How to produce a real 0V analog signal?

Status
Not open for further replies.
P

pavjayt

Member
Hello,

I am trying to make an analog signal from 0V-1V. By that I mean, I want to have a real 0V without noise in it. I can control the offset by using a offamp buffer, but I would like to control the noise output. I was able to reduce it a bit by using a capacitor to the gnd at the input signal, but not as much as I would like to reduce, currently the max and min noise is around 9mV to -11mV. What is an acceptable noise for a 0V signal?

I heard that the easiest way is to have a driver with double control (TTL + Analog). With the TTL signal, we can create a real 0 (whatever is your control signal).

Can someone enlighten me with a test schematic for this kind of driver?

thanks
 
what are you trying to do? what you are trying to accomplish will determine how much noise is acceptable.... and "-11mV" is not a correct way of characterizing noise signals, since noise is an AC signal. do you have 11mV of noise, or -11mV DC when you are expecting 0V?

also, show us a schematic please.... troubleshooting an unknown mystery circuit is like trying to target shoot in the dark.
 
Ok, I guess I need to be more clear. The situation is that we use an analog signal that varies from 0-1V to drive a modulator that controls a crystal. The modulator is so high in response that even a 10mV change in the input analog signal drives the crystal to a different position.

So the crystal can be OFF, ON(max) or anywhere in between depending on the drive voltage signal. Now, in order for the crystal to be really OFF, the signal must be a real ZERO with no fluctuations. So, I contacted the manufacturer about how they are producing a real ZERO signal and he replied me that they use a TTL+Analog control to produce real ZERO to measure its performance with respect to MAX (1V). So, the 0V not being a true 0 is having an effect on our results.

I will come up with a better way to explain this.
 
Hey, I was burnt by that too. If this is for calibration purposes only, you could use a relay for the 0V. Of course there are solid state approaches. You would still be dealing with thermal EMF's and Seeback EM's, but they are a LOT smaller.

Now, you might be able to use a D/A and a real zero to offset and zero.
 

Hello, this is not only for calibration purposes, if there is any leak more than the permissible 'ZERO' leak, it shifts the results. So, whenever we want to send 0V we want that to be a real ZERO. Could you please elaborate on what you suggested. Sounds interesting to try it out.

thanks
 
I like the D/A approach by KISS. coupled with a precision, low-noise opamp, you could get as close to zero as possible.
Of course, the opamp would have to be fed from a bipolar supply. Even rail to rail single supply opamps can only swing to about 50 to 100 mv of zero...and from what I read from your post, that is too much.
 

Sorry, I am not following D/A approach. The analog signal that we feed into the driver itself is derived from a D/A on a FPGA board which is then fed into a gain op-amp (which is driven at +/-5V Vss) to bump the gain by a factor of 2 and then is fed into the driver. I like the idea of using a relay suggested by KISS, but not exactly sure how I can implement that and how fast that switching would be. We need atleast 50ns response time.
 
How about then using an analog switch to ground (to a clean ground) the signal?
There are devices which have "on" resistances of 1 to 2 ohms.
 
Well, that changes the shape of things by a lot. I think you might have to create a real zero and then offset your final driver. Vos is temperature dependent.
Once the calibration is performed, then you don't need the real zero anymore.

50 nS is going to be really tough. The makers idea of using a TTL like strobe to output zero makes a lot of sense.
 

I am thinking of a circuit where I can switch the input signal to the driver so that when say from 0.1V and below it will replicate something like unplugging the cable to the driver while it passes through any voltage above 0.1V. Is it even possible or am I thinking about some magic here?

How can one use a TTL strobe to output zero?

thanks
 
ronv mentioned one way for strobing.

Look at the voltage/contrast application here: https://www.electro-tech-online.com/custompdfs/2013/04/digipots_jan09_slides__final1.pdf OP amp offset adjustments can be made with these devices as well as with D/A converters. So, as part of your self-test calibration sequence, you null out the offset.

Another reminder is to pay attention to quantization errors in your A/D or D/A arrangement. An easy was to illustrate this is pretend you had an A/D with 100 counts. If your measuring 1 bit, the error is +-100%; As you measure 99 counts the error is +-1%. So, a 1-5 V can be seemingly better than a 0-5V output. Your D/A can output 1-5V, but your final control element could subtract 1V from it to get a 0-4V signal. That 0-4V signal would be "better" than the D/A 0-5 V signal.

If you can control, measure and change that offset, you might be there.

My problem was, I needed a zero volt in to get the offset voltage. I had a system meter to actually measure the output. The system was "good enough" without the compensation, so it was left out. Without compensation, my output was like 50 pA larger on the most sensitive scale. AC measurements, were dead on and that's what we cared about.
 
Maybe you could post your schematic indicating where you need the zero volts?
 
ronv said:
Maybe you could post your schematic indicating where you need the zero volts?
Click to expand...

Here is the schematic of the system where I am trying to get a true zero input. I can replicate true zero by measuring the power at laser out when the Mod-In signal to the driver is unplugged, which is not ideal.
 

Attachments

  • Schematic.pdf
    12.7 KB · Views: 155
Raising the Control gate to 5v forces the output to ground. I chose a ZVN mosfet but I am sure there are better.

 
WTP Pepper said:
Raising the Control gate to 5v forces the output to ground. I chose a ZVN mosfet but I am sure there are better.

View attachment 73240
Click to expand...

I guess the In and Out in your schematic is the analog signal in my schematic that's connected to the driver. Not sure about the Control input in your schematic. Could you please explain about that? I am willing to try anything at this point to get this issue resolved.

thanks
 
In audio amps, it;s customary to use a FET that will short the imput to ground for the power-on mute system. as long as your circuit can stomach a short on the line, then that's a reasonable way to get zero.

A FET to ground at the input you want to be zero.

Another option is an analog switch or MOS relay.
 
I would be tempted to auto zero the offset through the op amp. Every time the DAC is loaded with zero, sample the output of the op amp and feed it back as a compensating voltage to provide a true zero. Would require a little thought to do properly though.
 
Thought 2: You could sample the output of the op amp at 0 on the DAC, invert it and add it to the output of the op amp. I am assuming that you are compensating for offsets rather than broadband noise. The compensating op amp could be low offset, and lower bandwidth, assuming again that it is offsets that you are trying to compensate for.
 
Status
Not open for further replies.

Similar threads

O
  • Question
Reading Digital + Analog Data from Sensors / Bus
2
Replies
26
Views
3K
Tony Stewart
P
How to change a omron E5CN 220vac to a E5CN 24vdc
Replies
10
Views
1K
patrick-Belgium
P
P
  • Question
Linear analog filter with zero amplitude in stopband
Replies
14
Views
2K
Tony Stewart
O
  • Question
Connecting ADC directly to MCU
Replies
5
Views
1K
rjenkinsgb
N
  • Question
Make signal lights for the AC outlet
Replies
12
Views
2K
ZipZapOuch
Menu
Log in
Register
Cookies are required to use this site. You must accept them to continue using the site. Learn more…