Monday, 4 January 2016

Lowther OB Further Thoughts

You may have noticed that in my BSC, there is a capacitor, when the conventional BSC employs an inductor and resistor in parallel. I have used instead a conventional parallel notch filter to remove the baffle step hump. Now, the conventional theory is that the baffle step causes a 6 dB rise at a certain frequency and then above that point, it is generally a 3 dB rise. However, in the model, because the start of the rise of the baffle is around the point as the crossover, about 200 Hz, the 0 dB component of the BSC is part of the filtered part of the XO. Therefore, I reason that I do not need to use a standard BSC, because my baseline is the frequencies after the 6 dB hump, which is really only a 3 dB hump.

Combined with the room gain of the speaker, I am able to rune the Lowther without too much inline attenuation, beyond what the notch and crossover has already provided. Looking a 3R to equalise. However, with an L-pad, I should be able to dial that in as required.

Not sure if the above makes complete sense, except that once used in the model with the values, it produces a flat +/- 1 dB response, which is a pretty good start.

Further tweaking as yielding another set of values, this is primarily affecting the Lowther.

100 uF and 4.7 mH

3.3 R, 1.5 mH and 68 uF

Baseline 2nd order passive crossover

Pretty much where it was before. It should also be noted that the values are not the conventional formulas. If we did that, the following is the predicted response.

Crossovers are at 200 Hz for Alpha and 250 Hz for Lowther. The notch filter is the same as above.

Lowther = 28 uF and 14.5 mH
Which will end up using 27 uF and 15 mH

Alpha 15a = 9.4 mH and 68 uF
Which will end up using 10 mH and 68 uF

Conventional 2nd order passive crossover

The benefit of the above is that the series capacitor is smaller for the Lowther and the series inductor is smaller for the Alpha. At the expense of the higher parallel inductor/capacitor value.

More tweaking to do. But at this moment, it feels like the 'baseline' option is the right option to go with.

Lowther OB BSC Revised

Have checked with the unused components and found close enough values for the BSC to remove the need to buy new components. So the new values are:
  • 3.3R Jantzen Superes Resistor 10W
  • 100uF Solen Capacitor
  • 1.5mH Jantzen Wax Foil 14AWG Inductor
The revised simulation with the above values - can always upgrade them later if this concept proves to be a winner! No change to the XO components.

Revised frequency response with amended BSC.

Sunday, 3 January 2016

Lowther OB Continued

Did some more modelling, and started to apply a BSC to the design. From playing around, it seems that  2nd order passive cross over is going to be the easiest start, though I am playing with using the same components as that of a 1st order, so that I can add the parts to make it into a second order crossover. I have also played with turning room gain on and off. In either scenarios, the outcome is fairly similar, save the level of attenuation for the Lowther.

In both scenarios, I am using a 450mm wide and 1,000mm tall baffle, where the centre of the Lowther 800mm from the floor and 140mm from one edge. The BSC applied for this is with 3R 68uF and 1.5mH, all in parallel.

I have tweaked the passive cross over so that the values are now using 68uF and 8mH. The assumed resistance of the coil is 2.7R and the L-pad is 2R. The active crossover is 260 Hz for the Lowther and 150Hz for the Alpha. The charts are below. Both have the same speaker layout, room gain, etc.

Active crossover option.

Passive crossover option.
Part of this, I also wanted to see what would happen if I upgraded to a PM2a or PM5a, and they all perform very similarly in the modelling. I am certain that there will be small tweaks that can be done to optimize each driver, but the basic parameters are very similar.

Right now, the question is whether I should building the passive version, or finish building the Boozhoundlab crossover. I am thinking that the costs for the two are similar. Though I am thinking that the number of caps for the active is quite large (14 nos. in total), and the cost may be relatively high if one chooses the boutique caps. Also, long term, it will probably be easier to upgrade the passive components as part size is not a huge issue. Plus, I already have 1.5 mH inductors for the Beyma handy...

Parts for the Passive 2nd order:

68uF Capacitor x4 (Lowther BSC and XO)
220uF Capacitor x2 (Alpha XO)
1.5mH Inductor x2 (Lowther BSC)
8mH Inductor x4 (Lowther and Alpha XO)
2R Resistor x2 (Lowther L-Pad, though I will likely use the Fostex R80B that I have)

Parts of the Active 2nd order: (Excluding the PS and the JFETS)

0.1uF Capacitor x8 (Input and HP output coupling and XO)
0.033uF Capacitor x4 (XO)
10uF Capacitor x2 (LP Output)
Whole range of resistors and PS caps

Saturday, 2 January 2016

Long time update....

What can I say, it has been two years since my last post, and what has happened in that time? Work, family and renovations... However, in that time, I did manage a few audio things, though I have not had the opportunity to blog about it. Hopefully, this coming year will be better. Here is a short summary of what has happened.

  1. Got myself an Fiio X3 Gen2 for travelling. This has been pretty good, but I have not been very happy with its pairing with my Earsonics. I am still awaiting my LH Geek Wave...
  2. Received the LH Geek Out SFi. I think I got it correct. It has been pretty good overall.
  3. Purchased and built the Elekit TU-8300R. Been fantastic Running them in with JJ 300B, Siemens ECC801s and James output transformers. Enjoying it and dreaming of upgrading it Elrog 300B and Noguchi Finemet output transformers.
  4. Built (finally) an OB with my Lowther DX3. Really enjoying them, but lacking bass. Which leads into the main contents of this post.
So, it took me all of 2 hours to build the OB for the Lowther, it was made using 16mm MDF, and tilted back 10 degrees with the baffle 450mm wide and 900mm high. It is nothing special. But the music that it makes is great, and really focused me on finishing these OB. I was reminded that I got the Lowther and Alpha 15a in 2011... Anyway, I started doing some simulations in Basta and got three working concepts. The three concepts are:
  • Passive first order crossover.
  • Passive second order crossover.
  • Active second order crossover.
Overall, I got the three options to look relatively similar in simulation. Note that all T/S data is by the manufacturer's supplied data and not through any measurement. I have also provided a fourth option using a higher crossover point using 1st order - this was done to keep values low and to use the crossover point to tame the baffle step, and obviate the need for a BSC. See the last option below.

Passive 1st order crossover
The parameters for this design are:

Lowther DX3
Crossover 68 uF
BSC 4R, 68 uF and 1.2 mH
L-pad of 8R

Eminence Alpha 15A
Crossover at 8 mH
'Room Gain' turned on with default settings.

Frequency response for passive 1st order crossover

Passive 2nd order crossover
The parameters for this design are:

Lowther DX3
Crossover 68 uF and 12 mH
BSC 3R, 68 uF and 1.2 mH
L-pad of 2R

Eminence Alpha 15A
Crossover at 10 mH and 200 uF
'Room Gain' turned on with default settings.
Reversed polarity.
Frequency response for passive 2nd order crossover.

Active 2nd order crossover
The parameters for this design are: The source voltage has been modified to allow for a 'flat' frequency response.

Lowther DX3
Voltage set at 2.83 V
2nd order at 250 Hz Q=0.707
BSC 4R, 68 uF and 1.2 mH

Eminence Alpha 15A
2nd order at 150 Hz Q=0.707
Voltage set at 5 V
'Room Gain' turned on with default settings.

Frequency response for active 2nd order crossover

Passive 1st order crossover (high crossover point)
The parameters for this design are:

Lowther DX3
Crossover 18 uF
L-pad of 6R

Eminence Alpha 15A
Crossover at 2.7 mH
'Room Gain' turned on with default settings.
Frequency response for passive 1st order crossover - high crossover point.