Re: Piezoelectric Stage for nanometer level precision
ProfHankD wrote:
mawyatt2002 wrote:
ProfHankD wrote:
mawyatt2002 wrote:
Awhile back the surplus market made some Physik Instrumente piezoelectric devices available. The devices offer the potential for extreme resolution and high precision when operated "Closed Loop" with position feedback. The complete OEM Closed Loop controllers/drivers/amps are very expensive, so an effort began to develop a custom solution at a much lower cost. One solution is a full Custom Complete Closed Loop system adaptable to various piezoelectric stages (P601K, P603K for example), another was a Custom Controller to interface with the OEM P603K Driver/Amp which are available on eBay at reasonable costs. Both operate from a Raspberry Pi with a single +12V at <1/2 amp supply.
Pretty slick....
If I were doing this, I think I'd just wildly gear-down a stepper... perhaps to drive a movement extracted from a cheap micrometer (e.g., like this ). Less driver circuitry needed, but a bunch of 3D-printed parts for mounting.
Thanks,
Gears have all sorts of issues with backlash, moving surfaces, mechanical noise and so on, and suspect even quality micrometers can't achieve these levels.
The piezo ceramic elements expand or contract at the molecular level based upon electric field, so there is no mechanical surface to surface movement. The cleverly engineered "flexures" in the Piezo Stages are like bending gears without any noise, hysteresis or other effects that plaque conventional bearings, however they have limited movement range. The piezo elements are held in position at the ends by natural compression and don't move during general excitation. So in normal operation these Piezo Stages have no surface to surface mechanical movement.
I'm familiar with piezo motors/actuators. They're actually commonly used in lens focus systems. There are lots of different types of linear actuators out there. All good stuff. I just think they're a bit beyond the usual do-it-yourselfer level....
Yes agree, this is more like DIY on steroids. Needed something to keep me entertained since retiring earlier this year from an active career producing 30 issued patents (couple still pending), Chief Scientist/Engineer, Adjunct Prof, IC designer, and this certainly kept me busy
We've used the superb industrial THK KR26, KR20 and KR15 linear rails with precision 400 step motors for some time now. Even with the highly advanced custom controller developed (beyond the available commercial controllers), based upon the Trinamic devices, these piezo stages are at another level. Trying to figure out to measure these nanometer levels is another whole topic of it's own, especially considering without a big budget and all out-of-pocket for the entire project, that's beens solved also
You can see the THK KR20 performing with the custom stepper controller at 800X here at Mike's Labs, note the precision and smooth return to position after the 12000 um excursions!
https://drive.google.com/file/d/19wE-B-dMesEXNtEZ552h_VnpWOjA0n0B/view?usp=sharing
Now the effort moves to the software aspect, which is our weak link. We've included dual 12 Bit DACs arranged to create an effective 20 Bit DAC, also available is a 16 Bit core DAC which can also be augmented to 20 bits. These DAC's are used for position commands, a 16 bit differential ADC with 4 stage PGA is used as position readout (also supports a 18 Bit Differential ADC with PGA). Bidirectional I2C isolation between the RPi and Controller is also utilized since the analog levels involved are so tiny, anything could easily corrupt things. The feedback circuitry is so sensitive that a toothpick placed on the stage can be sensed, even tho the full scale output from the Strain Gauge resistive sensors is only ~1mv full scale. Having lots of experience with ultra-precision and ultra-low noise systems played an important role in achieving these levels without the usual extremely expensive precision components.
I have the utmost respect for the Physik Instrumente designers with their OEM Driver/Amplifers after proceeding thru this myself (mostly without knowledge of the PI devices since they are all proprietary (the "K" means "K"ustom) and PI would not provide any info. What's interesting about this is that the two approaches to the electronic driver/amps are completely different, the OEM approach is a classic complementary class A/B "audio" type amp utilizing power MOS devices and creating a complex compensation network to stabilize this with the highly capacitive (>5uF) piezo load. The strain gauge sensing is accomplished with an expensive hybrid Instrumentation Amp, followed by a host of other active filters and amplifiers. Very sophisticated and well done design from what I could gather looking at the P603K OEM Driver/Amp which became available on eBay after the initial custom closed loop design was completed.
Our approach was completely different tho, the driver amp is based upon a complementary transconductance stage with high output impedance, this topology was selected because it's more tolerant of the high capacitive piezo stage load. Bipolar rather than MOS transistors were selected because of the higher output impedance and transconductance. The feedback design centered around utilizing a semi-Instrumentation Differential Amp since the Strain Gauge impedance is low (1K), which was implemented with 1 precision op amp rather than 3. The entire feedback chain was absorbed into the single differential amplifier approach since speed for the macro intended use is slow. Still this didn't alleviate the 1mv peak signal level sitting on a 2.5000V bias, where AC coupling can't be utilized to remove the DC bias because continuous DC feedback is required. So this feedback differential instrumentation amp required very high common mode rejection, high differential gain, and low noise with good stability. Using the concept that high volume, batch processed, like environment, electronic components usually match well (basis for most all IC designs), standard low TC 1% resistors from adjacent tape & reel positions were used with very limited trim. This avoided the costly ultra-precision "matched" resistors in normal Instrumentation amps, and works beautifully!!
The driver amp is also not a conventional class A/B type, in that it operates as class B during transitions and returns to class A at steady state but not with both output devices operating class A, only 1 device operates class A in steady state. Because the load is almost purely capacitive, no output current is required in steady state and only the PNP is active with bias.
Both Piezo Driver/Amp approaches work very well, and as of now I'm not sure which is best. The Full Custom Controller/Driver/Amp is more flexible than the OEM Driver/Amp which are specifically tailored for each type stage. We had to add an strain gauge offset correction adapter to get the P603K Driver/Amp to work with the P601K Stages since they had more offset than the P603K Stages, the Custom Controller/Driver/Amp has offset correction built in, so no issues working with either the P601K or P603K Stages.
Anyway the macro community now has a low cost Custom Open Loop, and Two Closed Loop Piezo Controllers to select from, and all work very well indeed
Best,
-- hide signature --
Research is like a treasure hunt, you don't know where to look or what you'll find!
~Mike
