Clarion and McIntosh parts

I recently had somebody refer me to PacParts who turn out to be an excellent source of spares for Clarion and McIntosh car audio. Parts include laser pickups, spindle motors, tuner modules, LCD/VFD displays, knobs and more.

Unfortunately they only ship to the US, however there are many companies around who offer international parcel forwarding such as Shipito.

Clarion CeNET – a first glance

I’m keen on reverse engineering at least some aspects of the Clarion CeNET bus in order to hook other gadgets to my car stereo. CeNET is a proprietary interconnect used by Clarion on many of their car audio products over the last ten years.

It uses a proprietary 13-pin square DIN style connector that is not available off the shelf. The pinout is shown below (this was taken from a Clarion service manual).

Electrically, CeNET appears to be a form of asymmetric serial bus. There is some discussion online that suggests a 38400 baud rate. The same person also says that it appears to use some form of encryption which may make it difficult to hack.

My experience with CeNET stereos is that the display will flash SYS or SYSTEM if the battery is disconnected or a CeNET device is inserted or removed. They do this until the stereo is powered on. Perhaps an encryption key is negotiated during that period?

On the hardware side, closer inspection reveals the bus consists of a transceiver IC (CA0008AM) connected to one of UARTs in the host devices microprocessor as shown in the simplified diagram below. Note that there are also some pull-down resistors on the Tx and Rx pins, as well as a 68 Ohm termination resistor across the Bus+ and Bus- pins. I’m unable to find data on the bus transceiver but I guess it’s possible that other generic ICs could be used here. The bus itself floats at 2.5v and swings +/- 200mV during signalling.

I have capture some data samples from a stereo connected to a EA-1251B iPod adapter that emulates a 6 disc changer. I haven’t yet studied them in detail but you can download the capture files below.

You will need to download Saleae Logic to view the capture files (it’s free). The cables were connected according to the colours in the diagram below:

In all captures I powered on the stereo from cold (no battery) immediately, then switched the unit on after about 3 flashes on the display (4-5 seconds). With the EA-1251 samples I switched from radio to the CD changer at about 7-9 seconds. From here the stereo waits for the iPod to wake up (it thinks the CD changer is loading) before it starts playing the first track of the first ‘disc’ (playlist) at about 14-15 seconds. The song then continues to play until the end of the capture.

Update: I’ve also grabbed a couple of samples of a TV tuner and display.

When the TV tuner is selected via the TV display it places the head unit into aux mode (it displays Aux on the screen). This could be a very useful hack; to make a ‘proper’ CeNet aux input. In the samples above, all units were powered up from cold (no battery) immediately after beginning the capture. The head unit is switched on at about 4 seconds, then there is a bit of wait for the display unit to boot up. Once this is done I’ve switched to the TV tuner at about 27 seconds. Shortly after (about 35 seconds) I switch back to the head unit’s internal FM radio.

It turns out trying to capture the bus pins was a waste of time as they normally float at 2.5V. I haven’t been able to look into that side of things any further as my only oscilliscope died before I could check it.

Hopefully somebody else finds this useful. I’ll keep working on it as I get time, and post any updates to the site if i come up with anything.

Vertical collapse on Storage System portable PC

I recently stumbled across a junked ‘Picollo’ portable PC made by Storage System Inc. This unit had a 66MHz 486 Overdrive processor and 10″ Sony Trinitron SVGA display and no doubt cost a fortune in its time.

Unfortunately the chassis had been partly gutted, missing the outer case and a variety of parts. The 10″ monitor still appeared to be salvageable however so I gave it a try.

After connecting everything as it appeared to go I connected a PC and powered on the display. Sadly the it was completely collapsed vertically, although I could make out enough detail in the thin line that appeared to be sure that it was otherwise functional.

The deflection is handled by the board mounted on top of the CRT. After fixing loads of dry joints around the horizontal output stage I decided to focus efforts around the vertical deflection stage.

This is based around an ST TDA1675A IC. I grabbed the datasheet for the IC and discovered that the circuit in the display was quite similar to the test circuit shown in the data sheet. The scope confirmed that the sync input was working, as well as the ramp generator. Unfortunately there was no output.

Google searching for the IC revealed many forum posts identifying this as a common cause of failure, and that it was also sensitive to failed components elsewhere in the vertical deflection circuit. Many people suggested replacing all of the surrounding capacitors and diodes as well as the IC just as a precaution.

TradeTech listed the IC as being on back order which was not ideal. In the mean time I though it would be worth attacking the other likely culprits in case the IC was still OK.

I replaced all of the electrolytics in the vertical stage as a matter of course but it didn’t fix the problem. I also identified a Zener diode (ZD201) with the marking ‘4A3’. After some research I found that this is in fact an HZ4A3, a 3.7V Zener. The closest I could get was a 3.6V 1N4729 so fitted one just in case. No difference.

There was also a standard 1N4004 rectifier diode (D207) mounted under the heatsink. I replaced this and it immediately fixed the problem! The weird part is that the diode appeared to check OK out of the circuit.

The horizontal yoke connector also had some discolouration on the plug due to heating because of a bad connection. I cleaned up the pins in the connector the the PCB header before reassembling it.

So now I have a dinky vintage 10″ Sony Trinitron SVGA display and just need to find a use for it! 🙂

HTC Desire and Belkin car charger

I noticed that my HTC Desire does not charge at full rate off my Belkin Micro USB car charger, despite it being the same rating as the original HTC wall charger (5v, 1A). It appears that the Desire obeys the somewhat recently introduced USB battery charging specification that states devices may draw up to 1.8A if the D+ and D- pins are connected by a resistance of no more than 200 Ohms.

This post shows how I modded my Belkin Micro USB car charger to charge my Desire at the full rate of almost 1A.

Getting the charger open is relatively easy; use a small flat screwdriver to pry the two halves apart. Note that the faceplate is attached to the side closest to the power LED. If the screwdriver takes too much effort you can use a vise to gently squeeze the sides along the seam to help break the seal.

Once the seal is broken the metal pin on the front will probably ping out and land somewhere awkward like under your workbench, so be prepared!

In inside looks quite tidy and simple:

Now solder a 180 Ohm resistor across pins 2 and 3 of the USB port.

To get it back together simply put a few drops of super glue along the seam. I used Selley’s Plastic Glue which is exceptionally good.

After this the battery status (Settings, About phone, Battery) shows as Charging (AC) instead of Charging (USB), indicating that the battery is now charging at full rate.

Hauppauge NOVA-S-Plus in Lucid

I’ve been using a Hauppauge NOVA-S-Plus DVB-S card for a while now with MythTV running on Ubuntu Hardy.

I recently upgraded to Lucid and had lots of difficulty getting the card to work with MythTV. All of the required kernel modules were loaded as expected (cx88xx, etc) and the device tree showed up as expected:

ls -l /dev/dvb/adapter0
total 0
crw-rw---- 1 root video 212, 1 2010-10-12 23:07 demux0
crw-rw---- 1 root video 212, 2 2010-10-12 23:07 dvr0
crw-rw---- 1 root video 212, 0 2010-10-12 23:07 frontend0
crw-rw---- 1 root video 212, 3 2010-10-12 23:07 net0

Testing dvbtune gave the following message:

dvbtune -f 1159000 -p H -s 22500000
FD 7: fd_dvr DEMUX DEVICE: : Device or resource busy

It turns out the fix is actually very simple, it’s just not very obvious or well advertised:

sudo apt-get install linux-firmware-nonfree

Apparently the firmware blobs have some legal restrictions regarding their distribution so they have been placed into a separate package as of Karmic.

Make sure you reboot after installing the firmware package.

Lucid gconf fail

This evening I encountered a weird problem when booting up Ubuntu 10.04. I simply had a black screen with a small error window saying:

There is a problem with the configuration server.
(/usr/lib/gconf2-4/gconf-sanity-check-2 exited with status 256)

Considering that I hadn’t knowingly changed anything last time I used the computer this seemed a little odd.

The following also appeared in /var/log/syslog:

Oct 13 22:05:34 hostname gnome-session[450]: WARNING: Error retrieving configuration key ‘/apps/gnome-session/options/auto_save_session’: Failed to contact configuration server; some possible causes are that you need to enable TCP/IP networking for ORBit, or you have stale NFS locks due to a system crash. See for information. (Details –  1: Could not send message to GConf daemon: Process /usr/lib/libgconf2-4/gconfd-2 received signal 6)

Fortunately the fix turned out to be easy; it’s simply a permission problem with /tmp. To fix it simply run:

sudo chmod 1777 /tmp

Eee 900 Bluetooth mod

I finally decided it was time to kit out my ancient but trusty Asus Eee 900 with internal Bluetooth. I had an old micro Bluetooth adapter lying about which came apart very easily to reveal an exceptionally small PCB.

I stuck this to the motherboard of the Eee using a small piece of double sided foam tape, just behind the external VGA connector, and then wired it to the unused USB lines connected to the WiFi adapter PCIe slot.

Power was derived from the same 5v rail that supplies the external USB ports. I simply scraped some solder mask off the tracks and soldered the wires directly.

Fortunately once I put it all together again it worked!

Another reason I try to avoid Windows

I decided to replace the hard disk in my Dell Optiplex 755 with a larger model (500GB). Migrating Ubuntu was trivial; Windows Vista was not. Clonezilla did its best but I ended up running into all sorts of Windows BCD corruption issues. Most of the various repair tools on the Windows DVD failed with incredibly non-descriptive error messages.

After giving up I quickly discovered that Windows would refuse to even install, citing Windows is unable to find a system volume that meets its criteria for installing. This was even after zeroising the first few GB of the disk with dd. WTF Microsoft?

Even a BIOS update didn’t do the trick. (hint to Dell, et al.: DOS is dead. Please give us a modern, OS independent BIOS update mechanism)

It turns out that two things seemed to have an affect on this: my HP Photosmart printer with its built-in card reader and floppy drive setting in the BIOS. Disconnecting the printer and setting the floppy drive mode from USB (default if there is no internal floppy drive) to None cured the problem.

By this stage I had wasted enough hours so didn’t even bother trying to see if the clone would work this time around.

Of course all along Ubuntu behaves just perfectly.

A guide to hacking Japanese car radios

Do you have a Japanese import car that won’t tune to local FM radio stations? It’s a well known fact that Japan uses a different public FM broadcast band to the rest of the world; 76 to 90 MHz instead of the usual 87.5 to 108 MHz CCIR band. Sure you can use a band expander but these suffer from a number of drawbacks, not to mention that the radio does not display the correct frequency of the station you are tuned to.

Here’s a guide on hacking your radio to convert it to natively tune across the CCIR band. This post is still a work in progress so I may update it from time to time and add more photos as I get the chance. Read more of this post

McIntosh EF-1080I Information

I’ve had quite a few requests for more information about the McIntosh EF-1080I amplifier (often but incorrectly referred to as the EF-10801). This unit is actually made by Clarion and is supplied as stock equipment with many Subaru Legacy and Outback cars manufactured between 1998-2003. Read more of this post