Author Topic: Replacement for the PM9610 Prescaler for Philips PM6654 High precision Counter  (Read 20151 times)

0 Members and 1 Guest are viewing this topic.

Offline SaabFAN

  • Frequent Contributor
  • **
  • Posts: 733
  • Country: de
Btw. Here's the schematic for the newest version. :)

Offline SaabFAN

  • Frequent Contributor
  • **
  • Posts: 733
  • Country: de
I finally got around to building up the new version after I had ordered the wrong chips (MSOP instead of SOIC).

Results so far:
- I goofed up the input of the first prescaler (again) |O : C22 has to be replaced with a solder bridge and pins 2 and 3 of U4 have to be connected with a 1 kOhm resistor. I have soldered a 0603 resistor across the legs, but soldering a resistor to C21 and the solder bridge that replaces C22 works too.
- Frequencies up to 2.89 GHz can be measured at input-levels of -30 dBm.  :-+
- Frequencies lower than 290 MHz cannot be measured reliably and anything lower than 350 MHz requires at least -10 dBm to be measured reliably. :--
  My investigation so far points to 3 possible causes:
   - Signal Source. I'm using the ADALM Pluto in CW-Mode and the spectrum looks horrible with powerful wideband spurs close to the CW-Signal, as well as strong harmonics. The output has massive jitter which gets really bad at
     frequencies below 280 MHz. At 100 MHz the jitter takes up almost 25% or a full period.
   - Oscillations in the second amplifier: I noticed 1.2 GHz parasitic oscillations that become quite strong and frequency stable when a finger is placed over U2, the output-trace of U2 and the Bypass-Caps C4 and C5. This is rather
   surprising,
     since until now I never had any oscillation-problems with the BGA2869 Amplifier and used it in several projects without any problem. Even mounted dead bug-style on a vero-board it didn't oscillate.
   - A problem with the splitter (R2, R4 and R5 arranged in a Delta-Configuration) and/or reflections from the RF-Detector for the standby-Circuit.
  I've also noticed a tendency to count harmonics if the input-signal has a frequency of less than 100 MHz. Interestingly, I've had absolutely no issues when I fed the prescaler a 24 MHz-Signal from my FeelTech DDS-Signal-Gen...

So 2 steps forward (higher sensitivity, frequency-range extended), one step back (Problems with the Divider-Chips again).  >:(

Sorry @ geostep and everyone else who bought one of the new boards. I really thought I had nailed it this time :(

Any ideas on how to fix or how to tackle this problem?
My big problem is that I don't have a good working signal source for frequencies between 24 and 900 MHz.

Offline SaabFAN

  • Frequent Contributor
  • **
  • Posts: 733
  • Country: de
Okay, I managed to get it working with a DDS Sine-Wave from my AD9910-Module (part of my Signal-Generator Project - yes, it is kinda working :) ).

The changes I have done so far:
  • Connected Pins 2 and 3 of U4 with a 1k Ohm Resistor
  • Replaced C22 with a Solder-Blob
  • Removed U2 and connected the input and output-pads with a 2 Ohm 0805 Resistor (fits almost perfectly and I was able to solder it without my magnifying glas)
  • Removed C23

With these modifications, a clean Sinewave works down to 80 MHz with Input-Power from 0 dBm down to -26 dBm. At 50 MHz the Output was rather jittery and the count was off, but the prescaler isn't supposed to cover anything below 50 MHz anyway. That's what Channel A and B are for :)

The reason for the removal of U2 is twofold: First the amp is oscillating (now there's no output even without diode D7 disconnected if there's no input), and second it is completely overdriven unless the input-signal is extremely weak. The result of the amp being overdriven is a horribly clipped signal that has a ground bounce-component that triggers the prescaler, resulting in jitter and counting frequencies twice as high as the input-frequency.

Standby-Circuit is still not working, though. And I don't have a second LM258 here to replace the OpAmp. I might have damaged it by powering the board with just positive voltage and leaving the negative supply unconnected - The Output of the OpAmp isn't changing at all regardless of the input-voltage. It's just driven against the positive rail all the time.
Also the SY100EL33 chips I am using (Mouser didn't have the MC100EL33) appear to be recognizing only the rising flank of the reset-signal and are not level controlled (datasheet isn't really clear about it and surprisingly rudimentary in general).
So since the outputs stay silent even if there's no input-signal most of the time, I would recommend to just leave D4, D7 and U3 unpopulated.

Tomorrow I'll repeat the measurements at higher frequencies with my HP8614. I expect it not to go all the way to almost 2.9 GHz any longer due to the amplification of U2 missing from the signal chain and also to have less sensitivity above 2 GHz.
I was expecting to run into a wall due to the limitations of the original hardware on the counter-board, especially the COAC-ASIC by Philips, at about 2.3 to 2.4 GHz, but that part is apparently capable of a lot more than advertised :)

Offline uski

  • Regular Contributor
  • *
  • Posts: 249
  • Country: us
Any update on this project ? :)
Specifically regarding the latest mentioned issues
 

Offline uski

  • Regular Contributor
  • *
  • Posts: 249
  • Country: us
I am thinking to do a same project for another similar counter (PM6665/PM6669). Slightly different board but very similar architecture.

What does B and TEST GO lines do ? I'm having some problems understanding what they are doing with the RF detection stuff

Thanks
 

Offline SaabFAN

  • Frequent Contributor
  • **
  • Posts: 733
  • Country: de
I've found some time to work on this again.

The boards I have work pretty well with the workarounds I've detailed in the last post when I use pure sine wave generators.
The ADALM Pluto unfortunately is completely useless as a signal source as there's always some kind of modulation on the signal, so I have no means to test the device with sine waves in the frequency range 400 to 900 MHz until I have my Signal Generator working.

The latest Revision has only one RF Amplifier and I'm revising the Standby-Circuit in a way that the board doesn't need the negative supply anymore, so no more damaged OpAmps when I supply only the positive voltage during testing :)

The "B"-Signal and "TEST GO" disable / enable the Output of the Prescaler-Board.
"B" adds a Bias-Voltage to the detected RF-Voltage or subtracts it from the RF-Voltage - I'm not entirely sure anymore ;) In any case, it turns off the output of the Prescaler when the Prescaler is not needed (in the PM6654C this is the case when Channel C-Mode is not selected).
"TEST GO" on the other hand pulls the other input of the comparator in the standby-circuit to a value that ensures the Prescaler-Output to be on under all circumstances for testing purposes.

Offline SaabFAN

  • Frequent Contributor
  • **
  • Posts: 733
  • Country: de
I just checked the Schematic of the PM6654C Mainboard again.
The B-Signal is left floating on that board, so it's not actually needed. It either is used when installed in a different instrument, or just there for testing purposes.

The TEST GO-Pin connects to the "DATA OPTIONS"-Connector "BU108" on Pin 24, but is not used by any Option that is listed in the Service Manual. So it was probably only there for testing during manufacturing.



Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf