Contents:

A2DP Bluetooth audio receiver Enjoy BT-AU01

Review

Device teardown

Opening the device

Chips

Teardown photos

Inter-connector description

Research

Defects in the design

Short circuit on the connector board

Antenna design

Stability problems

Fixing defects

Fix ground short

Important note: limitation for power supply

Fix power supply stability

Fix HF crosstalk

Get rid parasite antenna on LED

Get rid parasite connector board antenna

Idle noise

Hack to silence idle noise

What was done

Summary

Related: A2DP Bluetooth dongle audio receiver (SK-BTI-005)

A2DP Bluetooth audio receiver Enjoy BT-AU01

Very small Bluetooth audio receiver is available from several dealers on Ebay (about $27 with power supply). I got my one from Forkey as Stereo Bluetooth Adapter Audio Receiver For Home Stereo Speaker in a pack with US wall charger. It is often referred as Enjoy BT-AU01 CCAB11LP4300T4. Its manufacturer is probably Enjoy Computer technology Corp. and it is referred there as Bluetooth Audio Receiver BT-AU01.

Review

The device performs worst of all Bluetooth audio devices I even seen.

Listen to the car navigation and music to hear problems.

Problems I have seen:

I suspected that my item is defective. But watching the video referred by the manufacturer and their Note 2, I believe, that the device is broken by design.

Summary: I do not recommend to buy this device. There are better and cheaper ones.

Device teardown

I decided to keep the device and look inside.

Opening the device

Opening the case was very complicated. It is fastened by a glue on the whole seal. I needed a sharp blade, lot of time and a bit of violence.

The connector board was fastened by a piece of double side adhesive foam.

teardown image

Chips

OVC3860There are two boards inside – the main two layer board with the OmniVision Bluetooth chip OVC3860 (datasheet), 24C08C 1kB serial EEPROM and one SOT-23 3.3V voltage regulator, and the single layer connector board with connectors.

Teardown photos

top view of Bluetooth board bottom view of Bluetooth board

top view of connector board bottom view of connector board

Adjusted overlay image of the Bluetooth board for The GIMP:

overlay image

Inter-connector description

Tracing down pins and testing function, I believe that pins are connected in following order:

Pin (from up)Function
1Left (?) speaker
2Right (?) speaker
3Power ground
4Audio ground
5Battery (?)
65V power

Research

The data sheet of OVC3860 does not contain reference design, but I found another product using it – module BLK-MD-SPK-A made by Shenzen Bolutek technology Co., Ltd. (a photo is available here, here, here, use case here, here, here, datasheet).

Defects in the design

Short circuit on the connector board

ground shortDesigner apparently used a capacitor-less design with a separate audio ground. Such design makes impossible to connect power ground and audio ground together. They either had to use separation capacitors or they have to make both grounds disconnected. They did nothing of it.

If you look at the detail of the connector board, you see a short connection between both grounds. It causes DC bias on output and overheating of the device. I even believe, that it can permanently destroy the module.

The fix is simple: just breaking the short.

Antenna design

Standard antenna design rules prohibit to place any parts near the antenna. This board violates it in several ways. It makes Bluetooth connection much worse.

parasite antennas on the boardcorrect board designWhat is even worse? All these parts and PCB traces work as a secondary antenna and caused massive HF crosstalks into other parts of circuit. The LED dimly shines even after breaking the wire. It means, that induces more than 2V of RF signal there. See the image on left, which shows parasite antennas: both LED lines and ground plate around the oscillator. See the image on right for the correct solution on BLK-MD-SPK-A. The antenna is completely separated from the rest of the board by the ground plate with sufficient number of through holes.

Also the whole connector board acts like a parasite antenna.

Stability problems

The module is unstable and makes ugly sounds during playing, idle, connecting and disconnecting.

Comparing both boards, it seems that developers of our board saved too much parts. There are unpopulated places for these parts on the board: Circles mean test points, rectangles mean resistors, ovals mean capacitors. It needs further research, which parts needs to be added to make the board stable.

Measuring the missing parts, it seems that most of them are filters. Without these filters, the sound part of the chip receives signal from antenna, and it modulates the audio signal. It causes the noise and pops.

Fixing defects

Fix ground short

Just by breaking the ground short on the connector board, device stops heating, the audio DC offset voltage disappears, but the noise and pops still make the device unusable.

fixed ground short

Used a sharp utility knife, I divided the ground copper to two parts: audio ground and power ground.

Had board in the original state, did the cut line, got fixed ground short.

After this fix, the power consumption decreases to 40mA when idle and 70mA and more when active. These values are much lower, but still much more than 26mA declared by the datasheet when active. The device does not implement low power stand-by mode.

Important note: limitation for power supply

After the fix, the device will work well with a dedicated power adapter. But even after the fix, it will be impossible to power it from the same device as the audio is connected to (e. g. the car adapter shown on the photo).

In case of galvanic connection between power and audio, you have to do one of following:

Fix power supply stability

The input linear 3.3V regulator has stability problems. It causes parasite oscillations. Depending on power supply and power cable, my one oscillated between 3.17V a 3.3V. Adding a 1µF capacitor to the input improves situation a lot.

Fix HF crosstalk

Get rid parasite antenna on LED

Fixing HF crosstalk is much harder job. I tried to attenuate unwanted antenna circuit with LED adding a 200pF blocking capacitor to ground on one side and 3.3nF on other side, then 110Ω resistor in series and again 180pF blocking capacitor to ground. It helped a bit, but not completely. The LED lines broke the integrity of ground layer, and some parts of the ground layer itself act as parasite antenna. Also some traces act as a capacitor together with another trace on the reverse side of the board.

top view of Bluetooth board fix bottom view of Bluetooth board fix

This fix improved the situation, but it is not fixed completely.

Get rid parasite connector board antenna

After getting rid the complete connection board, the device starts to play cleanly. No more pops, noises etc. It indicates that the connector board needs a complete redesign (not making U antennas from power and audio lines), or at least shield and/or CLC attenuation circuit on each pin.

device without connector board

Idle noise

But noise remains when it is idle or when it waits for stream.

Watching the power supply with the oscilloscope, there are significant short drops of the power voltage (about 0.05V). Also idle power consumption is much higher that the data sheet declares. It indicates possible short circuit somewhere in the design (GPIO output or something else). It seems, that this problem is probably not related to the HF transfers.

Interesting observations:

It is also possible, that the audio should be muted somehow when idle, but the muting does nit happen.

Hack to silence idle noise

I was just been informed about stability problems of the audio in the older release of OVC3860 used in BT-AU01. So I started to search for hack instead of design mistake. Adding a parallel resistor 470kΩ between audio reference input (pin 20 of OVC3860) and ground improves idle noise a lot.

What was done

fixes on the top side fixes on the bottom side

Summary

Here are recommendations for the manufacturer in the next version: