Help driving GSI 120DT Galvo motors

[thorson14]

Hi everybody,

I'm in a bit of a bind trying to find a proper driver board for some GSI G120DT galvanometer motors I recently inherited. My PI is wanting to drive them with some ThorLabs driver boards he had lying around but this has proved very difficult (and I'm not sure it's even possible-any advice would be appreciated for this route).

What type of GSI driver board would work for my G120DT motors? All I have been able to locate available is through Ebay and surplus sites. There are various listings but there is not pinout documentation included so I'm not sure whether they would work or not.

Listings I am currently interested in:
http://www.ebay.com/itm/GSI-Lumonics...-/221817172381
http://www.ebay.com/itm/KLA-Tencor-G...nKQUZBE09aJOCw
http://www.ebay.com/itm/like/1315886...=US&rmvSB=true

Any information or advice would be incredibly appreciated.

Cheers!

[mixedgas]

120DT is no longer supported by nearly anybody, except me..

I'll have some docs for you in the morning. I have a few amps, like ten or so, laying around.

I work for a university, I build instruments for a living, I know how it goes...

Nearly every other galvo on the planet now uses optical position sensing for the most part, and those are capacitive,

Therefore there is a structural difference in the amplifiers...

It helps if I know what your trying to do with the galvos.
Nothing you linked talks to a 120.

Steve

Hi everybody,

I'm in a bit of a bind trying to find a proper driver board for some GSI G120DT galvanometer motors I recently inherited. My PI is wanting to drive them with some ThorLabs driver boards he had lying around but this has proved very difficult (and I'm not sure it's even possible-any advice would be appreciated for this route).

What type of GSI driver board would work for my G120DT motors? All I have been able to locate available is through Ebay and surplus sites. There are various listings but there is not pinout documentation included so I'm not sure whether they would work or not.

Listings I am currently interested in:
http://www.ebay.com/itm/GSI-Lumonics...-/221817172381
http://www.ebay.com/itm/KLA-Tencor-G...nKQUZBE09aJOCw
http://www.ebay.com/itm/like/1315886...=US&rmvSB=true

Any information or advice would be incredibly appreciated.

Cheers!

[mixedgas]

PS,
There are some General Scanning CX-660 on Ebay... Those drive 120s, but they also drive 330DTs, so there can be soldering involved to change configuration components or repair them. The boards are identical, but about 10 cheap components get changed to configure them. Do not rush out to buy them just yet. The nice thing about 660, they come with the power supplies and a internal ramp generator.
!
You Have other options. Someone will now plunge in to tell you how obsolete G120s are, and to buy a cheap 200$ Chinese galvo set. Beware, because many Chinese galvo sets do not have the linearity for lab applications. Nor do they all have the high grade bearings for tasks like Confocal Microscopy... One should not always seek advise from those who do not understand PIs... In a way they are correct, but you probably need something like Cambridge Technology or Pangolin's ScannerMax.
!
Steve

[thorson14]

Hi Steve,
Thank you for your great information. I look forward to information on the amps when you get the chance. I will take a look at the General Scanning CX-660s.

Some context for my this project:
My PI inherited an old Nikon PCM2000 laser scanning confocal microscope. Last it was used, it worked well. The problem is this machine and its various components were moved around the campus a number of times, passing through various hands, etc. Somewhere along the way, the original computer that was used to interface with the laser scanning confocal system was misplaced (and my guess is it was salvaged many years ago). I have hit dead ends at every turn trying to locate this machine. I have been in contact with Aventeck (they own the rights to service these Nikon systems as Nikon stopped supporting them many years ago) and they say to run the original controller (housing the drivers for the G120DT galvos along with the PMTs) I MUST have the original computer; there is a special-made board built into these Nikon manufactured computers. There was an option to have a board custom built, but that got quite pricey.

Fast-forward a few months and I am now trying to build a custom laser scanning confocal system with some of the components from the original build. The last (MAJOR) hurdle is driving the G120DT galvo motors. I was given a set of ThorLabs driver boards for their GVS102 galvo motors. I was weary of their compatibility, but my PI pushed to keep on trying them. All I get is a quick short in power every ~10s. My guess has always been that position sensing is not working properly and that the mechanism of sensing between components is incompatible. Your post confirms my suspicions. So now I am looking for compatible drivers for these G120DTs. thanks for letting me know that the items I was looking at on Ebay are not compatible.

One option I have pitched is simply to buy a new set of galvos, but beyond the investment there is the problem of disassembly of the scanning head, something my PI is not fond of.

It should be noted that I have the original driver board from the Nikon control tower used to interface with these motors. I have no pin documentation and have little experience trying to reverse engineer something like this. I have attached a photo of the driver board from the control tower for your reference.


I look forward to your information in the morning.

Thank you so much,
Todd

[VJ AIWAZ]

Steve did i give you those 660's or the other GS rack i cant remember or are they still here

[mixedgas]

AIWAZ, They are in Silicon Heaven. The corrosion from the salt water was too deep to allow for repairs.

Steve

[mixedgas]

Taming PIs takes time, there are various methods involving various medications, waterboarding, whips, scourges, electroshock treatments, maces, slapping with wet trout, and telling the truth. I prefer the later. You will have to tell your PI that a subtle difference in sensor technology just wasted a month. Not your fault, most of what you needed to know is "Pre-Internet".

!
Please excuse the exclamation marks, this forum software version does not like my web browser.
!
I LOOKED AT YOUR PICTURES!
!
Everything you need is on the original Galvo Control board, and the servo loops are already tuned for the mirrors on the galvo shafts. See that line of four potentiometers labeled integrator gain, damping, sensor offset, servo gain.... That is the tuning for the closed loop position sensor and proportional, integral, derivative loop, for one galvo. These amps need tuned and matched to each galvo-mirror combination, and further tuned to one of three tuning schemes to shape the galvo's frequency response.

!
Tuning is not easy if you've never done it before, and the Lasershow tuning methods (We use a test pattern) all over the net, do not apply to you. Tuning is done by applying various square waves and ramps for confocal use. While watching the position and velocity test points using an oscilloscope, Then the velocities of the X and Y galvos must be matched by "detuning" one galvo in a Confocal or LSIM system.
!
If you use the factory amp, you do not know the volts per degree setting for the position sensor. Tuning would therefore be "interesting" and have to be done very conservatively. It is therefor best to just assume the factory amp has the correct tuning and use it if possible. If not, buy CX-660s and go through the GSI tuning procedure, which I have someplace.
!
NOW, THE BURNING QUESTION IS, ARE THE GALVOS OK?
!
If you attempted to use the position sensor feedback:
!
The Thorlabs cards are based off a Cambridge Technologies design. They are designed for a photodiode based position sensor in the galvo. Those photodiodes run in photovoltaic mode, they produce a pair of differential voltages. The G120DT has a position sensor that outputs two DC to 1o KHz or so differential currents superimposed on a few volts of 2-3 Mhz RF common mode voltage. Without the needed RF low pass filters with a controlled group velocity characteristic on the amp boards, you NEVER had a chance with the Thorlabs boards. They were overwhelmed with the RF on the front end.
!
Hopefully the overcurrent shutdown prevented major damage to the galvos. If the coils exceeded 1.25 amps for any length of time the internal magnets are demagnetized. It goes without saying, do NOT try that again.
!
How exactly, and I mean EXACTLY, did you connect the Black, Orange, White, Red, Violet and Brown wires from the G120DT feedback cable to the Thorlabs Amp?
!
The Cambridge style AGC circuit in the feedback loop will apply 15V at as much current as possible to cause the LEDs to light up. It has a way of sensing this, using a current source and feedback from the position detection circuit.
!
The GSI AGC and POSITION SENSOR RF OSCILLATOR needs NOT MORE THEN 12V at a few mA, which also worries me if you attempted to connect the feedback..
!
There should be a board near the microscope scan head that merges the signals from the two galvos, probably filters them to eliminate the RF, and sends it down one cable to the existing AMP board. Do you still have it and the cables?
!
U14 looks like it is a DAC chip on the factory driver board. That is where you would inject your control signals to position the galvos. At the DAC. The digital portion of the board can be disconnected. It grabs position data off a bus, and there look to be some timing circuits there as well, specifically the FPGA.
!
You hooked up an amplifier that is three times faster then the G120s can handle. It is also capable of sourcing perhaps twice the required current.
!
It will take work, but you need to test the G120s prior to using them again. Do you have access to an oscilloscope?
!
Do the G120s have removable mirror mounts?
!

You have a limited choice, send the factory board and galvos to a specialist like me, buy the used CX-660s and tune them, or try to reverse engineer the cables and pinouts of the existing factory board. In any case, you need some "Galvo Professional" level help.
!
I'm assuming your going to try the Confocal Freeware that is out there, and Labview for the controller hardware?
!
Document scans coming in time, starting today.
Steve

[mixedgas]

PINOUTS:

CLASSICAL G120 DB15M connector

!
!

Pin Number Wire Colour Signal Description :
!
1 Red Drive Coil (See notes 1 & 2)
2 Yellow Drive Coil (See notes 1 & 2)
3 Black Position output (-)
4 Orange Position Output common
5 Red Heater - T-Type (See note 3)
6 N/C No Connection
7 Red Heater - T-Type (See note 3)
8 White Oscillator supply voltage (+)
9 Green Drive Coil (See notes 1 & 2)
10 Black Drive Coil (See notes 1 & 2)
11 Red Position Output (+)
12 Violet AGC Output
13 Brown Oscillator supply voltage common
14 N/C No Connection
15 Blue Thermistor - T type (See note 3)
!
NOTES

1/ To wire drive coils in series, connect Black to Green and drive Red to Yellow. Series wiring is recommended for use in position sensing (closed loop) mode.

2/ To wire drive coils in parallel, connect Red to Green and Black to Yellow. Use parallel wiring only when driving the galvo in open-loop mode.

3/ Heater blankets are mounted only on the G120DT type galvos and are not commonly used in light show applications.

!
PLEASE SEE ATTACHED... "GSI CARE AND FEEDING" in two parts.... My name for it, not GSI's.

Steve

[mixedgas]

If you have a Kindle..

https://www.amazon.com/LASER-SCANNER..._bc#nav-subnav

Hardcopy is there, too...

Doesn't really cover G120s, but the rest of the content is worth it, especially concerning amplifiers.

Steve

[thorson14]

Thanks, Steve. You have so much great information on this topic. Over the past 24 hours you have helped me more than the last 4 months combined. I'm not sure that is an indictment on myself, or just a reflection that I was looking in all the wrong places (or like you said, the info I need is "pre-internet").

I do have a Kindle and Amazon Prime so I will be downloading the Laser Scanners resource when I get home tonight.

It looks using the factory board is going to be the most practical path at this point. My knowledge (and time) is limited with tuning of the CX-660, but it's reassuring to know that that option also exists.

First, I connected the ThorLabs board to the G120s per the diagram shown in the picture attached below. I had previously found the G120 pinouts through this forum (I'm pretty sure you were the one who linked the laser light show database but I might be mistaken there). Tech support at ThorLabs suggested the connection of the position sensor power to the oscillator supply voltage (white).


I sent very small voltages at very low frequencies with this setup and once I knew it wasn't working (~10 seconds), I stopped the test, disconnected the power and haven't tested that since. I have even taken the ThorLabs boards back to another lab since that is NOT the right route. Note that this connection was from the wires directly connected to each motor, not the DB15M connector. I DO still have the board that merges the two motors' signals and that gets sent out via a single DB15M connection [pictured below]. For my testing, I had disconnected this board and matched the correct pins via the Laser Light show reference guide (the colors were good enough to match to the right pins, but I wanted to double check the pins and everything).


I should have access to an oscilloscope; if I'm not mistaken, the MyDAQ from National Instruments can run as a simple oscilloscope. I have been sending inputs from a MyDAQ through LabView. There is a Laser Fluorescent Dynamics center on my campus that I have been in communication with and they have some really nice software built for custom-made confocal systems. I think we will be able to connect everything through an Iotech personal DAQ.

The galvanometer motors are equipped with scanning mirrors that look to be removable. [picture]



From my standpoint my brief checklist of to-dos/options:
1. Check the integrity of the galvo motors with an oscilloscope
2. Connect to the factory driver board (with help)
-reverse engineer connections here. I'm willing to try this with guidance.
-send factory board to a professional (you). Do you have any estimations on turnaround time and costs if I sent this to you by the end of the week?
3. Buy and correctly tune CX-660 board (I want to try to avoid the tuning due to my limited experience)


I will relay information and options to my PI today. A new confocal setup on some of our other brand new equipment is pending this scanner working, so these G120s are very high priority.

Thanks again for your immense amount of help.
Todd