Technical Topics



HP Z3816A GPS Frequency Standard Receiver

Z3816A front panel layout

Z3816A rear panel layout

The Z3801A has been more readily available from numerous sources, and appears to be more popular. The Z3816A is a newer version, and has some hardware improvements. It is very similar to the Z3801A, but with these major differences:

Front Panel
Serial Data


GPS RX Channels

Serial Port
Oven Oscillator
Power Required


10 MHZ
HP 10811
+27 or -54 VDC
Four LED
On-Off Switch
10 MHZ, 19.6608 MHZ
User Config.
MTI 260
27 VDC or 120 VAC

Expect variations from this chart.

The Z3816A took advantage of a newer Motorola GPS receiver that had eight instead of the earlier six satellite channels. Also, the serial port in the Z3816A is already set up for RS232.

In the Z3816A, HP used an MTI Series 260 OCXO instead of their larger HP 10811 double oven oscillator that was in the Z3801A. The MTI oven oscillator is a high stability/high reliability oven controlled crystal oscillator (OCXO). MTI states that it offers thermal stabilities from 2.0E-010 to 4.0E-009 over a 100 degree C temperature range, rivaling Rubidium atomic clock performance. They also mention the 260 Series has been used in many applications worldwide as a direct replacement for atomic clocks.

So, which oscillator is better, the HP 10811 used in the Z3801A, or the MTI 260 used in the Z3816A? Since a GPS
disciplined oscillator is compensated for long term crystal aging, all that remains is the temperature effect on frequency. The MTI oscillator design has better temperature stability, and is probably better on phase jumps, which is mainly what matters for GPS timing applications. The HP 10811 dual oven oscillator is a super low phase noise floor design, which is probably not important for most hams. This makes the MTI oscillator the preferred choice.

Since the Z3816A has these hardware advantages, many would choose it over the Z3801A. However, I doubt that many amateur users would be able to measure a difference in their respective performances. For most, the Z3801A performance is substantially better than what they used to have for a frequency or time standard.

Z3816A top detail

Z3816A Main Board Connector Wiring
for both DC and 115VAC input versions

Z3816A power supply pinout and typical readings data

Pins numbered right to left, starting with pin 1:

(1&2) Red: +5V
(3&4) Black: Gnd
(5) Yellow: +15V
(6) Black: N/C (no connection but DC version has a wire)
(7) Blue: -15V
(8) Purple: N/C
(9) Green: N/C

The 20-70VDC input version has corresponding pins and wires (one to one correlation).

The 115VAC input version has 10 output pins -- all the connections are the same except the -15V which is on pin 10 of the power supply (the main board is, of course, the same as above.)

The DC input version draws about 1.2A on the +15V at initial power up due to the oven startup load. This drops to about 0.8A after the oven reaches temperature. The 5V load is about 0.6A and the -15V current load is not enough to register on a bench supply meter.

(Thanks to Gil Porter and Mitch Janoff.)

Download Z3816A ROM binary dump here. (Use HEX editor to read.)

Antenna Alarm Problem on the Z3816A

A frequent problem with this receiver is that some antennas cause the internal antenna current alarm to activate. Although the antenna is actually working, the receiver reports an error.

The Z8316 uses a Motorola M8 Oncore GPS receiver. This version uses the antenna current info in the status word in the data messages.

A simple work-around is to simply shunt the antenna connector with an external resistor. Nominal values for antenna current are between 15 and 80 mA. Values both under and over these cause a under-current or over-current flagging, so the selected resistor for current drawing should be chosen not to draw more than 15mA so the alarm works if the antenna really becomes open. 680 ohms seem to be a nominal value, but you may need to adjust this depending upon your specific GPS antenna.

Have trouble getting the Z3816A
to communicate on the serial port?

If you have GPSCon software, please review the help file of GPSCon under the section "Using the Z3816A". For those that don't have this it is copied below.

GPSCon should automatically put the Z3816A into SCPI mode, and take it out of full duplex mode, each time it is started. This is not necessary for other models of receiver, but the ID string cannot be examined until the receiver is in SCPI mode. So GPSCon will examine the response to the *IDN after having gone through the sequence of sending

If the model is determined NOT to be a Z3816A, this fact is flagged in the registry so subsequent starts will not invoke the commands. If you wish to do this manually, procede as follows: When the Z3816A is first used, it is possible that it is in "time of day" (TOD) mode. To get it out of this mode, press the "Stop" button. Then find the command "PTIM:TCOD:CONT 0" in the drop-down list box. It will be at the bottom of the list because it has no leading colon. Press "Send" to issue the command.

The second task is to ensure that the receiver is not in full duplex mode. This is not essential but it does reduce the amount of serial traffic. Find the command ":SYST:COMM:SER1:FDUP 0" and send it by pressing "Send". Hopefully the receiver is now in a state where it can respond to normal commands with the "scpi > " prompt. Press "Start" and the polling sequence should commence.

Additional notes by Craig McCartney, WA8DRZ

  1. -15 VDC from the power supply is not needed: On both of my units there is no -15VDC on the blue wire in spite of the fact that the p/s board silk-screen mentions this voltage. When I found this on the first unit I thought I had a bad power supply. I troubleshot it to the point that I found the components for the -15 output were obviously missing and were never populated on the circuit board by the factory. Apparently the Z3816A main board (at least the two that I have) does not need this voltage. The rear panels of both of my units are marked 20-72V DC near the power input connector (an XLR type normally used for microphones). When I fired up the first unit on the bench I used 24VDC. After a while, I tried it on 48 VDC and it worked just fine. I have not confirmed operation over the entire range of 20 to 72 volts, however.
  2. I was surprised to see that MTI oscillator is a 5.0 MHz unit, not 10.0 MHz as I expected. It is clearly labeled "FREQ 5.000000E+06". The MTI part number of mine is 010395-0558. However, the 10 MHz output on the rear panel is really 10 MHz and not 5.0 MHz.

Copyright © 2001-2016 Bill Jones, K8CU