Tweet
Once more, this isnt really specifically a heavyweight question, but we dont seem to have a technical weight-non-specific sub-forum.
One or two people will have been expecting this from me...
I appreciate that I can probably find this out on the web, but a) the experts here are more reliable, b) its more fun to torment you, and c) this might come in handy for someone else. Besides, people here are more adept at sorting out my misconceptions.
People may be aware that I have plans to do nasty things to try to send a digital signal over a conventional analogue transceiver, by encoding data in virtual stick positions at the transmitter end and decoding the pulse length on the other end.
The point of this is that radio modems seem to be slowish, expensive, and relatively non-standard, whereas buying the cheapest analogue setup I can find and doing nasties to it strikes me as simpler (given that I have no qualms about messing with embedded processors, and have evil plans for complex control mechanisms anyway). At this stage its only a thought experiment, so assume the phrase dont run before you can walk has already run through my brain.
So... a few questions have sprung to mind.
The first, which may render some others irrelevant, is: how does the channel encoding work? Is it handled outside the the transceiver electronics? This splits into two sub-questions:
1) how do multiple control channels get encoded (Im presuming theres a control pulse followed by a data pulse [Im with it enough to know that the data pulse is length encoded], or all the potential outputs are cycled through round-robin), and
2) how does each channel reach the transmitter - as a voltage level, or as a pre-encoded pulse length?
Im sure this is all somewhat transmitter-dependent, but if these things mostly work the same way then it would be nice to know. Obviously, from the point of view of the silly games Im trying to play, a pure on/off channel is ideal (with a vast amount of error-correction and synchronisation logic to sort out the mess, but thats my problem), but I dont know that I can access the hardware at that level.
The same problem applies to the decoder, obviously - does a receiver split the signal to multiple control lines, or does this happen inside the speed controller/mixer? Put it another way - is a transceiver with more channels going to make my life more complex than one with fewer, or will I be bypassing that logic? If the receiver doesnt try to separate the control signals then I may not be constrained by splitting the signal into channels. I understand the signal goes from the receiver to the speed controller as a pulse length, not just as a voltage (I was *that* confused recently) so life could be harder.
On a related note, is a transceiver which claims to support a large number of positions per channel (10-bit accuracy was quoted to me) going to be any better than a cheap and nasty one, or is all the benefit in logic Im not intending to use?
On a practical point: what, therefore, should I be buying as my transmitter/receiver pair? Its quite possible that the cheapest and nastiest thing I can find (with decent power output) would actually be beneficial, and that I should certainly stay away from expensive modified helicopter kits, for example. (Im presuming I can sort myself out with the physical controls, and dont also need to hijack the joysticks). Are these things so expensive anyway that Im better off going the conventional radio modem route?
And finally, this is where it gets tricky: how about telemetry? Is the circuitry inside a transmitter too large/delicate to cope with being run inside a heavyweight? (After Ive removed as much as possible, obviously.)
What are the chances of running duplex on the same crystal - alternating between receiving and transmitting? (Accepting that I have to be very careful with synchronisation to avoid losing control, and that I have to do the obvious thing to make sure the robot doesnt start driving itself.) Would the cycle time make this unmanageable? Assuming I dont *have* to run with telemetry, would events let me use two crystals some of the time if enough were spare, or would the logistics of this make me even more spectacularly unpopular than do obscure hypothetical posts like this one?
As always, here to learn, and thanking you all for any feedback.
--
Fluppet
One or two people will have been expecting this from me...
I appreciate that I can probably find this out on the web, but a) the experts here are more reliable, b) its more fun to torment you, and c) this might come in handy for someone else. Besides, people here are more adept at sorting out my misconceptions.
People may be aware that I have plans to do nasty things to try to send a digital signal over a conventional analogue transceiver, by encoding data in virtual stick positions at the transmitter end and decoding the pulse length on the other end.
The point of this is that radio modems seem to be slowish, expensive, and relatively non-standard, whereas buying the cheapest analogue setup I can find and doing nasties to it strikes me as simpler (given that I have no qualms about messing with embedded processors, and have evil plans for complex control mechanisms anyway). At this stage its only a thought experiment, so assume the phrase dont run before you can walk has already run through my brain.
So... a few questions have sprung to mind.
The first, which may render some others irrelevant, is: how does the channel encoding work? Is it handled outside the the transceiver electronics? This splits into two sub-questions:
1) how do multiple control channels get encoded (Im presuming theres a control pulse followed by a data pulse [Im with it enough to know that the data pulse is length encoded], or all the potential outputs are cycled through round-robin), and
2) how does each channel reach the transmitter - as a voltage level, or as a pre-encoded pulse length?
Im sure this is all somewhat transmitter-dependent, but if these things mostly work the same way then it would be nice to know. Obviously, from the point of view of the silly games Im trying to play, a pure on/off channel is ideal (with a vast amount of error-correction and synchronisation logic to sort out the mess, but thats my problem), but I dont know that I can access the hardware at that level.
The same problem applies to the decoder, obviously - does a receiver split the signal to multiple control lines, or does this happen inside the speed controller/mixer? Put it another way - is a transceiver with more channels going to make my life more complex than one with fewer, or will I be bypassing that logic? If the receiver doesnt try to separate the control signals then I may not be constrained by splitting the signal into channels. I understand the signal goes from the receiver to the speed controller as a pulse length, not just as a voltage (I was *that* confused recently) so life could be harder.
On a related note, is a transceiver which claims to support a large number of positions per channel (10-bit accuracy was quoted to me) going to be any better than a cheap and nasty one, or is all the benefit in logic Im not intending to use?
On a practical point: what, therefore, should I be buying as my transmitter/receiver pair? Its quite possible that the cheapest and nastiest thing I can find (with decent power output) would actually be beneficial, and that I should certainly stay away from expensive modified helicopter kits, for example. (Im presuming I can sort myself out with the physical controls, and dont also need to hijack the joysticks). Are these things so expensive anyway that Im better off going the conventional radio modem route?
And finally, this is where it gets tricky: how about telemetry? Is the circuitry inside a transmitter too large/delicate to cope with being run inside a heavyweight? (After Ive removed as much as possible, obviously.)
What are the chances of running duplex on the same crystal - alternating between receiving and transmitting? (Accepting that I have to be very careful with synchronisation to avoid losing control, and that I have to do the obvious thing to make sure the robot doesnt start driving itself.) Would the cycle time make this unmanageable? Assuming I dont *have* to run with telemetry, would events let me use two crystals some of the time if enough were spare, or would the logistics of this make me even more spectacularly unpopular than do obscure hypothetical posts like this one?
As always, here to learn, and thanking you all for any feedback.
--
Fluppet
Oh what a life.....
Comment