Looking for ADC for simple 1Msps Propeller oscilloscope
Bean
Posts: 8,129
We have looked around and the best chip seems to be the MAX1377 because it is cheap ~$4 and reads two bipolar inputs at the same time.
http://datasheets.maximintegrated.com/en/ds/MAX1377-MAX1383.pdf
But I haven't been able to find any demo programs or examples of people using it on the web, which kind of scares me off of it.
Also, I'm not sure about reading two SPI inputs at 20MHz and storing the data in the hub.
Has anyone used the MAX1377/1379/1383 before ?
I was just wondering if anyone has any suggestions for an ADC ?
Here is what we need:
1Msps for two channels (could use two chips if needed).
Bipolar inputs (so we don't need a lot of preconditioning circuitry)
At least 8-bits (more is better of course)
Cheap (<$10 for both channels)
Any ideas ?
Thanks,
Bean
http://datasheets.maximintegrated.com/en/ds/MAX1377-MAX1383.pdf
But I haven't been able to find any demo programs or examples of people using it on the web, which kind of scares me off of it.
Also, I'm not sure about reading two SPI inputs at 20MHz and storing the data in the hub.
Has anyone used the MAX1377/1379/1383 before ?
I was just wondering if anyone has any suggestions for an ADC ?
Here is what we need:
1Msps for two channels (could use two chips if needed).
Bipolar inputs (so we don't need a lot of preconditioning circuitry)
At least 8-bits (more is better of course)
Cheap (<$10 for both channels)
Any ideas ?
Thanks,
Bean
Comments
parallel interface its easy to read this faster (if you have 10 pins spare).
Last time, I was looking for ~100 MHz and got samples of this chip:
AD9283
There is a 50 MHz version that is one channel and $5...
The AD9201 look good, but from reading the datasheet it doesn't allow bipolar inputs.
Rayman, that looks good. I'll have to see if I can afford that many pins.
Thanks guys,
Bean
ADS8329 ADS8330
1Msps is fairly light, and many micros offer that now.
If it is for a scope, you might want some logic too, for trigger etc, and so something like this
http://www.digikey.com/product-highlights/en/psoc-4-arm-cortexm0/50179
gives a solution for 1 : $2.53, and just $24 for a Board
OR you might want to jump in speed, and Digikey suggests the MAX1192 is Dual, parallel, 22MSps for $4.38 @ 100
and ~20MSPS seems a better match to Prop peak sample speeds.
Surely parallel capture is the best, if the purpose is to build a scope, and there probably two capture paths, with 8 bits of ADC and a 9th trigger/digital bit would seem to fit ?
Or, if you want to be brave, TI has some interesting AFE parts that seem cheap
http://www.ti.com/lsds/ti/data-converters/ccd-cmos-products.page
Digikey has stock of
VSP5621RSLR Texas Instruments IC AFE 16BIT 50MSPS 4CH 48VQFN
4973 100 : $4.3133
The 4 channels at 16 bits is appealing, but the 50Msps is a little creative, as the sustained read rate given is
CMOS 4-bit
modules with 8MB RAM are available cheap.
in scan mode all 16 inputs can be sampled cyclically
- oh - but this is not bipolar ...
I'm not sure what you mean by bipolar ? - but the Data says this
Absolute Voltage Range 0~AVDD Volts
So it can do differential, but it does not go outside the rails.
fc(ADC) ADC conversion frequency 12-bit resolution - - 80 MSamples/s
http://www.nxp.com/products/microcontrollers/cortex_m4/LPC4370FET100.html
The 282kRAM, 64K ROM, and various cores and peripherals almost come for free.
The ADC seems to have FIFO and DMA, and can loop via a descriptor state engine + 14b Timer, as it reads.
As Smart ADCs go, this looks quite good
I think the Eval Board is $20, which seems a good price.
Well yes, a very good price - is there a link to the eval board?
I was hoping to get a device that could measure negative voltages (below ground) without the need for a negative reference, but I don't think such a thing exists.
I was getting confused by the output showing negative values, but I can see now that that just means that the negative input is greater than the positive input, NOT that the positive input is below ground level.
I think we will just use a charge pump to get a negative voltage, then use an op amp to buffer the signal, then to the standard propeller delta-sigma type ADC.
Bean
You can always use a summing circuit to allow you to read voltages below ground with a regular ADC.
You just add a fixed positive voltage to the voltage you want to measure...
High bandwidth opamp in differential configuration to mid-rail reference is how
I'd proceed I think.
DigiKey has stock at a $28 price. Just search their site for OM13054.
Also worth noting...a board with a TI version of the ARM Cortex M4 is available for $12.99 right here. It has two 12-channel 12-bit ADCs. This board works great with the free TI IDE (which full disclosure compels me to point out is a big download - but not as big as Quartus!).
That is the problem, with the scope probe disconnected it reads as having a DC voltage on it. So you need an op amp to get a low impedance signal.
Bean
http://www.ecircuitcenter.com/Circuits/opsum/opsum.htm
Only problem here is that the input impedance would be they show a 10k input impedance. There is no DC level on the tip of the probe though, other end of resistor is kept at ground... Maybe you could do this with the regular 1 Meg input impedance of oscilloscopes, not sure...
If you want No DC and a classic '1M to ground' Probe load, I think you have to use some negative voltage.(negative current)
It can be quite low current, if it is applied to the bias side only, but then it needs to be well regulated.
With std parts it is probably easiest to do a negative opamp rail and use the opamp PSRR.
I have seen scopes that split DC/HF paths, and use low offset precision parts for DC and HF parts for > 1Khz, but that needs a splitter and re-combine so bumps the parts count.
It is not easy finding well-priced, Scope-performance Opamps.
The Red-Pitaya scope uses a 14b ADC to avoid needing so many scope ranges and they scale to 600mV/1M so a Probe is 6V, which covers most bench needs. Seemed like a good approach, and I like a scope probe that can be a proper voltmeter.
As jmg said, I don't see any way around having a negative voltage supply to a buffer amp if you want 1meg input impedance.
But, there are simple charge pumps around that can take +5v and make -5V for you...
Just found a triple opamp with integrated charge pump:
http://www.analog.com/en/all-operational-amplifiers-op-amps/operational-amplifiers-op-amps/ada4858-3/products/product.html
It's for the "video" frequency range, which I think is what you're looking for...
Is the 'scope to be portable, battery powered? The ground is then flexible and could be center point between the two batteries, or it could use a single battery and virtual ground IC such as the TLE2425.