8-Channel DAC?

JonnyMac · 2011-11-02 12:24

I'm working with a customer who is using a 4-channel DAC on his product and now needs to bump to 8. The DAC we're using now is an I2C device and not easy to address (no external address pins -- has to be done via code); in an ideal world we'd find an 8-channel, 12-bit dac that uses I2C or SPI (we have both in the project).

Suggestions?

zoopydogsit · 2011-11-02 12:29

Ooops, you want DAC not ADC.
Ignore this - sorry.
If you need 8 channels muxed then what about the MAX147B?
http://pdf1.alldatasheet.com/datasheet-pdf/view/72837/MAXIM/MAX147BC/D.html

StefanL38 · 2011-11-02 13:30

just a short google search with 8 channel DAC came up with this.

http://newscenter.ti.com/Blogs/newsroom/archive/2010/11/16/ti-s-12-bit-8-channel-i2c-interface-dac-targets-multi-channel-high-density-applications-554530.aspx

best regards
Stefan

frank freedman · 2011-11-02 13:36

JonnyMac wrote: »

I'm working with a customer who is using a 4-channel DAC on his product and now needs to bump to 8. The DAC we're using now is an I2C device and not easy to address (no external address pins -- has to be done via code); in an ideal world we'd find an 8-channel, 12-bit dac that uses I2C or SPI (we have both in the project).

Suggestions?

Hi Jon,

I am getting ready to play arond with a couple of new (for me) chips. One of them is the TLV5610 from TI, it is a 12bit 8 channel DAC w/ SPI. QSPI, and Microwire serial inputs.

http://www.ti.com/product/tlv5610

The part TI sampled to me is in 20-SOIC format. it just may fill the bill for your client. If you have not already done so, you might want to check into creating a my ti login to the TI site.

Frank Freedman

JonnyMac · 2011-11-02 13:38

Thanks, Stephan, I did not come across that in my own searches

Leon · 2011-11-02 13:38

It's actually a DAC, which is what he wants!

Phil Pilgrim (PhiPi) · 2011-11-02 15:32

These, from DigiKey:

None is exactly cheap.

-Phil

Bits · 2011-11-02 15:39

Digi key is the place I go for all the ICs. Just type i2c in the search bar or click the link

frank freedman · 2011-11-02 16:17

Leon wrote: »

It's actually a DAC, which is what he wants!

Yes it is actually, the part of the note saying ADC was a typo for DAC!!!!!

frank freedman · 2011-11-02 16:23

No DACs are cheap, not TI nor Analog Devices. Nat Semi is now part of TI. Digi-key = sticker shock, and element whatever newark has become does not look much better....

Bits · 2011-11-02 16:27

Frank

You actually think Digikey is expensive or is that post up there interpreted wrong by me? I am stunned if this is true since I purchase almost 90% of my inventory there mostly due to prices and it saves me a lot of time. The other 10% is the propeller / ftdi from parallax

frank freedman · 2011-11-02 16:50

Actually only in reference to DACs. It is all relative. I have found that I end up checking both of these as well as a couple others depending on price and availability. Seems odd that these are so much higher than say some of the more advanced mcu chips.

Frank

Sorry about the digression JM/

Bits · 2011-11-02 17:53

Yea I agree with that! DAC and ADC can get pricey and quick, not to mention some of the extra needed supporting hardware. No need to be sorry Frank, my words did not reflect my intent well I guess my DAC was off by a few bits!!

RS_Jim · 2011-11-02 18:51

Jonny
Check out analog devices, I believe they are touting some 8channel devices.
JIM

frank freedman · 2011-11-02 19:03

On another note with who has what, I just remembered that a dual channel sample from TI had what looked to be the BB logo. Did TI also take over Burr Browns operations as well? I do what PhiPi and Bits do if I am not getting a sample device, google w/ part or digikey to start. Avnet is good for some off the wall stuff as well. Avenet does some good seminars in their field offices as well called their Speedway series. I will be at one for TI products in phoenix on 9 nov.

@ Jon they may have one also inSan Diego or Los Angeles.

Frank

Leon · 2011-11-02 20:25

TI took Burr-Brown over a long time ago.

pjv · 2011-11-02 21:05

Hi Jonny;

I presume there is a Prop driving things, and if your speed requirements are modest and load is light, then are a bunch of simple software RC DAC's not suitable? Ten bit granularity should not be difficult, and at low speeds surely 12 bits is quite possible. (counters using an 80 MHz clock are not much good due to horrendous switching transients).

Seems simple to me, or am I missing something here?

Cheers,

Peter (pjv)

JonnyMac · 2011-11-03 00:05

Thanks, all. Interestingly, I'm more of Mouser than Digi-Key guy -- funny how we all land on our favorite sources.

@PJV: No, RC DACs will not work in this project; I/O requirements are already stretched and we're using SPI and I2C devices to manage the IO requirements we have. We've go lots of options to look at though, as everyone pointed out, none is inexpensive.

pjv · 2011-11-03 07:34

Hi Jonny;

Not trying to beat a dead horse here, but.... being stretched for I/O would only impact your application if you did not have sufficient pins. Not the same case as being out of Cogs or Cog cycles due to I/O processing. If this latter were the case then a muti-threading approach would likely help; running multiple drivers in a single Cog. If pin count is your problem then you are right, R/C DACs cannot work.

Cheers,

Peter (pjv)

JonnyMac · 2011-11-03 07:41

Peter: We don't have enough I/O pins nor cogs to support RC DACs. And, if I'm honest, it would be trying to beat a dead horse trying to get my mind to wrap around your multi-threading code <grin> -- try as I have, I just don't seem to have the capacity to grasp it (I'm an actor, not a brain surgeon). Pins, precision, and the final output (driving 4-20 circuits) suggest we should stick with a single-chip solution. I'm already using and sharing SPI and I2C busses; something that drops in where our present 4-channel device presently sits is what we're after.

MacTuxLin · 2011-11-03 09:58

Not sure if it makes sense but how about using a 8-channel i2c switch like TI's PCA9548A ? Still using same prop's i2c to comm with 2 x 4-channel ADC using 2 channels of this switch?

frank freedman · 2011-11-03 17:24

Solved or still looking?

LoopyByteloose · 2011-11-03 17:44

This may seem a bit of a distraction, but does one really need a DAC or can the right PWM output replace it?

@JonnyMac
Most multithread code is time slice. So, if one can accommodate divide by 2, divide by 4, divide by 8, and so forth and still get good timing results (in this case divide by 8), it shouldn't be too hairy.

Of course, pjv might take it much further and not have simple time slices. After all, he has really mastered it all

AND, I suppose there are advantages to having the DAC buffer the level of the output between updates.

Bits · 2011-11-03 19:51

Loopy Byteloose wrote: »

This may seem a bit of a distraction, but does one really need a DAC or can the right PWM output replace it?

In my experience PWM wont always work. There are times (seems like every new design) where you need the processing time and PWM can take that time away. Another factor is impedance matching, multiple channels, voltage, temperature and noise.

Jonny

Here is a 16bit 16chan DAC from maxim , uses SPI among others. looks nice imo.

"allowsmultiple values to be loaded into memory in a single,
high-speed data burst"

Beau Schwabe · 2011-11-03 23:12

JonnyMac,

Just thinking out loud here... without using any more I/O's, you might be able to use two of the 4-Channel DAC's that you are already using/familiar with. The 'trick' is to leave one DAC just as it is, and have the other with the SDA and CLK lines swapped. I've done this before where the I2C address was hard coded and it worked perfect.

pjv · 2011-11-04 07:35

Beau.... you sneaky guy ! So clever and simple. I guess we should have expected this from you.

Cheers,

Peter (pjv)

JonnyMac · 2011-11-04 10:58

Will mark solved -- decision was made to live with the four channels on the board.

8-Channel DAC?

Comments