ebay u-blox LEA-6T GPS module teardown and initial test

[uncle_bob]

So, Bob, you're in favor of the cheap modules? At first it sounded like you were against them. Re-reading your prior post again, I can see how it could read the other way. I agree that $8-30 vs. $280 is quite a premium for buying directly from uBlox.

Hi

I'd much rather buy the real thing for about $40 or so....If I can't get them, then the "no name" parts would be my second choice. Buying in low quantity from uBlox would be a distant third.

Bob

[bitseeker]

Ah, OK. I'm clear now. Thanks, Bob.

[DrTune]

It's my decidedly uninformed opinion that this $8 V.KEL module "doesn't suck".
Specifically I mean by looking at what's on the board compared to the hardware integration manual
https://www.olimex.com/Products/Modules/GPS/MOD-GPS/resources/UBX-G7020_HardwareIntegrationManual_-GPS.G7-HW-10003-_Confidential.pdf
..there's a number of things in there which are more expensive than the "Cost optimized circuit" reference design; specifically

(absurdly large pic of the antenna input circuitry here

)
a) External LNA and what I assume is a SAW filter
b) TCXO - couldn't identify the brand but looking at the use of only XTAL_I pins it's not a regular xtal, they don't seem to be using the LDO_X_OUT internal LDO though, looks like a nonpopulated stuffing option and they're using the inductor near the marking "G" to filter some other power supply, probably the 3v3 LDO on the board)
c) Has an RTC xtal (as opposed to not having one)
d) Has 512KB SPI flash which isn't required, and
e) some sort of supercap for battery backup
f) the DC-DC converter option is in use (w/inductor etc),

looks like it wouldn't be hard to switch it to USB not UART by populating a couple of 0402's that have pads on there.

Don't know what the performance is like in absolute terms but I was surprised that they didn't cut half as many corners as they could have, given that it's very nearly the cheapest GPS module out there.

[bitseeker]

That's a very good first impression. From skimming the docs, it seems comparable to other modules, though it's a nav/position GPS, rather than a timing one. How much does that matter for occasional reference adjustments, though?

Maybe I should get one before they decide to remove "non-essential" parts.

[DrTune]

There's really an embarrassment of cheap GPS riches out there right now; choose a timing-rated LEA-6T for $30 or a mere $8.50 for the V.KEL (which is the next generation chip and picks up other sat systems like GLONASS);  both with 10Mhz (actually up to 24Mhz) clock output. (Two on the 6T, and two on the V.KEL if you can manage to get a wire soldered on the right pin of that QFN).  Stupendous value IMO. If you can't wait for China shipping the V.KEL can be got for $10 with 3-day shipping in the US (from Banggood, there's where I got mine, I have some more coming from China)

I've got three of the 6T's and am going to acquire a small bag full of the V.KELs (I have a GPS project in mind). 

It looks like the VKEL can be run at quite low voltages (assuming the TCXO doesn't drop out); the module datasheet says 3.3-5v but from the chip datasheet it looks like it will go a bit lower. It's pretty awesome for my purposes that the VKEL can even do autonomous position logging to its SPI flash.

When I have a sec (maybe in a week or so) I'll head-to-head the VKEL and the 6T time outputs. My hunch is that when configured the same they're both equally good.. but we'll see..

[bitseeker]

Is the back of the VK module the antenna?

[DrTune]

Yep, that's all that's on there. Like any other ceramic antenna..

[bitseeker]

Very good. That's what it looked like, but I wanted to be sure. Thanks!

[Robaroni]

I'm wondering how these $30 LEA 6T discs hold up to the elements. I'd like to make a new case and run it outside my lab window but we get hellacious cold tough winters here where it can get to 20F or 30F below. The simplest solution is to put inside the window but of course I'd get more coverage outside.

I just checked and the PCB will, with a little machining, work inside a 2" PVC pipe so I might build a holder to see how it goes.
Rob

[uncle_bob]

I'm wondering how these $30 LEA 6T discs hold up to the elements. I'd like to make a new case and run it outside my lab window but we get hellacious cold tough winters here where it can get to 20F or 30F below. The simplest solution is to put inside the window but of course I'd get more coverage outside.

I just checked and the PCB will, with a little machining, work inside a 2" PVC pipe so I might build a holder to see how it goes.
Rob

Hi

I would make sure it is well sealed. You would do much better to deploy an antenna and keep the electronics indoors.

Bob

[texaspyro]

I got in one of these pucks off of Ebay from a US seller (3 days, $40).  It works fine.   I wired it to one of my RS-232 level shifter / regulator boards  (https://oshpark.com/shared_projects/YPvKgMYa).   The pucks speak 9600:8N1 serial without having to mod the USB power connection.  Also you can drive them with 3.3V instead of 5V.   

Attached is a plot of the thing in action.  It's plotting the sat count,  PPS sawtooth,  PPS accuracy,  and fractional clock bias.  The plot in the lower left corner is a signal level map (but with not enough data collected yet to show much of interest)

[texaspyro]

BTW,  the firmware in these units supports the RAW data message so you can do carrier phase processing.

I suspect the reason that they are using these high-end timing receivers in a simple drone oriented device is that they bought a bunch of them as surplus items from a defunct company or discontinued product line... and when they are all used up, there will be no more of them with the 6T modules in them.

[texaspyro]

I got in a couple of the V.KEL Ublox based boards in.   They seem to be based upon a Ublox-7 chip... hardware id reports 00070000.   Performance seems good.  I am testing it indoors away from a window.  It's tracking 14 sats at the moment (including 3 SBAS sats).  It does not support position-hold or survey mode like the Ublox timing receivers do.

Putting it through the Lady Heather ringer, I noticed that the Timepulse 2 output is limited to 1000 Hz.  It will accept any value (it accepted 1 GHz) for the PPS1 output,  but my freq counters are tied up at the moment, so I don't know what the actual maximum is.

The firmware only has the new CFG_TP5 message, it does not support the older CFG-TP message.

[Ebel0410]

Hi
It seems that it is now not easy to purchase cheap u-Blox LEA-6T or NEO-6T .
Just ordered a GPS module with LEA-6T from aliexpress, 25$ (the same described here at the begining of that thread), where you can find NEO-6M from less than 10$.
Keep you posted.

[uncle_bob]

Hi

The "6" series parts have been out of production for a while now. The 8 series is on the market at low prices. I can't see a really good reason to go with anything other than the newer parts.

Bob

[cdev]

Unfortunately the NeoM8T and Lea 8T are consistently priced at $65-70 or more each.  For that though you get both timing and RTK, plus, they support that use, so you can use it as the base for a multi-constellation RTK capable system. The "T" models have two timing pulses, and its adjustable.

Hi

The "6" series parts have been out of production for a while now. The 8 series is on the market at low prices. I can't see a really good reason to go with anything other than the newer parts.

Bob

[uncle_bob]

Unfortunately the NeoM8T and Lea 8T are consistently priced at $65-70 or more each.  For that though you get both timing and RTK, plus, they support that use, so you can use it as the base for a multi-constellation RTK capable system. The "T" models have two timing pulses, and its adjustable.

Hi

The "6" series parts have been out of production for a while now. The 8 series is on the market at low prices. I can't see a really good reason to go with anything other than the newer parts.

Bob

Hi

If you are willing to pole the device, all of the timing information is there on the < $20 versions of the 8 series.....

Bob

[cdev]

Could you explain what you mean a bit more?  I would love to be able to build a GPSDO out of a cheap ublox module. There is a library on github that might be useful.

Hi

If you are willing to poll the device, all of the timing information is there on the < $20 versions of the 8 series.....

Bob

The M8 brings multiple GNSS constellations to the table for better navigation accuracy/fewer cycle slips..

 But, I think that they may still derive their timing just from one GPS system?  They may use multiple systems to tell the GPS what time/date and where it is at the beginning.

I think having multiple systems introduces some variations as to naming of satellites and NMEA sentences.

There may be advantages to using multiple GNSS systems for timing I don't know about but my gut feeling is that picking one system to standardize on would be the only way to go lacking some NTP-like setup for satellite systems that could sanity check all of them and reject "false tickers".

(You could likely with Ublox M8 or other newer GPS receivers now have several GNSS receivers each with its own NTP instance that used it in its refclock -so basically each with a different GNSS system as its timing source.)


They might disagree by some fixed offset or it might vary?

[uncle_bob]

Hi

If you pick the right part, they have the pps output and the sawtooth correction information. The T versions will output sawtooth once a second without any "prompting" from the user. The non-T versions need to have a request sent to get the sawtooth correction data. Of course if you are not using sawtooth, then there really is no difference at all ....

The Neo-M8N specs are online. Mr Google is quite happy to dig them up. The PPS output is on pin 3 .....

Bob

[cdev]

It's great the way that you can extract the data by multiple means. I really do need to get some more experimenter friendly ublox HW.

Do you know the name of the relevant messages so I can graph it to see what is happening there? I love being overloaded with useless information.

[cdev]

Have you ever measured the stability/jitter of its 1KHz ouput?

I got in a couple of the V.KEL Ublox based boards in.   They seem to be based upon a Ublox-7 chip... hardware id reports 00070000.   Performance seems good.  I am testing it indoors away from a window.  It's tracking 14 sats at the moment (including 3 SBAS sats).  It does not support position-hold or survey mode like the Ublox timing receivers do.

Putting it through the Lady Heather ringer, I noticed that the Timepulse 2 output is limited to 1000 Hz.  It will accept any value (it accepted 1 GHz) for the PPS1 output,  but my freq counters are tied up at the moment, so I don't know what the actual maximum is.

The firmware only has the new CFG_TP5 message, it does not support the older CFG-TP message.

[uncle_bob]

Have you ever measured the stability/jitter of its 1KHz ouput?

I got in a couple of the V.KEL Ublox based boards in.   They seem to be based upon a Ublox-7 chip... hardware id reports 00070000.   Performance seems good.  I am testing it indoors away from a window.  It's tracking 14 sats at the moment (including 3 SBAS sats).  It does not support position-hold or survey mode like the Ublox timing receivers do.

Putting it through the Lady Heather ringer, I noticed that the Timepulse 2 output is limited to 1000 Hz.  It will accept any value (it accepted 1 GHz) for the PPS1 output,  but my freq counters are tied up at the moment, so I don't know what the actual maximum is.

The firmware only has the new CFG_TP5 message, it does not support the older CFG-TP message.

Hi

You use the 1 pps output for a GPSDO ( 1 pulse per second). The jitter is mainly caused by the way the pulse is quantized. That's where the sawtooth correction comes in. It's in the 10 ns range. For a reasonably quiet GPSDO, you run filters with time constants out past 500 seconds or so. That pretty much puts you into the digital filter world .....

Bob

[texaspyro]

Have you ever measured the stability/jitter of its 1KHz ouput?

Nope...

[texaspyro]

Do you know the name of the relevant messages so I can graph it to see what is happening there? I love being overloaded with useless information.

Lady Heather will do that...

[mark03]

I'm still curious about this comment (above) that the non-T versions are almost the same as the -T version in the 8 series.

I don't know the correct terminology here, but I think it used to be the case that the timing firmware would support a process where you (1) run like a normal GPS for some day(s), averaging to get a super accurate position, then (2) switch into a timing mode where the 3D position is taken as a *given*, and all the module solves for is time.  This was supposed to cut down on jitter, I believe, by letting you tell the GPS engine that it is firmly affixed to an immovable object.  (Obviously neglecting plate tectonics here...)

Can you do this on the non-T modules now?  Or was this never as useful as it was made out to be?