Summary

I’ve seen discussion of the frequency response of the MC(S) standard (non-HiFi) sound system, with some suggesting that the rear channels didn’t get full bandwidth signals. Since I’m looking at subwoofers, I decided to find out by testing my 2008 MCSa.

The sound going to the rear speakers is severely cut. Compared to a 2KHz reference level, it’s down 3dB at 100Hz, and down 30dB at 20Hz!

The frequency response between 100Hz and 2000Hz to all channels is not flat. Response from 2KHz up is pretty good.

I’ll summarize the rear channel response in one word: sucks. Here’s a picture that shows how much it sucks. Remember the Y axis is a log scale.


Test Setup

The test setup consists of a HP 3314A Function Generator for the signal source, a Tektronics DM505 multimeter monitoring the source, with a Tektronics DM501A multimeter and DC504 dual channel scope monitoring the output. A pair of 10 ohm 1% 10 watt precision noninductive power resistors in parallel provide a 5 ohm load. As shown in Fig. 1, the Tek kit is mounted in a TM504 4-slot rack in the office, an a TM506 6-slot rack in the car.

I made the tacit assumption that the existing 15 watt per channel amplifier in the non-hi-fi sound system is a bridge topology, as that's the only way to get 15 watts in an automotive environment without an internal switching power supply driving the amplifier final stages with something higher than 12 to 14 volts. In a bridge amplifier, essentially two amplifiers are fed 180 degrees out of phase, each driving one leg of the speaker, allowing the speaker to swing rail-to-rail (or close to it) rather than rail/2 in a non-bridge (single amplifier) design. The implication for making measurements is that both leads going to the speaker/output are floating. That's easy with the DM501A multimeter, and the reason for using the dual-channel SC504 scope in Chan 1 - Chan 2 display mode.

Note that this is lab-quality test equipment, and while it’s 20 years old, it still provides lab-quality performance. This equipment is professionally maintained and calibrated.

HP 3314A Function Generator

The 3314A provides sine, square, triangle, and other waveforms from 0.001 Hz to 19.9 MHz and will drive up to 10 V into a 50 ohm load. Frequency and output amplitude are set digitally. The 3341A generates tones, pulses, sweeps, arbitrary waveforms, and will even play the Hallelujah Chorus if you hold down the right buttons on power-up. For this test, it’s used to provide sine waves at a constant amplitude. You set the amplitude, and that’s where it stays, across the frequency range. For these tests, the 3314A is driving a 600 ohm resistive load in parallel with whatever the amplifier provides.

DM505 Multimeter

Used to monitor the 3314A output, just to be sure. 3 1/2 digits, reading within 1% from 20 Hz to 20 KHz.

DM501A Multimeter

Used to monitor amp output. 4 1/2 digits, true RMS, reading within 1% from 20 Hz to 20 KHz, 0.6% 40 Hz to 10 KHz. Also displays in dB, which is quite useful for this application. When reading in dB, accuracy is 0.5 dB from 20 Hz to 20 KHz in the -15 to +20 dB range. Reference 0 dB is 1 mW into 600 ohms (0.746 V). We’re interested in relative dB here.

DC504 Scope

Used to monitor amp output. Dual channel, DC - 80 MHz. For this application, the “Chan 1 - Chan 2” display mode is used, which is what you need for monitoring bridge topology amplifiers, as neither speaker line is at ground level. Triggering driven by the synch output of the 3314A.

Test Methodology

The picture shows the test setup measuring one channel of a 15 watt per channel stereo amplifier (bridge topology) driving a 10 ohm resistive load. Set the 3314A for 2 KHz, and adjust output amplitude for a reasonable undistorted output on the scope and multimeter, in this case +17.0 dB. Change frequency to 20 Hz and measure. Repeat measurements at 50, 100, 200, 500, 1000 Hz. Recheck measurement at 2 KHz, it should be the same reference level. Complete measurements at 5, 10, 12, 15, 16 KHz, then recheck 2 KHz one more time. My test amplifier is essentially flat, down 1 dB at 30 Hz, down 2 dB at 18 Hz where the DM501A starts rolling off. On the high end, it’s down 0.6 dB at 20 KHz, and down 1 dB at 27 KHz, visible both on the scope and the DM501A.



Pre-test car checks

Set input to AUX, iPod and other devices disconnected. Set fader, treble, bass, balance all to zero and verify. Connect 3314A to aux input. As noted, a level is set at 2KHz, and then measurements made at 20, 50, 100, 200, 500, 1000Hz. Recheck the level at 2KHz, then make measurements at 5, 10, 12, 15, 16KHZ, and one last stop at 2KHz to make sure things haven't moved. Here's the test setup in the car.



Test Front 1

Measured response across left front (door) speakers, response measured at 20, 50, 100, 200, 500, 1000, 2000, 5000, 10000, 12000, 15000, 16000 Hz.

Test Front 2

Replace 6.5 inch speaker (Polk db651s) with 5 ohm (20 watt) resistor. Measure response at same points.

Test Front 3

Remove resistor, measure response with just the 4 inch (OEM) speaker as load. Note that the graphs for Front 2 and Front 3 are essentially the same. The differences between the Front 1 graph and the Front 2 and Front 3 graphs show that the complex impedance presented by the Polk db651s and its (cheezy) crossover have some effect on the amplifier.

Test Rear 4

Measure response across rear 6x9 speaker (Polk db691). I used a slightly higher reference level (+17dB rather than +15dB) to make sure the resulting graphs from front and rear channels would not be right on top of each other.

Test rear 5

Replace rear 6x9 speaker with 5 ohm resistor and measure response.

Next Steps

The Excel spreadsheet with the raw measurements is available on request.

I'll be putting together a report in PDF form in the next day or so with pictures, text, and the spreadsheet with the data.

As stated previously, this sucks! It's also quite surprising (and disgusting). Someone had to work hard to get a 30dB cut in the low end going to the rear speakers! And why? Why would you cut the low end signal going to your largest speakers, the ones best able to reproduce low frequencies? Maybe this is a British car after all -- Lucas electrics?

Installing Subwoofers

It's obvious that any subwoofer subsystem must be driven off the front speakers.

Or... What would it sound like if I swapped the wiring, swapping front and rear channels? That would give full fidelity to the rear (6x9) speakers, and cut the fronts at 100 Hz. Might be something to play with.

Rewiring

If I could get my hands on a victim head unit I'd be tempted to take it apart to see if I could find the active filters in the rear channels and disable them -- or just to bring out line level outputs.

That's not likely. My current thoughts are to add a 5 channel amp (such as an Alpine PDX-5) driving the 4 current channels, and a subwoofer. Since the PDX-5 doesn't have high level inputs, I need a level converter. The JL Audio Clean Sweep CL441dsp comes to mind, with both front and rear channels being driven by the front speakers. The Clean Sweep would flatten out the response in the midrange. That means I lose the fader control, but it also means I get full fidelity to all the speakers.

 

Nice writeup. I'm glad to see you used a purely-resistive load for some of the tests - that shows that the poor low-frequency response isn't from the impedance of the stock rear speakers shooting up at lower frequencies. (this is important because it proves that simply replacing the rear speakers won't cure the problem - the *huge* low-frequency attenuation is coming from the head unit itself.)

 

Yup, that's it -- a purely resistive load. I didn't do the pure resistive load on the fronts because the 4 inch OEM speakers are a PITA to pull, and the test results didn't indicate I had those problems.

I was also careful to document the test setup with respect to running bridge amps.

If there's a problem with my test methodology, please let me know!

I don't have the equipment at home to do automated and continuous sweep testing (yet-but I'm watching eBay), which would be interesting. I also didn't track phase over frequency, but did notice that it went all over the place on the rear channels as I went below 100Hz -- there's an active filter in there cutting out anything useful below 100Hz.

Bleah!

 

"Bleah" is right! I don't know if it's an active or passive filter on the rear-channel outputs, but you're right - for some reason, MINI purposely crippled the rear speakers with a 12 dB/octave high-pass filter at a 100 Hz cutoff frequency. (100 Hz is only a little more than an octave below "Middle C" on a piano keyboard, so we're talking about losing some important frequencies here).

 

Thanks very much for post this info.

The rear speakers are behind a large plastic panel. My suspicion is that the engineers had problems with that panel buzzing from low frequencies. The easy way out is to cut the low freqs.

 

WOOWWW! So I'm using a Hi-Low output converter off the rear speakers for my single 10 inch sub. This whole time I thought the box was just not the right cubic feet. No freaking wonder I cant hear anything below 50hz. Now if we could just find and eliminate the 12db/octave passive crossover...

 

Unclemeat -- I'm surprised your sub is putting out much of anything! But that's the first thing I was going to try with mine, picking the signal off the 6x9 speakers. Not gonna do that!

Scott -- 12dB/octave would take up a lot of real estate on a PCB if done in passives; much easier with a pair of op amps, yes? And that would help explain the bump and slight dip in response from 200Hz through 2KHz.

The more I think about it, the easiest "fix" to try is to swap the fronts and rears at the back of the radio -- or some place equally central on the wiring harness.

Any takers? Any way to get to that set of connectors that doesn't involve completely dismantling the dash and center stack?

Bob K6RTM

 

Unfortunately, that's what it will take for the base stereo (HiFi option has a separate amp where access should be easier). However some people have done the dismantling and it seems to be straightforward, although a lot of steps and knowing which screws and fasteners to deal with. For example, see this thread. Also I have the BMW service manual DVD and I can look up and put together a PDF of the steps.

Your results match well with raz636's measurements in this thread. It is truly a bizarre way to design a stereo!

 

Gee, if I'd seen those plots, it would have saved me some work! And I thought I spent a lot of time trawling the site while waiting for my MCS to arrive!

I'm more an RF kind of guy -- 1GHz is the low end for most of what I work with anymore. But I'd bet on a single-stage sallen-key or vcvs in the rear amp signal path. But if I got to digging that deep into the guts, I'd be soldering in four wires to add my own connector for line level outputs! My preference at this stage is to consider the rear channels "broken" and recover signals off the fronts.

Yes, given the transfer function for the filters on the rear, we can compute an inverse. But even using a DSP, I suspect the real-world results will fall under the category of "not pretty."

I suspect Robin's hypothesis that someone "cured" bass-induced rattle in the rears by eliminating all the bass might be correct. Sort of like high amputation for a hangnail, though, not quite addressing the underlying problem.

Does the (so-called) HiFi system have the same degree of suck in the rear channels? I'm curious.

Bleah. And I still have to put the rear seat back in!

 

Wow brilliant bit of work. Do you think the cleansweep can clean up the signal from the fronts a bit? If I end up putting in a 4-ch amp I'll probably try reversing the front and rears as you suggest. I have a CS in my Mazda3 but the signals there are actually pretty linear so I could move it over to the MINI.

I wonder if these response curves change at various volume settings, many cars are known for that.

 

Ya know, 100Hz is a good cutoff when you have a true subwoofer. Maybe they meant to put it in, but forgot (it's attached to the spare. )

 

Nice testing info! Thanks. Do you think the rear channel cutoff is related to "just" the Aux input or do you think the CD and radio are also rolled off?

It sure looks like a subwoofer was planned by looking at the rear rolloff. Is there a sub output coming out of the headunit we don't know about? Does someone have a link to the electrical schematics?

 

Just for clarification, you are suggesting to tap into the 6.5" speaker wires, right? And not the 4" midranges....

 

I have not seen any sub outputs on the wiring diagrams but that does not mean that they are not there and just not used. Many of these type of items are designed by an OEM shop (HK, alpine etc.) and just packaged for Mini, BMW, Ford etc. With out having the actual radio schematic there is no real way to tell of there is an extra set of outputs that just are not used by Mini.

Maybe the next person to take one out can check to see if there are some unused connectors on the back of the radio.

 

per the wiring diagrams the 6.5" door and 4" door speakers are driven byt he same wiring. It is the 6x9 in the back that are rolled off.

So tapping into the sires for the 6.5" and 4" would give you the flattest frequency response.

 

I suspect you're probably correct. For a 12 dB/octave passive filter with a 100 Hz cutoff frequency into a 4 Ω speaker, you'd need one ~281 µF capacitor and one ~9 mH inductor per channel, which would take up quite a bit of room inside the head unit if it wasn't originally designed to hold them. I'd love to be able to trace the wiring from the head unit to the rear speakers though - just to make sure the inductor/capacitor weren't integrated into the wiring harness itself.

 

It looks like both receive the same signal. (test cases 2 & 3 in the graph)

 

I'd love to see the same curves for the HiFi system to get a feel for the crossover separating the fronts. It'd help with component speaker selection.

 

raz636 did measurements from a CD with similar results: https://www.northamericanmotoring.co...ad.php?t=12078

That's a very interesting idea. There are wiring diagrams here (use IE, doesn't work for me on Firefox) but they only show the wiring harnesses on the car, not component schematics. That's the level that service people work at, and probably tough to get more detailed technical info.

 

The wiring diagrams are posted -- look for R56RadioPower.png and R56Radio.png (and remember the colour codes are in German, so SW = black).

The 6.5 and 4 inch door speakers have separate lines which are connected in parallel somewhere in the middle morass; I'd love to know where connectors X256 and X9420/1, and X257 and X9242/3 are located, as that's where the parallel connection takes place. Anyone have access to one of the services with that level of detail?

Thanks to rkw for the link - http://www.mcaw.info/wds/mini/us/r56/index.htm - a website which lets you see where these things are. Unfortunately, it looks like it requires IE on a Windows box, but I can do that when I need to. The bad news is that X256 and X257 are the car door connectors, and the speakers are paralleled at the doors.

But the GREAT news is that connector X9331, which has signals for all 4 speakers, seems to be located between the driver's door and the firewall! This looks like fun! Can anyone dig in there before this weekend?

About the midrange humps... Those could be artifacts of inartfully designed tone controls. Or, they could be compensating for cabin resonance -- yes, a clean sweep should flatten out that part of the front channel response, but what you really want is to normalize for complete system response at the listener's ear, right?

Thanks again, rkw, for the pointer to the mcaw site!

 

Wow, I haven't drilled down that site in detail, and there is a lot of information!

Here are the locations of some of those connectors -- look under Component and signal information ==> Plug-in, comb-type and solder connectors ==> (number of the connector). Unfortunately some of those connectors are door connectors or on the door itself.

 

rkw--

yup, those are the ones, particularly X9331, which hopefully will be our friend. I stayed up too late last night drilling down into things (and not figuring out why I can't see the wiring diagrams, getting "Error on page." in the lower left of the IE window).

Looks like a "simple" matter of ripping out the interior to get to that connector. I'll have a whack at it as soon as I can.

Anyone know if the connector parts are available through BMW? Is a special pin removal tool required? I looked at the PDF I have for installing the after-market Nuvi nav, and it doesn't mention any special tools.

 

You might need to load the SVG plugin from adobe on the computer you are using today to see the diagrams.

I looked around hoping to find something like an option or unused sub output but no joy.

One last hope, when you looked at the rear speakers response did you look at the rear woofer connection or the rear tweeter connection? There are 2 connections to the rear speaker (if I read the non-HIFI diagrams correctly). One for the woofer, one for the tweeter...\\--edit...nevermind, I was looking that the HIFI connetions

 

Dan--

Thanks for the hint on the SVG plugin -- loaded that, and I can see the wiring diagrams!

For the rear tests, I clipped directly across the rear speaker, and for the last test, pulled the connector from the speaker and jammed my exquisite test adapter into the socket. Wish there were a pair of connectors back there!

The X9331 connector is looking better all the time...

 

Or they could be compensating for peakiness in the driver response, but, ahhhh, given the rest of the quality of this system, I'm guessing it wasn't that purposeful.