I think you might be on to something. When I perch the dongle on my shoulder using the USB extension cable, all the packet drop problems go away. But the instant I put it back on my desk, they reappear again. I would think the RFduinos have better range than this?
It's not about range, its about https://en.wikipedia.org/wiki/Electromagnetic_interference. This is why I recommended the USB extension cable from the start. Think about it. What's better: TV antennae inside the house or outside the house. Outside, because the receiver has less "stuff" in the way. We can apply this principle to the Dongle as a transmitter. This is exactly what I do because there are lot of metal objects AND electronics near my desk surface. I only held back from posting pictures because I don't want to show off my messy desk.
Cyton+Dongle Range is better than you or I expect, as I found with rigorous testing.
I use a USB extension cable and let the dongle hang gently over the top of my monitor, this elevates it from my desk and moves it almost entirely away from other objects that likely cause EMI.
Using the same logic, my wireless router is mounted up high on a wall near the ceiling.
Right, but I've tested this in areas with less EMI and the packet loss is still there. I even went into my bathroom where there is no electrical equipment or metal objects and still I see packet loss.
I still believe his RFdunios have some weakness. Winson, when you tried sessions in other locations (as previously mentioned), you always had packet drops, correct? What kind of metal objects (desks, cabinets, shelves, etc.) are in those locations? I would think that a laptop sitting on a wood or laminate table, with dongle directly plugged in -- would offer almost no potential metal surfaces blocking or interfering.
I've tried my bathroom, where I'm putting the laptop on my lap and sitting on the edge of the bathtub, and setting the laptop on a completely wooden table. I've attached a photo of the latter. Both locations still experience packet loss.
Actually bathrooms have MORE metal than most other room: water and drain pipes, some tubs and sinks are actual metal with porcelain / ceramic coatings. Did any of your tests occur in the middle of a large-ish room, with the laptop on a wooden table. Or sitting on a wooden chair with the laptop on your lap, etc.
From all these trials, it seems likely that one or the other RFduino is weaker than it should be.
Not to be nit-picking, but in the photo above, your laptop IS next to a light switch. All electrical conduits, switches, outlets, conduit runs in walls and floor and ceilings, have surrounding EMF fields. Field strength drops rapidly, usually substantially gone in 3 or 4 feet.
Still, I suggest emailing contact at openbci.com and asking how to proceed. A replacement dongle would be an easy trial.
What is also mysterious is why single packet drops are not being handled without spiking. My understanding was that recent GUI mods offered that feature. Something's odd there.
But your computer is a source of EMI, and in the picture it's next to a wall with electrical wiring. In the videos, I already see packet loss interpolation. Please try separating the dongle and computer as I have recommended.
Better setup: use usb extension cable, move the desk away from the wall, place the dongle in the wicker basket, elevated slightly.
Alright, thanks for the help Richard and William. I want to reiterate that some sessions where I just sit on my normal desk next to an AC power supply and plug in my dongle normally, I get no packet loss. But other sessions (with no change in my equipment setup) experience heavy packet loss. If the EMI was a problem, I would still experience packet loss in all sessions. It seems that as time goes on and I run through more sessions, the packet loss gets less frequent? I'm not sure. It's really hard to tell with a bug like this that comes up every now and then and disappears at other times. I've been able to go so long without contacting support (since January) because the spikes are so transient, so I mistaked them for "normal" fluctuations in EEG. But they've seem to have gotten worse recently.
This still sounds like EMI! Also, these spikes do not exist in the raw data. It's a result of filtering in the GUI.
In the playback files, there is simply a missing packet. When the GUI tries to account for this using packet loss interpolation, it shows as a small spike (Much better than Large Spike that was happening in previous versions).
Winson, just as an interim improvement for packet loss, have you tried Richard's suggestion, about using the extension cord, and 'dangling' the dongle over the edge of a monitor? So that it is some inches away from monitor or computer surfaces?
Okay, to alleviate concerns about EMI, I went to a wide open grassy field and tried there. Any solid object was at least a few feet away. I saw heavy packet loss. I also tried dangling the dongle inches away from the monitor using an extension cable. I saw packet loss. Attached are two photos.
@eegstud24 You read my mind! I had thought about asking you to try going outside. Thanks for trying this. Even though it didn't seem to help, it has ruled out serious EMI issues. Also, thanks for checking the battery voltage.
As a benchmark, I put on my Mark IV headset, and launched the development version of GUI v5 using Linux. I experienced 0 packet loss over the course of 2-5 minutes, and I tried moving my head around. Sometimes, the electrodes will press harder / ease away from the scalp when moving around, but this looks different. Packet loss will currently look like a small spike, and show a message in the console log.
FYI I have spoken with others about your case, and you should expect a reply from official OpenBCI support regarding next steps soon.
Hopefully, a replacement can be arranged, and I will save your previous video in a Troubleshooting folder I have on my computer to document weird issues like this.
Comments
But at the same time, I sometimes have perfect sessions (of > 100 s) where I just plug my dongle in normally and don't experience any packet loss.
It's not about range, its about https://en.wikipedia.org/wiki/Electromagnetic_interference. This is why I recommended the USB extension cable from the start. Think about it. What's better: TV antennae inside the house or outside the house. Outside, because the receiver has less "stuff" in the way. We can apply this principle to the Dongle as a transmitter. This is exactly what I do because there are lot of metal objects AND electronics near my desk surface. I only held back from posting pictures because I don't want to show off my messy desk.
Cyton+Dongle Range is better than you or I expect, as I found with rigorous testing.
I use a USB extension cable and let the dongle hang gently over the top of my monitor, this elevates it from my desk and moves it almost entirely away from other objects that likely cause EMI.
Using the same logic, my wireless router is mounted up high on a wall near the ceiling.
Right, but I've tested this in areas with less EMI and the packet loss is still there. I even went into my bathroom where there is no electrical equipment or metal objects and still I see packet loss.
I still believe his RFdunios have some weakness. Winson, when you tried sessions in other locations (as previously mentioned), you always had packet drops, correct? What kind of metal objects (desks, cabinets, shelves, etc.) are in those locations? I would think that a laptop sitting on a wood or laminate table, with dongle directly plugged in -- would offer almost no potential metal surfaces blocking or interfering.
I've tried my bathroom, where I'm putting the laptop on my lap and sitting on the edge of the bathtub, and setting the laptop on a completely wooden table. I've attached a photo of the latter. Both locations still experience packet loss.
Actually bathrooms have MORE metal than most other room: water and drain pipes, some tubs and sinks are actual metal with porcelain / ceramic coatings. Did any of your tests occur in the middle of a large-ish room, with the laptop on a wooden table. Or sitting on a wooden chair with the laptop on your lap, etc.
From all these trials, it seems likely that one or the other RFduino is weaker than it should be.
Not to be nit-picking, but in the photo above, your laptop IS next to a light switch. All electrical conduits, switches, outlets, conduit runs in walls and floor and ceilings, have surrounding EMF fields. Field strength drops rapidly, usually substantially gone in 3 or 4 feet.
Still, I suggest emailing contact at openbci.com and asking how to proceed. A replacement dongle would be an easy trial.
What is also mysterious is why single packet drops are not being handled without spiking. My understanding was that recent GUI mods offered that feature. Something's odd there.
But your computer is a source of EMI, and in the picture it's next to a wall with electrical wiring. In the videos, I already see packet loss interpolation. Please try separating the dongle and computer as I have recommended.
Better setup: use usb extension cable, move the desk away from the wall, place the dongle in the wicker basket, elevated slightly.
Alright, thanks for the help Richard and William. I want to reiterate that some sessions where I just sit on my normal desk next to an AC power supply and plug in my dongle normally, I get no packet loss. But other sessions (with no change in my equipment setup) experience heavy packet loss. If the EMI was a problem, I would still experience packet loss in all sessions. It seems that as time goes on and I run through more sessions, the packet loss gets less frequent? I'm not sure. It's really hard to tell with a bug like this that comes up every now and then and disappears at other times. I've been able to go so long without contacting support (since January) because the spikes are so transient, so I mistaked them for "normal" fluctuations in EEG. But they've seem to have gotten worse recently.
This still sounds like EMI! Also, these spikes do not exist in the raw data. It's a result of filtering in the GUI.
In the playback files, there is simply a missing packet. When the GUI tries to account for this using packet loss interpolation, it shows as a small spike (Much better than Large Spike that was happening in previous versions).
Moving forward, let's try a replacement dongle. Please contact OpenBCI support, and we will get you taken care of!
Winson, just as an interim improvement for packet loss, have you tried Richard's suggestion, about using the extension cord, and 'dangling' the dongle over the edge of a monitor? So that it is some inches away from monitor or computer surfaces?
Okay, to alleviate concerns about EMI, I went to a wide open grassy field and tried there. Any solid object was at least a few feet away. I saw heavy packet loss. I also tried dangling the dongle inches away from the monitor using an extension cable. I saw packet loss. Attached are two photos.
@eegstud24 You read my mind! I had thought about asking you to try going outside. Thanks for trying this. Even though it didn't seem to help, it has ruled out serious EMI issues. Also, thanks for checking the battery voltage.
As a benchmark, I put on my Mark IV headset, and launched the development version of GUI v5 using Linux. I experienced 0 packet loss over the course of 2-5 minutes, and I tried moving my head around. Sometimes, the electrodes will press harder / ease away from the scalp when moving around, but this looks different. Packet loss will currently look like a small spike, and show a message in the console log.
FYI I have spoken with others about your case, and you should expect a reply from official OpenBCI support regarding next steps soon.
Hopefully, a replacement can be arranged, and I will save your previous video in a Troubleshooting folder I have on my computer to document weird issues like this.
Hello Richard, thanks for your response. I appreciate you going out of your way. I'll keep in touch with OpenBCI support going forward.