Skip to main content
Topic: Alternate design (Read 324 times) previous topic - next topic

Alternate design

I have been thinking about this quite a bit.  Too much probably.  Follow this train of thought. 

  • The ADC is limited to 1v or 2v peak to peak but tolerant from 0 to 5v input. We set it to 1v peak to peak.
  • The opamp is tolerant from 0 to 5v input.
  • The DG449 is tolerant from -v to +v
  • The DG449's -v=-5v and the +v=+5v.

See attached PDF

This will allow an input to the DG449 to be 10v peak to peak.  If we have a 10:1 voltage divider between the DG449 and the opamp the peak to peak at the opamp would be 1v.  By using a scope probe with x10 and x1 you have a peak to peak input to the probe of 100v p-p and 10v p-p respectively.  We would use the OLS's 5v supply and generate a -5v on the wing using a charge pump like the LT1054. 

The two ranges would be set with the probe.  Assuming 10 divisions on the scope plot then each division would be 1v/div for x1 probe and 10v/div for the x10 probe.  To set the ground reference you ground the tip or like on my scope probes there is a switch for X1,X10,REF.

This would cut the cost dramatically and you would still have ac/dc coupling. 

McZ
"Too many projects, so little time"

Re: Alternate design

Reply #1
If we're down at +/- 5v (or even 8), can we use a (max) 4051 or even a 4066?
Got a question? Please ask in the forum for the fastest answers.

Re: Alternate design

Reply #2
Nope, the max4053 is just as expensive, and the 4066 is too slow:

http://www.maxim-ic.com/datasheet/index.mvp/id/1352
Got a question? Please ask in the forum for the fastest answers.

Re: Alternate design

Reply #3
Latest design files for the alternate design.  The cost for 1,25,100,1000 pieces drops to $29.78, $22.62, $17.84, $15.50 repectively.
"Too many projects, so little time"

Re: Alternate design

Reply #4
I just did a quick update to my spreadsheet and it looks like the alternative design shaves off ~$1 from the cost.

The LT1054 seems to eat up the cost savings from removing the switches. I haven't had a chance to compare datasheets yet but is there a cheaper regulator we can use?

Maybe one way to get the price down is to offer a 20Mhz version. (Assuming there are drop in replacements for a 20Mhz version that are cheaper)

Re: Alternate design

Reply #5
Ok I look at reducing the ADC sample rate and keeping the design intact.  TI has a 60MSps part the ADS830.  Same as the ADS831 but slower.  Also found a -5v charge pump for a lot cheaper. 

Cost are down to Qty1=$26.99, Qty25=$20.27, Qty100=$15.96, Qty1000=$13.17. 

Spread sheet is attached.  Green highlights are the improvements.  Yellow highlights are areas that we might be able to reduce further.

McZ
"Too many projects, so little time"

Re: Alternate design

Reply #6
Nice updates, thank you.

The opamp struck me as expensive too. I spent an afternoon looking at the opamp last week, but I wasn't confident enough to try a substitute. The datasheet says we can use a (more common=cheaper?) voltage feedback opamp instead of the current feedback one.

TI may stop sampling the 831 if there's a big after-market modding community :)

rsdio is going to have a fit about the unregulated USB voltage, but at least with this chip vout=-vin, so it will be symmetrical (my understanding was the old one was solid -5 for the whole input range).

References:
http://www.microchip.com/wwwproducts/De ... e=en010584
http://focus.ti.com/docs/prod/folders/print/ads830.html
http://focus.ti.com/docs/prod/folders/print/ads831.html
Got a question? Please ask in the forum for the fastest answers.

Re: Alternate design

Reply #7
The voltage opamp was slightly more expensive.  I tried that.  I will look for more alternatives. 
"Too many projects, so little time"

Re: Alternate design

Reply #8
Here is the spreadsheet with the lower cost opamp.

Not much blood left in this turnip.

McZ
"Too many projects, so little time"

Re: Alternate design

Reply #9
You're right about that. When I sort by price at mouser and eliminate all single-channel op-amps, this is the first one that would (appear to) work.

The mouser 1s price is $2.25:
http://www.mouser.com/ProductDetail/Tex ... 6zoA%3D%3D

The old op-amp was $5.50, but non-stock (53 @ digikey):
http://www.mouser.com/Search/Refine.asp ... =OPA2691ID

For availability and price, this is a big upgrade.

Here's key figures I pulled for the datasheet for personal reference:
Property OPA2830 OPA2691
slew - 560V/uS - 2100V/uS
gain bw - 100MHz - 2GHz
-3db bw - 290MHz - 280MHz
feedback- voltage - current
Got a question? Please ask in the forum for the fastest answers.

Re: Alternate design

Reply #10
Here's an updated eagle schematic with the single relay and -5volt supply.

I added back in the over-voltage protection diodes from the first design that I think were removed from the update.

This is NOT updated with the new op-amp.
Got a question? Please ask in the forum for the fastest answers.

Re: Alternate design

Reply #11
huh, try again.
Got a question? Please ask in the forum for the fastest answers.

Re: Alternate design

Reply #12
BTW I forgot to reply to a question you had some time back.  The reason the -IN is the input is to invert the signal from the opamp which inverts the signal from the relay.  Your schematics are correct on that input.

McZ
"Too many projects, so little time"

Re: Alternate design

Reply #13
Yup, I think I read that section and posted it in the thread about five minutes later :) My mistake.
Got a question? Please ask in the forum for the fastest answers.

Re: Alternate design

Reply #14
Where are we on this? 
"Too many projects, so little time"