Actions
Forum Thread Details
  • Replies 36 replies
  • Subscribers 2433 subscribers
  • Views 3436 views
  • Users 0 members are here
Related

Seeking Your Interest in Roadtesting a Digital Isolator Board

rscasny

Hi All.

imageI had a call with one of suppliers who is interested in roadtesting one of its digital isolator boards. We have not done much with digital isolators in the roadtest program. But given the need for circuit isolation and the safety they provide, I wanted the community to roadtest this product. In addition, another reason was the learning experience the roadtesters can provide community members on this product.

Here is some basic information on this board:

The evaluation kits are designed to evaluate the MAX22563−MAX22566, a family of reinforced, six-channel, galvanic digital isolators in a 20-pin SSOP package.

Two types of evaluation boards are available:

  • The MAX22565CAEVKIT# is fully assembled and tested, and comes populated with the MAX22565CAAP+ (Not used in this roadtest)
  • The MAX2256XAEVKIT# is a generic board that has U1 unpopulated, allowing to select a device from the MAX22563−MAX22566 family. This is the kit that will be used for this roadtest. The MAX2256XAEVKIT# comes with U1 unpopulated and supports the following digital isolators: MAX22563BAAP+, MAX22563CAAP+, MAX22564BAAP+, MAX22564CAAP+, MAX22565BAAP+, MAX22565CAAP+, MAX22566BAAP+, and MAX22566CAAP+.

imageEquipment needed for roadtesting:

You need some test equipment to roadtest the MAX2256XA EV kit:
• MAX22563−MAX22566 device if U1 is unpopulated on EV kit. (The roadtester and I will get together to order the desired device needed.)
• Two DC power supplies with output range of 1.71V to 5.5V
• Signal/function generator
• Oscilloscope
Note: The EV kit should be powered from two independent isolated power supplies with nominal output voltage in the range from 1.71V to 5.5V. To evaluate the electrical parameters of the device without any isolation between the two sides, a single power supply can also be used.

Element14 would not provide the test equipment.

Roadtesters would need to have the bench equipment for this roadtest. If you are an electronic design engineer, you likely have access to most of these equipment. This seems like a lot to ask, and it is. But I think this is a valuable roadtest to offer the community. The reviews the roadtesters write would not only be good learning experiences, but would also provide customers more than a datasheet to select a digital isolator when the need arises. 

Here's some documentation on the board:

Datasheet
Manufacturer's page

For Additional Information

Here's a couple of polls for which to vote:



  • Are the outputs of signal generators generally isolated from Earth ground? I think I have always assumed they are Earth-referenced. If so, I suspect that's going to affect isolation results when the other side is connected to an oscilloscope. 

    If this becomes a test, I am going to have a strong interest in seeing the results!

  • in reply to baldengineer

    My Multicomp signal generator signals are definitely earth grounded. I am about to start a Digilent Analog Discovery 3 road test - I will check out its ground situation - running from a battery operated MS Surface. However I also have a hand held battery operated signal generator/scope that obviously is not connected to earth ground. I even have a hand held DVM that has some signal generator capability.

    Although I have enough equipment, I am still sorting out what I might learn from this road test.

  • in reply to dougw

    There is no question that the Analog Discovery 2 and 3 are Earth-referenced. Nothing on either is isolated.

    The AD2's published schematics show that. (Digilent decided with the AD3 they would no longer share that information.)

  • in reply to baldengineer

    I have a mix of ground referenced and floating generators.

    This one is very handy: 

    image

  • I think the X kit without chip soldered is complication for RoadTest because it need RoadTester to have SMD soldering skills and can introduce several issues to board (cold joint, shorts, unconnected joints). It increases requirements on needed equipment even more and also some faults can affect board performance. MAX22563−MAX22566 differs in direction of channels. I think the MAX22565CAEVKIT with presoldered isolator with 5ch in one direction and 1ch in opposite direction is better choice for RoadTest. And if somebody for some reason need different channel configuration he/she can simply desolder the default chip efficiently turning kit to X variant and proceed with custom chip.

  • Reminds me of this: /members-area/b/blog/posts/debugging-the-turing-welchmann-bombe

    They missed a trick by not providing an isolated DC/DC 5V voltage regulator to power both sides so only one supply is needed.  That would have been useful even for an evaluation board but if someone wanted to lump out for one to participate, they could jury rig a power input for a standard wall wart.  

    The roadtest could be achieved without a signal generator if it's possible for the tester to connect to a piece of equipment that doesn't share the same earth ground, like the Bombe.  It would need to generate fast signals to do the roadtest justice though.

    One other requirement for roadsters is of course the ability to solder a 20-pin SSOP package.  From the photo, the position looks accessible to get a soldering iron in, but hot air gun would probably be more useful.

  • "Two DC power supplies..."

    Wouldn't a tester need a third to establish the voltage difference betwen the two sides? If you just let bench supplies float, small amounts of leakage will slowly pull the two sides together, and testing is then a bit meaningless.

    The MAX22563 can manage up to 200Mbps. So that also means a reasonably good oscilloscope and generator if you're going to look at the top end.

  • in reply to baldengineer

    I'm surprised that the AD2 and AD3 can't be powered just from the USB from a laptop, but it does seem that they are intended to be used with a separate 5V supply.

  • SMA connectors are not the most practical way to fit the hobbyist test bench. Perhaps geared toward a high end lab. In a modest home lab, I would expect the evaluation connecting an MCU and a peripheral, perhaps a serial port. So a simpler and pre-populated board with headers is a better fit.

    On the isolation voltage - most of our test benches do not have a way of having 50V or more to test the isolation. It’s another opportunity (and a risk) to get a supply that is going to 100V or so. 

  • My speed is a XR2206 signal generator and a USB scope. Good enough to play with, not enough to do a good test on corner points.