PCI-DIO-32HS replacement recommendation

[prcngwsas]

Looking for recommended replacement for an older PCI-DIO-32HS (6533 family) card. The computer it's installed in is being replaced and the newer computer has no PCI slots (typical in current computer environment) and needs a replacement.

The application is a simple DIO using the DAQ ( library ) running under LV version 6.1 ( yes it's really that old) and the application is also being ported to newer version of LV.

Having plug&play cabling is desired.

Thanks

[santo_13]

What do you mean by simple DIO? is it static or dynamic? if dynamic, what sampling rate?

There is no direct replacement available, it is going to be redesigning the solution from scratch (mostly) to meet the requirements.

[crossrulz]

1. You will need to convert the code to use DAQmx instead of what is now called Traditional DAQ. It is not a simple 1:1 conversion. I would not say it is difficult, but you do need to be careful.

2. I am a fan of the 6509 DIO boards. They have 96 DIO lines. There is a PCIe version (PCIe-6509) and a USB version (USB-6509). These are good for static DIO. If you need the high speed, there are several options depending on how fast you need the outputs to be. I'm just going to point you to the DIO products and you can whittle down from there: PCIe DIO.

[prcngwsas]

Thanks- I was aware of the switch from the traditional DAQ to DAQmx, thankfully the code was architected to make that a few simple VI's to update. I have used the 6509 in various instances over the years, looking at what specifications I can find on the PCI card(6533) and application requirements, it looks like the application will require the minimum hardware clocked DIO card. The application is implementing a restricted JTAG 1149.1 protocol interface to the UUT.

Thanks again for the response.

[santo_13]

---

@prcngwsas wrote:

Thanks- I was aware of the switch from the traditional DAQ to DAQmx, thankfully the code was architected to make that a few simple VI's to update. I have used the 6509 in various instances over the years, looking at what specifications I can find on the PCI card(6533) and application requirements, it looks like the application will require the minimum hardware clocked DIO card. The application is implementing a restricted JTAG 1149.1 protocol interface to the UUT.

Thanks again for the response.

---

Do you know the clock rate the JTAG protocol is executed at?

[prcngwsas]

Not fast << 2khz- the product UUT isn't modern, it's a re-spin of an ASIC design, circa 1992, so the implementation is slow compared to anything modern.

The application is a very simple implementation of 1149.1 and was running on a PCI-DIO-32HS by essentially byte vector output/input. The same "protocol" is also used from a 486/DOS pc using an antique DIO-96. The device UUT has a "custom" digital return signal other than the 1149.1 defined signals.

It should be possible to implement with a 6509-type card- that certainly is a less expensive option than a NI6534/5.

The application certainly doesn't need a dedicated JTAG controller either, the data sent/received is a few hundred bytes to set/inquire internal states of the UUT.

Thanks.

[santo_13]

If your UUT doesn't depend on very strict timing and can work well with slower software-timed edges, 6509 would do the job.

6509 only supports software timed DIO, this means the fastest you can toggle the outputs would be typically in the order of 1ms i.e., 1kHz

[prcngwsas]

Thanks for the feedback- looking at the software I think the 6509 could work via software timing as that appears to be the way it's currently coded. I am still digging into the whole mess. There was no documentation of this system so I'm having to reverse engineer everything. I don't really know why the original developer picked the PCI-DIO-32HS card vs. the equivalent (of that era- circa 1998) 6509. It may be that they thought a hardware timed card was required but then that turned not to be the case. It may also be that the hardware buffer in the HS card allowed simpler software as the current routine simply writes a vector buffer out port 0 and then reads port 1 for the results data. Not sure what the 6509 input buffer capacity is, still looking at the card specifications vs. the PCIe-6535 (equivalent HS card to the PCI-DIO-32HS)

More research will be required. 

[prcngwsas · Accepted solution]

Thanks for the feedback, everyone. After looking at the application and hardware requirements and discussing with NI application engineers the PCIe-6535 is the best option. While the application doesn't need the more complex features of the card two items push the decision 1- the card can operate at 3.3v and the UUT re-spin works best at that level. 2- the input buffering of the card (and it's PCI-6533 predecessor) is what the current application depends on. Not wanting to totally re-write the application DIO sections, the "easy button" is to use the HS card.

Thanks again.