Trimble ThunderBolt Timing GPS Receiver GPSDO

Trimble ThunderBolt GPS Timing reciever running

Background
Description
Antenna
Power Supply
Start Up
Problem
ThunderBolt Monitor Programs
    Tboltmon
    DSPMon
    Timekeeper
    Lady Heather
Hardware Monitors
K04BB Monitor
iCruze Monitor - iCruze Monitor 8051 PCB
IC Table
External Oscillator Option
Disiplined vs. Unlocked
Maintenance
    Drawing
References
Links

Background

Trimble has been making GPS receivers since the system has been in existence. They made the Trimpack which was one of the first GPS units used in Desert Storm. The U.S. then used the PLGR GPS receiver and now (2008) the DAGR. But these GPS receivers were intended for use mainly as navigation tools and only supported timing in the sense of frequency hopping radios like SINCGARS. The Motorola GPS receivers have had a precision time capability and the M12T+ is very good at that (Motorola has sold the GPS receiver operation and it's now part of i-LOTUS and is sold by Synergy.

There are two first generation Trimble Thunderbolt GPS receiver versions.
p/n 48050-61 for the metal cage 24 Volt
p/n 48051-61 for the board only that needs +5, +12 & -12 Volts
The board only in a metal box that needs +5, +12 & -12 Volts may have been a custom part.
The first generation ThunderBolt receivers went out of production in mid 2006 because of Rohas lead free EU rules.

The current lead free ThunderBolt is the "E" version. Although not a direct replacement for the first generation timing receiver, it's very similar.

The Thunderbolt is what's called a GPS Disciplined Oscillator (GPSDO). The internal 10 MHz oscillator has it's tuning voltage controlled by the GPS time. The 10 MHz is divided down to provide a 1 Pulse Per Second (!PPS) output. If the GPS signal is lost it goes into holdover mode.

This is huge improvement over the old lab crystal oscillators that needed to be manually tweaked (calibrated) every 6 to 12 months to account for crystal aging. With the GPSDO the corrections are applied automatically in real time as long as you have GPS reception. A GPSDO is an excellent way to have a lab frequency standard to drive test equipment that accepts external 10 MHz references signals like counters, spectrum analyzers etc. The 1PPS signal is a good way to make precision time interval measurements.

Description

The unit shown here is from TAPR and was part of the Time Nuts mailing list group buy. They also have been on eBay (thunderbolt gps) recently. It's the receiver only board in a metal box.

Antenna

The Trimble Trimpack GPS receivers require about 41 dB gain between the antenna and receiver and so will not work with the common 20 dB gain antennas. But the ThunderBolt can work with between 18 and 35 dB gain which covers most common antennas. The treatment of the antenna and coax loss is not very good in the ThunderBoltBook2003. There is a better discussion in the ThunderBolt E User Guide.pdf.

The antenna connector center conductor supplies +5 Volts to the antenna so should NOT be used to power the newer 3.3 Volt GPS antennas.

Power Supply

The metal cage p/n 48050-61 version runs from a single 24 volt power supply and inside has a power supply board that generates +5, +12 and -12 VDC to power the receiver board. I expect that this is a very clean power supply (but haven't seen data to support that yet).

The single board p/n 48051-61 and the metal boxed single board version require +5, +12 and -12 VDC. It turns out that any noise on these supplies (maybe not all three of them) degrades the performance of the outputs as shown on the Allan plots (TVB Thunderbolt PS Noise). That's why I'm using an HP E3631A triple output lab power supply as shown in the photo at the top of this page.
The currents drawn are:

mAmps	+5	+12	-12
Cold	224	678	2
Warm	246	144	2

YMMV

Note: The HP E3631A is a very poor choice of power supply because whenever there's a power brownout or blackout it forgets it's settings and when power is restored comes up with the outputs turned off. I have an HP 6236A power supply on order that has manual knobs and will retain its settings after a power failure.

Power-One HTAA-16W-A Power Supply

A triple output supply rated:
Out: +5 VDC @ 2A
+/- 9 - 15 VDC @ .4A
In:100-240VAC

All the wires need to be removed and replaced with an AC line cord and a cable & connector to mate with the Thunderbolt.

Connector Pinout

TVB Power Supply web page has pinout and connector info.
Photos below show real parts. The illustrations in the cataloge are generalized drawings.


Molex 16-02-0103 female contact pins	Molex 50-57-9406 6 position shell

Molex hand crimp tool No. 63811-8700 (Mouser 538-63811-8700) one each for $321.98.
The Radio Shack tool works.

Start Up

When the unit does a cold start as delivered from the factory it will:
(1) warm up the oscillator (maybe 20 minutes)

During this time it can start locking on satellites and down loading the almanac data

(2) do a self survey to learn the antenna position (maybe an hour)
(3) change from "start up" mode to "disciplining" mode

Recently a power failure shut down my HP triple output lab supply and when I powered up the Thunderbolt it did not know the time and did not find any satellites. Changing to a diferent antenna seemed to fix it and when the first antenna was reconnected all seems fine.

Problem

The Mode is (1) Power Up, Activity is (1) Osc Warm-up, DAC Voltage is 0.000000, BUT
all the indicators are green.
Thunderbolt Not in Normal Mode

ThunderBolt Monitor Programs

Tboltmon.exe

This is a Windows program that allows setting up and monitoring the operation of the ThunderBolt receiver. The program is not required and can be disconnected from the receiver once you are happy with it's settings.

The only yellow dot is the one for Leap Second Pending. When first run there are a number of yellow dots. Although there is a Tracking Status Window, it uses a text table to display the data, not a polar graph like in the TAC32 program for the Motorola GPS receivers.

The ThunderBolt has EEPROM memory for it's key settings, called segments in the monitor program. How the ThunderBolt boots up depends on the contents of these stored segments.

If the factory default condition then on power up a self survey is done. This requires some time operating in the 3D fix mode to get an average position. If a precision reference position is stored then it boots up without the self survey.

The timing Outputs area in the lower left displays the PPS error (-17.64 ns UTC) so it looks like there is a sawtooth correction available.
When the cursor is anywhere on the window and right clicked there are a number of options relating to the serial port connection.

DSPMon_V1-53.exe

Trimble DSP GPS Timing Monitor screen shot

This is the program that's used with the Thunderbolt E (although the manual for the E version has information about another program). Once you right click and set the serial port to COM 1 9600 8-N-1 it works with the non E version. The Timing area has three data fields compared to the Timing Outputs area of the Thunderbolt Monitor program the new field is called PPS Quant error 0.0 ns. When run with a non E version Thunderbolt this field is always 0.0.

General Note on Software

Each GPS manufacturer has their own proprietary binary data format and it tends to stay the same for all their receivers. As each new receiver is introduced it may also have some new commands but the prior commands tend to be supported. So the latest version of the software tends to work with all the prior receivers.

Timekeeper

This software keeps the PC clock aligned with UTC.

Lady Heather's Disciplined Oscillator Control Program - readme.html - download

This program has a number of graphical outputs.

	Prior to doing the screen dump I pressed t (time and temperature control), Z (time zone) and GMST (Greenwich Mean sidereal time), but something is wrong since the time display is not incrementing every second, more like once every 25 seconds!
	To get this screen dump from a full screen image I pressed <CTRL><Print Screen> then F11 to go back to a LH window. But that caused 3.0 beta to crash and bring up the report an error window WIN XP SP2. Click on the image (17:14:47) to see the full size version.

K04BB Monitor

A low cost GPS Monitor - a hardware monitor

iCruze Monitor

This is based on A low cost GPS Monitor by K04BB.

	Available from fluke.l Power Supply Voltage: 5.0 Red=+ It turns out that this is one of the newer (Dec 2009 and later) units that has an on board 78D05 regulator on the samll PCB near where the 4 wires are soldered to the board. These units need at least 7 VDC to work and will work fine from +12 VDC.
	A T10 Torx bit is needed to open the case. The lower PCB is the 2x20 LCD and the daughter board is the custom uC for the Thunderbolt. To the right and above where the wires are soldered you can see a group of 3 pads, two rectangular with a trace between them. The 78D05 is there on the other side of the board.
Since the display is on a cable it can be located where it can be seen not where the Thunderbolt is located. This makes the display much more useful. Something has happened, the display is now blank? It is dead! I think either static or applying 6 VDC to the input (adjustable lab supply) since there's no input regulator. Working fine on on +12 V input. When I turned the supply down from 6 to 5 V that caused the 78D05 to drop out, turning it off.	The top line displays either GPS or UTC time & Date depending on how the receiver is set. The second line shows: No Message if it's not connected to the serial port of the Thunderbolt, or alternating: MODE: Normal ACT: Phase Lock RX: Overdet Clk GPS: Doing Fixes TEMP: 32.45 DAC V: 0.4874
iCruze Apr 2010 Broken Again April 2010 - the iCruze Thunderbolt monitor no longer displays No Message or anything else. This is the display and is no related to the RS-232 input. Don't understand why.	The problem may be static zap to the RS-232 connector since I was using it without the cover installed.
Tunderbolt GPS iCruze Monitor 8051 PCB	The Voltage regulator (Q2) is a 78D05 and D1, D2 & D3 are 000 Ohm resistors which means the C8051F330 uC may be powered by a voltage that exceeds it's specs.
Tunderbolt GPS iCruze Monitor Replacement CPU Board	This is the newest CPU board. Wouldn't you know that I removed the header pins from the wrong PCB. The wire colors are my best guess.

IC Table

The following ICs are on both sides of the PCB. If you can help identifying them and/or supply a URL to data sheets please contact me.

ID	Mfg	Function	Marking	Date
F2	GI	GPS BP Filter	GI M37
U01	CY	128x8 RAM	62128BLL-70SC	0402
U02	Trmbl	GPS uC	F.W.=3.00 48051-30
U05	ON	Logic	74HCU04A
U06	CSI	Logic	93C865	0341C
U07	generic	512kx8 Flash Memory	29F400BB-120EI
U09	DS	Thermometer	DS1620 "0310D1" = rev D ( good) the "E" rev has poorer resolution
U10	MACOM	27 dB Low Noise Amp	AM50-002
U14	TI	Logic	74AC174M	35AYC8M
U15	LT	V. Ref	LT1021-5
U17	Linear	Quad Op Amp	LT1014 ISW N08316	0401
U18	I	RS-232?	232IBE	41Z1L
U19	TI	Logic	74AC04M	3CC7EVK
U20	TI	Buffer Amp	OPA692
U22	Xilinx	256 cell FPGA	XC5202
U411	Infineon	GPS RF Tuner	PMB3332	0342
U460	TI		7301Q 3CT
U461	ST	Logic	74VHCT32
U999	Sawtek	Filter	855242	009410
Y1	32.768k	RTC xtal
Y2	3.6864	CPU xtal

External Oscillator Option

Trimble ThunderBolt GPS Timing Reciever PCB Back side

When you look at the bottom of the PCB (left side in photo at left) in the space under the OCXO there are two hole patterns like for a SMA, SMB, SMC or similar solder in coax connector. That would be consistent with a 10 MHz signal to the PCB and a tuning voltage signal to the external oscillator. Note that one of the TSIP commands allows configuring the tuning voltage parameters and is also part of the Tboltmon program.

Disiplined vs. Unlocked

When the oscillator is being disiplined it typically has more phase noise than when it's free running for some time intervals. Each GPSDO is different, but typically for time intervals shorter than 100 seconds you get better phase noise by turning off the disiplining. For plots see TVB web page Locked vs Unlocked.

Maintenance

Others have reported that a bunch of "?" appear and the unit does not seem responsive to the computer, but has 1 PPS and 10 Mhz outputs.
These were caused by bad joints relating to the RS-232 connector or the RS-232 IC.

The RS-232 interface is only 3 wires.

Pin	Function	Voltage
2	Data Out	-11.3V jumps once/second
3	Data In	0
5	Ground	0

Drawing

This is a pdf version of an Autocad drawing. ThunderboltDwg.pdf It's a test of the capability.

Product Idea

A DB-9 connector and shell holding a PIC micro controller that monitors the serial data stream and lights a LED(s) to show mode of operation, 1 PPS heartbeat and fault codes.
Note that the DB-9F connector only has TXD, RXD and ground connected.
There are no other pins used so a jumper might be added internally to bring out the 1 PPS, 10 MHz and to bring out a power supply(s).
Let me know if interested.

References

ThunderBoltBook2003.pdf, 142 pages) has many missing Adobe bookmarks, so that method of navigation does not work, so for now use the Table Of Contents.
Tboltmon.exe is the PC program to monitor and control the ThunderBolt.

Links

ThunderBolt FAQ, Power Supply Connections, Power Supply Noise, Phase Noise Comparison, Locked vs Unlocked, on TVB's web site
Johm Miles Notes on Trimble Thunderbolt performance and modifications web page confirms the external oscillator jacks.

Back to Brooke's PRC68, Products for Sale, Navigation, Military Information, Personal Home page

20708 hits since November 08 2008 page created 6 Nov 2008.