a bit different than your usual commercially available PSU.

But the product we’re looking at today is the recently released OOR, a fully balanced headphone amplifier with a lot of interesting features.
The most apparent aspect of the OOR is its power, 8w @ 50 ohms, which for users of the Hifiman Susvara, is quite exciting indeed.

It also functions as a preamplifier, can run balanced or unbalanced outputs, and has the ability to fully integrate with the 4 terminal sensing of the HYPSOS power supply to in theory provide better transient power handling than would be possible with a standard PSU.

The OOR is a discrete design, rather than being based on opamp chips.

OOR internals, Photo Credit: Highfidelity.pl

A full video review of the OOR/HYPSOS combo is on the way, but for now lets have a look at the measured performance.


Test Setup:

– Audio Precision APx555 B-Series Analyzer
– Amplifier was warmed up for 6 hours before measuring.
– AudioQuest Mackenzie XLR interconnects
– Neurochrome Dummy Load
– Full reports containing additional data and test configurations are attached
– Exact analyzer/filter configurations for each measurement are detailed in the full reports
– All measurements shown below are taken with a 32 ohm load via balanced output unless otherwise specified

Full Reports:

Reports available here (These contain 4v, 700mv and 50mv test results, please check you’re looking at the right ones when comparing!):

(32 Ohm load for 4v/700mv tests, 12 ohm for 50mv tests)

Gain: 15.6dB (High), 5.6dB (Med), -4.3dB (Low)

Measured Output Impedance: 0.35 Ohm (Balanced)

The most impressive thing about the OOR is the power it can output. It’s a fairly compact amplifier, but this unit delivered up to almost 10W @ 32 ohm. (Official spec is 1.6w @ 300 ohm and 8w @ 60 ohm).

32 Ohm: ~9.5W
50 Ohm: ~8.3W
300 Ohm: ~2W

The performance of the OOR does change a bit depending on what load impedance it’s driving and how you’re using it, so this post might be a tad longer than usual.
Firstly, the OOR has a ‘Bypass’ option, which physically bypasses the internal volume control (and gain switch), and effectively turns it into a headphone power-amplifier.
To use it in this mode you’d need a preamp or other external volume control, but the OOR does measure slightly better in bypass:

1khz 4v input 300 ohm (Bypass):

1khz 4v input 300 ohm (Internal Vol Control Maxed):

Not a massive change or anything, but slight. Interesting to note and worth playing with if you have a preamp.
The output level is slightly higher in bypass but only by just under 0.4dB so it does not fully account for the 3-4dB improvement in THD+N

Now let’s look at the ‘usual’ tests:

1khz 4v input, 4v out 300 Ohm:

1khz 4v input, 700mv out (Headphone Level) 300 Ohm:

Extremely low harmonic distortion here! With harmonics below -135dB. The SINAD/THD+N value is limited by PSU noise it seems.
This PSU noise is due to leakage from the HYPSOS. It can be eliminated by using the stock power supply, but in doing so you’d lose the transient power correction capability of the HYPSOS, so whether the HYPSOS or stock PSU will give a better sound…you’ll have to try both and see which you prefer. THD+N at 700mV was roughly 1dB better with the HYPSOS.
Personally I would say that if using IEMs, use the stock PSU. If using power hungry headphones, use the HYPSOS.

This is part of using a linear PSU. You will almost inevitably have more mains leakage, but no ultrasonic switching noise. Whereas a switching PSU you can get almost no mains leakage, but will have ultrasonic switching noise.

Idle noise with HYPSOS. Some 50/150/250 hz noise but no added ultrasonic noise.
Idle noise with stock PSU. No 50/150/250 hz noise but some added ultrasonic noise spurs (which are also higher in level).

1khz 4v input, 700mv out (Headphone Level) 32 Ohm:

When we put a tougher 32 ohm load on, now we start to see a few harmonics, though still very low.

1khz 4v input, 50mv out (IEM Level) 12 Ohm, low-gain:

Performance for IEMs is quite good especially given the crazy power this amp can put out for planars and other hungry cans.
It’s 15dB or so behind the quietest of quiet amps like the SA-1 or some DAPs, so possibly not the best choice if you only/mostly use IEMs, but it will certainly handle IEM use just fine. This is not a ‘difficult to drive cans only’ amp.

1khz 4v input, 50mv out (Headphone Level) 12 Ohm, using stock PSU instead of HYPSOS:

THD+N / Frequency:

40khz AES filter
20khz AES filter


Linearity with bandpass filter
Linearity without bandpass filter

SINAD/THD+N vs output level:

IMD (SMPTE) vs Output Level:


Crosstalk is lower for a 300Ohm Load:

1khz Square Wave:

32 Ohm Load
300 Ohm Load


7 thoughts on “Ferrum OOR and HYPSOS measurements”

  1. Hi Golden, nice review. Looks like a great amp, but then I saw the price. Way out of my budget! About the review, I find it a bit confusing to know which I/O you use though. Do you have any idea about performance(penalties) when using SE>SE SE>BAL BAL>SE or BAL>BAL? This is something interesting to know.


  2. “This PSU noise is due to leakage from the HYPSOS. It can be eliminated by using the stock power supply[…]”

    “THD+N at 700mV was roughly 1dB better with the HYPSOS.”

    So wait. Getting rid of the mains leakage plus its harmonics by switching to the stock PSU will decrease SINAD at .7Vrms into 300Ω by 1dB? How does that work?
    Are you not limiting the BW for THD+N measurements?

    • All my THD+N measurements use an AES17 20khz filter.
      It’s likely that the dominant factor is simply the noise floor, not mains leakage/harmonics.

      Keep in mind that an FFT can be very misleading as to the true level of noise. I’m using 256k FFT plots so that will often show stuff ‘below’ the noise floor.

      This is why the band-limited RMS value is sometimes more useful (but of course provides no context as to what that value is made up of)

  3. Very helpful measurements ans comments, many thanks!

    With that amount of power available for power hungry headphones, we get near the point where dedicated speaker amps come in as well, don‘t we?

    I saw in your videos that you are using the Benchmark AHB2. I own the Dan Clark Stealth since a couple of days, and need/want to pair it with a very powerful amp – at its 23 ohms, would you say the Oor with Hypsos coming from the traditional headphone amp side is the better approach, or rather the AHB2 with its very low distortions at low loads as you‘d expect from a speaker amp?

  4. Short and simple question ahead of your review, as I’m going to go and try the Oor and Hypsos on Saturday – is this a worthy pairing with the Susvara?

  5. How well does the Oor pair with headphones with low impedance like the Meze Empyrean/Elite? With and without Hypsos?
    Any “real world” experiences from Oor users?

  6. Nice measurements and well done. I have had both the OOR/HYPSOS now for a week, and paired with the HiFiMan he1000se is absolute magic.
    Also, the Audeze LCD 5s and Meze are wonderful, but the he1000se adds an element of 4th dimension that I have yet to find with the others, now, awaiting balanced headphone cables from Peter at DoubleHelix Audio, that may be the solution. By far, the best headphone amp and power supply I have owned.


Leave a Reply

%d bloggers like this: