I am going to buy a 10k, 256 step MCP42010 potentiometer (http://www.jameco.com/webapp/wcs/stores/servlet/Product_10001_10001_241752_-1). I am wanting to know a few things before I buy.
256 steps means 256 equally spaced out values up to 10k, right? Does this include 0? Could you give me the first and last 5 values of the output of the digital pot, assuming I list all 256 values it can produce?
What is the accuracy of this digital pot? 10%? 5%? 1%?
http://www.microchip.com/wwwproducts/Devices.aspx?dDocName=en010497
Somehow I doubt it goes to 0. That's a dead short.
The link didn't really help. What's the tolerance?
So if it doesn't include 0, then the values would be 110000/256, 210000/256, 3*10000/256, etc?
Actually, the link does help. It points, among other things to a data sheet...
From Page 1:
The wiper position varies linearly
So yep, guess that means evenly spaced
From Page 13:
There is a nominal wiper resistance of 52? for the 10 k? version,
125? for the 50 k? and 100 k? versions. For the dual
devices, the channel-to-channel matching variation is
less than 1%. The resistance between the wiper and
either of the resistor endpoints varies linearly according
to the value stored in the data register. Code 00h
effectively connects the wiper to the B terminal
So your minimal resistance is 52Ohm
The wiper seems to have some strange behavior though if you plan on powering it with low voltages. Check out image 2-23. This should influence your accuracy, though I must admit I am not sure if I understand it 100%
Good luck with your project.
Regards
p.
Oh, that link was helpful! There's a formula in the datasheet, haha.
Why buy one digipot when you can sample it and probably get 3 each of 3 different values for free, and they usually throw in 2 to 3 day shipping too?
I would strongly urge folks to use samples as a last resort. The economy of the US (among other countries) is bad enough that companies are looking for any way to cut costs. If samples are being taken liberally by users with other means to acquire that product, guess what'll get cut?
Sometimes you just can't buy "one" of something, or the shipping from some vendor costs more than the project budget. In those cases, yeah, look at taking a sample. But if you can get qty 1 of a part, shipped, for $10 or less... I see no way to justify abusing the system.
My $0.02.
SirNickity:
I would strongly urge folks to use samples as a last resort. The economy of the US (among other countries) is bad enough that companies are looking for any way to cut costs. If samples are being taken liberally by users with other means to acquire that product, guess what'll get cut?Sometimes you just can't buy "one" of something, or the shipping from some vendor costs more than the project budget. In those cases, yeah, look at taking a sample. But if you can get qty 1 of a part, shipped, for $10 or less... I see no way to justify abusing the system.
My $0.02.
I totally agree. I'm not a company, just a high school student wanting to become an electrical engineer. I am going to place an order with jameco soon, and I'll buy it there. Samples are really only for people wanting to buy parts in bulk.
blake305:
I totally agree. I'm not a company, just a high school student wanting to become an electrical engineer. I am going to place an order with jameco soon, and I'll buy it there. Samples are really only for people wanting to buy parts in bulk.
This is completely backwards. Samples are only for people wanting NOT to buy parts in bulk. You will at most get 2 or 3 samples per type except for some limited exceptions for very, very inexpensive parts. If you want bulk (1000 resistors or caps, 100 595 chips, etc.) sampling will never work for you. But for a few digipots, I think sampling is a good way to acquire a few. Most of the companies that sell digipots will sample them and will send 3 of 3 or 4 values no problem. Also, if you want to play with a National/Analog/Linear mid-priced ADC (say a $20 model) they will be only too happy to send you - ONE. That's a lot better than paying Digikey $30 for a $20 part because you only want one and then $6 to ship it. (nothing against Digikey!)
As a practical example, I nicely asked AD a while back to send me 6 digipots (2 each of 10K, 50K, 100K) and they helpfully did. I love Analog Devices. GO ANALOG DEVICES. I hope I can make their small investment money well spent. Their sales last year was $2.8B and EBITDA of $1.08B. They know what they are doing. The are not acting against their own interests here.
P.S. I love Jameco. Great people. Digikey is a good place to go if Jameco doesn't have it - they stock about everything.
JoeN:
This is completely backwards. Samples are only for people wanting NOT to buy parts in bulk.
I think the point is that a designer can take a handful of them on good faith that, if they perform adequately for the application, they will follow that up with an order of 10,000 of them once the device goes into production. In this case, the cost of a few free samples is negligible. I've always assumed this is the primary reason for providing samples.
On the other hand, for a hobbyist, taking 12 parts a couple times a year doesn't cost AD / TI / Nat'l much, but in your lifetime, you may only ever provide them with $50 - $1000 worth of orders. There's really no incentive to continue that practice.
I agree, they know it happens, but I suspect that's why there are such low limits on what they'll allow. If it were just professional designers, they would probably be willing to ship several hundred dollars worth of samples, knowing it's likely to turn into a quarter-million in revenue later. $) The fact that we're even allowed to ask for one-off parts, I feel, is extremely gracious but not a given.
Sorry to turn this into a soapbox -- I really can't say what the executives of semiconductor companies are thinking. I just want to raise awareness that there's still no free lunch, and if it becomes a casual way to acquire parts, we may soon find it takes a sales account, tax ID, and a retainer contract to get a printed catalog. So, I suggest (and that's all) that folks tread lightly. 
Can I please suggest if you guys want to continue hijacking this thread, please create a new thread in a more appropriate area of the forum?
I really don't see how any of this discussion relates to the original question.
Thanks.
I agree... fully and I will be short. A similar thing happened with Analog Devices and the AD9850/1 several years ago... it's why AD changed their sampling guidelines about 5 years ago...
Doc