chip-scale GaN fets

Certainly. My mfg computer is offline for updates. It will take a couple of days to complete and get it back online. I will be happy to post pictures zoomed in and out.

The Mantis lacks zoom capability. I cannot emphasize how important this capability is for electronics inspection.

I have an old American Optical AO569 Stereo Zoom microscope that has served well for the past three decades. It has zoom capability from 7X to 30X.

You set the focus at the highest zoom level for the object you are viewing, such as a pcb. Once the focus is set, you can zoom in and out without having to touch the focus again.

I generally have it set to 7X. This gives a broad area view with enough detail to check for gross problems.

Whenever I come across something that needs closer inspection, I zoom in to whatever magnification is needed to confirm there is a flaw that needs fixing, or that it is just debris or some other harmless issue. Then I zoom back out and continue working.

When I got the Dino-Lite, I discovered this capability was simply impossible to achieve. To change the magnification, you have to physically move the microscope closer or further away. This changes the working distance, which could be a minor or major inconvenience depending on the setup. But a far more serious flaw is the image goes out of focus.

Generally, it was so far out of focus that you don't know which direction to move the focus knob to bring it back in focus. The USB 2.0 interface was far too slow to handle the high resolution image, so there was a significant delay before the image appeared at the new setting.

This meant overshooting the correct focus setting, and more delay in getting the image focused. By the time I got it focused again, I forgot what I was looking for.

Even when the unit was focused, the image was not as clean and crisp as I would have liked. It was obvious the lens was poor quality, and I was not convinced the CCD was up to the standards needed for the work.

It only took a day or so to determine the Dino-Lite was totally unsuitable for electronics inspection, and I sent the entire unit back.

The Canon handycam solved these problems. The optical zoom operated the same as the American Optical. All I had to do was find the optimum set of closeup lens to give the working distance I wanted, and the Canon took care of the rest. I can zoom in over a 20:1 range, and the Canon keeps the image in perfect focus. I do not have to change the working distance to zoom, and there is no noticeable delay in the image response. The image is very high quality, as you would expect from Canon.

My thesis is zoom capability is absolutely essential for electronic inspection. If you do not have it, you might as well forget about finding problems that require high magnification. These include hairline cracks in solder joints or components, tiny solder bridges, flecks of debris from dirt or machining operations, and so on. The problem is even more pronounced with fine pitch ic's where the separation between the pins is very small.

Since the Mantis lacks zoom capability, I would recommend against using it in our work. A decent handycam is much less expensive. A few closeup lens and lens adapter will provide very flexible working distance and zoom range, and a simple mounting arrangement provides convenient and easy operation. Most decent handycams have USB 2.0 output. If this is fast enough for your application, it is all you need. Otherwise there are a number of HDMI capture cards available, starting at $99.

Are you sure about that nanosecond? Sometimes I have to swing large capacitive loads around really fast, maybe these GaN Fets can be used. I don't like this "BGA-style" mounting method but if they perform it would be ok.

[...]

Let us know how it performs on the bench, where the rubber meets the road.

--
Regards, Joerg

http://www.analogconsultants.com/

I wouldn't be surprised. Fets speed is mainly limited by:

• how fast you can drive the gate. Here you have the least series parasitics as possible, coupled with low gate swing and low capacitance

• series source impedance which introduces feedback on di/dt and this design again has the lowest parasitics here.

I've had a close look at these devices for a 30W/27MHz class E amplifier where it would have been an absolute perfect fit. Unfortunately my loads can change a lot (SWR almost everywhere) and those marvels aren't avalanche rated and they specifically say you can't do such things. I had to do that with a small SMD power mosfet, but it's

1.5W loss against probably lower than 500mW. Grin...

--
Thanks,
Fred.

I haven't tried it yet, but I'm practically certain that it would sink, say, 100 volts at a couple of amps in under 1 ns. I do that with small mosfets already, and GaN is much faster. We experimented with the Nitronex packaged GaN fets and they were awesome... but expensive and, at the time, not very reliable. These EPC things have (brilliant observation!) no wirebond inductance.

The gate capacitance of the EPC2012 is about 120 pF and it needs maybe

3 volts of swing. Some TinyLogic buffers would drive that nicely, and one might expect 300-500 ps edges at the drain. I'll try that when I get time. I'll have to lay out a PCB.

Now all I need is a use for it.

--

John Larkin                  Highland Technology Inc
www.highlandtechnology.com   jlarkin at highlandtechnology dot com   

Precision electronic instrumentation
Picosecond-resolution Digital Delay and Pulse generators
Custom timing and laser controllers
Photonics and fiberoptic TTL data links
VME  analog, thermocouple, LVDT, synchro, tachometer
Multichannel arbitrary waveform generators

The Mantis never goes offline for updates! And it turns on instantly.

How about this pic:

(not taken through the Mantis.)

--

John Larkin                  Highland Technology Inc
www.highlandtechnology.com   jlarkin at highlandtechnology dot com   

Precision electronic instrumentation
Picosecond-resolution Digital Delay and Pulse generators
Custom timing and laser controllers
Photonics and fiberoptic TTL data links
VME  analog, thermocouple, LVDT, synchro, tachometer
Multichannel arbitrary waveform generators

What is your point?

John,

I have some points to offer for your consideration.

You are the thirteen year old in the back of the class that does anything to disrupt the class to gain attention.

You have a pathological need to be the center of attention, and will do anything to achieve it. Regardless of the consequences.

The designs you offer to the group at best do not work. I have pointed this out on numerous occasions.

At worst, your designs gain the criticisms of others in the group, including Phil Hobbs, Spehro and others. You completely ignore the cautions and proceed with your extremely risky approach.

It seems that extreme risk has been your main guiding light all along. With your disastrous engine control experiment that resulted in total immobilization of an ocean-going vessel, your response was it was someone else's fault. They changed the design.

This is completely irresponsible. You did not put in the proper checks to verify the system was functional. There were no limit switches anywhere in your design. You did not check the operation before leaving the dock. You did not protect the interests of your customer.

This is the classic indication of a juvenile approach to design. You assume everything will work perfectly, and if it does not, it is someone elses' fault.

Another example is a temperature controller you designed and sold.

When it malfunctioned due to improper input, you stated "it dutifully followed the commands."

That apparently destroyed some very expensive equipment. But it was not your fault.

I submit otherwise. You need to protect your customers against all possibilities.

You should have had preprogrammed limits to the maximum temperature and rate of rise. If some instruction exceeded these limits, you should have shut down in a safe mode.

You are in a position to destroy your customer's equipment. You must do everything possible to protect it. If it fails, it is your fault.

Your designs have built-in flaws. When these result in loss, your response will always be it was not your fault. Someone else made the mistake.

This is the response of a thirteen year old in the back of the class.

Mike

Don't rent my name for your flame wars.

Cheers

Phil Hobbs

-- Dr Philip C D Hobbs Principal Consultant ElectroOptical Innovations LLC Optics, Electro-optics, Photonics, Analog Electronics

160 North State Road #203 Briarcliff Manor NY 10510 845-480-2058

hobbs at electrooptical dot net

Your criticisms were valid. They were ignored.

Mike

Your diatribe is just mean-spirited carping. Don't try to rent my name for it.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs
Principal Consultant
ElectroOptical Innovations LLC
Optics, Electro-optics, Photonics, Analog Electronics

160 North State Road #203
Briarcliff Manor NY 10510
845-480-2058

hobbs at electrooptical dot net
http://electrooptical.net

Not renting. Public domain. Your statement was an accurate enginering assesment of a flawed design.

You were correct.

Mike

Got an example?

Way over 90% of the boards that I design work first time, rev A etch, direct tp production without prototypes. These are serious 6 and

8-layer things, not some 555 toy.

What in the world are you talking about? The only immobilized ship I ever rode on was the first Avondale LASH ship on sea trials, and the problem was mechanical, a really interesting lockup mechanism on a lead screw in a DeLaval steam valve. I didn't cause it, but I fixed it.

I'll explain it if anybody's interested.

Now you are halluncinating.

Only partly. The customer's software told me to go to 3000 degrees C, so I clipped that to the max specified temperature and went there. Fried a couple of $30K NMR probes that, it turns out, couldn't handle the specified max temp. I worked with one of the end users at a university, not my OEM, to discover what was going on. These things happen some times.

Hindsight is cheap. The customer lowered the max temperature limit of the controller after it became obvious that the probes couldn't handle

400C.

I've designed thousands of circuit boards so far. Of course some had problems. And most of those were my fault. This is a complex business. My customers are used to things going wrong. What they appreciate is honest analysis and quick, effective fixes.

We have the usual 2 year product warranty. But we have an unlimited, perpetual warranty for design defects and software bugs.

Show us some stuff you've designed.

ps - more people here like the Mantis than don't.

--

John Larkin                  Highland Technology Inc
www.highlandtechnology.com   jlarkin at highlandtechnology dot com   

Precision electronic instrumentation
Picosecond-resolution Digital Delay and Pulse generators
Custom timing and laser controllers
Photonics and fiberoptic TTL data links
VME  analog, thermocouple, LVDT, synchro, tachometer
Multichannel arbitrary waveform generators

Ditto for moi, that's all I have to say.

Best regards, Spehro Pefhany

--
"it's the network..."                          "The Journey is the reward"
speff@interlog.com             Info for manufacturers: http://www.trexon.com
Embedded software/hardware/analog  Info for designers:  http://www.speff.com

Impressive. Not chip-scale, but have you seen Cree's parts?

Unmatched, 4GHz, 28V, 6A, Cgs=3D19pF.

$219 @ 1, DigiKey. (Joerg just flipped!)

-- Cheers, James Arthur

14 amps for 19 pF is awfully good, but the pricing needs work.

My other gripe about these RF parts is how little DC data they furnish. No drain curves, no transfer curve. You sort of slosh RF into the gate and expect more to come out of the drain. And what's the point of a noise figure curve at 28 volts and 400 mA?

--

John Larkin                  Highland Technology Inc
www.highlandtechnology.com   jlarkin at highlandtechnology dot com   

Precision electronic instrumentation
Picosecond-resolution Digital Delay and Pulse generators
Custom timing and laser controllers
Photonics and fiberoptic TTL data links
VME  analog, thermocouple, LVDT, synchro, tachometer
Multichannel arbitrary waveform generators

I don't know what the pricing is like on these if you're not buying a million but they're pretty good parts (I'm using eight similar devices boost supply).

Rumors have it that newer military planes might be able to launch gigawatts or terawatts of impulse RF at enemy things, enough to fry their electronics good. Apparently capacitive discharge into wideband planar antenna tiles all over the skin of an aircraft. Sounds expensive.

--

John Larkin         Highland Technology, Inc

jlarkin at highlandtechnology dot com
http://www.highlandtechnology.com

Precision electronic instrumentation
Picosecond-resolution Digital Delay and Pulse generators
Custom laser drivers and controllers
Photonics and fiberoptic TTL data links
VME thermocouple, LVDT, synchro   acquisition and simulation

I had to go to the page below and click the link to get that page:

19pF should be *very* fast.

14A? Imax =3D 6A. I guess Idss=3D14A could be safe for pulses.

Buy two and get the quantity discount! We're counting on you to buy them and do something insane.

Yep.

-- Cheers, James Arthur