1
\$\begingroup\$

I'm try to build a cheap pressure transducer to use with my cheap Hantek 1008c scope for automotive measurements. When I started taking measurements and testing it, I saw a 40mv peak to peak, approximately 1kHz sine wave riding on the output signal. I think this is just noise from a cheap transducer, but I got curious as to how I might eliminate this noise.

I'm using this 30psi transducer ( and a 200 psi one ), plus a Drok 2203 buck converter for power. The transducer has three wires, power, ground and signal. It expects 5v, and linearly outputs from 0.5v at 0 psi up to 4.5v at 30 psi. It's ground, and that of the BNC connector which goes to the scope are both hooked up to the Drok's ground.

I've already checked the power supply and it's giving a clean 5v. I decided to try shunting the AC component to ground using a capacitor between the signal wire and 0v ref, and after experimenting found that a 0.1uF mylar cap worked best, although "best" only reduced the amplitude of the sine wave noise by about 40%.

So I'm curious how one could completely eliminate noise like this?

Here are a before and after. I tried inserting images, but couldn't so just linked to a 50 sec vid.

EDIT

Here is what it looks like at 5us, without the cap

enter image description here

and with the cap:

enter image description here

I guess there was some kind of aliasing going on at the higher timebases.

So is this 400 kHz sine wave just a defect in the transducer as I suspect, and is there anything I can do to get rid of it completely without affecting the rest of the signal?

EDIT 27/04/2018

First off, the 400 kHz 35 mv p-p noise was indeed coming from the buck converter. But, I also found that the transducer itself was producing it's own noise. I replaced the buck converter with a 4.5 volt battery and saw the following:

enter image description here

I got the same result when I inserted the RC LPF right on the output of the buck converter. The best LPF results were with a 47 Ohm resistor and 1uF cap. It reduced the amplitude of the buck converter noise to only 3mv, without noticeably reducing the signal voltage, which is completely acceptable.

I may use the 4.5 volt battery, or the buck converter with the LPF, either way I'll need to get a better quality pressure transducer.

\$\endgroup\$
4
  • \$\begingroup\$ Add some resistance in series with the signal wire and put the capacitor on the scope side after the resistor. You don't say what settling time or accuracy you need -- you're only limited by the voltage divider you're creating with the 1MΩ scope input. \$\endgroup\$
    – Daniel
    Commented Apr 26, 2018 at 9:13
  • \$\begingroup\$ @Daniel Won't adding resistance in series drop the voltage on the signal wire and reduce the accuracy of the signal? It's supposed to be 0.0075 psi / mv for the 30 psi transducer with an accuracy of within 1.5%. These transducers typically are supposed to have a response time of about 1ms. I only have a very basic understanding of electronics from an automotive technicians pov. \$\endgroup\$
    – Robert S. Barnes
    Commented Apr 26, 2018 at 11:47
  • 1
    \$\begingroup\$ Humor me here, and try setting your scope to a much faster sweep (microseconds or tens of microseconds) and see if anything interesting shows up. \$\endgroup\$
    – Spehro 'speff' Pefhany
    Commented Apr 26, 2018 at 15:07
  • \$\begingroup\$ @SpehroPefhany indeed, something interesting did show up. At 5us it shows up as a 400 kHz sine wave; see my pictures I added. \$\endgroup\$
    – Robert S. Barnes
    Commented Apr 26, 2018 at 17:09

1 Answer 1

Reset to default
1
\$\begingroup\$

Seems like you are seeing noise from the switching power supply, since 400kHz is a plausible switching frequency. As well as bypassing the output you can try connecting capacitors across the input to output.

A simple RC low-pass filter will get rid of most of it at your scope without affecting the millisecond-level rise time of the sensor appreciably.

If you are curious as to why (real) 400kHz noise would show up as a (not real) 1kHz sine wave, it is because of a phenomenon known as aliasing.

enter image description here

\$\endgroup\$
7
  • \$\begingroup\$ But if I disconnect the transducer, I get a perfectly flat 5 volts on the power wire. I only see the sine wave with the transducer plugged in. When you say "across the input to output" do you mean between the 5v ref and ground? Or between the 5v rev and signal? \$\endgroup\$
    – Robert S. Barnes
    Commented Apr 26, 2018 at 17:24
  • \$\begingroup\$ Between the power input to the switching supply and the output. There may be some grounding things going on between your scope and the power. I am sure that if you try a battery rather than the low-cost switching supply you will not see this issue. \$\endgroup\$
    – Spehro 'speff' Pefhany
    Commented Apr 26, 2018 at 17:44
  • \$\begingroup\$ Sorry that I'm asking so many questions, but when you say output, do you mean the output of the switching supply, do you mean the signal output wire coming out of the transducer, or to ground? From the RC LPF tutorials I'm seeing, it seems like I would either put the cap across the power supply output and ground, or across the signal wire and ground ( with the resistor in series before the cap at one of those places ). \$\endgroup\$
    – Robert S. Barnes
    Commented Apr 26, 2018 at 18:37
  • \$\begingroup\$ One other question, won't the resistor drop the voltage significantly? \$\endgroup\$
    – Robert S. Barnes
    Commented Apr 26, 2018 at 18:51
  • 1
    \$\begingroup\$ Not sure what filter configuration would be best without knowing a lot more about that DC-DC converter. RC filter will not change the DC voltage if your scope is high-Z input. For example a BNC scope input may be 1M\$\Omega\$ || 20pF, so 10K will have little effect (1%-ish). A x10 probe will be more like 10M so even less. \$\endgroup\$
    – Spehro 'speff' Pefhany
    Commented Apr 26, 2018 at 18:56

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.