the USB tester arrived today, so I had a play . dongle draws 20ma, module 50/60ma. The small battery puts out 4.5v on the 5v rail.
But everything seems to be working normally, except the module data was still corrupted. So I put my new 3.5 X glasses to use and had a poke around with the soldering iron.
Being able to see what you're doing is a big advantage!
I'm now getting solid looking data from a stationary balcony run. The data had 15 satellites, where as the GW52 only had 6. But the GW52 had much better accuracy data and less track movements.
So next trick is to take the two of them on a bike ride, to see what the dynamic data is like.
Seems that my M8N may be better than yours Pete, ---- now that the cap is the right way round, and the RX lead to the converter has a better connection.
It will be interesting to see what happens when it's moving.
So I guess once it's obvious which is the best way to go, it may be worth investing in a genuine ublox device.
Good to hear you got the stuff sorted out. Sounds like the battery is not up to snuff, but your module is.
I think even most fake NEO-M8 modules use a genuine ublox M8030 chip. They just don't have the extra stuff added that make the NEO-M8N more expensive. Some of that stuff we don't need, but other stuff (like Flash and working BBR) makes a difference.
I have gotten some pretty good driving results with a Beitian 880. The larger antenna compared to the 220 gives better data (lower SDoP, less spikes and artifacts). Recording to a micro SD card with Openlog works beautifully. But I have some doubts about the SDoP values, they may be a bit too optimistic. I have another Beitian module with a 25 mm antenna on order so that I can do side-by-side comparisons.
Some good news. I took the antenna from the new "NEO-M8" and put it on the old one, and now I have a working "NEO-M8" module. Seem the antenna on the old unit was broken. Maybe I pulled too hard. But now the working model gets a fix in a minute or so, and is up to 14 satellites indoor after about 25 minutes. That's on par with the best ones I have tested so far. Saving parameters to BBR and Flash works.
I also got a chance to test the BEITIAN 880-based GPS on the water today. Mounted on top of the helmet, it was able to track 22-24 satellites. That's pretty amazing (and 50% better than in driving tests). Typical SDoPs for single points were below 0.3. At 5 Hz, the +- numbers reported by GPSResults were generally better than for the GW-60, often about 50%. The ranges overlapped for everything I looked at, even 1 hour and nautical mile, where they get quite small. Visually, the ublox 8 and GW-60 doppler speed tracks look similar, with no obvious advantage to either device, but also no big artifacts.
I have this GPS connected to both the Bluetooth transmitter and the Openlog logger. I used it with an Android phone that announced the speeds (over the speaker, not much wind today). The files from logging on the phone and from the Openlog are identical. That was to be expected, but it's still nice to see that both the Android software and the Openlog logger worked flawlessly. Maybe I'm old-fashioned, but I like being able to pop out the SD card for file transfers.
I'll do some more tests comparing the different chips and antenna setups. The 25x2 mm antenna in the Beitian 880 seems to get more satellites than the 18x2 antenna in the 220, no surprise. It will be interesting to see how the 25x4 antenna with the 'NEO-M8N" does in comparison.
Here's a couple of pictures of the B880/BT/Openlog prototype:
There's a USB charger unit at the front right. The Openlog is in the front left, the bluetooth transmitter in the back.
The GPS module is quite fat, even though the antenna is 2 mm thin. I have the suspicion that the unit works so well because everything is in metal boxes to minimize RF interference. The setup just fits into an GoPro housing (the slightly bigger one for the original GoPro).
I still have a few ideas on how to improve the layout, but I got a warning 
from the Boss today that all these $10 purchases add up .. maybe she looked at the order history on Amazon or the credit card bill. So I may have to wait a while before I make the next prototype 
.
Very nice job Peter! 
So has the Neo8m module that you thought was a counterfeit, actually turned out to be a genuine one now you have swapped antenna? Or have i got that mixed up?
Perhaps we need to make a database of sellers that have counterfeit and genuine Ublox M8?
Woops mine is the V3, but I'll still evaluate it's performance, and see if it's usable.
I've just googled a V5 board and come up with some interesting results. here's a $38 one from Bangood. One of the most expensive, so I figured it's got a better chance of being good.
Not it still says 6M &m and M8N, also the capacitor/battery is the same way round as mine was, before I changed it. So if it's correct on this board, could be why mine was put on back to front. Or this board could also be wrong.
And here's the cct.
Note it's showing an EEPROM and a NEO-6M chip.
Who knows what would arrive in the post.
I'll have a look at some of the cheaper ones.
As far as I can see, all the rest show an EEPROM, and the battery/capacitor is shown in both positions.
Some even say it's the M8 chip with EEPROM backup, price varies from about $10 to 50$
You could consider getting the Beitian BN-220 for $10.20 US (www.banggood.com/BN-280-UART-TTL-Level-GPS-GLONASS-Dual-GNSS-Module-Solution-GPS-Module-p-1281188.html?akmClientCountry=AU&rmmds=buy&cur_warehouse=CN). I have not tested this particular chip, but got good results with the 220 (smaller antenna) and 880 (which also has a compass). On Amazon, they have one negative review blasting them about claiming to be a NEO-M8N, and apparently, they have since changed their description. Or maybe wait a couple of days, I should get mine today. The GPS pins seem to be the same order for the 280 and the 880, so I should be able to just swap it out to test it.
Select to expand quotedecrepit said..
I've just googled a V5 board and come up with some interesting results. here's a $38 one from Bangood. One of the most expensive, so I figured it's got a better chance of being good.
Check the comments. There's one that states "It's not genuine ublox module.", and another that states "U-center sees NEO-7". I'd stay away from that one.
Select to expand quoteboardsurfr said..
.....
The second module I bought had several of these issues: no Flash, no functioning BBR, and (probably) no power for the active antenna. Here's a picture from ebay:
Note the board is labeled GY-GPSV3, and supports the 6M. That is probably part of the problem: the ublox 6 supported EEPROM (which this module supposedly has), while the 7 and 8 do not. The module that works after switching the antenna is labeled GYGPSV5-NEO, and does not support the 6M.
When buying a module, make sure it's on a V5 board, not a V3. Having a database of sellers for modules that work would be great, but there seem to be hundreds of sellers, and just a few buyers on this forum.
These boards are obviously just an evolution of using the ublox chips. I don't think you will see any difference between the pcb of a good working authentic module and one with a ripoff module on it. Other than maybe if they install the cap incorrectly... Ublox seem to have kept the outside components the same as they have changed their model numbers, so it makes sense that the PCB can support different iterations of these chips.
The important parts when it comes to flash versus eeprom are in that module's 'can'. So you wont be able to tell from looking at it. Even then a good copy could have the correct devices but be poorly implemented or have substandard parts.
I think the only way of guaranteeing an authentic device is buying from a proper Ublox distributor, or buying modules from companies that buy from genuine Ublox distributors. China being what it is for copying you are never going to know.
Select to expand quotedecrepit said..
Woops mine is the V3, but I'll still evaluate it's performance, and see if it's usable.
I've just googled a V5 board and come up with some interesting results. here's a $38 one from Bangood. One of the most expensive, so I figured it's got a better chance of being good.
Not it still says 6M &m and M8N, also the capacitor/battery is the same way round as mine was, before I changed it. So if it's correct on this board, could be why mine was put on back to front. Or this board could also be wrong.
And here's the cct.
Note it's showing an EEPROM and a NEO-6M chip.
Who knows what would arrive in the post.
I'll have a look at some of the cheaper ones.
As far as I can see, all the rest show an EEPROM, and the battery/capacitor is shown in both positions.
Some even say it's the M8 chip with EEPROM backup, price varies from about $10 to 50$
No, that's a different board. It has SPI connections as well as serial. The capacitor is mounted correctly as there is no other pad on the pcb to mount it to. You could be right about why yours was mounted incorrectly though, but that's a scary thought as I thought most surface mount parts are placed by machine.
The use of Neo-6m chips in these circuit diagrams is normal. Its more than likely just because the support electronics are identical from that version of chip up until the neo-8n.
Please note how that code on the chip looks versus the one you had on your fake that looks like a QR code. Apparently the QR code is a sign that it almost certainly is not genuine.
I just found an interesting link that rang a bell: portal.u-blox.com/s/
It seems that the ublox chip can be put into a "state of confusion" where it has a really hard time getting a satellite lock. I saw this today when playing around with the Beitian BN-280 chip. First time, it started as expected, getting a fix within a couple of minutes, and slowly getting more satellites. Then I messed around with the cables a bit, causing a few disconnections and restarts. After a few of those, the module seemed unable to get a satellite lock for 20 minutes. It would just show 1-3 satellites in SVINFO, but never get more. I started up a very similar chip right next to it, and that one had no problem getting 8-9 satellites in a few minutes. But when I took the module outside, it got up to 18 satellites in just a few minutes.
This is very similar to the behavior I have seen with the "NEO-M8" modules. The satellite reception indoors seems to vary quite a bit here; the chip that got 8 satellites today has gotten 14 on other days. It's possible that some of the issues I had with "NEO-M8" modules were more related to the "bad signal confusion" state than to anything else. Although the second module still was useless (old firmware, no Flash, BBR not working), even if it gets a fix outside. What made this so confusing was that most of the time, getting a fix indoors is no problem at all. From the apparent (but not real) randomness of the issue, this looks like a software bug to me.
Select to expand quoteFormulaNova said..
I think the only way of guaranteeing an authentic device is buying from a proper Ublox distributor, or buying modules from companies that buy from genuine Ublox distributors. China being what it is for copying you are never going to know.
The only distributor they list for the USA is Digi-Key. They list the NEO-M8N for about $35 (although the price goes down to $13 each if you buy 100; www.digikey.com/product-detail/en/u-blox-america-inc/NEO-M8N-0/672-1008-1-ND/6150671). This seems to be just for the GPS chip, so you'd need to add an antenna, and possibly more. Using it is beyond my skill level.
Ebay have just tempted me with this.
Looks like the real deal, only 7m and 8m and no EEPROM.
It from Flushing NY. It'll cost about AU$55 to get it here. I'm very tempted, then we can compare the real deal with the fakes.
Just talked myself into it. Estimated delivery 19 June to 18 July. Looks like it comes with plugs and leads as well, means I don't have to resolder stuff to terminals If I want to change interfaces, while testing alternatives.
Select to expand quotedecrepit said..
I like the onboard usb as well, pity I've already bought the neo.
Go on! Buy another USB one, just to give a valid comparison 
They even do quantity discounts, so you may as well get 5 or 10 
In all seriousness their prices look okay, and if I want a module that is genuine, this is probably a good source.
Ordered. Shipping was $10, without tracking, so total price is about 2x of what a Beitian ublox 8 with a 25 mm antenna costs. Checkout was amusing, since the credit card payment pages are in French. It will be interesting to see how this module compares to the other ones.
Select to expand quotedecrepit said..
I also had the battery thought so I tried with the "official" pi power supply, same thing, It's an instantaneous shut down as soon as the new file is created, I don't even have the time to see it happen, but the new file is there when I reboot.
I ran into a similar problem today that may shed some light on this. I wanted to use the Pi to compare the BN280 to the BN880, so I used a serial-to-USB converter. When I plugged it in, the Pi seemed to reboot. A minute or two later, it started logging.
I verified later that just plugging the USB-to-serial cable with the GPS module into the Pi causes it to reboot. Plugging in one of my USB hubs also causes a reboot. In contrast, plugging in any of the three USB GPS dongles does not cause a reboot. Even putting the USB tester in between the USB-to-serial cable and the Pi helps sometimes to avoid the reboot. Apparently, these Pi reboots when plugging in USB devices are a well-known issue (for example, check www.raspberrypi.org/forums/viewtopic.php?t=23205).
Select to expand quoteboardsurfr said..decrepit said..
I also had the battery thought so I tried with the "official" pi power supply, same thing, It's an instantaneous shut down as soon as the new file is created, I don't even have the time to see it happen, but the new file is there when I reboot.
I ran into a similar problem today that may shed some light on this. I wanted to use the Pi to compare the BN280 to the BN880, so I used a serial-to-USB converter. When I plugged it in, the Pi seemed to reboot. A minute or two later, it started logging.
I verified later that just plugging the USB-to-serial cable with the GPS module into the Pi causes it to reboot. Plugging in one of my USB hubs also causes a reboot. In contrast, plugging in any of the three USB GPS dongles does not cause a reboot. Even putting the USB tester in between the USB-to-serial cable and the Pi helps sometimes to avoid the reboot. Apparently, these Pi reboots when plugging in USB devices are a well-known issue (for example, check www.raspberrypi.org/forums/viewtopic.php?t=23205).
It looks like you could probably avoid this problem by powering the GPS module directly from the 5v power from the powerbank, and that way the Rpi only sources the power for the usb to serial chip.
Select to expand quoteFormulaNova said..
It looks like you could probably avoid this problem by powering the GPS module directly from the 5v power from the powerbank, and that way the Rpi only sources the power for the usb to serial chip.
Not with the converter I have - plugging the USB-to-serial converter in reboots the Pi even if there's nothing connected to it.
Starting the Pi with the USB-converter + GPS plugged in also works only sometimes, and fails at other times. That also is a problem that's been described (www.raspberrypi.org/forums/viewtopic.php?t=183747). There's a possible workaround with unbinding and binding USB drivers, but it's hard to check out if the problems are random. The last 5 times I started the Pi up, it recognized the USB converter + GPS without problems.
I'll wait for the Drotek module to come, and see if that can be plugged in without causing a reboot. The module should also work with Android phones over USB OTG.
Certainly sounds like the way to go, but I'm not rushing in at the moment, I'll see how the American neo performs first.
Meanwhile I've started testing with the fake module.
GW52 and module on the back carrier of my bike, and watch on my right wrist.
(Peter, did you say you used GPSARPro?) I've just tried, and although I've selected "open UBX" all I get is rubbish)
So I've used GPS visualizer instead to overlay he tracks.
In the main the GW52 and the module are closer to my path than the watch.
So now I'll see how the speed and accuracy compares with GPSResults
So we have 9 to 11 sats for the GW52 and 15 to 17 for the module, but the GW52 shows better accuracy. Although if they use different methods, how meaning full is it?
There's also a time difference of 18seconds.
So it looks like this module should be usable with the PI, but if it doesn't hold it's settings, no good for blue tooth or the data logger?
Here's the full 2sec GPSResults report.
GW52
2s:
time m [knots]
09:50:25 20.8 20.171+/-0.256
09:47:13 17.6 17.127+/-0.096
09:48:09 13.0 12.649+/-0.287
09:42:36 12.8 12.478+/-0.131
09:51:22 12.8 12.408+/-0.091
09:41:43 11.1 10.802+/-0.133
09:48:34 10.9 10.635+/-0.152
09:43:15 10.1 9.863+/-0.092
09:48:23 8.4 8.147+/-0.079
09:48:21 8.0 7.814+/-0.097
09:49:17 7.9 7.632+/-0.121
09:46:46 7.8 7.622+/-0.095
09:48:42 7.8 7.606+/-0.319
09:43:00 7.8 7.554+/-0.097
09:43:10 7.7 7.530+/-0.105
09:43:39 7.7 7.527+/-0.094
09:43:37 7.7 7.460+/-0.103
09:47:55 7.7 7.460+/-0.135
09:48:45 7.6 7.431+/-0.151
09:41:32 7.6 7.390+/-0.141
09:43:42 7.6 7.369+/-0.100
09:42:17 7.6 7.359+/-0.131
09:48:51 7.6 7.354+/-0.139
09:42:11 7.5 7.264+/-0.108
09:43:03 7.4 7.216+/-0.086
average[2]: 34.538km/h 18.649knots
average[3]: 30.834km/h 16.649knots
average[5]: 27.718km/h 14.966knots
Module
2s:
time m [knots]
09:50:43 20.7 20.150+/-0.382
09:47:32 17.6 17.098+/-0.324
09:48:27 13.1 12.685+/-0.407
09:42:54 12.8 12.458+/-0.286
09:51:41 12.8 12.404+/-0.429
09:42:01 11.1 10.804+/-0.351
09:48:52 10.9 10.571+/-0.388
09:43:33 10.2 9.936+/-0.269
09:48:40 8.3 8.054+/-0.290
09:48:38 8.1 7.915+/-0.348
09:47:04 7.9 7.668+/-0.357
09:49:35 7.8 7.602+/-0.438
09:43:18 7.8 7.555+/-0.281
09:43:57 7.8 7.545+/-0.281
09:43:59 7.7 7.477+/-0.281
09:43:28 7.7 7.464+/-0.275
09:48:13 7.6 7.421+/-0.314
09:49:09 7.6 7.404+/-0.405
09:41:50 7.6 7.373+/-0.398
09:43:55 7.5 7.334+/-0.342
09:49:03 7.5 7.325+/-0.322
09:49:00 7.5 7.255+/-0.314
09:42:36 7.4 7.235+/-0.367
09:43:21 7.4 7.204+/-0.280
09:44:02 7.4 7.195+/-0.288
average[2]: 34.492km/h 18.624knots
average[3]: 30.825km/h 16.644knots
average[5]: 27.704km/h 14.959knots
Select to expand quoteboardsurfr said..
I ran into a similar problem today that may shed some light on this.
I remember now, mine was a silly old fart problem. I forgot your script turned wifi and TV off, when a GPS is detected. I interpreted this as a shut down!!!!! Took me all night to realise what was going on.
with the Pi turned off by the python script it draws 20mA, plugging the USB hub in ups that to 40Ma and the Pi boots up. I have a feeling that if the Pi is running and the hub is plugged in, it reboots. But plugging either just the module or dongle in hasn't been a problem.
Select to expand quotedecrepit said..boardsurfr said..
I ran into a similar problem today that may shed some light on this.
I remember now, mine was a silly old fart problem. I forgot your script turned wifi and TV off, when a GPS is detected. I interpreted this as a shut down!!!!! Took me all night to realise what was going on.
I know you posted this, and it's quite possible that this was your only problem. It really depends on what you plug in. Apparently, some USB thingies very briefly draw so much current when plugged in that the Pi reboots. Then, when it starts up again, it may or may not recognize what's attached to the USB port. This is an old weakness of the Pi, and also can affect the larger Pi models with a full USB hub.
Select to expand quotedecrepit said..
(Peter, did you say you used GPSARPro?) I've just tried, and although I've selected "open UBX" all I get is rubbish)
GPSARPro can open some .ubx files, but it has a number of different problems. The Pi logger program simply records everything, including a bunch of records at the start, before the GPS has a time fix, so the early points have dates like 1970, 2000, or 2030. That throws GPSAR off. I have contacted Yann and sent him an example file, but have not heard back from him after his first response.
Select to expand quotedecrepit said..
So we have 9 to 11 sats for the GW52 and 15 to 17 for the module, but the GW52 shows better accuracy. Although if they use different methods, how meaning full is it?
Comparing the numbers directly is (almost) meaningless since neither company states clearly what the numbers mean. They are definitely calculated differently. One might mean one standard deviation, the other two - we don't know. I plan to get some descriptive statistics at some point to get an idea, but that's not a simple project, since a few outliers can skew the results.
But check your GNSS settings! 15-17 satellites indicates that the ublox uses only 2 sat systems, probably GPS and GLONASS. I usually get around 20-22 when I also enable Galileo, which drives the sAcc numbers down a bit. You need firmware version 3 to be able to track three satellite systems, though; both of my fakes came with version 2.0, and I doubt I can update the second one since it does not have Flash.
Also, make sure to let your modules record for at least 15-20 minutes outside before you start the test, especially if the module still has memory problems. It can take that long to download all the data the GPS needs from the satellites. You can watch that in u-center, using the graphs and/or message views.Select to expand quotedecrepit said..
There's also a time difference of 18seconds.
That's a bug in GPSResults. GPSAR Pro also can have problems with the date and time. When I read the times in my software, the tracks are perfectly aligned.Select to expand quotedecrepit said..
So it looks like this module should be usable with the PI, but if it doesn't hold it's settings, no good for blue tooth or the data logger?
Right now, all need the module to be configured to 115200 baud to work. If your module reverts back to 9600 baud, changing to 115200 using the serial cable is easy with the Pi. With bluetooth or Openlog, you would not have a cable connected.
With the Pi, using the serial port directly instead of USB is also possible, which would allow changing the baud rate from the logger program. That's not possible when using a bluetooth connector or Openlog (which have their baud rates set separately).
Select to expand quotedecrepit said..
Meanwhile I've started testing with the fake module.
GW52 and module on the back carrier of my bike, and watch on my right wrist.
So we have 9 to 11 sats for the GW52 and 15 to 17 for the module, but the GW52 shows better accuracy. Although if they use different methods, how meaning full is it?
There's also a time difference of 18seconds.
So it looks like this module should be usable with the PI, but if it doesn't hold it's settings, no good for blue tooth or the data logger?
Here's the full 2sec GPSResults report.
GW52
2s:
time m [knots]
09:50:25 20.8 20.171+/-0.256
09:47:13 17.6 17.127+/-0.096
09:48:09 13.0 12.649+/-0.287
09:42:36 12.8 12.478+/-0.131
09:51:22 12.8 12.408+/-0.091
09:41:43 11.1 10.802+/-0.133
09:48:34 10.9 10.635+/-0.152
09:43:15 10.1 9.863+/-0.092
09:48:23 8.4 8.147+/-0.079
09:48:21 8.0 7.814+/-0.097
09:49:17 7.9 7.632+/-0.121
09:46:46 7.8 7.622+/-0.095
09:48:42 7.8 7.606+/-0.319
09:43:00 7.8 7.554+/-0.097
09:43:10 7.7 7.530+/-0.105
09:43:39 7.7 7.527+/-0.094
09:43:37 7.7 7.460+/-0.103
09:47:55 7.7 7.460+/-0.135
09:48:45 7.6 7.431+/-0.151
09:41:32 7.6 7.390+/-0.141
09:43:42 7.6 7.369+/-0.100
09:42:17 7.6 7.359+/-0.131
09:48:51 7.6 7.354+/-0.139
09:42:11 7.5 7.264+/-0.108
09:43:03 7.4 7.216+/-0.086
average[2]: 34.538km/h 18.649knots
average[3]: 30.834km/h 16.649knots
average[5]: 27.718km/h 14.966knots
Module
2s:
time m [knots]
09:50:43 20.7 20.150+/-0.382
09:47:32 17.6 17.098+/-0.324
09:48:27 13.1 12.685+/-0.407
09:42:54 12.8 12.458+/-0.286
09:51:41 12.8 12.404+/-0.429
09:42:01 11.1 10.804+/-0.351
09:48:52 10.9 10.571+/-0.388
09:43:33 10.2 9.936+/-0.269
09:48:40 8.3 8.054+/-0.290
09:48:38 8.1 7.915+/-0.348
09:47:04 7.9 7.668+/-0.357
09:49:35 7.8 7.602+/-0.438
09:43:18 7.8 7.555+/-0.281
09:43:57 7.8 7.545+/-0.281
09:43:59 7.7 7.477+/-0.281
09:43:28 7.7 7.464+/-0.275
09:48:13 7.6 7.421+/-0.314
09:49:09 7.6 7.404+/-0.405
09:41:50 7.6 7.373+/-0.398
09:43:55 7.5 7.334+/-0.342
09:49:03 7.5 7.325+/-0.322
09:49:00 7.5 7.255+/-0.314
09:42:36 7.4 7.235+/-0.367
09:43:21 7.4 7.204+/-0.280
09:44:02 7.4 7.195+/-0.288
average[2]: 34.492km/h 18.624knots
average[3]: 30.825km/h 16.644knots
average[5]: 27.704km/h 14.959knots
The numbers look good! Mostly within +- 0.03 knots, all that I checked within +- 0.1, which is very good for 2 seconds.
Select to expand quoteboardsurfr said..
Comparing the (error estimate*) numbers directly is (almost) meaningless since neither company states clearly what the numbers mean.......
*My addition for clarity- SQ
I think the companies do say what they mean, but they do not define it exactly as we would like. They are probably never going to explain 'exactly' how this is calculated, as it would be an intellectual property proprietary secret. I'm not saying this is a good reason, but it is what it is.Select to expand quoteboardsurfr said..
They are definitely calculated differently.
For the same reason stated above, we don't actually 'know' that the method used is different. It actually seems logically most likely that it is based on the same first principles, and is therefore either the same , or very similar. In fact, I think, for all practical purposes, even if there is a slight difference in the way they go about it, it is reasonably comparable.
I can think of one way it might be possible to do a direct comparison. That would be using the same method that Dr. Chalko used to validate the error accuray for the Locosys GW devices. But that would require Dr Chalko giving us access to his data and methods, and he has already said that he can't do this for the reasons stated above.
Another way might be to examine the software code (firmware) in the devices, but I would think that would be encrypted, or locked down in some way.
