ScanSpeak subwoofer 26W/4558T00
Hypex DS 4.0 plate-amp


Copyright 2015 © Troels Gravesen

   Go to on this page:

As you may have noticed, no subwoofers have ever been published on these pages and the aim of this project is to supplement a pair of smaller speakers with subwoofer(s). Quite often we see people buying a pair of very small speakers in the 3-6 liter range due to high-WAF or other preferences. Next we see a subwoofer tucked away up against the wall or in corners in order to help in the 30-100 Hz range. This may not exactly turn a mini into a maxi, and the minis not having a high-pass filter really can't be played louder than before.
There's a lot of energy in the 100-200 Hz range and neither the sub, nor the satellites handles this frequency band very well. We have a dynamic hole between the two, but a subwoofer may create an acceptable ooumph in the low range and for some that's enough. To have a better dynamic presentation we may need at least a 6" bass driver in the main speakers, maybe even an 8" driver, but time will tell, because I'm going to mate these subs with my Discovery W12, Ellam 98 mkII and AudioTechnology 3-Way Classic and hear what happens going from around 50 to 100 and up to 200 cm^2 membrane area for the main driver.

Otherwise there's nothing new under the sun here: A subwoofer featuring the well-known ScanSpeak 26W/4558T00 driver and the plate amp Hypex DS 4.0. Google the driver and you'll find a lot of applications. From my google search I find discussions on the significance of 22 mm versus 26 mm cone excursion, but these people are talking earthquake movie set-ups and large rooms. Not what I'm aiming at here, rather a decent response down to 30 Hz for general hifi application. Subwoofers and plate amps have been around for a long time and subwoofers even has a dedicated website,

So, kindly sponsored by ScanSpeak, I got two 26W/4558T00 drivers and I ordered a pair of DS 4.0 plate-amps.

Modeling the 26W sub soon makes it clear we're not going to use a ported cabinet. We need a decent port diameter and the port is going to be too long, really long. Making a slotted port soon adds one third to cabinet volume, so this was given up also. Using a passive radiator is basically the same as a port, only does not take up any volume. Downside of passive radiators is cost, but we do pay dearly in many other areas to reduce volume.

Subwoofers are rarely very sensitive, often 83-86 dB/2.8V but as we place them close to the front wall or even in corners we get some 3-9 dB room-gain, hence suitable for most main speakers. From simulation seen below you can see the impact of having the 26W some 30 cm off the floor, some 50 cm from the front wall and 180 cm to side walls. We get some 4-6 dB room gain, sometimes even more of in corners. If we didn't have room gain, many speakers would sound rather lean, even from an 8" bass driver. The tricky part for any loudspeaker constructor is to voice the speaker taking into account an estimated average room gain contribution in the low range, maybe except for very small speakers were we really do not expect much bass at all. Some people are disappointed from their 6" bass driver, which may have a decent sensitivity of e.g. 88 dB, but placed on a narrow baffle and into the room, it looses a lot of low-end power due to baffle step loss and lack of room gain.



26W/4558T00 subwoofers. Click image to view large

Download driver specs here.

ScanSpeak passive radiator 26W/0-00-00, download spec here.

Passive radiators kind of falls into two groups, those for general hifi aiming at 30-40 Hz bass extension and then dedicated subwoofer passive radiators where we are talking moving mass of 4-500 grams. Here almost 4 times the moving mass of the subwoofer unit. We're talking deep bass from small cabinets and the price to pay is efficiency.
Generally passive radiators deliver a little less bass extension compared to ported systems and a little steeper roll-off. In return better bass definition and transient response like a closed system. As always in hifi, some people love them, others hate them, nothing different from ported systems.
Passive radiators are certainly not cheap compared to a plastic tube in a ported design, but a port doing the same here as the passive radiator takes up considerable volume. A 12 cm port diameter here would be a little less than 1 meter long, a chimney! Seriously low-WAF!




Click images to view large

Download DS4.0 specs here.




My test cab was actually a little larger than these, 43x43x43 cm to make 50 L net volume for experimentation.
My final cabs will be 40 x 40 x 40 from 20 mm Baltic birch, heavily braced.
The compartment for the plate amp takes up some 7 liters, thus net volume close to 40 litres.


Box simulation vented

Above simulation of response from closed and ported enclosures @ 40, 50 and 60 litres.


What has to be taken into account when dealing with subwoofers, is the room-gain. The room is an amplifier and adds to the low end response as can be seen from above from placing the driver 50 cm from front wall, 30 cm from the floor and 180 cm from side walls. We get some 4-5 dB extra from such placement. The reason a lot of subwoofers work at all because efficiency is sometimes appallingly low.


Box simulation passive radiator 

Above and below UniBox simulation of passive radiator.

As can be seen we get a Fb of ~20 Hz from these heavy weight passive radiators.

Adding a second PR looks like this.


As always: Test cabs before cutting final panels


My test cab is 43 x 43 x 43 cm making a total of 57.5 litre from 22 mm MDF. The plate amp enclosure is 7 litre brutto, hence a net volume of ~50 litres minus drivers and bracing but I usually make the increase in volume from damping material make up for the volume of drivers and bracings. I used 30 mm egg crate foam on internal panels. From experiments on decreasing cabinet volume my final cabs will be 40 litres and with the 26W/0-00-00 passive radiator from ScanSpeak.



Let me state immediately: The scaling of the dials on the Hypex DS4 is of little use, if any!
The knob indicator runs below and above the scaling and we have e.g. EQ level split into 10 major levels for 0-6 dB. Not very useful.
So, whenever referring to a setting, I'll use a clock scale, e.g. 1500 is three-o'clock, OK?

50 liter cabs, Vifa passive radiator

Left: First thing to do is measuring the response with no equalisation added. Red is driver, green is PR unit.
Right: Here the two readings are summed in orange graph.
As can be seen, the box tuning is seriously low from the vintage Vifa PR. This a probably a ~400 grams default PR. No data available.

Left: Here the same as above, but with EQ set to "1500" and EQ level to ~5 dB. As can be seen we have a decent flat response down to 30 Hz and some -3-4 dB @ 20 Hz. Not bad at all, but remember, this is near-field measurement, not what we get in a room.
Right: Equaliser corner frequency set to 20, 30, 40, 45 and 50 Hz. Now, this shows the nonlinearity of the dial. First around 45-50 Hz things start happening. Purple 45Hz, blue is 50 Hz. Red is no EQ. At 20, 30 and 40 Hz not much is happening.

Left: EQ level set to 5 dB. Purple is bypassing low-pass filter.
Red = "1700" (120 Hz), blue = "1600", green = "1500", light green = "1200" and orange = "0900". So much for linearity of dial.
Right: Equalisation added and response curved for corner frequences "1700", "1600" and "1500". This is a very narrow range for significant differences in corner frequences.

Final test of cabinet was reducing volume to 35 litres.
Left: Nearfield response of 26W and PR. Very little difference at all.
Right. Summed response from 26W and PR (orange) in 35 litres cabinet, equalisation added.



Above the speakers participating in the test, Audio Technology 3-Way Classic, Discovery 12W and Ellam 98 mkII.

First of all: Setting up a subwoofer is anything but trivial, in particular with the Hypex DS4 where we have the option of bass lift and without measuring gear, it's even more troublesome. We may take near-field measurements of the bass driver and passive radiator and make a summation and tune the bass lift to a flat response down to some 25 Hz, but then we place the subwoofer up against a wall and have to do it all over again due to room-gain. Room-gain is very much part of the equation. The good thing is that room gain is almost linear in the range the subwoofer covers, so once tuned in for a specific pair of satellites, we only have to deal with level.
All tests here were made from a single subwoofer placed between the speakers about 0.75 meter from the front wall (little room gain).

First speakers on test were my Discovery 12W speakers. Small they are and my expectation was that this wouldn't work very well. I was expecting a serious lack of coherence due to the fact that the 12W doesn't move much air in the 100-200 Hz region. Trying out different points of crossover I ended up at 120 Hz, the highest point of crossover the DS4 allows. Now, there certainly is a lack of coherence in the low end but overall if lifts the performance considerably being able to hear an upright bass with some volume and some physical impact. When it comes to the overall enjoyment of listening to music the subwoofer certainly does its thing.

Next the Ellam 98 mkII. Here we have a bass driver close to twice the size of the 12W and things start sounding right. The point of crossover is lowered to some 60-70 Hz and overall coherence is significantly improved. Still, 95 cm^2 membrane area can't move air the way a 10-12" bass driver does, but adding a subwoofer we're certainly better off although the 100-200 Hz slam of bigger speakers is still missing.

The AudioTechnology 3-Way Classics feature an 8" bass driver, more than twice the size of the Ellam midbass and the distance between the AT-3WC and the Ellam is bigger than the distance between Ellam and Disc 12. All of a sudden we have to deal with getting the deep bass right and all DS4 settings had to be changed. Lower point of crossover and fine-tuning of bass lift as well. Actually the AT-3WC seems to favor 2 subwoofers as efficiency is rather high and a single 10" bass driver runs into trouble of not having enough output.
We need some musical source with some serious deep bass to fine-tune and having set up the sub with some trio jazz I threw in a DALI test CD and the subwoofer membranes were almost popping out of the speaker baskets! That's what I mean by not overdoing the bass lift. We may use this as a "loudness" function and all of a sudden find out we're having way too much low-end output. If we go for a single subwoofer, I would recommend a 12" subwoofer for the AT-3WCs.

Part II will cover the construction of the final two 40 liter cabinets.