Millisecond precise imaging

Discussion in 'STT Series (obsolete)' started by golgot78, Sep 19, 2016.

  1. golgot78

    golgot78 Cyanogen Customer

    Joined:
    Sep 19, 2016
    Messages:
    25
    Hello Doug and Adam,

    It’s works but…
    We then put the camera on a test bench (cf. description in attached powerpoint file) to check reproducibility of the accuracy and measure the trigger bias, it shows several problems:

    1. The triggering command sent to the camera is sometime ignored by the camera.
    2. The variation from one trigger delay to another is fluctuating a lot (of up to 60 ms). This is far from our +/- 5 ms need.

    Could you please consider this attached problem description. I think that we should schedule a teleconference meeting early next year with engineering teams to try to find a solution to our problem.

    In the meantime I which you all a very good Christmastime!

    Gautier
     

    Attached Files:

  2. Adam Robichaud

    Adam Robichaud Staff Member

    Joined:
    Sep 29, 2014
    Messages:
    764
    Location:
    Ottawa
    Our hardware engineer was out of office today, but he should be in tomorrow. I have a suspicion that this is largely related to the STF-3200's shutter actuation, but I need to confirm before making any promises.
     
  3. Doug

    Doug Staff Member

    Joined:
    Sep 25, 2014
    Messages:
    6,575
    Upon receiving the trigger command, the camera performs a sensor flush operation and then starts the exposure. This is the reason for the extra delay.

    When the exposure actually starts, the open collector Trigger Out signal is pulled low. If you can use that to resolve your timing issue, then that would be the simplest way forward.

    Otherwise we would have to look at customizing the firmware for your application.
     
  4. golgot78

    golgot78 Cyanogen Customer

    Joined:
    Sep 19, 2016
    Messages:
    25
    Hello Doug and Adam,

    We were aware of the extra delay you mentioned and we considered it. Beyond this bias that we could correct, the hard point for us is the important jitter described in our report.
    I don't really understand, in spite of your explanations, why the "Trigger In" way doesn't not give a very accurate and reproducible timing. Also I don't understand why the flush operation is for, just before an exposure. As Adam, I more suspect the shutter disc mechanism management.

    Anyway, as you suggested we will now focus on the "Trigger Out" option, hoping it will accurately tell us the actual open time.

    We'll modify consequently our setup and do some measurements for that early January and let you know.

    See you in 2017!

    Best Regards,

    Gautier
     
  5. golgot78

    golgot78 Cyanogen Customer

    Joined:
    Sep 19, 2016
    Messages:
    25
    Hello Doug and Adam,
    First of all, I wish you a happy new year!
    We managed a couple a days ago to determine the latency between the exposure command and the "real" exposure date by spying the I2C pin 7 (Open Collector Genl Purp Out). There is for us a 60ms bias between the switching state on the I2C output and the real exposure start (when the matrix starts integrating light).
    So a priori no need for you to adapt the firmware.
    Another question concerning the camera caracteritics: what is the limit external temperature the camera is able to bear without any damages ? Indeed, we plan an observation session in the coming days and external temperature by night is about -15°C to -10°C (5°F to 14°F)... We didn't finf any info on that topic in the camera user manual.
    Thanks in advance for your answer.
    Best regards,
    Gautier
     
  6. Doug

    Doug Staff Member

    Joined:
    Sep 25, 2014
    Messages:
    6,575
    Not to worry, the cameras are designed to operate to at least -30C ambient. In fact we have customers who operate them at -40C.
     
  7. golgot78

    golgot78 Cyanogen Customer

    Joined:
    Sep 19, 2016
    Messages:
    25
    Hello Adam,

    We are still working on it but we have the following observations to report:
    There is no output signal at all from PIN2 (Open Collector Trigger Out) which was the one we supposed to be working with Trigger Out option.
    In fact the one working with Trigger Out option is PIN7 "Open Collector Genl. Purp. Out". But its voltage changes from 0.68 V (when shutter is closed) to 0.01 V (when shutter is open).
    We expected it to range from 3.3 V to around 0.0 V.
    Could you please confirm the range we observed is nominal or our camera has something wrong?

    Thanks in advance,

    Gautier
     
  8. Colin Haig

    Colin Haig Staff Member

    Joined:
    Oct 27, 2014
    Messages:
    2,960
    Gautier, did you add an external pull-up resistor on the output?
     
  9. Doug

    Doug Staff Member

    Joined:
    Sep 25, 2014
    Messages:
    6,575
    Colin put his finger on it. These are open collector outputs - they need a pull-up resistor.
     
  10. golgot78

    golgot78 Cyanogen Customer

    Joined:
    Sep 19, 2016
    Messages:
    25
    OK, we will add a pull-up resistor because we don't have some in our current setup.
    But then could you please confirm that the voltage range we observed at PIN7: from 0.68 V (when shutter is closed) to 0.01 V (when shutter is open) is nominal?
     
  11. Colin Haig

    Colin Haig Staff Member

    Joined:
    Oct 27, 2014
    Messages:
    2,960
    YES.

    Electronics 101
    The transistor works like a switch, with an emitter pin and collector pin that are the switch connections. The emitter pin is connected to ground (0 V), and is one side of the switch. The collector pin comes out to Pin 7. Nothing else is connect to the collector pin - it is an "open" collector.
    When the shutter is open, the transistor is turned ON, and that closes the switch, connecting the collector to the emitter at ground level. So, you will see a voltage on the open collector of 0V. So, your measurement for 0.01V is normal.
    When the shutter is closed, the transistor is switched OFF. The collector is disconnected from ground. So, you won't see 0.0V any more.
    Instead, you will see some "random" voltage - it is "floating" above ground. If you connect your circuit to Pin 7, you will see whatever voltage your circuit puts to it.
    The solution - you connect a resistor (say 3.3k ohms) from pin 7 to a source of +3.3V (pin 9), the value at pin 7 will be close to +3.3V
     
    Last edited: Feb 9, 2017
  12. Colin Haig

    Colin Haig Staff Member

    Joined:
    Oct 27, 2014
    Messages:
    2,960
    Here's a simple schematic.
    SBIGtriggerCH.png
     
  13. golgot78

    golgot78 Cyanogen Customer

    Joined:
    Sep 19, 2016
    Messages:
    25
    OK! It's clear now. Many thanks!
     
  14. Colin Haig

    Colin Haig Staff Member

    Joined:
    Oct 27, 2014
    Messages:
    2,960
    I also think the way you are measuring the shutter is prone to error - I would use an electronic (hardware) timer, and read that from a computer. e.g. use a counter that uses the 10MHz reference from the GPS, and is gated on by the the trigger signal you send, and gated OFF by the pin 7.
    Your Raspberry PI's will have latency of their own, depending on what OS you are running and what the PI is doing. I'm sure you can minimize this by writing really really good code.
    Good luck!
     
  15. Colin Haig

    Colin Haig Staff Member

    Joined:
    Oct 27, 2014
    Messages:
    2,960
    Also, if you are using the Trigger out open collector pin, you will also need a pull-up resistor on that too.
     
  16. golgot78

    golgot78 Cyanogen Customer

    Joined:
    Sep 19, 2016
    Messages:
    25
    Hello,
    We are still in our quest for high accuracy time stamping of our pictures.
    For a ultimate refinement, we now have to consider the shutter speed.
    According to the User's Manual, the shortest exposure is 0.12 s. My understanding is that this exposure time is induced by the intrinsic shutter speed and then, I can deduce that the shutter speed is 0.12 s to travel the 14.85 mm of the CCD length = 123.75 mm/s in the sensor plane.
    Is my assumption correct? and do you have a better knowledge of this value?
    Indeed, if correct, the start time of the exposure of one side of the field is different of 0.12 s than at the other side.
    A subsidiary question: does the shutter speed is constant or is varying according to the exposure time set?

    Thanks in advance for your help!

    Gautier
     
  17. Doug

    Doug Staff Member

    Joined:
    Sep 25, 2014
    Messages:
    6,575
    The shutter rotation speed is constant. I don't know offhand what the exact rotation speed is.

    The shutter is even-illumination; it moves in the same direction when opening and when closing.
     
  18. golgot78

    golgot78 Cyanogen Customer

    Joined:
    Sep 19, 2016
    Messages:
    25
    Hello Doug,
    Thanks for you confirmation of the constant rotation speed.
    But please could you also tell me the value of its exact speed.
    Does this speed is immediately correlated to the minimum exposure time?
    I guess you probably made those measurements in the development of this camera/shutter.

    Thanks in advance,

    Gautier
     
  19. Doug

    Doug Staff Member

    Joined:
    Sep 25, 2014
    Messages:
    6,575
    The engineer who designed the STT shutter is no longer with SBIG. It would take some research. I'd suggest you simply take a video of the shutter rotation and measure it yourself.
     
  20. golgot78

    golgot78 Cyanogen Customer

    Joined:
    Sep 19, 2016
    Messages:
    25
    Hello Doug,

    We are not really equiped to make this kind of measurements right here. We made some attempts to determine shutter speed with a smartphone and slow-motion mode but the result is not very accurate...
    I guess you probably have better means to determine it on your side or maybe, as you suggested, make some research.

    Thanks in advance.

    Gautier
     

Share This Page