Simple Passive Harsch XO Using PTT6.5 and RS28F in a Waveguide

Jump to Latest
xrk971

xrk971

Founder of XSA-Labs
Joined 2012
etsy.com/shop/XRKAudio
Paid Member
Nice work Jimk04!
Looks cool with a bare phase shield

20211204_153604-jpg.1002604
 
  • Like
Reactions: 1 user
O

okankav

Member
Joined 2008
xrk971 said:
Hi Solve,

To get Harsch, you need 4th order electro acoustic slope. Note that Purifi has natural 2nd order roll off at 3.5kHz. Now apply electrical low pass 2nd order at 3.5kHz and you will get 4th order. Now apply 2nd order high pass on tweeter circa 3.5kHz. You may need EQ to shape it.

The best way to do this is to measure the raw, and use the DSP simulator in REW to test these out. Otherwise trial and error will take a long time and never get quite close. Even though it’s easy to change, remeasuring takes time. Whereas simulating in software is quick and precise. Then transfer the DSP settings to hardware.

You can also do this on Xsim and then transfer the filter to DSP implementation.

Measure on woofer axis as that gives proper delay. No delay is needed unless you like to listen with tweeter on top.

You will notice you have baffle step. So a shelving filter needs to be applied to bring up the bass. I did this with a resistor parallel to the inductor. Here, use a high frequency shelf about -5dB, Q=0.5, circa 900Hz on the Purifi.

The tweeter is very sensitive and will need significant attenuation to get to 84dB. Just plain padding resistor or level attenuation. It needs at least a -12dB/oct high pass though to keep distortion below 2kHz from being and issue.

Good luck!

If you look at your raw woofer data, note the natural roll off at 3k5. That’s the clue to use BW12 at 3.5kHz to get BW24 overall.


Alternatively, you can buy the relatively inexpensive parts for the passive XO and have a nice result that I worked to get right so you don’t have to. But if DSP active is the way you want to go then that’s fine.

XO:


The air core could be replaced by much less expensive iron core and probably sound fine. More bass actually as DCR is lower. P2P construction could certainly be used here.
Click to expand...
Hi, I have a pair of ptf6.5x08 nfa-01. How should I change the crossover values?
 
xrk971

xrk971

Founder of XSA-Labs
Joined 2012
etsy.com/shop/XRKAudio
Paid Member
It’s rare that simply halving or doubling caps and inductors for 4 vs 8ohm drivers will give the best crossover designed for one. It usually means a redesign as it is a different driver really. You can try it, but I would suspect it won’t be ideal. I don’t have one to try so can’t help you. it might sound good actually, you never know.
 
M

mcevoy

Member
Joined 2010
xrk971 said:
Hi MainFrame99,

Yes, I have a floorstanding design - it has been fully simulated in Akabak and CAD plans have been drawn, please PM me if you are interested in the full plans. It uses all 3/4in BB plywood and is 40in tall x 14.25in deep x 9in wide (external dimensions):
attachment.php


Without room gain, and after baffle step loss in 4pi space, the speaker has a nominal sensitivity of about 84 dB and a -3dB point is 40Hz and -6dB is 29Hz.

Here is the predicted SPL vs frequency - it is designed with a smooth rolloff that will integrate with a room's modes to give a nice response down to 30Hz:
attachment.php


Here is the predicted impedance curve which shows the tuning is at 31Hz:
attachment.php


It is designed with the PTT6.5 woofer on the top and the tweeter below it (RST28F with WG300 waveguide for example). If you want to use it with another tweeter, that is fine too but you will need to design your own XO or go with active DSP amplification.
Click to expand...
Love this design XRK! Could this be made about 32" tall? Do you have a measured drawing that I could see?
 
M

mcevoy

Member
Joined 2010
xrk971 said:
Thanks for that Ugg10. The design looks good and I think the 30in tall one is close to what Mcevoy was looking for
Click to expand...
Hi XRK. Thanks is close to what I’m looking for. Newbie question: Do you measuring line length down the middle of the line or the full line lengths? Down the middle of the lines I get 88” for XRK’s design and a few inches less for Ugg10’s design.
 
U

Ugg10

Member
Joined 2017
@mcevoy if you add up the red numbers (length of the red line) on my sketch then that is the line length, comes out to 88.5” so very close to XRK’s original and iirc I estimated using 3/4” wood (19mm). Yo can always add a plinth and feet/outriggers to get to your desired height. Just remember that the crossover is designed for your ears to be level with the woofer centre (note the sketch has the tweeter and woofer the Sharon way round for the Harsch version). Hope that helps.
 
U

Ugg10

Member
Joined 2017
@xrk971 out of interest did you ever try or model a combination of the PTT6.5 and the 10f8424 in a Fast/WAW combination, maybe with Harsch. So a cross between this speaker and your RS225/10f one. I did a quick model with published data (the higher sensitivity PTT6.5W-4 version) and a simple LR2 seems to work really well with a 500-600hz crossover point. Thoughts, I think this would make a nice compact speaker (16 litres BR or your stand mount TL), might also be nice with a SB MW16P-8 if fist is an issue with the Purifi.

Any thoughts welcome.
 
L

lstasz

Member
Joined 2016
Way back on page 4 almost 3 years ago. I wouldn't remember that long ago either. I didn't know if anyone would be looking at this thread since there was no activity since last December. I will be looking at the simulations on this thread to help me simulate with the Dynavox LW6004HR-N. I'm new at simulation so I have a lot to learn. I think you mentioned that the bass extension would not be as good and that a straight MLTL might work.
 
You must log in or register to reply here.
Top Bottom