BeoMotion wrote:

But this is not the main focus at the moment and you should always keep in mind that you need to have fun at tinkering and some basic linux and programming knowledge.

The only exception to this that I'd mention, is that I'm in need of as much data as possible to test my code and further my understanding of the masterlink protocol.

So if there are any members who don't have the required tinkering/linux/programming knowledge, but who would still like to contribute in some way to the development process, they might consider investing in one of the interfaces (and a raspberry pi) and simply leave it connected (to masterlink, and to their network / internet connection) and continue to use their B&O system as they normally would - I could then provide software which would log the masterlink data and upload it to a central location (or send it by email) overnight for use in testing code and analysis of the masterlink protocol.

There may be a small privacy concern for users with a BM5/BS5 in doing this, in as much as I would be able to see the names of the tracks you had been listening to in the logged data, but I promise not to tell anyone else or to criticise your taste in music 😕

Obviously at the point where we (and others) start to produce working applications for the device, you could simply redeploy the device for a more interesting use by simply installing a new image (power down the RPi, eject the memory card, copy a new image to it using a PC or mac, reinsert the card into the RPi and boot up again, voilà)

Martin.

What would the total cost be for buying the testing equipment be?

leosgonewild wrote:

What would the total cost be for buying the testing equipment be?

You'd need the following:-

• A Raspberry Pi 2 Model B - approx 30 euros
• Power supply for Raspberry Pi - between 4 and 8 euros (depending on model chosen)
• An 8GB Micro SD Card - less than 5 euros (but I'd happily post one preinstalled and configured to anyone who wants to use the data logging setup for a while, you can then copy a new image to it later when you want to actually do something useful with the box)
• A cat5e ethernet cable to connect you RPi to your network (or a USB WiFi adapter)
• half a cat7 ethernet cable to connect the 'BeoPi' interface to your nearest masterlink junction box
• The 'BeoPi HAT' interface card (or whatever it ends up being named) - see below.

The first three or four items can often be bought as a kit, but can sometimes work out cheaper to buy them separately.

An idea of cost for the Masterlink interface (BeoPi HAT) is the bit we are trying to determine right now.

Potentially it will be available either with or without a DAC, (without being cheaper) and either would suffice for the data logging but unless you intend subsequently to use it purely for MLGW type roles, rather than audio, you are likely to want the 'with DAC' version.

We should have at least a ballpark indication of cost in a few days once we know what quantity we are likely to want to produce. Broadly speaking the cost per unit decreases as the quantities increase, but it is not quite this straightforward as there are a number of assembly options:

• for 1 or 2 prototype units, it is possible to solder by hand, but this can be tricky with very small surface mount components and is very time consuming.
• for a small quantity of units, a stencil can be used to apply solder paste to the circuit board, components are then placed in position by hand and the entire unit is baked in an oven to melt the solder, but since the process is still carried out by hand it remains time consuming.
• alternatively, components can be purchased on a tape, and a pick and place machine does the assembly work, this tends to yield the best results but components then need to be bought on a tape and this can mean buying many more components than are actually required or requires the fabrication house having to add leader tape to each strip of components so that it can be handled by the machine (thereby increasing tooling costs) - that said, this tends to be the best option as soon as quantities are sufficient to justify the initial setup / tooling costs.
• each assembly method also has an impact on the design of the circuit board, when hand soldering for example, you would generally want to increase the size of the pads under each component but this would certainly not be desirable for pick and place assembly.

Kind Regards,

Martin.

Thanks Martin for your detailed description.

Regarding the price and quantity of the BeoPi HAT:
I wouldn't do more than 10 pieces manual. Even if you have stencils and an oven this is really time consuming and a hard work in general. We are talking of about maybe 30 very small components that each has to be placed accurately on each PCB.

The PCB manufacture I normally order my PCBs can do smaller production runs with their pick and place machines. 
For populating 20 boards we are talking of about 40 euro per PCB (manufacturing and populating) without the component cost.
When ordering 50 this will drop down to about the half. The PCB company would also take over the ordering of the components in both cases. 

For the components you have to add maybe another 20...25 euro (DAC version).

This are only very rough estimations. We will have a more precise calculation when the first prototype series is finished and working. 

BR,
BeoMotion. 

BeoMotion wrote:

Thanks Martin for your detailed description.

No worries, just wanted to make it clear that we weren't just evading the question and explain why it is so difficult to calculate precisely at this stage.

BeoMotion wrote:

For populating 20 boards we are talking of about 40 euro per PCB (manufacturing and populating) without the component cost.
When ordering 50 this will drop down to about the half. The PCB company would also take over the ordering of the components in both cases. 

For the components you have to add maybe another 20...25 euro (DAC version).

Great, thanks BeoMotion, so we're looking at around 65 euros for the interface with current levels of interest, perhaps dropping to sub 50 euros if demand increases.

Even adding the cost of a Raspberry Pi if you didn't already own one it's still in a price bracket comparable to say an airport express. I suspect (and totally understand) that at the present time this will appeal mainly to those who have already had some experience experimenting with the Raspberry Pi (the geeks amongst us!) but I can see this becoming more popular as applications are developed. For example, a 100 euro MLGW / BLGW alternative or a 100 euro Beomaster 5 alternative anyone?

BeoMotion wrote:

This are only very rough estimations. We will have a more precise calculation when the first prototype series is finished and working. 

Yes, and I would hope that given the complexities of calculating manufacturing cost that I've already outlined, nobody here will be too critical if the cost estimate changes a little, but I'm sure it helps to have a ballpark figure as otherwise people are unsure as to whether we are talking about a unit which costs 10 euros or 1000 euros and it is therefore difficult for them to determine whether they are interested or not 😕

Martin.

This would be groundbreaking. I would take 2!

I'm following the thread with the same interest. So, count me in for 2 PCB's.

Soldering/building will be no problem.

Another question. I have a first model RPI B version lying here. Is this one still useable? Or do I need to invest in the newer latest version.

Is the GPIO the same for both boards?

@steve77: Thanks for your interest! :-)

@beobuddy:
There are two versions of model B out there. The very fist ones doesn't have the P5 connector and therefor no access to I2S. If you have a RPi with the P5 connector you could simply desolder or cut off the analog TV connector to make the BeoPi HAT fit.
For using I2S digital audio you would have to run 3 wires from the P5 to the appropriate pins on the 40 pin connector of the BeoPi HAT. 
Remember that the screw holes in model B are different and the HAT would be quite loose.  Thanks for the interest!

Just want to add that i would take 3 boards. 🙂

Regards

Espen