Building a satellite

[ShredBird]

1 hour ago, Julian5 said:

 

Unfortunately, EMCORE sold their photovoltaic company 2 years ago and looking at other companies like SolAero they seem to have 7,500$ minimum orders.....

Bummer.  You can find some prebuilt solar boards from a company called Pumpkin and some other off the shelf parts sellers, probably not cheap though.

[djneo]

This is super cool. i always wanted to also design one and just have it hanging on the ceiling.

 

 

My original idea was to build a nano sat like http://www.delfispace.nl

 

i visited the dutch ESA facility and there i've met the person behind that project and got to hold and play with the Delft-C3 testing bed. (as the real one was in space ofc) and he explained how it worked.

 

He also told me that the cost of the satellite was pretty low. and the launch was the most expensive part

[Julian5]

On 22/11/2017 at 12:03 AM, ShredBird said:

Bummer.  You can find some prebuilt solar boards from a company called Pumpkin and some other off the shelf parts sellers, probably not cheap though.

Yeah I found a company that sells gallium arsenide sells by individual cell and they seem to be cheaper. Worst case scenario I found some sketchy gallium arsenide cells on aliexpress that could work.

 

11 hours ago, djneo said:

This is super cool. i always wanted to also design one and just have it hanging on the ceiling.

 

 

My original idea was to build a nano sat like http://www.delfispace.nl

 

i visited the dutch ESA facility and there i've met the person behind that project and got to hold and play with the Delft-C3 testing bed. (as the real one was in space ofc) and he explained how it worked.

 

He also told me that the cost of the satellite was pretty low. and the launch was the most expensive part

Yeah its a great project! Definitely very fun! And the satellite cost is indeed low compared to the launch.

[ionbasa]

More interested in the bulk energy storage, I do R&D on lithium battery packs. I saw a battery holder in the middle. I'm assuming 18650 cell?

 

You may want to insulate the cell if possible. Space is pretty harsh on batteries, some (if not most) satellites that use thermoelectric generation also use some of the spare heat to keep cells at a relatively safe temperature. CubeSats obviously can't use such technology, but insulation of some kind helps in the long run.

If you ever plan to launch, check around the major Universities, most have clubs dedicated for these. Multiple Universities will chip in together to reduce the cost of launching CubeSats up.

 

If you ever need any recommendations, feel free to ask me. 

[Julian5]

12 hours ago, ionbasa said:

More interested in the bulk energy storage, I do R&D on lithium battery packs. I saw a battery holder in the middle. I'm assuming 18650 cell?

 

You may want to insulate the cell if possible. Space is pretty harsh on batteries, some (if not most) satellites that use thermoelectric generation also use some of the spare heat to keep cells at a relatively safe temperature. CubeSats obviously can't use such technology, but insulation of some kind helps in the long run.

If you ever plan to launch, check around the major Universities, most have clubs dedicated for these. Multiple Universities will chip in together to reduce the cost of launching CubeSats up.

 

If you ever need any recommendations, feel free to ask me. 

Yeah it was indeed a 18650 and I did start experimenting with nichrome wire to warm the battery when on the "cold" side. However, on this satellite, I will not be going to use a battery and it will be purely relying on the regulated voltage from the gallium arsenide cells. 

 

Oh and thanks for the support!

[Julian5]

So I just finished designing a first version which I do believe is full of issues (obviously being my first pcb) Due to legal issues and ease I have decided to use a 433mhz tX based on a 433mhz resonator. This PCB does not have sensors right now and relies on a 4v 100ma 30mm x 70mm silicon Solar panel which I managed to find for cheap. I also believe the 16mhz crystal is not wired correctly and Tbh if anyone could just give it a check if would be nice. Anyways here are the pics of the PCB and the parts: 

 

Spoiler

 

 

 

It is based off these schematics 

Spoiler

 

 

U1 - ATMEGA 328P-AU

U2 - 0.1 uf Cer

U4 - 22pf Cer

U5 - 22pf Cer

U6 - U7 - 1k R

U8 - XC6206P332MR

U9 - S8050

U10 - 2SC5537

U11 - 1uf cer

U12 - 433mhz resonator 

U13 - 7pf Cer

U14 1uf cer

C1 - 7pf

C2- 47Uf  R3 - 20K R4 - 27k 

Any help on checking this pcb would be greatly appreciated.

@ShredBird @ionbasa 

 

[straight_stewie]

On 11/4/2017 at 9:38 AM, Julian5 said:

I would like to make the smallest satellite possible with as many sensors as possible and maybe even a camera, I will also design and manufacture the PCB with SMD components. So really its just a tiny PCB with solar cells, a chip (will most likely use an atmega) and a bunch of sensors, however, my ultimate goal is a CMOS camera.

1 of the primary considerations when using a camera on these is not power usage (the solution to that is simple, sort of). The biggest problem is that if the camera accidentally gets a direct look at the sun at high altitude, it's done for. 

I'm not too sure of what the best solution for that problem would be, but I do know a potential solution for the energy issue. Instead of having enough power to run all the sensors at once, you really only need enough power to run one or two sensors at a time. You can poll the sensors in a round robin fashion, only turning them on one at a time to get the measurement, and then shutting them back off immediately after the measurement. Consider leaving the altimeter on always and doing the round robin with only the other sensors, that way you can tie your readings to altitude more accurately.

[ionbasa]

8 hours ago, Julian5 said:

So I just finished designing a first version which I do believe is full of issues (obviously being my first pcb) Due to legal issues and ease I have decided to use a 433mhz tX based on a 433mhz resonator. This PCB does not have sensors right now and relies on a 4v 100ma 30mm x 70mm silicon Solar panel which I managed to find for cheap. I also believe the 16mhz crystal is not wired correctly and Tbh if anyone could just give it a check if would be nice. Anyways here are the pics of the PCB and the parts: 

 

  Hide contents

 

 

 

It is based off these schematics 

  Hide contents

 

 

U1 - ATMEGA 328P-AU

U2 - 0.1 uf Cer

U4 - 22pf Cer

U5 - 22pf Cer

U6 - U7 - 1k R

U8 - XC6206P332MR

U9 - S8050

U10 - 2SC5537

U11 - 1uf cer

U12 - 433mhz resonator 

U13 - 7pf Cer

U14 1uf cer

C1 - 7pf

C2- 47Uf  R3 - 20K R4 - 27k 

Any help on checking this pcb would be greatly appreciated.

@ShredBird @ionbasa 

 

Are you having issues flashing bootloader code on the ATMega? The 16Mhz oscillator is correctly wired in your diagram. See below:

 

One question though:

Why aren't you using the SPI pins on the ATMega? IIRC on some chips you can program on the Digital IO pins, which looks like it should work with your schematic, correct?

[Julian5]

6 hours ago, ionbasa said:

Are you having issues flashing bootloader code on the ATMega? The 16Mhz oscillator is correctly wired in your diagram. See below:

 

One question though:

Why aren't you using the SPI pins on the ATMega? IIRC on some chips you can program on the Digital IO pins, which looks like it should work with your schematic, correct?

I have not ordered the pcb yet and just wanted to make sure, my only concern is that I am using this smd crystal and it has 3 pins, looking at my arduino nano on my table it seems like its wired up differently. However I cannot find a schematic of the arduino nano.

As far as the SPI pins are concerned  I just followed the schematic and as long as it works I dont have any issues with it. 

12 hours ago, straight_stewie said:

1 of the primary considerations when using a camera on these is not power usage (the solution to that is simple, sort of). The biggest problem is that if the camera accidentally gets a direct look at the sun at high altitude, it's done for. 

I'm not too sure of what the best solution for that problem would be, but I do know a potential solution for the energy issue. Instead of having enough power to run all the sensors at once, you really only need enough power to run one or two sensors at a time. You can poll the sensors in a round robin fashion, only turning them on one at a time to get the measurement, and then shutting them back off immediately after the measurement. Consider leaving the altimeter on always and doing the round robin with only the other sensors, that way you can tie your readings to altitude more accurately.

For now the camera idea is set aside due to power restrictions, however when I do come to it I have already started thinking about cycling the components and having sleep periods etc. And for directional purposes I considered electromagnets but they are irrelevant on this scale of a satellite, thus the dual sided solar panels.  

[marten.aap2.0]

2 minutes ago, Julian5 said:

However I cannot find a schematic of the arduino nano.

is this what you are looking for?

Spoiler

 

[Julian5]

Just now, marten.aap2.0 said:

is this what you are looking for?

  Hide contents

 

Unfortunately no, Im looking for a schematic with the traces from the atmega to the 3 pin crystal since I believe one of the pins is connected to ground or something like that.

[straight_stewie]

2 hours ago, Julian5 said:

Im looking for a schematic with the traces from the atmega to the 3 pin crystal since I believe one of the pins is connected to ground or something like that.

Unfortunately the diagram (<-- Click me!!) is a PDF and not an image.

However, the 16 mhz crystal is connected to pins 7 and 8 on the Atmega. The crystal is also connected to ground through one of it's own pins. In order to know which pin is which on the crystal you'll need to find the datasheet for your crystal.

[Julian5]

51 minutes ago, straight_stewie said:

Unfortunately the diagram (<-- Click me!!) is a PDF and not an image.

However, the 16 mhz crystal is connected to pins 7 and 8 on the Atmega. The crystal is also connected to ground through one of it's own pins. In order to know which pin is which on the crystal you'll need to find the datasheet for your crystal.

Oh ok! I have it now, will modify the PCB later. 

[ionbasa]

1 hour ago, straight_stewie said:

Unfortunately the diagram (<-- Click me!!) is a PDF and not an image.

However, the 16 mhz crystal is connected to pins 7 and 8 on the Atmega. The crystal is also connected to ground through one of it's own pins. In order to know which pin is which on the crystal you'll need to find the datasheet for your crystal.

 

24 minutes ago, Julian5 said:

Oh ok! I have it now, will modify the PCB later. 

Good catch. Didn't know you were using a 3 pin oscillator. All those diagrams have it as a two pin oscillator. Hmm Which part number for the crystal in particular? Some 3 pins require GND, V+, and a clock out. So attention would need to be taken when wiring it. One such example is the MAX7375.

[straight_stewie]

8 hours ago, Julian5 said:

Oh ok! I have it now, will modify the PCB later. 

While we are on that subject, you may want to make sure that you have protection circuits for each pin on the Atmel chip. Depending on what altitude you are planning on floating it at, the decreased atmosphere (and thus higher ambient radiation) could become an issue. This is a nifty article about the specific issues, titled 10 ways to destroy an arduino. Specifically, methods 3, method 5, method 6, and method 10.

[ShredBird]

I haven't looked it over in detail yet, but one comment I have is to make the top plane of your PCB a power plane and the bottom plane a ground plane.  This is common practice in PCB design as it simplifies getting power and grounding to everything and it also serves as board wide capacitor that can filter your power.  A quick internet search on PCB design best practices should give you some things to consider.

Also, you keep mentioning you are not going to use a battery and I have to reiterate that is not a good idea.  As your satellite does not intend to have any type of attitude control it will likely tumble in orbit.  It won't be a useful satellite if it's turning on and off constantly while it's spinning.  You may also have to confirm that your satellite has turned on and detached from the delivery vehicle successfully for launch requirements.  I don't think anyone will consider launching your satellite if you don't have its mission/power plan laid out.  No one wants to just put junk out there with the hopes that it will work, you will have to prove it.

[straight_stewie]

5 minutes ago, ShredBird said:

Also, you keep mentioning you are not going to use a battery and I have to reiterate that is not a good idea.  As your satellite does not intend to have any type of attitude control it will likely tumble in orbit.  It won't be a useful satellite if it's turning on and off constantly while it's spinning.  You may also have to confirm that your satellite has turned on and detached from the delivery vehicle successfully for launch requirements.  I don't think anyone will consider launching your satellite if you don't have its mission/power plan laid out.  No one wants to just put junk out there with the hopes that it will work, you will have to prove it.

While I agree with your point about not using a battery, it's not that kind of satellite. It's essentially just a weather balloon payload designed to the same specs as a cubesat. I would venture to guess that if OP had the contacts necessary to delivery a cubesat into orbit he wouldn't be coming to us for help, he would be going to whomever his industry contacts are.

[ShredBird]

48 minutes ago, straight_stewie said:

While I agree with your point about not using a battery, it's not that kind of satellite. It's essentially just a weather balloon payload designed to the same specs as a cubesat. I would venture to guess that if OP had the contacts necessary to delivery a cubesat into orbit he wouldn't be coming to us for help, he would be going to whomever his industry contacts are.

That's fair, in any case, just file it under "things to consider if it were a satellite".

[Julian5]

15 hours ago, ShredBird said:

I haven't looked it over in detail yet, but one comment I have is to make the top plane of your PCB a power plane and the bottom plane a ground plane.  This is common practice in PCB design as it simplifies getting power and grounding to everything and it also serves as board wide capacitor that can filter your power.  A quick internet search on PCB design best practices should give you some things to consider.

Also, you keep mentioning you are not going to use a battery and I have to reiterate that is not a good idea.  As your satellite does not intend to have any type of attitude control it will likely tumble in orbit.  It won't be a useful satellite if it's turning on and off constantly while it's spinning.  You may also have to confirm that your satellite has turned on and detached from the delivery vehicle successfully for launch requirements.  I don't think anyone will consider launching your satellite if you don't have its mission/power plan laid out.  No one wants to just put junk out there with the hopes that it will work, you will have to prove it.

Thanks for the feedback, I will look into making each side a ground and power plane. 

 

On the topic of batteries, I have considered using an array of capacitors for short-term storage of electricity. The idea is also to use a launch device that launches it with the panels facing away from earth. What would be a way of confirming the satellite has been deployed? Something like an accelerometer could do the trick?

[ionbasa]

6 hours ago, Julian5 said:

Thanks for the feedback, I will look into making each side a ground and power plane. 

 

On the topic of batteries, I have considered using an array of capacitors for short-term storage of electricity. The idea is also to use a launch device that launches it with the panels facing away from earth. What would be a way of confirming the satellite has been deployed? Something like an accelerometer could do the trick?

Look up Maxwell Ultracapacitors if you want to go that route. They're a bit  pricey but may fit your need.

 

If  you want something to verify that the satellite has been deployed from the launch vehicle an accelerometer should work, look up IMD (Inertial Monitoring Device), they come in  3,6, and 9 axis DOF (Degrees of Freedom). A simpler solution may be to use an altimeter. You could also use a Low power GPS  receiver on the launch vehicle and satellite to verify telemetry data.

[straight_stewie]

7 hours ago, Julian5 said:

On the topic of batteries, I have considered using an array of capacitors for short-term storage of electricity. The idea is also to use a launch device that launches it with the panels facing away from earth. What would be a way of confirming the satellite has been deployed? Something like an accelerometer could do the trick?

Would it be possible to simply have a wire that's attached between the satellite and the launch vehicle that breaks (but stays with the satellite) when the two separate? Then it's as simple as generating an interrupt on separation.

[Julian5]

13 hours ago, ionbasa said:

Look up Maxwell Ultracapacitors if you want to go that route. They're a bit  pricey but may fit your need.

 

If  you want something to verify that the satellite has been deployed from the launch vehicle an accelerometer should work, look up IMD (Inertial Monitoring Device), they come in  3,6, and 9 axis DOF (Degrees of Freedom). A simpler solution may be to use an altimeter. You could also use a Low power GPS  receiver on the launch vehicle and satellite to verify telemetry data.

I will look into the capacitors, I think I will use the accelerometer anyways for monitoring purposes. I wont use GPS since its very challenging to get the appropriate certifications for launching.

 

12 hours ago, straight_stewie said:

Would it be possible to simply have a wire that's attached between the satellite and the launch vehicle that breaks (but stays with the satellite) when the two separate? Then it's as simple as generating an interrupt on separation.

Theoretically you could use the wire as a switch and then maybe even use it as your radio antenna if it had a specific breaking point for that frequency, However I believe the accelerometer is a more extensive and precise way to measure the launch without any "External" and mechanical components.

[straight_stewie]

3 hours ago, Julian5 said:

However I believe the accelerometer is a more extensive and precise way to measure the launch without any "External" and mechanical components.

The accelerometer is a great tool for data gathering, and even building an INS unit. However, I don't think it's the best tool for detecting separation. There's a reason that real spacecraft almost always use wire runners and/or conductance to determine stage breakup/separation. If you wish to use an accelerometer you have to write an algorithm that knows the difference between any outside influence during launch, a failure, and a purposeful separation. With the wire loop you just have to determine if it still conducts or not. You can even generate low or falling edge interrupts to avoid the polling problem. 

[ShredBird]

Your satellite is not allowed to be powered on until it separates from the launch vehicle.  Usually that interface already exists as part of the payload launcher design.  If you talk to a potential launcher they will give you the details of that interface.  If you have deployables you generally have to wait for a minimum amount of time to deploy them for the safety of the launch vehicles and other payloads.

 

A satellite in orbit will always experience total darkness for some amount of its orbit as it has to pass in the shadow of the Earth.  Your satellite with no attitude control will tumble and flip through orbit.  One method to alleviate this is to place rods of magnetic material along a single axis of the satellite.  Because the magnetic moments are aligned along axis, they will naturally align to the Earth's magnetic field.  This will give you a predictable orientation in a passive manner, you would be able schedule a power plan with your batteries.

[Julian5]

4 hours ago, ShredBird said:

Your satellite is not allowed to be powered on until it separates from the launch vehicle.  Usually that interface already exists as part of the payload launcher design.  If you talk to a potential launcher they will give you the details of that interface.  If you have deployables you generally have to wait for a minimum amount of time to deploy them for the safety of the launch vehicles and other payloads.

 

A satellite in orbit will always experience total darkness for some amount of its orbit as it has to pass in the shadow of the Earth.  Your satellite with no attitude control will tumble and flip through orbit.  One method to alleviate this is to place rods of magnetic material along a single axis of the satellite.  Because the magnetic moments are aligned along axis, they will naturally align to the Earth's magnetic field.  This will give you a predictable orientation in a passive manner, you would be able schedule a power plan with your batteries.

Yeah I was indeed thinking about the attitude control and instead of using electromagnets just static neodymium magnets would work. Is there any previous examples of this being used before? And would it be more beneficiary to use large capacitors for short term storage or smoothing or to use small rechargeable cells for when entering the dark side. I also think a magnetometer could be a useful to figure out the current orientation of the craft.

 

5 hours ago, straight_stewie said:

The accelerometer is a great tool for data gathering, and even building an INS unit. However, I don't think it's the best tool for detecting separation. There's a reason that real spacecraft almost always use wire runners and/or conductance to determine stage breakup/separation. If you wish to use an accelerometer you have to write an algorithm that knows the difference between any outside influence during launch, a failure, and a purposeful separation. With the wire loop you just have to determine if it still conducts or not. You can even generate low or falling edge interrupts to avoid the polling problem. 

Yes that is true, maybe I could combine the radio antenna with the separation wire or use a very thin copper wire.