Author Topic: Advantest TR5825A Frequency Counter  (Read 909 times)

0 Members and 1 Guest are viewing this topic.

Offline pigrew

  • Frequent Contributor
  • **
  • Posts: 535
  • Country: us
Advantest TR5825A Frequency Counter
« on: January 27, 2018, 10:13:58 pm »
The Advantest TR5825 is two-input (100MHz, 1 GHz) frequency counter from the mid 1980s. Its standard model comes with a 5 MHz OCXO, custom LSI ICs, and has an option for an external keypad which can enable math functions.

One interesting feature of this counter is that the LF path can be switched between direct and reciprocal counting. This is disguised as a "square wave" and a "sine wave" mode on the front panel. In my experience, both work well for reasonable-amplitude signals, and I've not yet found a reason to use one mode or the other.

The unit presented herein has the standard-precision OCXO and GPIB card. I purchased it through eBay for about US$60 in parts-only condition.

There are two separate power switches. The front panel is a hard-switch controlling the front panel and counting circuitry. The one on the rear can be used to turn on the OCXO while the front-panel switch is off.

The OCXO is holding its frequency pretty well, to within a few ppb of a GPSDO over a week. One weird thing is that the counter reads about 30 ppb low right after startup (even with the OCXO warm). It takes about five minutes for the counter to fully warm up. This seems weird, as I'd expect the crystal oven to be the only thing that would require warming (at least for the LF path).

The only major issue, as it turns out, is that the keypad had failed. The switches are of the same design as the Racal 1992 counter. They have a rubber spring contact that becomes brittle with age. Note that the switches the Advantest are larger than the Racal switches, and contain square plungers. I repaired them by replacing the rubber bits with the spring contact and button of the ubiquitous 6mm square tactile switches, plus a bit of Kapton tape. It isn't perfect, but it's working well enough for the counter to be usable.

Another defect is that that there's a 13 kHz whine coming from the unit, which I've not yet been able to isolate. It is even if the front-panel power switch is off, so I think it must be related to the power supply or the OCXO circuits. I don't know if the unit always was like that (perhaps the design engineers went to too many rock concerts?), or if it's something age related.

Why is there a grid-pattern on the panel in front of the digital LEDs? It's distracting and weird.

« Last Edit: January 28, 2018, 12:56:54 am by pigrew »
 

Offline pigrew

  • Frequent Contributor
  • **
  • Posts: 535
  • Country: us
Re: Advantest TR5825A Frequency Counter
« Reply #1 on: January 27, 2018, 10:16:52 pm »
The Mostek MK38P70 CPU is a beautiful ceramic DIP with a piggybacked EPROM.

There is a switch to change the multiplication ratio of the HF path, and a mysterious button on the back of the front panel PCB. I'm scared to press it.

The HF board has a few ceramic SMD capacitors, quite modern for the 80s. I guess they were used due to better RF performance than THT.
« Last Edit: January 28, 2018, 01:01:33 am by pigrew »
 

Offline pigrew

  • Frequent Contributor
  • **
  • Posts: 535
  • Country: us
Re: Advantest TR5825A Frequency Counter
« Reply #2 on: January 29, 2018, 12:56:07 am »
I took an additional photo of the GPIB card. It has another MK38P70 processor connected to an Intel GPIB driver (P8291A) and a DAC.

The board has tantalum and ceramic capacitors (no electrolytic), and its own 10 V linear regulator (probably for the DAC).

I hadn't intended on this being much of a repair thread, but I should note that there is significant noise on the 5 V regulator at 26 kHz. I think the noise from the power supply is related to that. My current plan is to blindly replace the electrolytic capacitors on the power supply board.
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf