How long should a project take?

I have been a few months on a project. ( as well as a couple of other simpler projects for the same customer simultaneously )

It is multi faceted, in as much as I am overlaying some text and graphics onto a Hi Definition LCD display ( obviously it has to be synced to the incoming timebase, which was the first step )

This is partly micro based (Arduino software ), and partly 74HC chips to deal with the video speed .

I used one micro ( which I plug into the Arduino to program ) as the video controller , it deals with the vertical switching, the incoming wireless commands to say which number/graphic to display, and the enqueing of the commands in an interrupt routine , and the dequeing in the main loop.

A second micro handles an audio announcer, using the Wavehc library, to announce the same number displayed.

I have set it up so that when the announcer is busy, the incoming commands will be queued, and when its finished announcing, it will dequeue the next number , and display and announce each queued number in turn.

I made the mistake of telling the customer that I had the prototype working, and now they want four finished units by Wednesday !

Apart from trying to correct the mistakes on the rushed pcb, I now have problems ( I think ) with the software.

Several guys on this forum have been very helpful getting the software sorted out for me ( I only started Arduino late last year )

What is an average time to get such a project to production?

Boffin1:
I made the mistake of telling the customer that I had the prototype working, and now they want four finished units by Wednesday !

Personally, I think your clients are being very unrealistic thinking you can go from prototype to multiple finished units by Wednesday. You're going to have to use some diplomacy here; what does your contract with them say for a timeline? You do have such a contract, right?

You might be able to stall, tell them you'll demo the prototype to them on Wednesday, to make sure that you have met the objectives set out in the contract, then work out a time schedule based on that meeting (depending on whether you have to go back to the drawing board or not) to ramp up to produce four "finished" alpha devices (I don't know what they are expecting this device to look like - full finished and styled cases with custom PCBs, or simple cases picked up off-the-shelf?).

Kinda a "stall" tactic, but if you phrase and handle it right, letting them know you are trying to deliver the best product for them (while sticking to your contract), they should appreciate it (hopefully).

Contract? Whats that, this is Africa :slight_smile:
I have been consulting to these people for 8 years now with nothing in writing.

I made a prototype for this project last year using cmos chips for all the video side, and an micro for the audio ( using the wavehc library reading from a SD card )
They had a customer interested about 3 months? ago, but moved the goalposts by asking if I could do it with a function that queues the incoming numbers.

I said I could , but not with cmos logic chips, I would have to switch to a micro for the video. ( I also had some other projects to design and get into prototype stage during this time)
Thanks to this forum I managed to get the software working, and finally got a perf board proto running a couple of weeks ago. I designed the pcb over Xmas/new year, and received the boards last Friday ( all the factories close for summer holidays here over Xmas )

Meanwhile the company has promised delivery of four units by the 10th Feb!( the pcb just fits in a plastic box so that bits OK )

The more they ask me when the final one will be ready, the more I tell them there is no way of knowing, I said I hope by end Jan, but I always qualify this with if all goes well. (They are non technical people and do not understand how difficult it is to give a time-line for R&D )

One director suggested I should have ordered all the components first ( like before I started designing it ! )

Of course, with the rush I made some mistakes on the track layout, and I have had a couple of OC tracks ( probably from all the flexing and poking around fault finding, breaking the pad/track)

And now I have a problem that the audio side just wont work at all, so I have to write a quick sketch to test that on its own.

And I still have to tweak the video switching circuit ( which i have on a daughter board as I knew it would need some experimenting )

I have had all 4 boards populated, so once I sort out the bugs, it should work out over the weekend, but I don't need this pressure - I am a Great Grandfather after all :slight_smile: I should spend my weekends fishing and drinking beer !

the problem with the audio micro was a lazy crystal, I stuck an LED ( and res) on pin 13 and loaded the chip with old faithful "blink.pde" and nothing happened, changing the xtal restored the blink... You can't see it on the scope as the probe stops the oscillation anyway..

Boffin1:
You can't see it on the scope as the probe stops the oscillation anyway..

You need to throw away your probe, it puts too much load on a circuit. I routinely probe micro-style oscillators with some 30yo no-name passive probe; in most cases, oscillator keeps working while being probed. Also, you can try to probe through a piece of thin paper and see if it helps.

or I could connect the probe through a 3.9pF cap that I have in the drawer, its got 22pF caps to deck anyway.

As a general rule of thumb, I find that once I have a prototype running, I'm only about a third of the way done with the project. That could just be me though.

(They are non technical people and do not understand how difficult it is to give a time-line for R&D )

That applies to a lot of managers I have worked for as well and they were supposed to be technical.

but moved the goalposts by asking if I could do it with a function that queues the incoming numbers.

I said I could ,

That was your mistake, I know you are trying to be helpful but that is what causes 90% of the problems with projects. At that point you should have said it is like starting the whole project over.

I find that once I have a prototype running, I'm only about a third of the way done with the project.

I would agree with that especially if you are going into production. At one company I designed something and had a working prototype in three days and it took them a year to get it into production.

Great thread for all those thinking of getting into the design biz... :smiley:

I have been here so many times (over 40 years doing this) that I now own a T-Shirt factory so I can churn out the T-Shirts faster..

I hope you got those units working... hehehe

We used to have a "two, two and a half" rule.

Everything would...

Take twice as long as you thought
Cost twice as much as you thought
Earn half as much as you thought

Take this and use it well grasshopper :slight_smile:


Rob

Remember when you were a kid in the car on a road trip.
Are we there yet?
Developing New stuff always takes more time than reproducing existing technology.
Good Luck

How long will it take?
My answer is usually "How long does it take to catch a fish?"