Low cost portable signal generator

[i_am_fubar]

Hello there, I have developed a project that I have been in need of for a while, and am considering taking it to production. I'm trying to get a feel for the market before I push ahead with it as the start up costs (membrane cutters...etc) are quite high.

The product will be a digital signal generator capable of outputting PWM waves of varying frequencies and ratios. It will be programmable through a decimal membrane keypad with an LCD screen showing the exact frequency and ratio that you are getting. The specification is below:

Power: 9v battery or external transformer
Size: 92mm x 66mm x 28mm
Frequency: around 2KHz to 2MHz
Duty cycle: 0% to 100% (resolution dependent on frequency)

I am also considering giving it the ability to read in a PWM signal and display frequency and duty cycle.

Estimated cost would be around £70 - £80 ($140 - $160) initially.

*NOTE*
This item is NOT FOR SALE, I would simply like to know who would be interested in buying one if they were.

Any comments or questions welcome.

[mechatronicsman]

Hi,
while a signal generator is a very useful equipment I think you are aiming too low on the specification side for that high cost. E.g. here is a commercial one with same spec commercially available for $109 http://www.ldbwebtronics.com/newfusige0vc.html, but with possibilities for much lower frequencies.

For $150 I would at least require a frequency range of 0.1Hz to 100 MHz. 2MHz max are too low, e.g. for driving embedded micro controller boards.

Also computer control possibilities (changing settings trough a USB/RS232 interface, for automated test support) would be a very good feature, significantly increasing the limit for what it would be worth paying.

Hope this gives you some idea

-bjorn

[i_am_fubar]

Oh, thanks for the feedback.

Ok, so, on the flip side of the coin, what would you be prepared to pay for the functionality I've described?

Size restriction is my enemy here, as I'll be designing it to be as small and portable as possible.

I could probably have a much wider range of output frequencies, though you would loose the ability to have anything other than a square wave (no PWM).

I could possibly put on a mini B USB connector for an interface. And the primary intention of this device is to create PWM wave forms for switch mode power supplies, though increasing that for micro controllers would be useful.

I could also put in an analogue input so that the output ratio on the PWM, or the frequency are modulated to that value.

As to cost, the estimates are for a batch run of 25 units. if I did 100, I could probably get the price down to $75ish.

Think I shall run all my ideas through this forum, develop the products that people want, not what I think they want

Sam.

[mechatronicsman]

Originally Posted by i_am_fubar Oh, thanks for the feedback. Think I shall run all my ideas through this forum, develop the products that people want, not what I think they want Sam.

Good idea A point where may people take the wrong path of trying to estimate a customers need.

Small and portable is a plus. Also getting it with membrane switches would make it a good choice compared to make one myself using simple keyboard-type switches, or an AVR starterkit or similar.

I would say if you have 1KHz to 2MHz PWM support, and 0.1 Hz to 32MHz or higher square wave (50% dutycycle), adjustable amplitude, and a serial interface (USB? UART type) to change frequency by computer control, in a small form factor with easy to use interface it would be well worth the $75.

Any one else having inputs?

-bjorn

[i_am_fubar]

This will be, depending on its success. The first of many products in a range I have in mind. Including a computer based oscilloscope, power supply, wireless Data links, analogue sig-gen... etc.

Membrane keypads are nice, have started using them at work. but theres a £300 set up charge ($600ish)

Output was going to be only 5v TTL for direct connection to logic.

Next question, for a few extra dollars, would it be worth having a back lit LCD display as opposed to a reflective one?

Also, out of the following list, which would be most desirable?:

• Digital signal generator
• RS232 analyzer and generator
• Analogue signal generator
• Computer based oscilloscope/spectrum analyzer
• Computer controlled analogue and digital I/O interface box (with C libraries)
• Universal motor driver (stepper, DC, servo) computer connection.
• Pulse timer/Chronograph

[henrikw]

Well, I'd go for:

• Computer based oscilloscope/spectrum analyzer

Definitely if you can/want to compete with poscope

Henrik

[i_am_fubar]

The functionality they've managed for the given cost is remarkable. I doubt I'll be in a position to compete with that scope for quite some time. I can foresee the development costs being quite high. Though I'll be sure to give it a stab at some point.

Maybe worth making some sort of computer controlled universal lab center including all the functionality in my list.

[henrikw]

Well, it doesn't necessarily have to contain all the functionality, I often find myself in a position where a simple scope with autosettings and two channels is quite enough.

If you can't compete on functionality, go for the K.I.S.S. principle

I would love to see your work and appreciate that you run your thoughts by us.

Best regards

Henrik

[i_am_fubar]

I've got a lot of ideas, and a lot of designs (I think my archive is up to 50 now). I just lack the time and money to implement most of them.

Once I sort my site out, I will be putting projects on it, as well as selling some of them.

Keep your eyes on www.SamRhodes.com (There are a lot of dead links, in fact, only the mill is up there at the moment, sorry about that.)

And KISS is nice and easy to do. Will keep the simple scope project in mind.

Sam.

[mechatronicsman]

Originally Posted by i_am_fubar Output was going to be only 5v TTL for direct connection to logic.[/LIST]

Trust me.. with that limitation you will loose 70-90% of potential customers.

5 V levels are phasing out quickly. Less and less embedded electronics/MCUs support that now. 5 volt is mostly used in industrial control and white goods appliances today.

And TTL 5V levels are definitely useless in most applications. A TTL level only guarantees 2V out on logic high. And a CMOS integrated circuit (most ICs today are CMOS) need 70% of VCC for detection as high. I.e. CMOS high level need to be at least 3,5V to be guaranteed to be detected as high. So if 5V only be sure the drivers are rail to rail drivers, not TTL drivers.

This is one of the reasons many people struggle driving circuits directly from the parallell port of a PC, as old PCs have TTL signal levels, not suitable for directly interfacing modern CMOS stepper drivers for instance.

So to be usable to other applications than CNC and industrial control applications you need at least to support 3,3V levels in addition to 5V. Preferably adjustable amplitude between 1,8 to 5V.

-bjorn

[i_am_fubar]

I didn't realize all that. Thanks for the info dump.

I did mean full rail to rail 0v,5v outputs. Wasn't aware that TTL was so grey.

I guess it wouldn't be difficult to add a variable amp at the output to give 0.5v - 8.5v.

I think I may need a bigger box.

Sam.