Hivi 3.1A DIY Speaker With Sehlin Mod Review

[direwolf08]

How did yours turn out? So many possible ways to go to finish them

Pics

 

[Cal_Cobra]

Pics

Love the veneer! I did a similar theme, but more of a ying to your yang. I used red oak with a mission oak finish, it's dark, think vintage Stickley with patina I went for a vintage high end retro look. DId the speaker panel and rear panel in satin black, and it came out really nice.

I applied the 5th application of finish yesterday on the veneer. Tomorrow night I'll wet sand/steel wool the veneer, then go into polishing and wax mode. Then I can finally install the speaker drivers, and give them a listen.

 

[Scoobis]

Sehlin Mod with a twist - literally​

One thing that immediately struck me and did not set well with me was the position of the inductors thinking there must be a lot if signals being induced between them. Some quick testing and measurements verified this. I spent the last weekend injecting signals into the coils at various phases between them and finding the best fights and angles to bring the interactions own as much as I can. in some cases lower than my equipment could effectively measure.

Sorry there is no before and after measurements, I didn't even have plans to post this. but, thought that some of you might find it interesting and might want to try this too.

The mounts were 3d printed, and I've posted them on Thingiverse HiVi Inductor Mounts Hivi 3.1a

any questions or comments please post.

I've had this it since 2019 but stopped the build after building the cabinets due to this, kinda shelved it until this past weekend. - I will be finishing them up now and be giving them a listen, .. id love to AB those against ones without the mounts to see if the difference is audible.

a photo from another angle.

 

[Scoobis]

So I've been a bit delayed in finishing my speaker build...by...5 years. the foam first turned yellow within a week or so of it being unboxed and sitting out... and is now brown after being out for 5 years - it used to be white, and soft but is kind of feeling a bit crispy now when pressed. I'd like opinions on if I should still use it, .. or maybe use some polyfoam to replace it, or go with polyfill etc?? i know Scott had commented in his crossover design that it was based around using the foam and did not know how things would change with different infill. ... any thoughts on this?

I do have some new poyfil high-density polyfoam and seems extremely similar except for its blue/green color and it is about 1/4 in thicker. see pics below ... white paper included for white balance / color comparison.

 

[Cal_Cobra]

So I've been a bit delayed in finishing my speaker build...by...5 years. the foam first turned yellow within a week or so of it being unboxed and sitting out... and is now brown after being out for 5 years - it used to be white, and soft but is kind of feeling a bit crispy now when pressed. I'd like opinions on if I should still use it, .. or maybe use some polyfoam to replace it, or go with polyfill etc?? i know Scott had commented in his crossover design that it was based around using the foam and did not know how things would change with different infill. ... any thoughts on this?

I do have some new poyfil high-density polyfoam and seems extremely similar except for its blue/green color and it is about 1/4 in thicker. see pics below ... white paper included for white balance / color comparison.

I don't think I'd use it. I used the foam that came with mine, it worked out fine as far as installing it goes (although the cutout template they have is flawed, easier to just measure the sections that need cut and do it yourself).

I think because it's enclosed, without the ability of light to hit it, once installed, it'll be stable.

 

[Cal_Cobra]

I completed my Swan HiVi 3.1A DIY build, and as luck would have it, my WiiM Ultra arrived the same day I completed my speaker build. I have them staged on my kitchen island while I test them before relocating them to their final dedicated listening space.

I must say I'm most impressed with the SQ on the speakers I ended up using only 3 of the 18 crossover components from the kit, going all out for the full Sehlin crossover upgrade to mod #6. Thanks for sharing the design, it wasn't an expensive upgrade and the speakers sound terrific! I was previously listening to a pair of the infamous, love them or hate them, Sony SSCS5 3-Way bookshelf speakers (I have the Niel Blanchard crossover upgrade components, just need the time to do another speaker upgrade haha ). Most of my time was spent on the speaker cabinets. I veneered them with red oak on the sides and top, classic satin black on the speaker baffle, rear and base and they came out great. The veneer has five coats of finish, buffed out with an orbital polisher

When I did the initial speaker testing/burn-in/shake-down, I had them powered by the Aiyima T9 Pro, and I though they sounded great. That is until I hooked them up to the far superior, far more powerful FX Audio L07amplifier (200W+200W@4Ω), and that was a pivotal experience! With the new amplifier these speakers really came to life, they sounded better in every way, far more punch, wonderfully detailed sound stage, crystal clear, even when pushed to demanding levels, and they can take some juice! It really highlighted how a more powerful amplifier isn't just about being loud, although it certainly can get loud.

The Swan HiVi speakers with the Sehlin crossover upgrade are really quite amazing!

For anyone either interested in, or are planning to do this, don't waste four hours of your life watching the Z videos on YouTube, instead find the Swans Speakers Facebook page, and watch all nine of the videos on building the Swan HiVi 3.1A DIY build. Aside from the crossover upgrade, these videos tell you everything you need to know to build the kit, with sound, no nonsense info from an informed experienced speaker designer and builder. This is his kit, so nobody really knows it better.

Ditched the Aiyima T9 Pro and WiiM Ultra (has issues w/Amazon Music playback) and ran the speakers from the FX Audio L07amplifier and a SMSL DO200 PRO DAC, and it sounds stunning!

 

[eksine]

For the love of God don't paint it the way he did with latex and a roller like it's drywall. The way he glued it up really messy and without checking if the corners match up bothers me. On my kit all the side walls were warped, I don't see how masking tape is strong enough, I used duct tape. And I'm not an expert but a long time ago I took a basic soldering class, you're supposed to mechanically twist the wire so that you're not relying on the solder alone to form the bond because solder can break with vibration. Both are minor but even if he's the original speaker designer his craftsmanship is questionable at best. At the very least the paint job is sketchy. Watch some car painting videos , that's the right way. That's my opinion

 

[Scoobis]

I noticed in the HiVi video, it seems they were originally using some kind of polyfil type batting.
i also noticed in a M3A repair video, they used a Polly batting in that and not foam. not sure it makes a difference but its interesting.

since my foam has basically started to disintegrate, i am in search of what to use ... foam or Poly fill etc. i had actually considered fiberglass for a moment, but found out that's not so great for the bass response in a ported enclosure plus the health hazards.

i am on the opposite side of the fence on the finishing... I actually wanted that kind of roller stippled look, I almost went with duratex, to each there own i suppose.

 

[Cal_Cobra]

I noticed in the HiVi video, it seems they were originally using some kind of polyfil type batting.
i also noticed in a M3A repair video, they used a Polly batting in that and not foam. not sure it makes a difference but its interesting.

since my foam has basically started to disintegrate, i am in search of what to use ... foam or Poly fill etc. i had actually considered fiberglass for a moment, but found out that's not so great for the bass response in a ported enclosure plus the health hazards.

i am on the opposite side of the fence on the finishing... I actually wanted that kind of roller stippled look, I almost went with duratex, to each there own i suppose.

I noticed in one of the Swan videos that their kit in that video came with both the foam as well as the poly fill batting, so i suspect at one point it was both.

These are DIY kits, some folks will leave them bare, others will paint w/rollers, Duratex, Rust-Oleum spray paint, etc., and others will go all out with a show paint job or veneer the cabinets. The Swan videos showed how to use the black faux wood vinyl cover, paint and they also show the DIY3.1 speakers on their FB in various veneers. None of these speaker cabinet finishes will change how the speakers sound Do what works for you.

 

[jayapple]

Has anyone tested HIVI audio?
HIVI has some fame in China, but most of them imitate Dynaudio, especially now that Dynaudio is produced in China, some accessories are the same.

I am curious about their brand sound effect.

 

[Scoobis]

there's actually a video on youtube where a guy puts a signal generator on one inductor and an oscilloscope on another with a speaker and he shows that unless you have inductors perpendicular to each other they cause electromagnetic interference. with the board layout here unless it's 12 inches apart they will interfere. you need to rotate 2 inductors on this board 90 degrees vertical to do it properly and spacing it apart more would be better. those caps and resistors also cause interference believe it or not and they are too close. also I really don't trust that board to properly conduct current. I bet you there's too much resistance , like a restricted water pipe that's undersized. I would use wire instead and ditch the printed board.I'm thinking of using the entire box as the board, meaning place the parts all over the place to increase their distance. I thought about an external crossover too, but not sure if it's worth it.

The info that is floating out i=on the internet on what good and bad positions are for coils is incomplete at best. you would be surprised how sensitive the null is at some positions and angles. Also, some of the angles that are supposed to be bad, can actually work with little to no inducted signal between the coils, but the angle and heights are sensitive and change over distance too.

you can look at my mounting of my coils on my recent post which drastically reduces the inducted signals between the coils. according to the internet wisdom, this should not work but it does.

there are nulls at certain angles / positions and distances. totally worth doing the oscilloscope and signal generator tests but you need to be able to inject more then once signal and be able to alter the phase tog et a really good idea what's going on between all the coils. obtaining nulls on multiple coils at once is like assembling a jigsaw puzzle were the peace's are all the same color

i wish i had done a video of me making the mod and the tests for the coil mounts now. a difference of 1 degree or a Height different of 1 mm can skyrocket the inducted signals. that is why i used 3d printed mounts so the angles would be reproducible by others.

totally interesting stuff,

 

[eksine]

The info that is floating out i=on the internet on what good and bad positions are for coils is incomplete at best. you would be surprised how sensitive the null is at some positions and angles. Also, some of the angles that are supposed to be bad, can actually work with little to no inducted signal between the coils, but the angle and heights are sensitive and change over distance too.

you can look at my mounting of my coils on my recent post which drastically reduces the inducted signals between the coils. according to the internet wisdom, this should not work but it does.

there are nulls at certain angles / positions and distances. totally worth doing the oscilloscope and signal generator tests but you need to be able to inject more then once signal and be able to alter the phase tog et a really good idea what's going on between all the coils. obtaining nulls on multiple coils at once is like assembling a jigsaw puzzle were the peace's are all the same color

i wish i had done a video of me making the mod and the tests for the coil mounts now. a difference of 1 degree or a Height different of 1 mm can skyrocket the inducted signals. that is why i used 3d printed mounts so the angles would be reproducible by others.

totally interesting stuff,

this is the video I mentioned. so you're saying he needs to test out various signal frequencies to do a proper test?

 

[Scoobis]

hmmmmm..... ok finally got the speakers assembled and..... well...there is something not right to my ears.. i went straight for mod 6 which might have been my first mistake, assuming that's the sound i wanted - listening on paper instead of hardware.

There appears to be a bit of a hole / holes in the audio spectrum. i immediacy noticed this on the first song i played. AC/DC you shook me all night long. The hi hat that goes along with some of the drum beats is almost inaudible. other symbol sounds are fine. I have double / triple checked the crossovers, i even tried changing the phase of the mid just to see if it was a phase issue. and it persists.

i have done an a/b the speakers with several other speakers and its even more obvious.

before i start diving into this more, I thought it best to ask the community about this. see if i am alone or not.. both my speakers do exabit this, it doesn't appear to be a bad driver or bad componenet. but its totally one of those things that, once i noticed it i cant un hear it.

heck, even on my tv speakers you can hear the high hat clearly

 

[GGroch]

Before i start diving into this more, I thought it best to ask the community about this. see if i am alone or not.. both my speakers do exhibit this, it doesn't appear to be a bad driver or bad component. but its totally one of those things that, once i noticed it i cant un hear it.

heck, even on my tv speakers you can hear the high hat clearly

One common issue when building the kits with original crossovers was failing to scrape the varnish off the induction coil ends before soldering them. Getting the cap polarity wrong might also cause this kind of thing.

 

[Scoobis]

One common issue when building the kits with original crossovers was failing to scrape the varnish off the induction coil ends before soldering them. Getting the cap polarity wrong might also cause this kind of thing.

thanks, i did check that the continuity was good on coils, .. i did a good job on getting the varnish off, and even tinned the leads before i even insert and soldered them, so good connectivity there.. I did a second continuity test from the pads though the coils when things didn't sound right just in case, just to be sure. also double checked all the right values were in the right positions. my first thought was what did i mess up ... totally not saying i haven't. made a mistake,. but if so, i have I've not caught the error. i will go though it all again anyway.

all the caps are non polarized, so, polarization is a non issue in this particular case.

 

[Ktacos]

sounds like a driver may be wired in the wrong polarity. I generally measure each driver with the others disconnected to make sure everything is wired in the correct polarity.

 

[Scoobis]

sounds like a driver may be wired in the wrong polarity. I generally measure each driver with the others disconnected to make sure everything is wired in the correct polarity.

I have to admit that was one of my first suspicions as well, it really does sound like some kind of cancelation .... I will double check the wiring, and also do some tests to verify the polarity of the drivers is actually marked correctly.

i already verified the woofer using a current limited power source to show positive excursion, so i know its correct. i can not do that on the other drivers... i can likely test the rest by driving them with the same signal and see when they cancel vs add. work my way up comparing woofer to mid and mid to tweeter. unless somoni has a better method.

 

[Cal_Cobra]

View attachment 383547 hmmmmm..... ok finally got the speakers assembled and..... well...there is something not right to my ears.. i went straight for mod 6 which might have been my first mistake, assuming that's the sound i wanted - listening on paper instead of hardware.

There appears to be a bit of a hole / holes in the audio spectrum. i immediacy noticed this on the first song i played. AC/DC you shook me all night long. The hi hat that goes along with some of the drum beats is almost inaudible. other symbol sounds are fine. I have double / triple checked the crossovers, i even tried changing the phase of the mid just to see if it was a phase issue. and it persists.

i have done an a/b the speakers with several other speakers and its even more obvious.

before i start diving into this more, I thought it best to ask the community about this. see if i am alone or not.. both my speakers do exabit this, it doesn't appear to be a bad driver or bad componenet. but its totally one of those things that, once i noticed it i cant un hear it.
heck, even on my tv speakers you can hear the high hat clearly

I tested AC/DC you shook me all night long on a pair of Sony SSCS5 3-way book shelfs and I heard a bit more on the hi-hats vs the Swans, but the Sony's are known for sizzle mode trebble, so that's not shocking IMO. That said, I the Swans hi-hats replication on that track was spot on.

 

[Scoobis]

I tested AC/DC you shook me all night long on a pair of Sony SSCS5 3-way book shelfs and I heard a bit more on the hi-hats vs the Swans, but the Sony's are known for sizzle mode trebble, so that's not shocking IMO. That said, I the Swans hi-hats replication on that track was spot on.

Which version of the crossover are you using?

 

[Scoobis]

this is the video I mentioned. so you're saying he needs to test out various signal frequencies to do a proper test?

no, i don't think frequencies are that critical to test or demonstrate induced currents between coils. .

It is hard to include everything in such a video. i will give one example: The horizontal coil, if it is placed so the center height of it is centered with the heights of the other coils, there is a null point that can be exploited for lower or even null the interactions at some angles. the potion and angle is more sensitive the closer you get. Think of it in terms of 3 dimensional space not just the two dimensions with the pcb being a fixed plane.

Think of a null location this way...if a coil is in a location that can induce noticeable currents but is positioned so one half a coil is in the field that induces a current flow in one direction and the other half is in a field that can induce a field in the opposite direction, the two induced currents will cancel each other out, and the net result is no induced current flow though the coil. so no induced signals between the two (works both ways).

depending on the angles and distances, the position null point can be very small.

These are things you can see an experiment with if you can inject signal(s) and see them on a scope to see the interactions.

those simple charts for positioning inductors, at a high level i don't disagree with, but, its just not the full story and is a bit too drummed down. the videos you posted dives deeper and that is very cool. but if you play around with it you yourself you soon see there is more to know and things you can do that seem to fly in the face of the conventional wisdom.

A statement that made in the video that you can not completely get rid of interactions without them being way far away is summing you will find there is an exception to that where if the coils are at a point were the fields can cancel each other out in combination with the angle at which the magnetic field crosses at angles that can only produce week currents at best. so the net is there is no interaction. problem with this kind of setup is that the position for it to be zero can very sensitive. Again i am just saying that there is always more to know not that the info presented, not that it is in any way wrong. but unless your going to dive deep and test each crossover, then using angles / positions that are less sensitive to height and angle changes are a wise choice.

i am unsure if i have enough coils laying around to do a video demonstrating all this. it be nice to just add to the info already demonstrated.. it would be a long video though..

on frequency changes....changing frequency changes the inductive reactance and will change voltage / current that gets applied. its best to stick with a single frequency for the testing so your doing apples to apples when it comes to the induced signal A frequency sweep when you think you have things set is worth while just co confirm things.. same goes for sweeping phase angle when ejecting signals into multiple coils.. you have to consider that the signals at the coils Weill not be in phase with each other so when testing multiple coils that are postpone close enough that the there within each others magnetic fields, that is summing to consider.

Take the hivi crossover as an example, the coils are way to close. so without some carful positioning, your goin to have a lot of interactions. So i made a lot of use of finding the null points to drastically reduce interactions. but with 4 coils, its a bit of a jigsaw puzzle tying to get them all at zero interactions, so there is a little compromise. in one example raising one coil by 2mm made a huge difference in the amount of current / voltage being induced.

i am sure there are other combinations of positions that would be just as good, but that's what i came up with.
.
i totally agree though, if it is an option, distance is your friend.

if you have a spare hivi crossover board, you should try out the mounts i made , do measurements in comparison to the conventional mount positions and with the new mount positions, ... also change those mount angles and watch the changes. i think you will find it very interesting and informative, and a bit counters intuitive but at the same time might clarify some things that help you in future layouts

oh, if you do this, load down the "receiving" coil with a resistor, i've used 8-60 ohms with good results so whatever you have on hand is likely ok, just avoid using a high value resistance.. it helps prevent picking up noise. and also insure your dealing with actual induced signals that make a difference not low current stuff that shouldn't drive a speaker anyway. it is easy for a test setup to pick up noise from the signal generator without such a load that is not induced thought he pol interactions, so defiantly use a load resistor across the receiving cool.

Also it helps to have a scope that can go down to 500UV because when you get down to nulling those coils out, the signals will be down in the noise. i got to a point where the signals were so low, that i was getting more signal from the wires radiating than then from the coils. When you get things down that low, you pretty much met your goals anyway.

for frequency i picked frequencies that had a reasonable inductive reactance ... again in that 8-60 ohm range seems to work well for me. i happened to match the resistance to the inductive reactance when testing .. that made fore easy math for me to get a good idea of the actual amount of signal being induced, but, .. you don't totally doo not need those kind of exact measurements so for back of the matchbook testing its not critical at all as your really doing comparative analysis anyway. position A vs position type things.

hope that helps explaining my observations

if you wanna dig deep into this, maybe we can pick this up on its own thread and post results there etc.