Other Parts Discussed in Thread: ADS1274, TLV431, , THS4521, TMUX6104
This thread has been locked.
If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.
Sure, I had a quick look and have the following major comments :
- I see you are trying to use TS3A switch to select biasing current between 2 and 16mA. this has basically two issues:
> the switch selected is 4.6V switch. it can not be used at the current path at which voltage goes beyond 21V (node pin U22/1D for example)
> if you replaced the MUX with one of the high voltage MUX's like MUX36xxx or MUX61xxx the switch resistance is in the range of 100 Ohm and would definitely change the accuracy of the output current (which is not critical anyway, but need to be considred) for most sensors current +/-20% should be ok but you need to consider that.
- using same switch for MUXing the different attenuations
> again, TS3A is low voltage switch, and must be replaced with high voltage MUX/Switch
> when choosing high voltage MUX, remember that most MUX's are (1:4 differential MUX) MUX's (while TS3A is 2 [1:4 SPQT switch]). you can use separate MUXs or check if MUX61xxx family has this configuration already.
- Schematic also has the attenuation output driving Diff-amp input directly.
> This will not going to work. output impedance of the attenuator migh go to 60kOhm, while your input impedance of the diff amp is 1kOhm. so you are driving relatively low impedance with high source impedance and you get very high attenuation. you need to put a buffer between the attenuator and the diff-amp.
I can't go through the whole schematic as it requires some time, but that what I have from first sight.
I had a quick look according to available time, and here are some comments:
- the gain stage has a very high gain (1.1M/2k), the gain stage (as it is fixed) should be designed so that it has maximum Vout (peak to peak) less than VDD33. possibly you would design with 1.5V peak on top of Vcm=1.65V. check your maximum input levels and possible attenuation levels to achieve that goal. I don't know if you really need the different attenuation. if not needed just go with single setting and remove the switch and the extra resistors.
- you have 2 circuits generating Vcom. one on page 3, and another one on page 4. you can remove the circuit in page 4 as the ADS1274 has internal Vcom generator already.
- I didn't try the current source (using TLV431) with switch in the feedback. not sure how that would affect startup if switch is open or if the circuit would latch indefinitely in the wrong state. I would recommend to simulate the circuit or build a quick prototype to check if it working the write way.
Another important point:
- you are using 5V AVDD, means your common mode is 2.5V. that matches your reference which is 2.5V. then your full scale input is -2.5V to +2.5V. you are powering the THS with 5V. So far so good, your THS should generate +2.5/-2.5V as it has a gain of 1 (not 2 as in TIDA-01471 design), input should be +/-2.5V around 2.5V Vcom. this is not possible with the choice of U12 power supply of 3.3V. you either need to power it with 5V (still it will be hard to get +/-2.5V output. you need to keep some headroom by adjusting the attenuation properly), or increase the gain of the THS to 2, and then you are able to receive +/-1.25V input, which allows the use of 3.3V for U12. However, in the latter case, you need to generate another Vcom2 to bias the U12 (and all other LMVs) and U5 (and all other THS's pin 1) which is 1.65V (for this you can use the auxiliary circuit in Page 4). note that the 2.5V Vcom is still needed for the pin2 of the THS differential drivers.
By the way, my needs are as follows:
- You need to plan signal levels at all stages, usually as a block diagram showing which common mode and signal swing would appear at each node in case of maximum input level. you would then add the supply of each stage, using datasheet information you can check if those signal levels are ok for each stage or not. if not you would need to change gain/attenuation settings or power supply for specific stages if applicable.
Two levels are given and fixed, those are the maximum input swing (+/-10V in your case), and the full scale of the ADC, which is determined by the ADC ref (in your schematic, it's +/-2.5V). the signal chain should ideally map the input to the FS level at the ADC input, so 10V should appear as 2.5V at the ADC input. as the input is differential, then each output of THS4521 would swing +/-1.25V around the Vcom of 2.5V. THS4521 input though is single-ended, and should swing +/-2.5V around the Vcom of 2.5V. that means input to THS is 0 to 5V. looking at schematic, this is not possible if you are using 5V supply for the LMV unless you make sure maximum signal is not exactly 0 to 5V, rather 0.2V to 4.8V for example (by adjusting attenuator), and changing the gain of the THS4521 to maintain the output to be fullscale level for the ADC.
- my comment about headroom means: if you use 5V for THS and design for +/-2.5V swing at each output, then probably it wouldn't work as there is always some limit to the output swing which (in case of rail to rail) is about ~100mV or 200mV from the supply rail. that means you need to set the gain to a lower value to avoid clippling, distortion, and errors.
- you mentioned you must use attenuation, I still don't know why variable? you have +/-10V input, then adjust a fixed attenuation to get the levels you need at the input of the buffer. Note: variable excitation current has nothing to do with input range. input range is solely determined by the sensor.
- In place of MAX309ESE I would recommend using the new TMUX6104.
I was on a business trip last week, so I haven't replied to you for a week.
- I using MAX309ESE because it's cheap.
- How do I adjust attenuator to make the ADC at 0.2V to 4.8V in my schmetic?
- I use attenuation because my input range maybe +/-10V, +/-5V or other values.
Again, this is a signal level planning that you need to do.
- start from the ADC, determine maximum single ended input in normal operation (should match max input, or a little bit lower)
- work back stage by stage, determining for each stage the gain/attenuation. and calculate input/output peak to peak voltage. ensure for each stage that the device is powered by the proper supply, and the input/output can swing to the specified limits without clipping or distortion (most rail to rail amplifiers needs some head room to the rails and don't actually go to the rail value), common mode level is also determined here
- finally adjust the attenuation to deliver the required levels from input +/-10V or whatever input levels.