gotmc / ivi Goto Github PK

If a single driver is going to handle multiple models, I'll need to implement one or possibly multiple of the following new functions:

1. Have specific new functions for each supported model, such as NewE3631A(inst) instead of New(inst)
2. Have a NewQuery(inst) function that queries the instrument and confirms it is supported.
3. Have a New(inst, model) function that specifies the model of the instrument to create.

In all of these, how do I handle passing in an option to reset the device? Is that always a boolean that is provided? What if I want to specify NewE3631A(inst) and I also want to query the device to ensure it actually that? I could simply mandate that a query is always performed.

I plan on standardizing all HP and Agilent equipment to be Keysight since that is the name now being used by the company.

Section 6.16 Initialize in IVI-3.2: Inherent Capabilities Specification states that the Initialize function has an IdQuery boolean parameter that queries the instrument for its ID (*IDN?) and verifies that the IVI specific driver supports the particular instrument model.

Right now all SCPI command strings include \n as the command line terminator. What if an instrument doesn't expect a newline as the SCPI terminator? For the Prologix GPIB, the line terminations are stripped by the GPIB controller and then a terminator is added based on the ++eos command setting.

One option is to include a terminator that can be set for the instrument instead of just hardcoding \n. Ideally the IVI driver has the proper line terminator for that driver. Although, I don't know if the line termination could need to change for the same piece of test equipment based on the interface used to communicate (i.e., GPIB vs LXI vs ASRL vs USBTMC).

The DS345 driver leaves the internal trigger set to single when configuring a burst currently.

The signature for the New function should be changed from New(address string) to New(address string, reset bool) so that users can select whether or not to reset the device when creating a new IVI instrument.

The inherent methods that call *IDN? should only call *IDN? once and then cache all results—mfr, model, s/n, etc.

As far as I can see DC loads are not covered by IVI.

Would you accept a PR anyways?

I guess there should be "VOLT" in the voltage function and CURR for the current function?

Refactor code and add functionality to the IVI driver.

Need to have timeouts for connecting to devices using VISA or directly using LXI or USBTMC. This might be something that needs to be added to the other packages.

Should we replace the Instrument interface with io.ReadWriter interface? Stated differently, do we really need the StringWriter and Querier interfaces?

type Instrument interface {
	Read(p []byte) (n int, err error)
	Write(p []byte) (n int, err error)
	StringWriter
	Querier
}

type StringWriter interface {
	WriteString(s string) (n int, err error)
}

type Querier interface {
	Query(s string) (value string, err error)
}

I think the only reason I'm using the StringWriter and Querier interfaces is to handle automatically adding the command termination. Is that true? What about using the Prologix GPIB controller? Would that work without the StringWriter and Querier interfaces?

Even though the IVI Foundation hasn't defined a standard for DSAs, I need to create one based on the 35670.