Access Time of PROM, SRAM, and DRAM on an HP 3588A Memory Board

[Smoky]

I have a question about the memory board in an HP 3588A Spectrum Analyzer.

There are 16 unpopulated locations where I plan to experiment and add identical DRAM.

What's on the board are ten PROM chips that have a max access time of 200ns.

There are two SRAM chips that have a max access time of 120ns.

And the 16 pcs of DRAM have a max access time of 100ns.

Is there any benefit to replacing the two SRAM chips with ones having a max access time of 70ns? or would this cause instability?

Thanks 

[Smoky]

I'm going to go out on a limb here and take a stab at the answer.

I went to the CPU board of the 3588A and I found a chip, a TMS320C25FNL, a digital signal processor.

I read through the datasheet and it's described as having a 100ns cycle time.

I'm guessing that a SRAM with a an access time any faster than 100ns would be no benefit.

Am I even warm

[srb1954]

I'm going to go out on a limb here and take a stab at the answer.

I went to the CPU board of the 3588A and I found a chip, a TMS320C25FNL, a digital signal processor.

I read through the datasheet and it's described as having a 100ns cycle time.

I'm guessing that a SRAM with a an access time any faster than 100ns would be no benefit.

Am I even warm

No. The TMS320C25 DSP chip probably has its own separate memory immediately adjacent to it for fastest speed. This memory has to be much faster than 100ns, typically 30ns, if the DSP is running at a 10MHz clock.

The memory on the PCB pictured is likely to be only for the main application processor, which will run at a slower memory cycle time than the DSP. The cycle time will be determined by the main processor clock so just putting faster SRAM chips in is not going to speed things up. You can generally substitute faster SRAM without any issue but these SRAM chips appear to be battery backed so it is important that the replacement chips have sufficiently low standby power to avoid the battery going flat prematurely.

HP designs of this era often stored critical configuration and calibration data in the battery backed RAM and if this data was lost the instrument becomes uncalibrated or even unusable or severely curtailed in functionality. It would pay to check the battery voltage while you have PCB out.

[Smoky]

Thank you SRB1954 

The battery was replaced due to the original having lost its power and it probably means that all calibration data was lost before I took ownership.

Even though the analyzer passes its internal calibration routine, it's probably off.

The relays can also generate a calibration fail message and someone recently was able to adapt newer relays to resolve the issue. I have those adapter boards and relays ready and waiting to be installed.

I do not plan to replace the SRAM but I intend to add the additional DRAM.

The service manual has a step-by-step procedure to calibrate the analyzer so the future still looks bright for this machine