Pico Rucksack
Active version

Darius Kubarth - April 2008

The Pico Lino 2 is a small monitor design based on a full-range driver by Vifa released by the German DIY magazine Hobby-Hifi: [http://hobby-hifi.de/Archiv/07/05_07/05_07.html]
The driver is 45Eur, wood (MDF) is 20Eur, other parts (damping, cap, inductor) are 15Eur, makes around 75Eur per box!!! You can't go wrong with that, can you?
And most important of all, they sound fabulous! Sound stage imaging is superb, you really "see" the instruments and singers in front of you just like you would expect it from a good full range driver. The Vifas don't reproduce the high frequency details like a 300Eur tweeter does, so don't compare the Pico Linos with 5000Eur monitors, but they're still very transparent and fine sounding. Don't let the small size fool you, these transmission line speakers easily go down to 50Hz. Unfortunately the maximal sound pressure level becomes lower with deep bass as cone displacement is bigger at low frequencies and might overextend the little drivers. Nevertheless, they are definitely loud enough to make you feel uncomfortable and if you still need more loudness, you can release them from heavy bass by using a crossover and a subwoofer.

Stepping ahead

Well, but there's always something to improve...
The original design utilizes an inductor, a cap and a resistor in parallel to attenuate the mid freq range, which is normally over emphasized, but I decided to build my own amps (gainclones) for these cuties to make them "active" and simply included the EQ network into the amp using a gyrator instead of an inductor. I myself dislike the idea of having big capacitors and inductors in the signal path for a simple reason: The cheap ones lack decent performance, the expensive ones hurt my pockets.

Previously, when I started the project I had the plan to place the amps on the backs of the loudspeakers and so I gave the amps the name "Pico Rucksack". Rucksack = backpack in German, by the way.
But unfortunately the amps became to big for the small Pico Linos - mostly because of the huge power transformer.
I kept the name, though.

Two different notch filters with different Qs and center frequencies were suggested by Hobby-Hifi, I myself decided to go for that one as it better matched my demands:

It's no rocket science to figure out the center frequency of the filter and its Q to replace it with an active and gyrator based circuit. Unlike a real inductor, a gyrator has no core to saturate and to cause any kind of distortions. Surely some applications might rise the demand for "real" inductors to introduce "color" to the sound, but that's definitely not the case here.
Some calculations later I came up with the following component values to replace the "real" inductor:

The inductor is completely gone and the two used capacitors are way smaller than the 15uF one originally thrown in. Smaller capacitors are not just cheaper, but also known to perform way better.
The original plan had a significant drawback: The high frequencies passed the capacitor and the low ones the inductor. Well, especially high frequencies are known to get blurred by cheap capacitors and especially low frequencies cause distortions in passive magnetic components like transformers or inductors. Now only a small fraction of the mids passes our replacement circuit. The amount itself depends on the boost or attenuation applied to the specific frequencies. It fully cancels out with no EQing applied.

The main difference compared to the original circuit is the gyrator being in series with the capacitor instead of parallel. That makes the whole thing a bandpass, not a notch-filter. The circuit is supposed to serve as a frequency dependent load to shunt to ground an oamp's positive input or the feedback path and thus to control its gain around that particular frequency.

One of my design goals was to end up with a no-compromise amp, might be overkill for the Pico Linos alone, but why go for anything less than perfect?
I started the project with the idea to avoid caps and inductors and if you pay close attention to the schematic, you will noticed there's not just the inductor gone, but also there is not a single cap in the path. Yes, it's fully DC coupled. The upper -3dB edge is around 0,5MHz, by the way. Megalomania at its best!

Due to the very wide frequency response care and attention must be payed to the component placement to avoid oscillations at the upper bands and hum at the lower ones. The line input (high Z, 50kOhm) should be placed at the most opposite corner away from the mains transformers not to pick up any radiation. The rest of the signal path is kept at lowest possible resistance for best performance and SNR, for that reason a good oamp with the capability of pushing signals into 600 Ohm lines is mandatory. A NE5532 will do, an OPA2604 is IMHO opinion a better choice.

The PSU

The PSU is almost quite right out of the text book, absolutely nothing fancy here. It delivers +/-30V for the power oamp and +/-18V for the EQ circuit.

The first prototype

Due to the simple design and very common components, the first prototype got knocked together in some very few hours - plus the countless hours on the shitter to come up with the first concept drawing, but that's a different story.

The whole PSU, the balanced line receiver and the EQ circuit were placed on one single breadboard. Most of the real estate on the board is occupied by the coolers and the huge 4700uF caps. Actually 95% of the real estate is the PSU.
The power oamp is better off being attached to the enclosure for thermal reasons. The enclosure acts as the cooler. To avoid oscillations the passive components surrounding the oamp should be placed as close as possible to the oamp itself. It's highly recommended to solder them directly to the oamp pins. It might not look too tidy, but since it's inside the box, nobody will notice.

Boxing day

A cheap 25 x 10 x 16 cm metallic enclosure seemed to be the perfect housing for the power oamp, the EQ circuit, the PSU and the mains transformer. The original plan was to place the enclosures on the backs of the Pico Linos. Unfortunately the mains transformer makes the whole backpack too heavy for the Pico Linos, they tend to pass out to the back. That was when I decided to make the Pico Rucksacks stand alone.

(The ugly surface is just some protective foil to be removed)

(The enclusere pictured above doesn't have the front plate's LED, yet)

Setting up the right EQ curve

There's not much need to change the center frequency of the filter circuit, it perfectly matches the Pico Linos' frequency response. But if you still need to, a potentiometer instead of the 1,5k resistor in the gyrator circuit is a good point to start with. It will allow you to move the upper edge of the frequency dependent load. The lower frequency is determined by the 47nF cap. Unless you find a variable cap, replacing it might be the only way to change the frequency.
Let's focus on setting up the right amplification factor instead of the frequency as this a real must.
Hobby-Hifi suggests to attenuate the mids by around 2dB. You figure that out by comparing the frequency charts with and without the filter circuit or by doing the math and some calculations on the schematic.
I myself found -2,5dB to get me the most pleasant results in my listening room and position. But most likely I wouldn't get the difference between -2 dB and -2,5 dB, anyway.
First, it's handy to have a way to find the 0dB position before looking for the -2,5dB position. If you've have chosen a precision potentiometer with many turns, it's hard to find the 0dB position in the "middle". Depending on the test equipment you have at hand, two approaches are possible:

  1. Connect a function generator to the amp's input and monitor the amp's output on your oscilloscope. A 1kHz square wave will have it's 1st harmonic round about at the frequency we mostly process: 2,22kHz. Any setting away from 0dB will deform the square wave and make it look quite odd on the screen. Turning the potentiometer till it looks like a perfect square wave will quickly uncover the 0dB position
  2. The LM3886 power oamp amplifies the signal by a factor of 23 (=2,3k/100R in the feedback path) or 27,14dB. The EQ circuit itself is unity gain. It's logical that feeding a -27,14dBm sine wave into the amp will result in 0dBm at the output if no EQing is applied. A sine wave at 2,22kHz will change it's level depending on the potentiometer's position and only become 0dBm at the output if the EQ is set to 0dB.

The second approach can also be easily used to set the desired attenuation for the frequencies around 2,22kHz by turning the potentiometer until the output level lowers by the desired amount of dBs.
Trial and error is by the way another approach to set up the EQs. Just listen and turn the potentiometer till it sounds good to you in your listening room, listen again, turn again. Works for sure.

The final touch

Realising it might be good to have an indicator, if the amps are powered up or not, simple LEDs got placed on the frontplates of each one of the amps. I had heaps of blue super-bright LEDs in my workshop in stock, because I love those blue lights on my studio gear, so I went for these and hooked them up between 30V and 0V. A 15k resistor in series with the LEDs gets them to go really low on current and nicely glow. Good WAF.

First impression after listening tests? Outstanding!!! And way better than the original version with that LCR thiggy and a conventional amp - took two minutes of A/Bing to realize that.
Best of all, the EQ curve can be freely set depending on the listening room and taste. Well, actually the mids can not just be attenuated, but also boosted. You can't do that the passive way, can you?

For sure it was fun to build it and it was worth it. The final costs were around 120Eur for the Pico Linos (I got the two drivers for under 70Eur) and the Pico Rucksack was 70Eur per channel. 70Eur is for sure slightly more than you pay for an inductor, a cap and a resistor, but you get a better sound, a no compromise amp and heaps of fun along the way. That's worth it, isn't it?
Hope you enjoyed the reading and can successfully build a Pico Rucksack, too.

Building the Pico Lino 2 monitors

Inner guts...

Before and after smoothing the primer

Ready for the lacquer...

Ready for the driver...

Finished