($40 GPSDO working well now, my mistake)

[pigrew]

So now its possible the optimum voltage may be lower than 12 volts.

If it does well with a well-regulated 9 volt supply that would mean I can use a decent transformer containing well-regulated linear wall wart that I have.

The two regulators (LT1963 and LT1765) on the board both list 25V as their maximum input voltage. I'm not sure about the LDO (LT1963A), but my guess is that it's powered by the other regulators.... I expect that it'll be happy with anything between 8V and 14V, but I've been running it at either 9V or 12V, and it seems OK. I've been running mine off of a bench supply (66309D) until I get around to making a case.

But, I'm very unsure about a case. How much ventilation should it have? I don't want the board to cook, but I don't want to expose it to air conditioning, either.

For the commands, one confusing thing is the number of satellites. Both $STATUS and $EXTSTATUS return a count. However, sometimes one is higher, and other times the other is higher. I'm not sure about the difference. The one in $STATUS is usually larger.

Another interesting thing is that when parsing commands, it doesn't look at all of the characters. For example "$GETAB" and "$GETA" both return the "$GETA" message. It seems to only look at the first few characters of the command name.

[cdev]

One count of tracked SVs may include the tracked SBAS satellite(s), one may not?

At one point I also saw a "$WSTATUS" or similar. Have you see that?

The two regulators (LT1963 and LT1765) on the board both list 25V as their maximum input voltage. I'm not sure about the LDO (LT1963A), but my guess is that it's powered by the other regulators.... I expect that it'll be happy with anything between 8V and 14V, but I've been running it at either 9V or 12V, and it seems OK. I've been running mine off of a bench supply (66309D) until I get around to making a case.

But, I'm very unsure about a case. How much ventilation should it have? I don't want the board to cook, but I don't want to expose it to air conditioning, either.

Good question! Most GPSDO's Ive seen don't have conspicuously large ventilation holes.. Do they not have any? I don't remember seeing them. 

But shouldn't the oven be somewhat self contained?

>For the commands, one confusing thing is the number of satellites. Both $STATUS and $EXTSTATUS return a count. However, sometimes one is higher, and other times the other is higher. I'm not sure about the difference. The one in $STATUS is usually larger.

[cdev]

Were it not for the shipping costs, some old servers (in rackmount cases) sell for really little now because of their power consumption to performance ratio. If one lived near the seller, it would likely be possible to get a 19" rack case, perhaps even with an LCD, by buying some old server and removing the server part to leave the case.

The problem is, the shipping cost likely makes that uneconomical with ebay.

I think I may put it in a PCB box or even just a plain white cardboard box lined with copper and soldered, (that makes it easy to cut out the hole for the LCD - making it the right size) to get it out of harms way, for now.

[pigrew]

Yes, the photos of cases I've seen online rarely have ventilation, though I can't believe that this is a good idea. They must run quite hot. I'm tempted to 3D print a case with ABS.  I don't have other sensitive things running, so I don't think that it'd be so useful for me to shield the case. As for a power supply, I'm torn between using a DC wall-adapter and buying a AC->DC converter off of eBay.

If I hold my hand on the trueposition board's OCXO to cool it down, it will sometimes go into holdover mode, so it does need some insulation. My guess is that a case with a small number of holes would work well (just enough of a case to prevent drafts). Even then, I think that it is likely drifting less than my Racal 1992 04E counter does (I see a shift in it of a few ppb with a day->night temperature swing).

I've never seen the $WSTATUS message. My above post documents all of the commands and messages that I know of. Do you have a different firmware version than I do?

[I also got a Trimble UCCM module, but it's receiver seems quite deaf. I only get 2-4 satellites with it. It could be aged too much. The trueposition boards have better reception, but confusingly only report 8 channels even though the receiver should support 12.]

[cdev]

The receiver may be using some/all of those channels for its local SBAS system. Many receivers also have another config switch where they can be told to use the SBAS sat(s) in resolving the position as if it was a GPS satellite.

Two PDFs I found on the Furuno GT8031 (if you don't already have them) are attached.

[pigrew]

The receiver may be using some/all of those channels for its local SBAS system. Many receivers also have another config switch where they can be told to use the SBAS sat(s) in resolving the position as if it was a GPS satellite.

Two PDFs I found on the Furuno GT8031 (if you don't already have them) are attached.

The docs say that it should be able to have 12 normal channels + 2 SBAS, so it doesn't make sense that the GPSDO processor only reports 8. Though it's possible that it is actually using more and just not reporting them. We could sniff the GPS module's TX pin to see what it's saying, and try to decode it's messages, though I'm not planning on doing that unless we can find a GPSDO command to directly send messages to the GPS module (at which point the GPS module communication could become useful).

[cdev]

The long rows of contacts might hold some additional signals?

This was also interesting, I thought.

https://blog.regehr.org/archives/794

There probably are better ways, I would think, though, to measure timing down to the individual cycles.. with GPIO-containing hardware.

[pigrew]

My command list has been slightly updated:

• I finally received a $WSAT response. I'm assuming it's for WAAS satellites used for SBAS. It's very rare that I receive that message.
• My guess is that $STATUS is a count of tracked satellites in channel 0-7, and $EXTSTATUS reports channels 0-9. $STATUS had a max value of 8 last night, and $EXTSTATUS had a max value of 10. Sometimes one is higher than the other, but always within two counts.
• The unit seems to only process one command per second. If more than that are sent, it'll ignore most of them.
• I don't know the meaning of $GETP  [returns -1 255] and $GETA  [returns -1]. I'm not planning on trying to set those variables without knowing what they are.

[cdev]

"almanac" maybe for "$GETA"? 

Many GPS's let you download the almanac from their memory.

[pigrew]

Is it just my board, or do all of them have a huge reverse-biased diode (D9, SMCJ5.0A-TR) beneath the antenna port? At -5V, it should be about ~5nF. Shouldn't that nearly completely short out the RF signal? I'm confused, very confused.

I'm tempted to remove it, and see if signal strengths increase. How is it even working? What am I missing?

EDIT: And I did remove it. The C/N0 did not significantly change, and it seems that the reported signal statistics are very imprecise. Every time I power it on, the reported signal statistics have change by a few dB*Hz. Perhaps the diode has so much inductance that it doesn't matter to RF?

[texaspyro]

Here is a list of (most of) the Trueposition commands (besides the ones sent automaticallY.  Lady Heather now implements these.  There are a couple of others...  not recommended you mess with them.   I managed to blow away my unit serial number.

      send_true_cmd("$RESET");        // warm reset unit
      send_true_cmd("$FACT");         // factory reset unit
      send_true_cmd("$GETVER");       // get receiver version
      send_true_cmd("$GETA");         // get attenuator value
      send_true_cmd(out);             // $SETA %d - set attenuator value
      send_true_cmd("$GETPOS");       // get position
      send_true_cmd(out);             // $SETPOS %ld %ld %ld - set position (lat lon alt in same format as $GETPOS)
      send_true_cmd("$GETDELAY");     // get gps delay
      send_true_cmd("$GETBDELAY");    // get board delay
      send_true_cmd("$GETP");         // get pot values (also has a $SETP command)
      send_true_cmd(out);             // $SETDELAY %d  - cable delay
      send_true_cmd(out);             // $SETBDELAY %d - rf delay?
      send_true_cmd("$UPDATE FLASH"); // update flash memory settings
      send_true_cmd("$PROCEED");      // get out of boot mode
      send_true_cmd("$GETVER");       // get receiver version info
      send_true_cmd("$PPSDBG");       // gives temperature at 1Hz (alss, value only updates every 5 secs)
      send_true_cmd("$KALDBG");       // Kalman filter debug messages - gives temperature at 1Hz
                                                       // but value only updated every 5 seconds.
      send_true_cmd(out);             // $SURVEY %ld - do a self survey for %d hours
      send_true_cmd("$PROCEED");      // get out of boot mode

[pigrew]

      send_true_cmd("$PPSDBG");       // gives temperature at 1Hz (alss, value only updates every 5 secs)

$PPSDBG is very useful, but I've not been able to convince myself that it outputs a temperature.  It outputs, for example, "$PPSDBG 1187155826 1 29.87666e3 30 5 0 0 0.0".

Yes, the value is around 25-29, but I've found it seems very dependent on the timing state. I've been using $SETDELAY to adjust the time-offset, it has a very large effect on the various parameters. I'm not sure about the exact meanings, but I don't think that my unit is returning a temperature for any of these parameters. The floating point number seems to be a roughly proportional to the frequency offset in parts-per-trillion. It seems to trend towards about 29e3 on my unit. My current thinking is that it's the OCXO tune voltage.

Another thing to note is that this also contains the status-state (same as in $STATUS), but it updates much more frequently, so I'll use this status instead of the $STATUS status. The meanings of the numbers are the same.

I'm in process of updating my earlier post with some of these details.

[cdev]

$PPSDBG 0 turns PPS debug off

Also the float numbers on mine (however just turned it on now, so still warming up) are around 37...  flag after it is 0-3, then an integer, neg or positive, then two single digit integers, then at the end a short float.

  Going to bring the output into a spreadsheet program when I have a bit more time later to make graphing the various flags easier.

I am hoping there are some more signals on the unit to use, a 10 MHz square wave would be useful, also a pp2s if it has one..

There may be other UARTS..


Uploading image which shows location of potential test points on my unit

[cdev]

With the sentence based stream of data coming out of the GPSDO every second or less, it might make the most sense to simply save each of the command outputs into their own tables in a database (a different table for each type of data) and queries made against it every second to determine what gets written to the display, LEDs, whatever. The tables would be maintained in a temporary storage and once they reached a certain size, which wouldnt have to be very large, but it could be configurable, the older entries would be purged. Only the most recent entries would be displayed.

This would lend itself well to a platform like the Raspberry Pi that has a bit more power than Arduino.

Since so little old data really needs to be saved the database would be tiny in terms of RAM.

[pigrew]

The OCXO control voltage is reported as part of $PPSDBG. The big floating point number seems to be the ADC count, in units of 6.25e-5 V. This may make sense if Vref=4.096 since 4.096/2^16=6.25e-5.

V ~=6.25e-5*floor(PPS3-0.045)

The $KALDBG parameters seem unintelligible to me. KAL3 somewhat tracks the OCXO control voltage, but not always. Many of the parameters reset to 0 when the GPSDO locks (at startup or the end of holdover). Kal5 may be the time derivative of Kal3. Kal5 looks like a smoothed version of PPS4.

I was also having trouble with my CTS board maintaining lock. The PPS4 (PPS phase?) parameter would drift upwards away from zero and eventually it would go into holdover. I ended up executing a $FACT command, and it started working properly. Perhaps with my screwing with the board, it got a bad gain value somewhere. I have not tried sending the $UPDATE FLASH command, as it seems like it automatically writes settings to flash occasionally?

[texaspyro]

With the sentence based stream of data coming out of the GPSDO every second or less, it might make the most sense to simply save each of the command outputs into their own tables in a database (a different table for each type of data) and queries made against it every second to determine what gets written to the display, LEDs, whatever. The tables would be maintained in a temporary storage and once they reached a certain size, which wouldnt have to be very large, but it could be configurable, the older entries would be purged. Only the most recent entries would be displayed.

This would lend itself well to a platform like the Raspberry Pi that has a bit more power than Arduino.

Since so little old data really needs to be saved the database would be tiny in terms of RAM.

Lady Heather basically does that.  It has a circular buffer plot queue (defaults to 3 days at 1 second per entry).  The plot queue supports up to 14 different values.  Most of the interesting data from the device gets saved in the plot queue (and written to a log file).  The plot queue can also be dumped to a log file.  The plot queue view can be adjusted, scrolled,  filtered, etc.  You can display various statistics (min, max, span, average, RMS value,  etc) and display a linear regression trend line or remove a trend line from a plot.  Also display the derivative of a plot.

And it runs on a PI.

[cdev]

Is LadyHeather on the TruePosition ready for release?

Lady Heather is really pretty amazing and I would love to be able to use it on the TP.

I just am poor and can't afford to run even cheap color screens all the time  (since it seems to get more accurate the longer its on, I plan to do something like that with the TP.)

Can one run LH in some kind of text mode where it could work with small displays and read from/write to some push buttons and LEDs?

---

also, sort of an unrelated question.. It seems as if so far I have never seen it give me a "time figure of merit" of 1 and I have only seen it get to 2 briefly. Its actually getting better, 2 is new, for a while it only had gotten to 3.

 I am guessing that being in a case will likely be helpful to improving that figure.. that the effect of a real case around the board would make it more immune to (dramatic) changes in ambient air temps stable and improve that TFOM.  Its a bit less than a meter away from an air conditioning unit whose top is also serving as the ground plane for its antenna. Its also close to a fan which runs much of the time.

The GPSDO is not currently in a case at all, its just bare.

Also, with the eclipse coming up I wanted to point you to http://abelian.org/vlfrx-tools/ which is a VLF receiver for PCs that runs on linux and it contains a host of different functionality which I need to spend some time with to figure out.

It includes some timing-related programs including one which presents a novel method for getting data from the PPS pulse using the audio card input which might work with almost all hardware and give you a way to use the pps "pops" with LH that would not be dependent on running on RPI or a machine with a real serial port.

I don't know. But it might be worth trying.

[texaspyro]

If run Linux/RPI/macOS compiling from source is easy.  Most Windoze users have problems (well, duh).  If you can build the current v5.0 distro from ke5fx.com,  you just need to copy the *.c* files from the X11 .ZIP and do a make.

No well in hell will Heather ever run on a screen smaller than 640x400.

[cdev]

Yes, LH works quite well now with my Venus GPS's- thank you!

So I am sure it will compile.

On my desktop I could use the RS232 to hook the GPSDO up to my PC directly, including sending it PPS via the DCD pin.  (currently I am using an arduino's built in FTDI USB-UART chip to connect to it.)

[cdev]

LH would be the ideal UI for it, I agree.
There is a lot of data that doesn't lend itself so well to a tiny text display, or even a tiny graphic display.

[texaspyro]

It works well with the RPI 800x480 touchscreen.  It's still a little small for my tastes... more, yeah more, that's what I want... more.   I tweaked up the display code to work better on smaller screens.  And added a touchscreen keyboard.

I've been thinking about having some instrument cases built that would have space for the RPI screen and enough extra front panel space for connectors, etc.   Rear panel would have a cutout for an IEC power plug and maybe a small fan.   Maybe some cutouts for some panel mount USBs and ethernet.  You can build a lot of useful stuff with a RPI and a touchscreen.  I'm surprised nobody seems to sell a similar RPI box.

[cdev]

You should mock it up in a cardboard box and that way you can try out a bunch of different layouts.

A good selling point would be if you internally had the mounting particulars set up so it would be compatible with a bunch of different popular SBCs. Not just the RPI.

Spend some time with it and use it and change your mockup a cople of times to see if you might prefer it some other way. Once you find it works best for you then go for it.

I think the only way to do something like that is as a labor of love where you say to yourself upfront that you are unlikely to make any money and might even lose money.

That enables you to do it the way you want it and only then you often do make money.

If you can accept upfront that you may not, then you can do it and have fun with it.
But its a jungle out there as far as profitability.

Very few businesses are so inherently profitable that luck doesn't play a very large part in their success.

[texaspyro]

I'm thinking about a place like Front Panel Express.   They CNC custom panels and have a line of modular cases where they have corner/side rails with slide in panels that set the enclosure size.  They have a nice designer program that lets you layout the panel / enclosure.  They can engrave, etc the panels and paint the enclosure to your tastes.   Their panels are fairly reasonable, but I don't expect a decent setup would be less than $100.

Another company like that is Cam Expert.  They don't have the fancy designer program, but are cheaper.

[cdev]

The RPI definitely has a place in industrial settings. People save so much on the computer they might be willing to spend more on a case. But once they spend a certain amount, then they will start thinking, why not just buy an industrial computer in a rack. Thats not such intelligent thinking because the Pi has a great many advantages but, its the way most people think. That said, you still likely have a lot of room there. I am particularly cautious when it comes to money these days. People always need computers, though, and there again, compelling arguments can be made for Pis. Now if they could only make it work with SATA drives and have gigabit ethernet capability!

[cdev]

How does one get it into $KALDBG ?

Quote from: pigrew on Today at 09:07:59

The OCXO control voltage is reported as part of $PPSDBG. The big floating point number seems to be the ADC count, in units of 6.25e-5 V. This may make sense if Vref=4.096 since 4.096/2^16=6.25e-5.

V ~=6.25e-5*floor(PPS3-0.045)

The $KALDBG parameters seem unintelligible to me. KAL3 somewhat tracks the OCXO control voltage, but not always. Many of the parameters reset to 0 when the GPSDO locks (at startup or the end of holdover). Kal5 may be the time derivative of Kal3. Kal5 looks like a smoothed version of PPS4.

I was also having trouble with my CTS board maintaining lock. The PPS4 (PPS phase?) parameter would drift upwards away from zero and eventually it would go into holdover. I ended up executing a $FACT command, and it started working properly. Perhaps with my screwing with the board, it got a bad gain value somewhere. I have not tried sending the $UPDATE FLASH command, as it seems like it automatically writes settings to flash occasionally?