Insane Hydraulics
Bold and audacious blog about fluid power
Today I will be looking into Wika A-10 pressure transmitters - the model that I've used for many years. I'll show you what these transmitters carry inside, we'll talk about how they work, and I will show you another pressure transmitter model, that will probably take the place of the A-10s in my future projects.
So... Industrial pressure transmitters. Normal people use them in hydraulic systems to read and communicate pressure, the most "audacious" techs stick them into DIY gadgets for wireless diagnostics.
I've used many brands and models over the years and ended up using the Wika A-10s the most - simply because they seemed to have the best quality/price ratio for most of our applications. Unfortunately, during the last couple of years, the A-10s started to fail quite often. To the point that our clients started to look for alternatives.
That's how business works, right? When something starts to suffer from recurring failures, you move to the next competitor - nothing strange about it. But since I like to open everything (especially after it failed) - I decided to give the broken transmitters a more thorough inspection and asked our clients not to bin the failed sensors, but rather give them to me for "autopsy" purposes.
After a while, I gathered 2 x 25bar 4...20 mA models, and 8 x 400 bar 4..20ma. 10 in total, a nice round number to begin playing. So, I opened them all up, and peeked inside:
First - the 25 bar model.
As you can see, the construction is very simple. There is a stainless steel membrane in the middle, with the thin film sensing elements forming the Wheatstone bridge legs visible under the clear coating. Unfortunately, I don't have a microscope to zoom into the pattern (a good reason to get me one, by the way). Then you have the tiniest wires connecting the sensing elements to the PCB pads, and a Wika-labelled non-leaded chip.
Both of the sensors that I got had parts of the measuring bridge with infinite resistance - the obvious reason why they wouldn't work. I tried wiggling the tiny wires to see if any of them were loose - but they all seemed to be well connected, which means it's a thin film failure, so there's nothing I can do here. I pulled one of the boards out and here's how it looks from the other side (the current driving NPN transistor and some passives):
I must say that I hate it when OEMs put their brand names on chips. That's not cool, Wika! However, if you google for a bridge sensor interface IC in a 36-pin VQFN package, you quickly come across the likes of TI PG300, PG305, etc.. and without going into much detail I'd say that a lot of things "add up" here (at least the pins 17 and 18 come from the bridge output). I should look into it when I have time. Maybe I can use it for something else. In any case - the data sheet is quite educational. It's impressive what these tiny chips can do - provide a complete front end for the resistive bridge, incorporate the 24-bit ADC and a 14-bit output DAC, and digitally compensate for the offset, non-linearity, and temperature! With all this progress, why don't people live on the moon yet?!!
Anyhow - to the 400 bar models now:
I was surprised to find that all 8 employ a different chip - a ZSC31050FIG1. Once again - the data sheet is super informative. No "fancy labeling" this time. Here's what I found:
One sensor had the bridge "blown up" - infinite resistance across all leads, and the transparent glue over the bridge elements even changed color. Not sure what happened. Another sensor had one of the tiniest wires disconnected from its pad. I could actually make it work when I pressed it onto the pad with a small screwdriver. I am guessing this may be an assembly issue. These wires are very easy to damage.
The other six transmitters had all the wires connected, and the bridges measured about 6.1-6.5 kOhm across one diagonal, and 5.5-6 kOhm across the other. Two of them got "permanently stuck" outputting about 3.4-3.5 mA, and the other four worked in the sense that their current output changed by precisely 16 mA from 0 to 400 bar, but their zero point was way off the 4.000 mA ( I measured 3.540, 3.379, 3.689 and 6.051 mA respectively). Once again - I am not sure why this happened. I guess this could be solved with re-calibration. Another thing to look into when I have time.
And here's the contender for becoming my new go-to pressure sensing solution - the IFM small type PT pressure transducer, in this particular case the PT5000, which is a 400bar sensor with 4..20mA analog output.
I tested it thoroughly. By the way, I forgot to show you my test setup (Pressure Maker 2 at work!):
The accuracy of the sensor is better than its specs, across all the range! Of course, this can (and most likely will) change with time, but then again any future offset can be dealt with. The resolution is impressive as well. I was taking the current readings with my Brymen BM869s multi-meter, which measures 4...20 mA lops with 0.001 mA accuracy, and I can also measure the lower current range with 0.0001 mA accuracy - and I was pleasantly surprised to discover that the 400 bar sensor can reliably sense pressure that I can create with my mouth (about 0.1 bar on my best day). I could even modulate it and see the current change. This means that the DAC in the chip the IFM used has at least a 14-bit resolution, which is great (especially for those who use such pressure sensors for diagnostics).
We'll be putting these bad boys in service soon and see how they fare long-term. They are actually even cheaper than the A-10s, so hopefully, this one will be a win for all. (Except for Wika...)
Here's what I learned: