AVS Forum banner

Custom MiniDSP 2x4 Balanced build

17K views 74 replies 21 participants last post by  Sam Ash 
#1 · (Edited)
**Prologue**

This thread is fairly lengthy, but I wanted to take the time to explain the decisions that led up to the need for this build. Stay with me as I promise it’s worth the read, or at least I hope it is. I have done my best to document AND link to the parts used for this project, but over time some links may become broken. Please refer to the item numbers and google them to find additional suppliers if needed.

I have tried to break this thread up into sections. The Backstory & Problem, Solution & Project design, and then Build pictures. Jump to the last post if all you care about is the pics.

**Backstory**

“Call me Ishmael”. Lets start with a quick introduction about me. I’m a *bit* of an OCD person, who works in IT for a living (go figure right).

*Edit* Ok, I’m let’s be honest, I’m a MASSIVELY OCD guy who overthinks everything to the chagrin and annoyance at times of friends and family. I don’t start a project or buy anything without researching it till I drive myself batty, and then sit indecisive wondering if I am making the right decision. I’m sure reading this thread in its entirety you will come to the same conclusion.
*Back on topic*

I have been slowly acquiring DIY subs with the goal of increasing the ULF in my large room. I’m a basshead at heart and have always appreciated tight musical deep bass.

Problem: Need MOAR ULF!!!!!

Research: How do I get more ULF

In order to increase the ULF in my room I realized that I either need large ported boxes tuned low (10-15 Hz), or sealed boxes and a TON of power. The pro’s and con’s of Ported v Sealed sub were researched ad nauseum and eventually a consensus was reached. Mainly it was based on the fact that my wife isn’t a fan of large ported boxes, I didn’t have a ton of room to build one even if I wanted too, and if I did build one it would have to be smaller than I wanted resulting in a compromise in performance which I couldn’t accept (see there’s that OCD coming out). That left me to look at sealed subs of a large cone size with a ton of power thrown at them.

Sub Build: Determining the right sub and enclosure size

Looking at subs, it was fairly easy to see, given my research, that a lot of cone surface was needed to get where I was hoping to go. 12 and 15 inch subs weren’t going to cut it without 4 or more drivers, which was cost prohibitive. 18, 21 & 24 inch subs were viable options, with 2 likely getting me “close” to what I wanted. 21 & 24 inch subs would require large boxes, even sealed, given their circumference. Additionally the cost for them tends to be somewhat steep for the obvious reason that most are in the “made to order” category.

Another factor is that I lacked the woodworking tools to build a box, so ordering a flatpack, or buying a custom made one was a MUST. The largest kit available at an affordable (to me) price was only available for an 18 inch driver, which made the 21 and 24 inch subs not a viable option IMHO.

That led me to look at 18 inch, and after again a ton of research in WinISD (again there is that OCD) it was determined the best driver option was the Dayton Ultimax 18. It also helped to know that someone like Mark Seaton was using the same stock driver in his builds in a similarly sized box.

At the time of purchase the intent was to buy 2 of these kits total, however for reasons I won’t get into that didn’t happen. I bought a single Dayton Ultimax 18 flatpack kit from parts express and got it built with the help of a friend. A second one of these never materialized, but what did happen is that at a later date an additionally DIY sub box was purchased which consists of 4 x 12 inch Infinity drivers. When these drivers as a single system, wired for 4 ohms were modeled in WinISD, they had simulated performance very similar to the Dayton Ultimax that was already in use.

That left finding the right amp.

Researching the “right” amp FOR ME:

In researching amp options, it was quickly apparent that the most used option was a pro amp as opposed to a plate amp. The benefits of this are that more powerful amps can be purchased in the future without any modifications to the sub itself. So plate amps were out.

In reading up on pro amps in DIY builds, lots of folks were very happy with DIY subs paired with nuke DSP model amps. The issue’s I saw with the iNukes were largely the fact that they are IMHO ugly as heck, and the fans are loud enough to almost necessitate replacing the stock fan. Realizing that there were mainly issues for me, and not ones most people encountered as the iNukes are intended for pro use. In the pro audio realm the super bright LED’s make it easy to see the display at night, and the fan noise is minimal as they are normally placed far away from the people and speakers. There are always problems when you take an item not intended for a consumer HT setting and attempt to make it work in one. For 99% of folks replacing the stock fan, spray painting the front and covering the lights with tint would have solved those problems. But again my OCD kicked in. Not a fan of modifications that void warranties, and not trusting myself to do a decent enough job (for me) on painting it a different color, that left both QSC and Crown amps as possible options. QSC was eliminated due to again high fan noise, leaving just the Crown.

Crown amps had a very good reputation, overall and had a fairly decent aesthetic to them. They also allowed for RCA OR XLR connectors to be used, as well as speakon AND regular speaker wire. To me this was a big plus, giving me lots of options. They had just released the Gen 2 models, and being budget minded, after seeing that the differences were small and IMHO not worth opting for the newer model, the decision was made to go with a XLS Series 1 amp. Additionally, after seeing the XLS2500 bench tested, its specs were fairly inline with the manufactures published specs. This put to rest the main unknown about Crown amps in a HT setting, which was if they had implemented a 12db per octave filter in the amp starting at 20 Hz. You could read responses online from folks stating one way or the other, and I even contacted Crown myself and was told there was one there. The bench tests proved otherwise.

Problem: no programmable DSP in the Crown XLS

The problem with the Crown XLS series was that it was missing the ONE feature the iNuke DSP line had that I considered of value, and that was a programmable software based DSP. While XT32 in my Denon did a decent job of calibrating subs, The OCD basshead tweaker in me, wanted to install some shelf filters to attempt to lower the F3 of my boxes, or just add some sub 20 Hz boost. Neither of which the Crown had by default.

Solution: Aftermarket DSP, which to choose

That flaw lead me to look at existing options to add DSP to the equation for not a lot of money. The options I saw were the Antimode, SMS-1, MiniDSP and AS-EQ1. The Antimode and AS-EQ1 were removed from the running as they are mainly automatic EQ’s and wouldn’t not allow much customization. The SMS-1 lost out in that it just didn’t have nearly the feature set that the MiniDSP had.

The MiniDSP I needed was the 2x4 balanced. The Mini-DSP is the REAL reason for this thread.

**Backstory complete**
 
See less See more
  • Like
Reactions: eng-399
#3 · (Edited)
**Build Picture**
Below are pictures, in sequence of how the build proceeded:

Chassis holes milled out


Feet added:


Rear panel partially assembled and color coded:




Test fit:



From left to right: RCA & XLR Inputs (only one set used at a time) | XLR Outputs | Ground lift switch | IEC plug



From left to right: Blue LED backlit Power Button (not in this picture) | USB programming connection



Powerbutton installed:



MiniDSP 2x4 balanced in a box:













When Rich started to get ready to install the MiniDSP in the enclosure we spotted a problem.... Can you guess what it might be....


The solution was to remove the plug part and solder directly to the connections. We also decided to strip out the stock connector and use the included adapter to allow for easier install and wiring of the backlit LED in the powerbutton.





Then it was time to remove the MiniDSP board from its enclosure. 4 screws and a push later and we started to expose the board.



As a sidenote, the enclosure itself is VERY NICE and made of good quality stuff. I'd highly recommend purchasing the enclosed kit, even if you only need the board.



The wiring is starting to get there....



Pretty much done



And we have POWER!
 
#4 · (Edited)
The BEAST LIVES!!!!!!





 
  • Like
Reactions: eng-399
#5 · (Edited)
Now we just need to
  1. Ground the board


**Last Note**

I welcome all questions and comments on this as I go. My intent was to try and keep the "write up" portion of the thread at the top, and then let any dialog on it after the fact be posted below to help folks in the future who care mainly about the parts/build aspect and not as much the back and forth.
 
#7 ·
Thanks. I initially wanted to wait till everything was all done before posting it, but since AVSForum lets you edit threads ad infinitum, it made sense to post it since most of the work is done and its mostly just WIP pics of the implementation that are left.
 
#12 ·
If I for some reason in the future I have to sell this, I can sell the MiniDSP separate from the enclosure and the in a box version sells better and was all of 12 dollars more...
 
#16 ·
Cool, I love DIY stuff like this, especially since XLR gear tends to be super expensive.

My own project ideas involve buying several XLR balanced DACs and building a custom enclosure, with DSP ideally, and 16 channels. But the main problem is getting 16 channels of audio in digital form. I know HDMI 2.0 can do up to 32 but there's hardly any commercial HDMI DACs out there and the only ones I could find were 8 channel and HDMI 1.4 only. Other options would include USB-based or network-based like Dante but again, it ends up being not that much more expensive to just buy a Yamaha or Marantz pre-pro and get full Atmos decoding (and most importantly, upscaling) too.
 
#17 · (Edited)
Very nice job! FWIW, I have upgraded the MiniDSP units in my two systems, along with a Raspberry PI/HiFiBerry streamer, with the iFi iPower wall warts, and got an audible improvement in sound quality.
Heck, with that box, you may have room for a real linear PS. Since you're goin the extra mile...:)
 
#18 · (Edited)
*Update* (Also added to the top section)


When Rich started to get ready to install the MiniDSP in the enclosure we spotted a problem.... Can you guess what it might be....


The solution was to remove the plug part and solder directly to the connections. We also decided to strip out the stock connector and use the included adapter to allow for easier install and wiring of the backlit LED in the powerbutton.





Then it was time to remove the MiniDSP board from its enclosure. 4 screws and a push later and we started to expose the board.



As a sidenote, the enclosure itself is VERY NICE and made of good quality stuff. I'd highly recommend purchasing the enclosed kit, even if you only need the board.



The wiring is starting to get there....



Pretty much done



And we have POWER!


Now we just need to do 2 things.

  1. Ground the board
  2. Ensure it sync's to the program on a computer

I'm 100% NOT SURE how to ground the board. We stumbled across this thread on the MiniDSP site which seemed to indicate we should use brass standoffs rather than nylon.

Any thoughts appreciated....

Please note: We did not ground the XLR's to the chassis.
 
#21 ·
Great build, so clean!

Sent from my SM-G930W8 using Tapatalk
 
#23 ·
Did you add up the total cost on your bill of materials? When I was trying to get my rat's nest of equipment organized, I ended up using gaffer tape to tape my Minidsp to the top of an amplifier to secure it. I really like your solution. Looks very professional. Just wondering what the total cost ended up at. Sorry if I missed it.
 
#24 ·
I haven't tallied the final cost yet for the materials. If I had to hazard a guess I'm under 200, above 150 for the parts, not including the MiniDSP. But to me, the end product will be worth it as it will fit in perfectly and allow me to use regular RCA or XLR cables and not be something that is easily moved because of its lack of weight.

Since all the parts are links to the actual products, and I've only really bought stuff from Parts Express, Douglas Connection and Digikey, you could easily price out the parts cost yourself.

Once everything is all said and done I will likely do that just for giggles, but as of now I'm not tracking it because I told myself my goal was to stay around 200 for the enclosure top to bottom and I'm sure I will be under that.
 
#29 · (Edited)
We went with the second diagram in this picture....





Great work Dan I enjoyed watching this come to life. Hope you don't mind if I copy this project for my own room[emoji51]. The project looks very clean and professional. Great work!
Thanks, but my buddy deserves the credit. He did the work, I just did the legwork on layout and ordered parts. And it means a lot coming from you as I've enjoyed watching your builds.

ONE THING I would have done is go with a slightly taller enclosure that what I did. They make a 2U version of the enclosure I bought. Not a lot more price wise either.

You are more than welcome to copy the design, that was the main intent of this thread. Build a better MiniDSP mousetrap ;)
 
#28 ·
Great work Dan I enjoyed watching this come to life. Hope you don't mind if I copy this project for my own room[emoji51]. The project looks very clean and professional. Great work!
 
#35 ·
Something worth mentioning is that if you keep it disconnected from power long enough it will need to be "flashed" with the software first or it wont work.

So dont freak out if you go to hook it up and get nothing. Make sure to connect it up and sync it and your good to go.

I also suggest hooking it up and NOT connecting the outputs first to allow you to view the input and output gains to set your jumpers accordingly.

I have my input set at "high" (2VRMS) and the output at 4VRMS.
 
#37 ·
Weird. Well maybe it was because it was connected to two different computers that caused the settings to not be remembered....

Here is what happened

1. I received and programmed it with a highpass @ 10hz and a lowpass @80hz.
2. I sent it to my buddy Rich who did all the connection soldering.
3. After connecting it all up he connected the device to his computer and sync'd the settings between his PC and the MiniDSP.
4. The MiniDSP settings I had put in were not there.

Between me sending and him getting and installing it was maybe a week or two max.....
 
#39 ·
I was wondering but hadn't spent the time to dig in super deep to find that out yet.
 
#40 · (Edited)
I thought I'd update this a bit. I still haven't dug super far into the software yet, but I've added what amounts to some "bling" to the unit. I wont get into a debate on if these make a audible or measurable difference, but they help with my personal OCD (which this project was all about resolving anyway).

In the simple category I purchased a 10 foot USB 2.0 USB-A to USB-A cable, so I could, you know program the unit. I realized I didn't have one initially so to program it I had to pop the top, disconnect the cable we added, use the cable MiniDSP sent me, program it, and then put it al back together... Big PITA lol. Thankfully this arrived today, so MAKE SURE YOU HAVE ONE, assuming you go with the USB-A side on the connector facing out, instead of the USB-B side (which would let you use a standard printer cable to program).

As to the actual "Bling", I'd previously purchased a set of 50 of these RCA noise stopper plugs from Apollo AV. They are cheaper than the Audioquest or Cardas versions and IIRC they are actually who makes those, so these are just their version which your paying manufacture cost without the markup for a name.

They also sell XLR versions, so I purchased some of their male and female versions to fill in the unused holes on the MiniDSP & the Crowns (since they are bridged).

They sell these on eBay (duh thats where the links are from), as well as Amazon.

I also stumbled into a pair of Wireworld Series 6 Equinox 1.5m XLR cables at a decent price, so I snatched those up as well. I am still going to need to get a decent set of RCA cables for the AVR -> MiniDSP connection and I'm looking at either the Audioquest Black Thunder II or BlueJeansCable LC-1's. I've run both before and like both.

The LC-1's spec's (design page here) make it an interesting cable, both for subs and also for mains, and its very cost effective. Is it 20 cents a cable and Monoprice, umm.... no. Is it 200 dollars a set, no. I also liked the terminations as they were snug enough to not pull off easy, but also didnt exhibit the death grip I've found with some of the Monoprice RCA's in the past (they pulled the ground shield off 3 different inputs on one of my amps). A set of 6 footers would run me about 50 bucks plus a couple in shipping to my door.

The Black Thunder cable I ran was another well built cable, price was decent (I got it used) and I felt the terminations were quite nice and easy to connect and disconnect. I wont get into what audible changes I feel it made as then we get into a cable debate and get way too off topic. However buying new, a set of 1.5 or 2m of these cables runs into the 160 dollar range, which is a bit pricey for... you know... sub cables lol...

**Slightly Off Topic** For those of you looking for some decent speaker cables at a very affordable price, I do suggest BlueJeansCables speaker cables. I really like they way they sonically weld the cables so you NEVER need to worry about the termination coming loose, or pulling out if you tug too hard on them. And for the price they aren't that bad if you dont want to deal with making your own DIY cables and terminating them. Assuming of course you dont need like 100 foot runs. Front 3 speakers with the AVR nearby would be fairly cheap IMHO. Surround cables could get pricey so those may be worth DIY'ing. I also like them because the Belden cable they buy comes from good old Richmond Indiana, so I'm supporting a local Hoosier company :D
.
 
#41 ·
A friend turned me onto this video regarding the RCA/XLR caps I felt worth posting here given it shows some quick measurements of them

 
#42 · (Edited)
Next job: Tackling the actual EQ which is what this was built for. I started posting in Matts multiple subwoofer thread, and have moved those posts to this thread below, and added commentary to them if needed..


@mtg90

Here is my plan of attack for tomorrow. Let me know if I've got it wrong. I've NOT redone XT32 since getting a second Crown amp and adding the MiniDSP. Given your previous suggestions I'm going to follow it and run a single cable to the Mini so XT32 sees both subs as a single unit, and EQ the two through the MiniDSP to set distances, etc....

So here goes:

1. Turn off XT32, set the crossover on my mains as high as possible. Set the LFE as high as possible.
2. Turn off my 5 channel amp
3. Run REW on each sub independent of each other
4. Keep running REW sweeps while playing with the distances (after measuring the actual distances) to get them to sum right
5. Use REW to generate the EQ filters for the two subs, apply EQ filters into the MiniDSP
6. Connect a single sub cable from the MiniDSP to the Denon
7. Rerun XT32
8. Use REW to measure the sub to main crossover integration
9. If needed adjust the distances in the Denon
10. Apply any additional EQ in the MiniDSP such as shelf filters and verify results via REW.
I've started but getting hung up..... I'm trying to import the REW .txt file into my MiniDSP balanced 2x4. Each time I try I get this message: Your biquad in your input file is not correctly formatted.

I'm using the 2waySubAdv plug in.

Any thoughts?

PS here is what it looks like after playing with the delay on the Dayton v Infinities. I'm POSITIVE I can eliminate that 80hz one with either EQ, or distances (MiniDSP wont let you go higher than 7.5MS delay)

Now keep in mind this is without ANY XT32, or MiniDSP filters to help which should help make a difference too...

So I realized AFTER I posted that graph... XT32 was still on, so I had to redo ALL MY MEASUREMENTS.... I don't have anything showing those, but I did get somewhere close.

One thing I've realized is a slight pitfall of the MiniDSP is that you can only set the delay on each channel to a MAX of 7.5MS which may or may not work well for everyones situation. In my case I think I need more, as adding in more delay on the Dayton may even out that 80hz dip...

Currently the fronts are set at 250, and the LFE of the sub is also 250. The fronts are actually OFF in that measurement.

I am moving from letting XT32 handle BOTH subs, so using a MiniDSP to EQ the two together (and apply EQ), then let XT32 see it as a single sub and set distances, etc....

Then once THAT is done, I will adjust the distances to try and see what I can do about it, if anything...

But I'm stuck right now since my REW filters wont load into the MiniDSP. So until I get that done, I can't redo XT32 and anything else I want to do.
Apparently the MiniDSP site is NOT UP TO DATE. When you generate the filters and go to export them, you need to select Send filter settings to equalizer NOT send Export filter settings as text

Got it working.... got some measurements.... not sure if I helped or hurt......

Will post them later tonight (still feel like junk and have like 60 diff measurements).

My thinking is that I may have actually done it backwards.

Here is my potential revised order:

1. Connect both subs to Sub 1 & Sub 2 respectively
2. Run XT32
3. Zero out the distance settings in XT32
4. Use MiniDSP to set distances by measuring via increments in WinISD
5. Generate MiniDSP Filters
6. Apply MiniDSP Filters
7. Apply house curves

I will post my graphs tonight and see....

I may also just need to find a way to get Matt and Mike down to help me out one day lol.....
Here are my final measurements from last night. I have today off so I may try to run each sub independently off the Denon so I can add additional delays via the distance, or I may just sit here and do nothing.... because I'm still home sicker than a dog lol...

This is the subs only... I think the "After" also is with the sub crossover at 80hz, hence the drop off there compared to the other measurement...



This is the right channel and sub comparison. This is why I think I need to do XT32 with each sub run off its own channel, THEN adjust via the MiniDSP. Besides some low end boost, there really isnt a ton of difference between them....

 
#43 ·
Nothing wrong with trying different scenarios. My argument for running xt32 second is just trying to free up audyssey processing to areas of the bass not well addressed by PEQ.

IS your hope that audyssey will adjust the delays better? You can run audyssey to get delays, then turn off the correction, setup the PR'S, and then rerun audyssey. You could also use the values audyssey comes up with for distance.

I agree you should run each sub off its own audyssey channel. Audyssey can align the two subs and does so in a sophisticated way that is computationally the same as my earlier suggestion. It aligns the impulse response for the subs to each other and to the mains.

I've never needed or found a need for 7.5ms of delay or more other than when the dsp I was using had its own delay. I have found issues where the dsp had inherent delay that made it so the mains were off and needed more delay. I would have expected audyssey to account for it but my measurements of the impulse response showed it was far off.

Your 80hz null looks much lower Q and less deep. The rest of the nulls seem to be high Q and may be hard to treat.

Is this setup now so the main speakers are running at their proper crossover level and subs with the proper low pass? Is this still with the 250hz crossover?

Are you applying boost to get rid of the 80hz null or is that all changes in delay. I'll post my last measurement. I need more work too.


Sent from my iPhone using Tapatalk
 
#45 ·
Nothing wrong with trying different scenarios. My argument for running xt32 second is just trying to free up audyssey processing to areas of the bass not well addressed by PEQ.
That was my thought as well, however I'm not sure in my application thats the right call. But that was also why I did things the way I did yesterday.

IS your hope that audyssey will adjust the delays better? You can run audyssey to get delays, then turn off the correction, setup the PR'S, and then rerun audyssey. You could also use the values audyssey comes up with for distance.
My hope is that by being able to use delays in BOTH XT32 (via distance) AND the MiniDSP, which is limited to 7.5MS. The distance it tells me 7.5MS equals is 2.58M or 8.5 feet. My subs are FARTHER APART from each other than that, so I cant see if adding MORE delay to it would be helpful, which given how the 80hz deal was slowly going away as I played with the delays in the Mini (after using the ones from XT32 originally)

I agree you should run each sub off its own audyssey channel. Audyssey can align the two subs and does so in a sophisticated way that is computationally the same as my earlier suggestion. It aligns the impulse response for the subs to each other and to the mains.
I think thats where I am as well after looking at my responses.

I've never needed or found a need for 7.5ms of delay or more other than when the dsp I was using had its own delay. I have found issues where the dsp had inherent delay that made it so the mains were off and needed more delay. I would have expected audyssey to account for it but my measurements of the impulse response showed it was far off.

Your 80hz null looks much lower Q and less deep. The rest of the nulls seem to be high Q and may be hard to treat.
The 7.5MS delay was actually taken by looking at the XT32 distance for my Dayton and converting it to meters, then seeing how many MS of delay equaled that distance. In this case, 7.5MS delay was actually LESS than the distance in feet XT32 used for calibrating the dayton.


Is this setup now so the main speakers are running at their proper crossover level and subs with the proper low pass? Is this still with the 250hz crossover?

Are you applying boost to get rid of the 80hz null or is that all changes in delay. I'll post my last measurement. I need more work too
The measurements of my subs in the AFTER is with the crossover at 80hz and my mains also at 80hz, NOT 250. The before measurement in the SUBS ONLY graph is WITH the crossover at 250.

I'm gonna do a couple things around the house and then likely redo everything I did yesterday with the subs each running on their own channel lol....

Then after that I will see what I need to do with the two subs delay wise...
 
#44 ·


Here is my last run. You can see all the evidence of various room modes. I have axial modesat 25hz and 47hz and you can really see their effect. I have an 80hz Null, like you, that is not an axial mode in my case. It could be SBIR, but it also doesn't match anything other than a ceiling reflection (but seems too deep to be a ceiling reflection). No amount of delay adjustment seemed to fix it, but having said that, I am not optimally setup at this point either. I sold my minidsp due to noise and only have DSP for two of the three subwoofers and have to rely on a phase switch and analogue PEQ for now. That 113hz notch does match the rear wall reflection but its oddly so high Q that it looks more like a resonance.
 

Attachments

This is an older thread, you may not receive a response, and could be reviving an old thread. Please consider creating a new thread.
Top