Vectortrex

iTrader: (1)

I've been fighting with this for a few weeks and am pretty much ready to throw in the towel. I would have already done so if not for the work involved in "de-bosing" these cars.

My FD came with a bad head unit, so I swapped it out with a fairly high end Alpine unit. I installed the crunchfield adapter everyone mentions, and everything works. However...

It sounds awful.

The bass mix is inconsistent across volume ranges, the mid highs stomp all over everything at driving volumes, bass representation is all over the place. Mids are flat out gone, and I can't nail down highs at all.

All I can think is that I've done something horribly wrong, or the OE head units have out of the box EQ settings that correct this.

Has anyone dealt with this before? I just replaced the center speaker and amp after discovering it was dry rotted (the cap had come off and was rattling around, and the amp was buzzing) but this didn't improve audio quality in any obvious way outside of shifting the sound stage toward the middle of the car.

Andre The Giant

iTrader: (21)

The factory radio has been tuned for the correct response across the range. Most manufactures do this when designing the car. So I would say that is the cause of most of the difference in your setup.

Montego

iTrader: (1)

My first FD had a perfectly working BOSE sytem and it was glorious.... Many young people made the mistake of thinking the wave tube is a bass tube. Nope, and because of it they were very disapointed and the bad reputation of bose blows came about.

Anyway the Bose head unit emitts a passive signal (unlike your average head unit). What that means is that each bose speaker has an amplifier because it is expecting a low level signal to come through. Now stick an active signal instead (your average headunit) and overlolad occurs. On top of that, the BOSE system works at 1Ω impedance instead of the 75Ω impedance that an average system runs at. What that means is that there are mismatch issues as well (a mismatched impedance results in degregation in signal quality). Hence the crutchfield adaptor comes into play as it is supposed to fix all that. Impedance match and attenuate the signal. Now add age to it and it is no wonder a lot of these systems don't sound good.

Since I liked my original FD bose system I also tried to replicate it on my second FD that came with non-bose. The end result was that yeah it worked, but not at the level that I was expecting. So my sugestion is that it is just not worth tackling it. Not only because of all the mismatch impedance issues but your bose speakers are 20+ years old and IMO age is very likely the culprit in your setup. Since you already have a high end head unit just invest in some good speakers and a decent amplifier and you're set.

my $0.02

Vectortrex

iTrader: (1)

The center speaker also has issued despite having a brand new driver and amp (Straight from Bose).

I think I can repair enough of the old head unit to make it functional, so I may try swapping it back and to see if the quality is significantly different. The door drivers and wave drivers should also probably be inspected for dry rot if it still sounds bad, so I guess I'll need need more data before I can make a reasonable choice on how to proceed.

I was hoping to find someone else who'd managed to get good results out of the stock bose system and an aftermarket deck though. I see a lot of people installing them on the forums, but I've never seen any notes on audio quality being different.

Natey

iTrader: (19)

I kept it stock looking at least. The headunit is an alpine, which is obviously not OEM, but the door speakers and Bose snake are intact, but with a aftermarket speakers and amp. A lot of people put gauge pods in the center dash speaker...we're not watching movies and don't need surround sound anyway. At least not as badly as we need to monitor boost and water temp.

Like Montego said, every speaker in the OEM set up needs it's own amp. It's much easier to swap out the speakers and sub. That way it's all matching, easy to work on later and sounds great too

newbbe

I have an FD with bose system which seems to be in good condition (nothing egregious with its sound, no blown speakers). It doesn't sound that great imo. Perhaps they sounded a little better on release, maybe capacitors are getting old.

newbbe

A few thoughts:

- Montego is right the amps are by the speakers. You could easily overdrive them with a different head unit. How does it sound with the volume on the head unit **** turned very low? If ok at that setting, could be related.
- Usually an amplifier's input has a higher impedance (easy load) than a speaker (difficult load). Montego do you mean the speakers are 1 ohm? That could be but the amp is between them and the head unit...
- Not sure where the 75 ohm comes from? Most speakers are 8 or 4 ohm impedance.

Anyway vectortrex, I'd suggest a couple things:
- listen to each speaker individually, with all the other ones unplugged. Then if there is a problem with one of them, you will find it.
- you already changed the head unit, so unless cost is the issue, i'd consider just changing out all the speakers and such. Take out the snake, wrap it up nicely and store it, get regular nonamplified speakers all around, get some hatch space back. BTW if you do this you may need to run larger gauge speaker wire to the speakers.
- as i type this, i'm perusing the wiring schematic in the service manual. There are separate outputs for the waveguide assembly on the factory head unit. This MIGHT mean that there was crossover circuitry in the head unit. If so without that is going to sound crappy. If that hypothesis is true, you could get an appropriate crossover, and run your head units output to it, and its high output to the front speaker stuff, and the low output to the waveguide. The crossover point would be fairly bass-ish I imagine and I'm not sure exactly where. Also it's possible these filters for crossover are in the speaker amps -- in which case this wouldn't be a problem. This could be figured out i suppose listening to one speaker at a time at low volume for frequencies that are outside its range. Or by opening the amp and looking at the circuitry. BTW if this theory were right, routing a full bandwidth signal to the front door speakers might give them too much bass and make them distort.

good luck

newbbe

hi,

I'm looking at the factory "service highlights" manual. Says a couple interesting things.

First, it says the frequency range of the waveguide is 40-200Hz. So, if a crossover were required (unclear) somewhere near 200 might be appropriate. If you listen to each speaker separately and it's the waveguide that sounds bad, maybe that is the issue.

Second, it says the door speakers are "full range". So my guess is they don't need a crossover? The center speaker (if i'm reading this wiring diagram right) is driven off the same head unit output as the side speakers, so it probably has a crossover circuit in its amp. So in that regard it should be ok.

Here's a pic from the service highlights. As montego mentioned most of the speaker drivers are 1 ohm.

Vectortrex

iTrader: (1)

The center speaker is driven from the door speakers, and is used for imaging. One would assume it also has a built-in crossover since the doors are full range, and there's no way for the head unit to target it directly.

The wave also appears to have a built-in crossover, as I don't hear any high frequency noise from them.

The quality does indeed get worse at volume, but this is always at fairly normal listening volumes. FDs aren't the quietest cars in the world, but I wouldn't consider it "cranked".

Speaker impedance is entirely irrelevant unless you plan on replacing the speakers with aftermarket parts. Since they're driven straight from the amps, the head unit doesn't need to care (outside of not overdriving the amps)

I'm going to try adjusting the adapter a bit tonight, see if I can get better results by relying on it less (and making sure all of the outputs are balanced). I'm not expecting miracles out of this, but I'd like it to at least sound roughly as good as the stock setup on my Civic. I feel like that's a realistic target for a premium audio system even if it is a bit old.

newbbe

agree

good to know.

so which speaker subsystem sounds bad? if they all sound bad, i would suspect one of two things:
1) something with the head unit. could be tested by connecting it to some other speakers, heck bookshelf speakers, and seeing how it sounds to confirm it is "ok".
2) something with the wiring or the various subsystems connecting back in parallel for some reason. this might be tested by connecting just a single subsystem to the head unit, and if they all sound good in isolation, something odd is happening with the wiring or combining. do all drivers sound 'bad'?

Vectortrex

iTrader: (1)

OK, I got it sounding pretty good now! Here's what I had to do:

1) The bose adapter was NOT set level across all outputs. Check and reset all.
2) Alpine deck has a factory EQ buried deep in the "paramic EQ" submenu. This is separate from the genre EQs and bass/treble adjustments. It's also really weird. Reset to flat.
3) I'd moved the fader back a few clicks to drive the wave a bit louder. don't do this. Reset to center.

What I noticed after re-reading a bunch of documentation is that there's a very intentional overlap with the door speakers. If you drive the doors at different levels (left right) then you get inconsistent center stage imaging, but only on the highs. It's really uneven too since only part of the highs are mixed into the center, so **** sounds really weird. If you drive the wave higher than the doors(fader), you lose a bunch of mid/low mid representation and clarity (the wave is good at bass but bad at precision; the doors are the opposite).

Basically, the whole thing is very carefully tuned to work under the expectation of a flat, even signal across all outputs. The second you start deviating from that **** gets weird.

I've not had a chance to test while driving, but I tested up to "this kinda hurts my ears" in the garage and got good results across all volumes. Going to call it a win for now.

Worth noting the "Simple EQ" (bass and treble adjustments) seem to be a much safer alternative if you want to tweak these without further damaging the sound stage.

tl;dr version: Check your damn adapter.

Montego

iTrader: (1)

Yes but you already looked that up

Correct & my mistake. I'm an RF Engineer, and 99% of my work involves 50Ω systems... The other 1% involves 75Ω systems. So my mind naturally went to 75Ω as the odd ball impedance.



That is correct the bose waveguide has no melody. It is purely a low frequency device.


Actually it is completely relevant. Mismatched impedance decreases maximum power transfer. What ends up happening is that power gets reflected back into the head unit. Not that big of a deal at audio frequencies but a real killer at high frequencies.

This pictural will help explain what I am talking about:



The dark blue line is the signal being innjected, the light blue line is the standing wave resulting from impedance mismatch, and the pink is the power being reflected back into the system due to mismatching of impedances. The yellow signal is what is actually being transferred minus all of the losses . The better the match the better the power transfer will be.

Another way to look at it is water going though a hose:



The mismatching causes the water to flow back into the faucet.

One of the most diffucilt things to understand is that:
Impedance ≠ resistance even though both read in Ohms.
Where impedance = R±JX Ω While resistance is just R Ω

For example this is an impedance matched system:



Now if you take a ohm meter and hook it up across the 75Ω coaxial cable leads you will not get a 75Ω reading. You will get a short instead. However, if you hook it up to a network analyzer and plot the impedance on a smith chart you will see the cable has an impedance of 75Ω.

When mismatches occur a matching circuit has to be developed:



Which essentially fools the incoming signal to see a matched impedance. Confused yet? Yeah it took me little bit to actually understand that concept myself.

Sweet! glad you got it working.


The thing about the bose system is that all speakers work in unisome to bring out the melody.


Refer you back to this:

newbbe

thanks. this helped me understand this better:

electronics.stackexchange.com/questions/93230/why-do-we-not-care-about-matching-the-input-impedance-of-non-rf-amplifiers

Montego

iTrader: (1)

Interesting page... not too accurate on the impedance match vs resistance though.

^^ That is a real bad example and a common misnomer
Impedance ≠ resistance even though both read in Ohms.
Where impedance = R±JX Ω While resistance is just R Ω . A simple test will unprove the paragrah above: Instead of a 10 km cable grab a 1 meter cable at 50ohm of impedance. Hook up one side to ground the other to the battery. It will short out and it will not yield the expected 200mA. (I don't mean to harp but I can't tell you how many times I have tought other Electrical Engineers who don't seem to grasp 50 ohm impedance vs a 50 ohm resistance)

But in any case like I said earlier the frequency mismatch is not that much of a deal breaker at audio frequencies. Reason being is wavelenght. The larger the wavelenght (lower frequencies) the more forgiveness that it yields for mismatching.

The best way to really think about the power transfer is to think of plumbing and the current of water rushing through it. If a pipe diameter changes all of the sudden (a mismatch) you will get back wash (return loss), turbulation (ripple) in the flow. A matching circuit is the equivalent of a matching pipe that gradually changes the pipe diameter to reduce any ill effects in the water current.

Well anyway I hope that this makes sense

newbbe

agreed.

the part i found helpful was the final post there by snowcat. although i can't vouch for its correctness i am no expert on analog electronics.

newbbe

you mention water hose, heck that makes me think about exhaust systems. those have reflections of importance and the notion of "impedance" there likely apt? think of a horn loudspeaker for example that's an impedance matching device. i think.

Vectortrex

iTrader: (1)

Again, the speaker impedance is entirely irrelevant. The head unit will never see the speaker. It doesn't know, it doesn't care. You could change the speakers to a higher or lower impedance (assuming they would survive) and the head unit would know the difference, nor would there be any change in signal integrity between the head unit and amp. You can even disconnect the speaker entirely, the head unit will be none the wiser.

The reason for this is that the head unit is only going to see the input side of the amplifier. A quick check of my old dead center amp I have shows that part being a fairly typical ST TDA7256 H-bridge amp, which has a datasheet-listed input impedance of 50k ohms.

So no matter what speaker you install after the amp, the head unit will always see a 50k ohm impedance input.

Yes, there could be reflection between the head unit and amps (However this probably does not matter in our application due to frequencies and cable length) but this effect does not in any way scale with or relate to the speaker. You would also be hard pressed to find a car audio head unit with a 50k output impedance. I would suspect the OEM clarion unit wasn't matched either, though I've not been able to locate any data to confirm or deny this.

Montego

iTrader: (1)

wow and with an attitude to boot... Are you actually trying to argue sinusoidal wave propagation? The only reason you got your system working is because you bought the components that others told you to get. The fact is you do not have a clue on why it worked.


You literally lost me at this:

First of all the data sheet states that it is an input resistance. Nevermind that it is an IC which does not particularly mean the headunit sees that. The problem is that you are assuming that all amplifiers (bose and non-bose) have the same input impedance just because they share an IC (an impedance matching circuit fixes that quite easily btw). Now according to bose back in the day, the actual input impedance differs. I don't remember exactly by how much but they were adamant that it wasn't going to work (this is before the scosche adaptor). Yes I called them and they were even kind enough to provide me with a circuit schematic to help me out. Well it did work but it didn't provide the sound quality that I was looking for.

Given that no adaptor existed at the time, I made mine work via an attenuator to bring down the signal levels coming from the head unit. Like I said it worked but the sound quality wasn't as crisp as I wanted it to be. After discussing it with other coworkers (RF Engineers) we concluded that I was having return loss issues due to mismatching and resulted in signal degragation. Which really surprised us due to frequencies involved but hey there it was.

I'm done with this. You are going to have to excuse me but I'm not going to argue signal degregation with anyone that doesn't have first hand working knowledge on this very topic. So believe what you want.

Vectortrex

iTrader: (1)

I'm not trying to be rude; if I've come across that way then I'm sorry.

I don't understand how a two stage amplifier might pass the impedance of its load back to the input. The block diagram on that particular chip doesn't go into details on the specific circuit (beyond that it's a class AB amp), but the input side seems to just be driving the pre-amp, which is later fed down to the power amps.

This could easily just be me not understanding audio amplifier circuit design, but I've been unable to find any good data on amplifier input impedance other than finding ratings for other hardware, and "best practices"

In short: I'm not really well versed in audio circuit design, so if you're confident then it's probably safe to assume I'm just wrong.

newbbe

I'm no expert but I think that is correct. The amp by the speaker will bridge the signal.

I didn't really understand the rest of that post.

Pondering more, clearly there are many cases with audio where we care about impedances -- antennas (duh!), tube amps, phonograph, digital connections, electric guitar pickups, to name a few. Probably not this scenario though.

Glad it's working, enjoy.

newbbe

Unless Bose did something funky with the input impedance on those amps (it is a closed /proprietary system after all). Or unless the wires they used have a lot of capacitance for reason. Unlikely?

newbbe

sorry i missed this interesting comment earlier. hmmm.

Montego

iTrader: (1)

It's alright it’s hard to convey emotion through the keyboard

This is my bad. You are correct the output and input impedance are invisible to each other. I take things for granted since I do this everyday and failed to elaborate. To me it is second nature as it is a typical practice in amplifier design to impedance match to whatever the requirement is. Also given the output impedance being different in the bose system, logic suggests that so is the input.

I will repeat myself so it is all context:
At the time that I decided to tackle this on all I had were assumptions and therefore decided to call BOSE and get the real scoop. They provided me an engineering schematic and also confirmed that the amplifier ran at different impedance. Even though they stated that I would be hard pressed to make it work I decided that at the audio frequencies the mismatch of impedances would not really make any significant distortion. And come on let’s face it, bose wanted to sell me a headunit… Anyway ignoring the mismatch at the headunit, all I needed was an attenuator to bring down the signal power so I wouldn’t overload the bose components. Hooked it up… Not bad… Not excellent though and I was looking for excellent. The only culprit that we could come up with was return loss between the headunit and amp. It was the only thing that fit.

Very important:
Now I called scosche because I was very curious about this and what they sell for a whopping $40 bucks and all it is a fancy attenuator. If I remember correctly mine was like 9 bucks. No matching whatsoever on either devices. But in all honesty I wasn't talking to an engineer and the person behind the phone couldn't really elaborate in detail but lets just operate on what we know.

Taking everything into account I can only conclude the following:
My initial assumption of mismatch not making that much of a difference was correct (the new scosche adaptor proves it). But leaves the question on why mine was subpar. So this is what I have:
- Scosche adapter provides a much cleaner signal that my $9 one. Now at these frequencies and what the human ear is actually able to differentiate I don’t see how it could make that much of a difference. However cheap components = cheap results.
- The attenuators are actually on par with each other but the problem is somewhere else in my system (even though I troubleshot it to death)
- There is no problem with my system, no issue with the attenuator and what is not excellent to me is excellent to others (beauty is in the eye of the beholder scenario).

Vectortrex

iTrader: (1)

Bose using lower impedance speakers shouldn't suggest a different than typical amplifier input impedance, which is almost always high in audio applications. We know the inputs are line level, so we should be following SOP for amp design. We want high impedance to reduce loading on the head unit. If we had low impedance inputs, we'd need to add needless power handling circuitry to avoid overloading the output stage (at which point we might as well just amplify in the head unit)

At least, this is my understanding after doing a bit of research into it. Standard "could be wrong, I know nothing of this filthy analog stuff" disclaimer applies.

For attenuator, it might be worth cracking it open and looking at the topology. There are a number of ways to skin that cat, and the cheapest (single inline resistor) approach is not ideal. (Though realistically I'm not sure how noticeable even that would be)

Really the biggest weakness (in my opinion) of the stock system is that you don't get stereo imaging of the highs. I'm not sure if it was bose's decision or Mazda insisting on it to cut costs, but since all of the high end of the sound stage gets crushed into mono, you will notice a difference if you A/B test it against pretty much any song that uses any kind of stereo effect in the high ranges.

I'm still debating whether or not I'm bothered by this, but I'm considering dropping the stock center and installing some tweeters in the dash corners instead. Biggest thing holding me back is I don't want to modify the dash for it, and running the wires would be a pain even with surface mounts.

As a bonus I could use the center speaker hole to move my gauges somewhere more visible.