The different projects


Final Chapter of this Voigt saga

I have measured the final verision of these Voigt loudspeakers. I measured at 1/2 meter distance in front of bass driver, midway and on axis of tweeter horn. There is a dip at 1-2 kHz and reversing the phase of the tweeter did nothing to improve the dip (data not shown)

If I separate the tree pairs the good pair matching up to 5 kHz is clearly seen.

Near field responses of bass driver and pipe opening is closely matched and I thought that I leave as is as good enough. Time to listen

Compared to Tyrland/Stridbeck TL-6 (think IMF TLS80 they are more sensitive and the the last octave of bass is missning.I am not too keen on those tweeters so it is time to get rid of those pipes and get back to my big baffles.


I used the nonmatched 1 and 4 horns for the actual loudspeakers. The serial resistor is a 15 ohm and the parallell is a 4W 8 ohm. For a more high power application a higher rated resistor would be better. The serial caps are 2.2µF vintage electrolytical bipolar 50 Volt types. That is hardly the limiting factor in this setting.

When measureing the impedance the pair matching is very good and the tuning is at 40 Hz, I will make some measurements of frequency response as soon as it works with the WAF.


Piezo horn crossover

Counting from the tweeter, I added a 8 ohm serial resisor, this does not affect the resistance in the working range 2-20 kHz but protects the amplfier against low impedance way higher up in frequency. I then added a 8 ohm resostor across the driver to get a 8 ohm resistance. With this setup I could try different capacitors to see if I could equilize the the lift arount 4 kHz and get a flatter frequency response. I tried the horn with no crossover at all, then adding a 3.3 µF cap, a 2.2µF and finaly a 1.65µF.
This is how the responses turned out
If I normalize it to the  level at 10 kHz it looks like this
Both graphs show that the smallest cap is overdoing the eqalization, It appears that 2-3µF will give the flattest response. Compared to no filter I get about 4 dB reduction of level below 4 kHz. I get a -6dB point somewere between 3 and 3.5 kHz so this might work out with my Peerless drivers. With a bit of luck I can run them flat out otherwise I will fiddle with a 6 dB filter and Zobel to tweak the low pass filter. I will only use second hand electrolytic caps, fancy plastic film are wasted on this setup especially the tweeter.


I have four Motorola KSN1016 drivers to use. For frequency measurements I screwed them to a sanding block and put them over the corner of a loudspeaker to get as much free space as possible.

The impedance measurements grouped  the drivers in two pairs 1&4 and 2&3. (I numbered them in random order as I saw no distinct markings on the drivers to separate one from the other,). The very high capacitive impedance shows how they can be used without any crossover (but then you can not adjust the frequency response at all). I have read that you can add a 8 ohm resistor across the driver and then it will behave like any other 8 ohm driver when using serial capacitors for cross over
Then some frequency measuremnets with Omnimic on axis about 1/2 meter distance.
The results looks quite bad. Especially for 1&4 that are all over the place and unmatched. 2&3 are at least matched and also slightly smoother. A closer look show that the level increase about 6 dB from the dip at 8 khz to the low end peak around 4 kHz. If I can get a 6 dB/octave filter starting below 8 kHz it should result in a flat respons down to 4 kHz and then a rapid drop of. I will make some test soon.
In the time domain all four drivers have some severe ringing around 4 kHz.

If I supress the output at 4 kHz by 6 dB relative to the output above 8 kHz those ridges should go down as well.


A late update:

The open baffle has been sidetracked yet again. This time by a generous gift by Rockolga who gave me two Voigt pipe speakers. An updated version from the Danish High Fidelity designs from the 70s.

They had cutouts for 8” drivers (good) and E-V T-35 horn Tweeters (not so good). I have some really good Peerless 8”  861832 drivers that work well in Voigts and then I also have some piezo tweeter horns that fitted in the horn cut outs after some elbow grease.

The speakers were undamped apart from having 175 gram of BAF fiber stuffed in the closed end of the pipe.  Based on B Börjas article I stuffed 80 gram BAF in the first section and 45 grams as a wedge around the first bend and then behind the driver in a wedge like fashion (after removing the initial 175g that is). It worked out quite well. According to Börja I should have added rock wool and fiber glass wedges here and there. As I am not to keen on working with fiberglass I used some 20mm thick sheets of dense BAF (Biltema) and used that to cover the walls in the section of the bass driver.
The upper partition is seen to the left in this picture

I then measured with omnimic  just in front of either the driver or the pipe opening

Driver (black undamped pipe and blue trace with 80+25 gram BAF.
The damping at 200 Hz is atributed to the BAF, a thing to note is the weak dip at 50 Hz. Usually the pipe gives a sharp dip in driver output at the fundamental resonance of the pipe. The finding suggest a leak somewhere. (Later I added some plumbers putty behind the Piezo horn and the potentiometer)

Pipe opening
The BAF reduces the peaks by about 5 dB. That the pipe peaks in output around 60 Hz indicates that something is wrong. It should be below 50 Hz.

The sealing and the BAF sheets lowers the tuning by 10 Hz from 50 to 40 Hz. Good But the dip at 40-45 Hz is not as strong as I would like to have it. (lower two panes)

The impedance looks like this
I still suspect some leak but it is quite good and close to expeced values.

Time to look into the tweeter issue.


With deadlines coming closer I finaly have made some progress with Sleipnir. The Whardedale SFB that got me going has several features I do want to change. Insted of having one baffle as a slate I intend to use a three baffles: one for the four bassdrivers with a lot of bracing, standing straight up. One baffle for the midrange driver, angled slightly upwards and with no bracing but with bitumen lamination like the Spendor BC-1. And lastly a third panel angled 45 degrees upwards for the tweeeter
This is a prototype with only two bass drivers I have tried out how to mount them and also cutting out holes with a router, this time with some scrap material below and also to round of the cutout with a router bit.  I am not sure how to place the vertical side panel, well inside the front panel edge as the SFB or covering the edges of the baffles. The finish will be very vintage looking...

All alnico drivers now, I plan to use 16mm MDF and rely on laminations and extensive bracings rather than having 2x25mm MDF,

Even this small baffle lowers the Fr of the drivers.


I added a 4.7ohm serial resistor and a 3.2 ohm paralell  resisor and remeasured with both phases tried out.
The orange one is the the response with the Seas H86 tweeter in the right phase. Reversing (blue) is far inferior and the brown line with no tweeter.  The response as measured on axis from the tweeter at about 75 cm distance shows a decent respsonse. Flat to 10 kHz and a falloff above that without any nasty peaks. Padded down by 10 dB the tweeters are well protected from abuse.

I also tried adding a zobel filter to the woofer to see if it changed the filter resoponse but it is more or less the same so I will not add one. Time to wrap up this one by finding some screws and assemble the speakers, and move on to the next project.