Products > Test Equipment
HP 53310A Modulation Domain Analyzer -frequency microscope-
bozidarms:
Oh my God,
i was many times on that site and didn't see the service manual -
I'm getting old :-//
Tnx and regards
TurboTom:
Okay, I got the local oscillator board repaired again, but what the heck, this was quite a job...
Actually, there wasn't any particular component really faulty, it just appeared that the VCO module (Vari-L SVCO-1124) has slightly drifted over time towards higher frequencies. This must have shifted the frequency range that could be covered with the phase control loop filter just out of the required range at the low end. I measured that at 625MHz I couldn't get a lock but at 630 the PLL would lock perfectly stable. My approach was now to just shift the voltage at the VCO control terminal lower by about 300mV (there was enough margin there) without affecting the loop filter characteristics otherwise. In the path from the last OP27 to the VCO there's a voltage divider of 464 ohms in series and 1960 ohms to ground, resulting in a divison factor of 0.81. I simply replaced the 1960 ohms resistor by another divider of 2.2k to ground and 18k to the negative rail (-15V) of the opamps.
This actually did the trick and the PLL again locks over the full frequency range without a problem.
The option 31 is quite an improvement over option 30 (which is more or less included and consists only of a frequency divider for the input signal). Since option 31 is a complete down-converter, the instrument maintains the same accuracy as on the lower frequency channels up to 2.5GHz. Now my MDA needs to be "touched up" a little and finally I'll try to install the memory extension, this should be possible since there are sockets installed on the main (A1) board and also some jumpers labled 16k / 64k. When the chips that I think will fit drop in, I'll give it a try...
Anyway, if anybody "stumbles" across a component-level circuit diagram for the instrument, I'll be highly interested...
Cheers,
Thomas
TurboTom:
Here's how to install option 001 (extended memory) in a 53310A that came without it (if anybody's interested to upgrade his MDA):
You need 4 Cypress CY7C194-25PC (or Hitachi HM6708P-25 or Motorola MCM6208P25 or equivalent) 64k*4 SRAM chips (DIL24 casing) and one 2.54mm jumper.
Disconnect the connectors from the CRT display board and lift the board in its guide a few centimeters so the sockets lalbeled U68~U71 become accessible. Insert the SRAM chips in the sockets. Move the two jumpers XW2 and XW7 from the position "16k" to "64k". Place a jumper on XW8/SEL1. All these jumpers are located in close proximity of the SRAM sockets and can be easily accessed with the CRT board lifted.
After this, slide the CRT board back down and reconnect all the connectors. That's it, now your MDA should report the extended memory option installed.
Have fun,
Thomas
bozidarms:
As i have just said,
it took a short time, and 53310A is not only fixed but also upgraded.
Thomas you are truly Turbo, awesome!
Regards
maxwell3e10:
Someone on eevblog suggested the 53310A to me and I like them so much that now there are 3 in the lab!
But now I have a retro-computing challenge. There is 53305A phase analysis software that can still be downloaded from Keysight website. It works by getting continuous time stamps from 53310A over GPIB. From the time stamps it can calculate phase deviations and other parameters for a periodic signal. The software is circa 1994 and only runs on a 32-bit machine in 16-bit mode. It's designed to use a GPIB-AT card, so I haven't tried to interface it with the instrument.
The problem is that the 53310A programming documentation does not have any info on measuring time stamps. If I open the software with a hex editor I find many undocumented commands, such as CONF:XTIM:TST (timestamp?) and a whole bunch more.
I haven't tried to figure out the undocumented commands yet. Perhaps someone has an idea how to do it efficiently, maybe use a decompiler or a GPIB sniffer on a really old machine.
It would be good to recreate a modern version of this software. Not many frequency counters can measure continuous time stamps and the ones that do cost many times what a used 53310 does.
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version