Modifying OEM head unit for un-eq’d line out

[ean611]

funny you should mention Nak and McIntosh unclemat, I had a scare yesterday where my old Eclipse CD8053 looked dead, turned out to be a lose wire, and those were the only two radios (McIntosh Mx406, Nakamichi MB-X) that I could have used as a replacement.

 

There are good looking radios that don't look obnoxious, you just pay for it.

[utc_pyro]

Easy Soilder Points:

I fixed the line out points. The Panasonic keeps changing things between the years! Attached is the updated version. They did something a little different this time were the L and R channels are combined together, and then those are mixed again some were down the audio path to the sub-amp. This probably gives better stereo separation then the old design.

 

Again, sorry about the grainy cell phone macro pic quality. I'll get the E-PL1 in the spring for the full system install.

 

Pre-Amp:

Some thoughts on my pre-amp for this thing: I did my math wrong before, the max one can get (using a rail-to-rail op-amp) is 2.828Vrms using the stock 8V radio power supply. I'm not sure if that's going to be high enough, I may need to use a dc-dc converter to get a negative voltage rail. Also most of the "good sounding" opamps I've come across are NOT rail-to-rail, and use older bipoler technology, thus 2V out is optimistic.

 

This may not be all THAT bad though, seeing as a lot of aftermarket head units put out that much power, but it will be more susceptible to noise. What I do have at my disposal is a Yaohus DY 12D1212-2w dual output DC-DC converter (bought for a previous project). This can actually be used produce a true +/- 12V regulated output, giving the potential of an 8.4V audio output. This is inline with most aftermarket EQ's, crossovers, ect can provide.

 

Only thing that may be bad about this DC-DC converter is noise and ground loops with the onboard radio (uses 4v as a floating ground).

 

Also I found a good op-amp design refresher: focus.ti.com/lit/an/sboa092a/sboa092a.pdf

[ean611]

utc_pyro, 2V is more than fine for output if it's clean and crystal clear. I wouldn't suggest over designing to get higher output, I would mount your Rockford 3Sixty.2 near the radio to get it to 5V, or use a separate line driver to get it to 8V.

 

If you put the 3Sixty.2 near the radio, it'll do your line driving for you with much less work on your part, and no issues with DC-DC converter noise.

[utc_pyro]

Update, I got the circuit designed (I think). I decided to just patch into the power circuit on the radio it's self as it has the proper 8V+, V-Ref, and S.GND ready to go around the volume control chip. Though one could re-create these in the off-board circuit, it'd just add more complexity (and noise) then necessarily. V-Ref on the radio is created by a resistor network and IC372 (voltage follower op-amp), so it should provide a better reference ground. The easiest place to get this is Pin1 of the volume control IC. VCC is right next to it at Pin 36, and there are a lot of S.Gnd points on the bottom of the board to get.

 

The biggest thing that I'm unsure of is handling ground loops external to the circuit. I've added a second coupling capacitor for ground into the circuit. This should prevent ground loops from forming between the head unit and the processor, keeping noise out of the system.

 

In the simulation I used a 10K output impedance to match the input of an 3sixty.2, and it came out flat for the entire range I'm looking at. I'm not sure the impedance of the Audison bitsOne that I'm looking at now, but it should be close. It also gives a ~2vrms output before clipping, so that SHOULD be high enough to run any external processor. The OPA4228PA is able to hit within 1V of ether rail according to multisim, so even with rail-to-rail opamps it wont be much better.

 

Any feedback would be appreciated on this. I'm going to start looking into a place within the radio to stash the circuit board.

[ean611]

Update, I got the circuit designed (I think). I decided to just patch into the power circuit on the radio it's self as it has the proper 8V+, V-Ref, and S.GND ready to go around the volume control chip. Though one could re-create these in the off-board circuit, it'd just add more complexity (and noise) then necessarily. V-Ref on the radio is created by a resistor network and IC372 (voltage follower op-amp), so it should provide a better reference ground. The easiest place to get this is Pin1 of the volume control IC. VCC is right next to it at Pin 36, and there are a lot of S.Gnd points on the bottom of the board to get.

 

The biggest thing that I'm unsure of is handling ground loops external to the circuit. I've added a second coupling capacitor for ground into the circuit. This should prevent ground loops from forming between the head unit and the processor, keeping noise out of the system.

 

In the simulation I used a 10K output impedance to match the input of an 3sixty.2, and it came out flat for the entire range I'm looking at. I'm not sure the impedance of the Audison bitsOne that I'm looking at now, but it should be close. It also gives a ~2vrms output before clipping, so that SHOULD be high enough to run any external processor. The OPA4228PA is able to hit within 1V of ether rail according to multisim, so even with rail-to-rail opamps it wont be much better.

 

Any feedback would be appreciated on this. I'm going to start looking into a place within the radio to stash the circuit board.

 

 

Sounds good. Are you mounting other stuff on PCB or what? I'm very interested to know how this sounds when you're done.

[utc_pyro]

I'm just mounting the circuit as seen above (duplicated four times) minus the voltage divider. There'll be two sterio pots for F-R gain settings, and four RCA jacks as an output. Add three extra capacitors for noise supression on the power rails and I'm done.

 

I'm thinking about getting a PCB fabed to fit inside the case were the sub amp would be, but I'm not sure if it'll fit yet. If not, I'll go in a case strapped to the back of the radio.

[ean611]

I'm very interested in how sound quality comes out, though I doubt it will make me go back to OEM radio from my Eclipse CD8053.

[mikeislord]

This thread is a great read. Made me terribly miss EE stuff.

 

You're right about one thing utc_pyro, doing stuff like this will definitely keep your mind sharp. I'm only a few years out of college and I feel retarded from the daily grind. 3 1/2 years and I feel like I only use 25% of the brain power I used in college, which still wasn't most of it. /sigh

[dschultz]

If you can make it small enough, tack it to the side of the tuner's metal case possibly...

That's how the Leptronics Aux-in board is mounted.

[utc_pyro]

Eh, maybe it wont fit. Pulled the radio out and the area is smaller then I though it was. A board could be made to fit here with the physical connectors and all, but I'd need to use surface mount components and that's a PITA.

 

So it'll be a shielded box zip-tied to the back of the radio. Simple enough I suppose.

[bravonoj]

/subscribe - very cool, and good work!

[Guest]

Any updates?

[utc_pyro]

I'm just about done with the PCB design, I actually just got back on it this weekend. Hopefully I'll get it done tomorrow and post it up for review along with a bill of materials, but VSLI/Advanced Digital Design homework is calling.... Why the heck did I decide to go back to gradschool when I already have two jobs?!

 

The design has changed a little bit from the above circuit. I was playing with the idea of doing a proper low frequency roll-off with it at ~10hz, but it makes the circuit too complex. I have removed the ground coupling capacitor as the 10uf Polyester caps are WAY too big to put more then needed in. I'm just praying that if I ground this thing properly I wont get any noise. Also the opamp has been changed from a fully bipoler design (OPA228 series) to a JFET design (OPA134 series). This will give a higher max output voltage to lower the noise floor. Lastly, I'm changing "R6" to a 2.2K resistor, giving a max gain of 10x (20db). That should be plenty, but I have no clue what the stock output voltages are.

 

I just ran across people doing this to Mazda radios tonight.

 

http://www.msprotege.com/forum/showthread.php?t=123620193

http://www.msprotege.com/forum/showthread.php?t=123618904

http://www.msprotege.com/forum/showthread.php?p=2669100#post2669100

 

Might be useful info for people wanting to do this but don't want to design there own audiophile grade pre-amp.

 

Speaking of that, I'm going to have to order more then one PCB when I have it fabbed. If any one else wants to go in with me on them, hit me up. Prices get lower the more I order, same with the components.

 

More to come....

[utc_pyro]

Ok, got the PCB designed for the amplifier board. One could use a line drive for this and skip what I'm doing here, but this will give a cleaner signal then most off the shelf designs.

 

The PCB's are going to cost $22 each + shipping for an order of three, and go down with quantity. If any one wants one, let me know in the next two days.

 

2x BI Technologies 20K audio pots. P092N-QC15AR20K. $3.08ea

2x 8-pin dip sockets. 1825093. $.47

2x Kycon RCA jack panels. KLP42X-61-RW/RW. $2.11

4x WAMA 1uf Polyester Film capacitor (C4), input coupling. MKS2-1.0/63/5. $1.10

4x WAMA 3.3uf Polyester Film capacitor (C5), output coupling. MKS2-3.3/63/5. $3.68 each

1x Tyco Terminal Block (power). 284510-3. $2.00

1x Tyco Terminal Block Header. 284539-3. $0.78

2x Burr Brown/TI OPA2143 opamps. $5ea (ebay)

 

Still trying to pick the following, but it should be easy. They are all standard:

6x 100K metal film resistor, 1/8 watt

4x 2.2k metal film resistor, 1/8 watt

4x 100ohm metal film resistor, 1/8 watt

1x 5mm spacing input buffer capacitor (C1), 16V, low ESR Aluminum Organic Polymer

2x 3.5mm spacing "vref" buffer capacitor (C2), 10V+, low ESR Aluminum Organic Polymer

2x 3.5mm spacing, opamp buffer capacitor (C3), 16V, low ESR Aluminum Organic Polymer

 

Hum... Maybe an off the self line driver IS the answer. This is getting into the $80 each range. Though it would be easy to lower cost with cheaper components.

[Zimm]

How difficult would this be to install if we were to buy one off you? I can solder but I mostly only do wire splice type stuff, nothing on a PCB (yet?).

 

How much to you anticipate the total price to be? $80 seems reasonable but much more than that and it gets iffy.

 

I would be interested in this but need to know if I can feasibly install it and not burn down the head unit in the process.

 

Thanks.

[utc_pyro]

I'd sell the first batch of pcb's at cost and give the part number list for anyone to pick up at mouser. I'm looking at ways to make them cheaper though. This pre-amp would make most audophiles smile ear to ear, but I'm not sure if it's worth going THAT high quality for an in-car application. I need to dig back out my radio and see what sort of caps it uses.

 

Assembling a PCB is fairly easy, and there arent any particularly heat sensitive componets on this board. Also there are only 7 solider points needed on the bottom of the radio one needs to hook it to. One shold be able to do the install in two or three hours.

[ean611]

utc_pyro.

 

One issue, while your preamp may make an audiophile smile ear to ear, it doesn't add a high grade D/A

[utc_pyro]

I got a bunch of PM's about this, so I'll just post it here... I'm going to make a second version of the board that's smaller and simpler for people to install them selves. More importantally to most of you, it'll also be cheaper. $40 a kit sound better? Nicer electrolitic caps shoud be just fine as the rest in the signal parth are. What you'll be saving in THD+N by bapassing the amp stage and LOC will be a LOT greater then that two little caps add to it.

 

On the loudness disable, I think it's going to require a little more work then I thought. I'm going to have to sim it in P-Spice tonight, but I think the volume will be out of wack.

[Guest]

I am definitely interested. I'll get the pics of 09 HK headunit for you soon.

[Zimm]

Sounds good on all this. I think I would be able to do this assembly so long as I go slow.

 

As for the SRS WOW bypass, can't we just turn that off if we don't want it? I know I can in my '07 Spec.B. Honestly, I leave it on as it makes the radio sound quite a but better.

 

Basically what I am interested in is a setup that will give me clean line outputs that do not go through the amp eq system. Honestly I would probably only hook them up to my sub right now but may go to a full amp system later. I would just like to have that option and also get proper bass response for now.

 

Thanks for you efforts on this, much more than I would know how to do.

[utc_pyro]

The SRS wow (JRC 2195) is a separate chip that's also used as an input selector, before the fade/balance/volume control chip this work is focused on. I'm not going to mess with that, as It's the loudness compensation that one cannot otherwise turn off.

 

One can install the line-out amp without disabling loudness. It would be an optional second step.

[Zimm]

I see. I would be fine with disabling any non-defeatable loudness and would probably want that. I just figured since the SRS WOW can be turned off it wouldn't need to be bypassed.

 

Either way, it sounds like you know more than I on this stuff so I would defer to your judgement.

 

Keep at it and thanks for tackling this.

[Zimm]

So one other question. Would one need a RF 3sixty.2 for this setup or can you just run the outputs of this unit straight to the amps?

[ean611]

Zimm, the point of UTC_Pyro's work is that you would not need a 3Sixty.2 to go between the HU and the amp.

 

That being said, I would highly recommend one, as I have an aftermarket HU (Eclipse CD 8053) going into my 3Sixty.2, and the equalization is an incredible aid.

[utc_pyro]

As ean611 stated, you wont need one. The amp will step the signal up to the point that any amp can take it and go. I just bough an 3sixty.2 beacause I like the control over my audio... I have a 24-band DSP EQ just for my

 

I should have the final version of the V1 board posted when I get home from work (making up time). I needed to tweak a few things about the power capacitors and how the power-input is layed out. I'm going to order the first three boards tomorrow.

 

After the V1 boards are in and built, I'll design the V2 "for everyone" kit board. The V2 boards will be the cheap, simpler, build-install-and-go type board. Electrolitic instead of polyester capacitors, no expensive pots (fixed gain), cheaper opamp, no RCA inputs, ect. Perfict if you're just intalling a sub, or a cheaper system.