Simple multi-channel serial ADC (8-ch)?

You don't have to, we will do it for you.

nly

With a couple of MUXs, we can deliver 16 channels at that price.

.=2E

$5 for 1K is no problem for a 50MHz ARM.

nes,

t I

That's why you need hardware averaging, or low pass filtering.

Hard to believe that. Where are you in?

It usually takes a couple of days to learn a uC, most of the time dealing with I/O. If your contractor can't learn a new uC, look for another one.

We can deliver the source code for the program and Jtag downloader via USB. You just need a 6 pins header (0.1" or 0.05" pitch) on your board.

nly

.=2E

nes,

t I

Analog devices has 8051 version of their microconverters

I have used them and with no bits toggling with a DC input, so noise is not a factor, but I also have used 10 bit AVR's and had only a single bit toggle, with 4x average you can get 10 bits without noise, and you don't need any external crystals. In any case you can do A/D converstions in sleep mode if your layout is not great.

LTC1090 or LTC1290 ?

- Henry

Hi Jörg,

$ 3.98 from

May be but this is only a example. Guess your very special AD wouldn't be in stock too.

May be you want to use a very cheap ATtiny with a 4067 as Mux? I'm sure this isn't to beat. for low data rate you can create your own quad slope AD using the cheapest controller to be found.

I saw the sample results from 1 Hz full scale sine wave without a plane or such special features just given to the minikit sampled there and transmitted this online with 2 kSps over UART (UART-USB-Bridge - Yes powered from USB! but with onboard 3V3 regulator). It looks fine without missing codes, only a few toggles around the last bit. In my view a very good result. I didn't expect more than 9 bits.

It is a ARM7 controller and the only thing you need is the datasheet (handy vs 100 pages) and little knowledge of Ansi C. There are good examples from analog and around in the net. But If I were you I would use a cheap one and make my own dual or quad slope with a few 4051 mux in front of it as I remember you want to measure DC signals and mostly you design as cheap as possible for mass production

Marte

I can do it for around $3, sans external mux. With an ADC chip ;-)

Works as long as you don't have slower patterns in there. Seen that more than once on uC (we even ended up having to shut it down during acquisition).

Sacramento, CA. No problem here to find talent in any field or uC. But I have clients in more remote locations.

I might take you up on that some day. Just not for this project, the architecture is nearly done in analog/mixed but without any MIPS horsepower. There is another reason for not having a uC. The board needs to be super duper quiet 90% of the time when data acquisition takes place. There are some low level RF signals up to 100MHz that want to remain undisturbed.

Err, I clicked on it. Seems they don't have 1000 but only 490. They want $4.52 per chip if you buy them all.

I am going to use something like the TLV1548, around $3-4. Plenty in stock, there are many similar ones. If the SW guys don't like the initialization chat I might spoof it like I described before. But maybe they don't mind.

Slope might be a bit too slow because I need eight channels and SW guys usually prefer to read those in one fell swoop.

That is indeed quite good.

Or use a cheap 1-channel ADC and a HC4051. That's really hard to beat.

Hi Jörg,

I would prefer a µC so you can implement the control of the 4051 with the same timen than the ADC and the SPI/I2C or whatever else. Have a buffer, possibility of averaging... and the noise is almost from peripheral clocks, you don't need this with several µC. They can be very quiet, I guess more quiet than a discrete ADC with external oscillator pin.

Marte

They also need a data-in sermon. Looks like older 5V devices, not too much stock, no TSSOP and very expensive. But thanks anyway, I'll see if LTC has more modern ones.

Yes, probably a uC could work here if clocked slowly. It doesn't have to do much. But there is one more pain left (don't I always find a hair in the soup?): Programming it. Beats me why companies like TI keep their SBW details a secret. They also do not provide a simple download tool. By simple I mean something that can be clicked from a PC and then automatically download into the uC. The user interface should consist of an *.exe file that gets double-clicked from Windows Explorer, no more. Currently you have to use one of the programming suites for that. I use IAR and I am comfortable with it but a biomedical engineer who does mostly mechanical and chemistry stuff won't be. I can't expect my clients to learn a design suite to program my boards. At least not the smaller companies among them.

The analog/mixed solution I am designing now will not have that programming hassle. Somebody stuffs it, somebody tests it (probably me), then build a few dozen more, a tech tests those and then plops them all into the system rack.

Hi Jörg,

Don't tell such things. I'm really not the biggest programmer on earth, but I guess this may be the work for max half a day. For a programmer with practice it will be half an hour.

SBW?

I remember for ADUC there is a complete docu around, I guess it is general for the ARM7 derivates. At least LPC and ADUC can be programmed with a freeware tool available on the net (linux based, because the manufacturers supports only windows based programming tools ;-)

That should be easy too. You only have to spy on the UART while programming and send this sequence to the controller :-)

The mspgcc guys may help here. I guess there is a downloader with command line support there too.

Flashing is a very easy tool. this may be done by every person having the head not only on the neck to prevent the rain:-)

The time you search around will be longer than to program this little code, isn't it?

Marte

P.S. few years ago I thought similar. No µC if ever possible. But I see, the times are changing...

Coding is easy, I meant the download procedure (for people other than us EE types).

SBW = Spy-by-wire, the download scheme for the new MSP430 F2xxx family. Some of those don't have a built-in UART anymore. SBW is a nice two-wire link. Would even be nicer if it was one-wire, maybe some day.

If it has one. Why not use what's already there (SBW)? Why do they make that so hard by not releasing the innards of the protocol? I understand if someone doesn't want to disclose grandma's famous chocolate cake recipe but SBW can be reverse engineered if someone really wanted to. They might as well open it up which would increase sales.

Not for SBW, AFAIK.

Sure they can learn it but that takes time. It's the same as someone asking you to diagnose a modern car where the engine began to stall at times. I am sure you could figure it out but first you'd spend many hours learning because this ain't grandpa's old motor anymore. That's why they made those little connectors under the dashboard and a small hand-held diagnostic wand.

Nah, it's pretty much done. It'll be a SPI converter. Either an 8-ch or a 11-ch depending on price and stock levels.

I have nothing against uC, sometimes they are the tool du jour. But other times you can achieve a very robust circuit without. One that can be produced until long after this particular uC model has gone obsolete ;-)

Well, we are getting close to a simple downloading solution for ARM. We have a gdb server running on the target and jtag flash the chip via usb.

We can remotely gdb the chip, or have a friendlier program (WinGdb?) to download it. The remaining work is to figure out how to interpret the AXF file. The goal is to:

Compile in Keil (perhaps WinArm eventually), strip AXF into binary image. TCP it to the target (or programming box), flash it with usb jtag.

Exactly as ordered. First function is to read and write the flash remotely. This will work with any FTDI 2232 based usb cables. We currently have a target running Linux 2.6.16 and an usb Arm. If your target is on static IP, you can remotely debug/download it across the country.

That's what is needed. Now if this would be available for a smaller uC such as the MSP430 I believe it would certainly increase its market penetration into areas that have gone without any uC so far. IOW not chasing the competition but truly new markets.

IIRC from the Yahoo MSP430 forum there is a guy working on it but last time he was still looking for someone who would take on the PC side software. I don't want to spoil the broth here but, sorry, Linux isn't really an option. If you take me as an example none of my clients uses Linux. All Windows :-( .... ok, one MAC.

I thought about a uC here after the discussion with Marte but will forego it again this time. This board needs good radio silence most of the time. The only way to achieve that within a few usec is to have an external clock that is gated and then it'll all become quite esoteric.

Hi Jörg,

AFAIK this is similar available for the older MSP430 series as well as the ATtiny.

It shouldn't be a big problem to port to win32.

But here you need to flicker around with the external clock while measurement are in progress while with a µC you can have the quiet while measurement and have the external clock only while transfering the data.

Marte

pe

the

cks,

more

why would your clients be programming your boards?

Hmm, that was the first question I fired off at the TI seminar. My question was whether there is a simple Windows-based code dump routine that would let non-programmers do a bootload. The answer was "nope". Later I asked again but the answer didn't change :-(

Also, nowadays one really wants to embrace the F2xxx family (usually SBW-only) because older ones didn't even have a proper brown-out reset, among other things. By the time you add all that stuff you might as well go analog.

It would be tough for me. Me and Windows frequently clash ;-)

Sorry, I must have expressed myself without clarity. It's not data that is transferred but small signals that are extracted from other signals by that board. Pretty much down to the noise level. Then these are sent to other boards in analog fashion. During that time there cannot be any clocks and the pauses are very short. Just a few usec.

While there are those few usec between measurement periods that isn't quite enough time to start a MSP430 which is already one of the fastest uC to come out of low power sleep mode.

If you run SPI the whole clock scheme is under system control. I can request SPI to stop running after x microseconds no matter whether or not some task is completed, then resume after y microseconds. Probably this can also be handled by a uC as long as it is a fully static design. We did that in the early 90's with one from the 89C51 family. It was a very similar situation but in that case there was no choice because we had to run rather complicated communication with a connected device. Plus that client had expressed the desire to talk to this board with a Qume terminal (remember those?).

Because they are contractually their boards :-)))

But seriously, suppose they'd want a minor mod. With a proper downloader I'd do it, send it to them and tell them to connect the board, then double-click modgizmo.exe or something like that.

Without, it goes something like this: Do the mod, go to the SWA or Jetblue web site, book flight, book hotel, book rental car, fly x hours, rent car, crash in hotel bed, drive to client, program boards, drive back to airport, get in traffic jam because some knucklehead speeded and spun out, miss last flight...

ure

lope

h the

r,

ocks,

more

Yes but compare that to the mess that would occur if your customer wanted a mod to your serial based stand alone A/D board (it would be a HW mod)

There is no reason why you can't consider a CPU based A/D a piece of unprogrammable hardware. Once you program it, put it on the board and make no accommodations what-so- ever to reprogram it. The serial stand alone A/D doesn't have that feature, and it appears it's acceptable to your customer, so why add it.

Yes. How about a Lear Siegler ADM3?