Fnirsi SG-003A & SG-004A Signal Generators

[Aldo22]

aside from the incorrect revision output, not much is working right off the bat.

Which parts are you referring to?
I played around with the Python demos a bit today and made some changes.
It worked perfectly for me.
Of course, I didn't test anything out of the ordinary.

Code: [Select]

> python3 example_usage.py

1. Reading current input signal configuration...
Current input signal configuration:
  Signal Type: 2 - Voltage

2. Reading current output signal configuration...
Current output signal configuration:
  Signal Type: 1 - Current

3. Reading output switch status...
Output enabled: False

4. Changing output signal to voltage mode...

5. Turning output ON...

Setting Output Value: 0.2
Current Input Value: 0.2031
Setting Output Value: 0.3
Current Input Value: 0.3028
Setting Output Value: 0.4
Current Input Value: 0.4044
Setting Output Value: 0.5
Current Input Value: 0.5013
Setting Output Value: 0.6
Current Input Value: 0.6019

[Martin72]

I tested this first with my Modbus Poll software.
That’s why I suggested to Fnirsi that they use the standard function codes specified in the protocol standards, so that standard software like Modbus Poll can be used to its full potential.
Right now, I’m only using the test board with a simple input console.
If I can’t get any further with that, I’ll take a look at the Python program.

[BillyO]

In a nutshell, how does this compare to the LB06?

I guess no one that has one of these also has an LB06.  Maybe I'll buy both and compare.  Who knows, I done dumb shlt before.   

[Aldo22]

In a nutshell, how does this compare to the LB06?

I guess no one that has one of these also has an LB06.  Maybe I'll buy both and compare.  Who knows, I done dumb shlt before.   

I only have the SG-003A (the SG-004A has more features, is more accurate, and more expensive), but based on the information available, I’d say these devices differ quite a bit from the LB02/LB06.

Here are a few main differences I’ve noticed:

The LB0x models can’t measure and output at the same time because they only have two ports. The SG-00xA models have four ports.

The LBx models have a basic LCD display, while the SG-00xA have a color display that can also show graphs (attached photo).

LB0x models can only output 0–11 V, while the SG-00xA can output 0–24 V.

The SG-00xA models are calibratable.

The SG-00xA models are programmable (e.g., increase the voltage by 100mV every second).

etc.

As I said, I’m not personally familiar with the LB0x models, but I’ve noticed these differences (without guarantee).
I’m sure the LB0x models also have certain advantages. For example, the LB02 is more accurate than the SG-003A, but SG-004A is also accurate.
The SG-003A is the least expensive of them all.

I think you have to compare the specifications carefully and decide which features are more important to you.

[Martin72]

I tested this first with my Modbus Poll software.
That’s why I suggested to Fnirsi that they use the standard function codes specified in the protocol standards, so that standard software like Modbus Poll can be used to its full potential.
Right now, I’m only using the test board with a simple input console.
If I can’t get any further with that, I’ll take a look at the Python program.

I've made some progress already.
My testing is based on Fnirsi's datasheet regarding the protocol and the Excel spreadsheet from the post:
https://www.eevblog.com/forum/testgear/fnirsi-sg-003a-sg-004a-signal-generators/msg5966599/#msg5966599
However, the table contains errors; I'm working on a correction based on tested results.
As I said, all of this was done using Modbus Poll's “Test Console.”
Here's what I've managed to do so far:
Read the firmware (even if the value is wrong—that's on Fnirsi)...
Read and write output values; read and control the output switch status.
Bit by bit, I'm working my way through Fnirsi's data protocol.
That way, you understand it better than if you just used a ready-made program, and you can modify it as needed.
Here is a link that converts the float values to hexadecimal, which you can then use in the data telegram.
https://gregstoll.com/~gregstoll/floattohex/

If you're using a terminal program that doesn't automatically generate a CRC, here's a CRC calculator tool:
https://www.lammertbies.nl/comm/info/crc-calculation

Martin

[KungFuJosh]

I've made some progress already.

TestController supports ModBus. If you really like this device, you could make a TestController driver definition for it.

[Martin72]

If you really like this device

 

Right now, I'm actually only interested in the device because of its Modbus interface.
The reason is that we (the company where I work) also offer a Modbus interface for some projects, and I had developed a test device (a Smart Display acting as a Modbus master) so that my colleagues wouldn't have to waste time counting bytes...
Of course, we use a standard Modbus protocol in accordance with industry standards; the Fnirsi device appeals to me precisely because it doesn’t have that.

[KungFuJosh]

Of course, we use a standard Modbus protocol in accordance with industry standards; the Fnirsi device appeals to me precisely because it doesn’t have that.

“It's the imperfections that make something beautiful. That's what makes it different and unique from everything else.”
― Bob Ross

 

[Aldo22]

My comparison above with the LB0x gave me an idea:

Is it possible to plot a diode characteristic curve using just two wires with the SG-003A without any additional equipment?

Yes!

- Output: V, Input: mA
- Programmed voltage rise
- Curve display

In the image: Yellow is voltage, blue is current.
The graph is a bit cramped because I chose a 0–4 V range; it could be made to look nicer, but you can see the point.

You might not "need" this, but an LB0x definitely can’t do this. 

[Kurt_!]

They ran out of yellow paint for my SG-004A?!

[Kurt_!]

I have another question.
My device arrived with a 60% charge.
I almost fully charged it.
But...it took a very long time!
A 5-volt, 2-amp charger was included.
It works (tested).
Unfortunately, the device only draws 1-2 watts.
My Ugreen charger charges at 10 watts.
The SG-004A is supposed to charge at 5 to 12 volts.

Is it normal for it to take so long and charge so weakly at 5 volts?

[Filippo52]

I can comfort you.
The included charger...it sucks.
But the supplied USB-A to USB-C cable sucks even more.

Just change the cable to significantly increase the wattage, up to 5W.
If you use a power supply capable of negotiating according to the USB-PD 3.1 standard, the device will negotiate 12V and you can get about 10W.

If you look at the power supply ripple on the oscilloscope, you'll want to change it too.

Filippo