Robin 2.5 Tower (HiVi M5a)

The Robin 2.5 has been a long time coming. It started out as the Robin, followed by the Robin II, then the Robin IIa. All were 2-way designs, based on the HiVi M5a. That was in 2010.

As in most 2-way speakers with smallish 5.25″ mid-woofers, bass is not their strong point. The Robin 2.5 addresses this weakness head-on. Another M5a was added to reinforce the bass region.

All 3 drivers share the same box internally. There are no internal dividers. The lower M5a acts very much like a woofer in a 3-way system. The M5a above it takes care of the mid-bass and mids. Since both the mid-woofers are the same and share the same box tuning, the lower M5a will seamlessly strengthen the bass of the one above. In a nutshell, that’s the concept of a 2.5 system.

For the highs, I chose Vifa’s Dual Ring Radiator tweeter, the XT25TG30-04. This tweeter is a fine performer for $30. I could have used a tweeter double the price but it’s not in my nature to have a tweeter that cost more than two mid-woofers combined.


(Fig 1) HiVi M5a Frequency Response (No Smoothing, mic at 0.5 meters onaxis, 5 msec gating). Disregard measurements below 500Hz.

HiVi M5a Frequency Response

Let’s start with the M5a. Fig 1is the Raw response of the M5a. No filters in line. Only 500Hz and above is of importance. Below that, my room is affecting the measurements.

What I’m looking for is the cone breakup. HiVi’s M5a cone is made of a composite of Aluminum and Magnesium. The advantage of metal cones is that they behave more like a true piston. However, they do have one weakness and that is they “ring” like a bell when they breakup. For the M5a, it appears the cone starts to go out of control at 3kHz. The first peak is recorded at 4.3kHz, second peak at 8.6kHz.

Ideally, I would cross at 750Hz~1kHz but that’s not possible. The M5a has to hand over to the tweeter and 1kHz is impractical. With the Vifa XT25TG30, lowest is about 2.2kHz.


(Fig 2) HiVi M5a with 12dB filter (No Smoothing, mic at 0.5 meters, 5 msec gating)

Crossing the M5a

Fig 2 is the M5a with a 12dB filter with adjustments made to suppress the cone breakup. Note how tamed the peak at 4.3kHz is. I left the peak at 10kHz unattended because it is 30dB below. I doubt it’ll be a nuisance in use.


(Fig 3) Vifa Frequency Response of XT25TG30


(Fig 4) Vifa XT25TG30 RAW Frequency Response (No Smoothing, mic at 0.5 meters on axis, 5 msec gating)

Tweeter Reality Check

Fig 3 is from Vifa with the XT25 mounted onto a IEC Baffle. Note the flatness from 1kHz to 10kHz. Fig 4 is with the XT25 mounted onto the tower with a baffle width of 8.5″. See the notch at 2.8kHz. That is caused by the edge of the baffle. Is there a cure for this? Well, the easiest is to have a baffle of 1 meter x 1 meter. I’m sure we can all agree that’s not very practical. The other solution is to leave it and use it to our advantage.


(Fig 5) Vifa XT25TG30 RAW (Black), with 12dB filter (Red) with filter and LCR (Blue)

Fig 5 shows 3 plots. Black is the Raw response of the XT25, Red is with a 12dB filter and Blue is with an LCR added to flatten the impedance peak at 500Hz. As can be seen, there is not much difference with the LCR.

Robin 2.5 hp_lp_12db

(Fig 6) HiVi M5a x2 Low Pass and Vifa XT25 High Pass (No Smoothing, mic at 1 meter XT25 axis, 5 msec gating)

Fig 6. Brown trace is of the two M5a combined. The Red trace is of the XT25. Disregard the measurements below 500Hz. It is no longer gated. But it does reveal some interesting points. The huge suck out at 150Hz is caused by “floor bounced” with the mic at 1 meter. It is absent when the mic is 6~8ft away. As we move further down to 50Hz, we can see the lower M5a reinforcing the bass. In Fig 2, the 50Hz from a single M5a is about 15dB lower than the mids whereas in a 2.5, the bass is level.


(Fig 7) Robin 2.5 Frequency Response (No Smoothing, mic at 1 meter on XT25 axis, 5 msec gating)

Robin 2.5 Frequency Response

Fig 7 is the Frequency Response of the Robin 2.5. Microphone is 1 meter away, on tweeter axis. No Smoothing applied. So no glossing over response plot. Warts and all will be shown.

The region of interest is from 500Hz onwards. First thing you’ll notice is it is not flat. Well, this speaker is not for professional recording purposes. A ruler flat response is not required or even desired. It is for my listening enjoyment.

Remember earlier I mentioned about taking advantage of the notch in the tweeter. See the dip at 3kHz. That’s from the tweeter. Some may say that’s a flaw. I look at it as a natural EQ. Because of that dip, the Robin 2.5 doesn’t suffer from excessive sibilance. Our ears are most sensitive from 2kHz to 3kHz. That gradual downward slope from 1kHz to 3kHz helps to smooth out recordings that are too bright. Treble is restored after 3kHz. Note the gentle down slope from 4Khz to 20kHz. I deliberately tuned the Robin 2.5 this way. It sounds more pleasant when playing loud.


(Fig 8) Robin 2.5 Crossover Frequency (No Smoothing, mic at 1 meter XT25 axis, 5 msec gating)

Fig 8 is with the XT25 wired in reversed phase. A sharp notch is observed at about 2.9kHz. The slopes on either side are not symmetrical but it doesn’t affect the sound. I’m more interested in how well the drivers are crossing.


(Fig 9) Robin 2.5 Distortion

Robin 2.5 Distortion

The distortion sweep (Fig 9) did not reveal anything bad. 2nd and 3rd harmonics are about 50dB below from 1kHz to 10kHz. Not going to hear any distortion.

With this plot smoothed at 1/6th octave, the action of the bottom M5a is more obvious. We can clearly see the bass from 100Hz to 50Hz is now amplified.


(Fig 10) Robin 2.5 Impedance & Phase

Robin 2.5 Impedance and Phase

Fig 10 is the impedance sweep of the Robin 2.5. Overall, the Robin 2.5 is not a difficult load. Minimum impedance is 4 ohms at 200Hz. At 8kHz, it’s 3.5 ohms. It should not present problems with properly engineered power amps that are rated at 4 ohms.

Sensitivity is quite high, about 90dB. Low power amplifiers like tubes and chipamps can be used as long as they are rated to drive 4 ohm loads.

Speaker Placement

To get the best out of the Robin 2.5, put the Robin on stands so that the tweeters are at ear level. These speakers need breathing space, so keep side and back walls at least 3 ft away. Listening distance between 6ft~10 ft.

Sound Quality

Clear. Transparent. Depth. These are the first impressions on hearing the Robin 2.5 . Voices and instruments jump out. No listener fatigue. No “shouty” mids and unnatural bass.

All for a grand total of $87 per speaker (two M5a=$54, XT25TG30=$33). Unbelievable.

Crossover & Box Dimensions

If you’re interested to build the Robin 2.5, contact me for the crossover component values and box dimensions. This design is not for commercial use.