I don't use Odysee, so what was the magic thing Dave supposedly said?
Could MagicDAQ's isolation be inside the can? (like motherboards with built-in Ethernet phy components)
Could MagicDAQ's isolation be inside the can? (like motherboards with built-in Ethernet phy components)
Go to magic DAC website. It's Data Sheet does not specify a Hipot isolation voltage. It only states that the USB cable FG (Frame Ground) has a 1 megaohm series resistor to the MagicDAQ DGND / AGND.
See Quote 1:
The USB cable GND is not electrically connected to the MagicDAQ DGND / AGND. This ensures that measurements are not affected by the computer the MagicDAQ is connected to.
See Quote 2:
1M Ohm resistance between USB Cable GND and DGND/AGND
They do not understand the wording about USB isolation and what this supposed to mean in our field of test equipment.
Take their words at face value and you will get nailed...
All Dave has to do is continuity measure the USB GND pin, (not the usb connector GND, but the GND power out pin.) and any GND on the IO connectors. If it is anything less than 1 megaohm, then you know the USB is not isolated at all and 'Magic DAQ' might end up with some very displeased customers and potential fires or damage to your valuable PC if you take their word at face value.
I did not see true Optocouplers and proper HV clearance spacing on that PCB in Dave's video. In fact, the clearly visible 1M resistor next to the USB connector with a small parallel cap looked like an 0603. Clearly not designed to isolate any high voltage.
Anybody know where I can find the PCB edge clip thingos that are on the ceramic hybrid at 50:40?
I know there's castellation and clip on SIL/DIL lead frames but I'm looking for something that's surface mount and ideally has a bit of stand off. Looking closer the hybrid Dave is showing might have also had pins but they were cut off. Castellation is probably cheaper in large quantity but its a pain for smaller runs and rapid prototyping due to the increased cost and lead time for a castellated PCB. Would also prefer to have some additional standoff so I don't have to worry about a keep out area on the PCB beneath where the hybrid is being mounted or not being able to use the bottom layer of the hybrid. Not being through hole is also a strong want to avoid penetrating through and disrupting underlying layers.
I found these
https://leadframes.co.uk/surface-mount-leadframes/ which are pretty much exactly what I had in mind but they don't seem to be stocked anywhere?
Wouldn't this do?
https://www.ebay.com.au/itm/ADUM3160-B0505S-1500V-USB-to-USB-Voltage-Isolator-Module-Support-1-5Mbps-12Mbps/113799737310?hash=item1a7efe2fde:g:kTMAAOSwIsddFOC1
1500v usb isolation.
You cannot advertise USB isolation built into a product & then expect the user to purchase a separate third party device to get the function. (Not to mention will that module pass ~2watts of 5v power from side A-B...) Ok, say that MagicDAQ had that IC on the PCB. Well, that IC appears to have a Hipot isolation voltage of 1500v and a slow speed limit of 12megabit. Why hasn't MagicDAQ built for and specified that, or their own chosen isolation IC's isolation voltage capabilities?
Based on the writing in MagicDAQ's documentation, it would seem at one point they may have had the USB FG connected directly to their GND which may have introduced a ground loop interference because of poorly wired local test hardware on their side. Sort of like in the old days, we used to cut the third GND pin on our Amiga video editing systems to counter a 60hz hum bar scrolling through our video.
You cannot advertise USB isolation built into a product & then expect the user to purchase a separate third party device to get the function.
Oh, I see. Sorry. I was just trying to solve the problem not realizing the issue is the specs.
Anybody know where I can find the PCB edge clip thingos that are on the ceramic hybrid at 50:40?
[img]
I know there's castellation and clip on SIL/DIL lead frames but I'm looking for something that's surface mount and ideally has a bit of stand off. Looking closer the hybrid Dave is showing might have also had pins but they were cut off. Castellation is probably cheaper in large quantity but its a pain for smaller runs and rapid prototyping due to the increased cost and lead time for a castellated PCB. Would also prefer to have some additional standoff so I don't have to worry about a keep out area on the PCB beneath where the hybrid is being mounted or not being able to use the bottom layer of the hybrid. Not being through hole is also a strong want to avoid penetrating through and disrupting underlying layers.
I found these https://leadframes.co.uk/surface-mount-leadframes/ which are pretty much exactly what I had in mind but they don't seem to be stocked anywhere?
https://www.digikey.de/products/de/connectors-interconnects/contacts-leadframe/416
McBryce.
Thank you!!
Edit: Unfortunately none of the surface-mount type I am searching for.
Regarding the hybrid circuit: I have used ceramic hybrids a lot in my telecom circuit design days. Mostly the ones from Mitel (long gone; absorbed into Microchip after several rounds of acquisitions). It saved doing a ton of wheel-re-inventing analog design work and by having the modules mounted vertically they also saved quite a bit of space. I wish I had some pictures...
Anybody know where I can find the PCB edge clip thingos that are on the ceramic hybrid at 50:40?
[img]
I know there's castellation and clip on SIL/DIL lead frames but I'm looking for something that's surface mount and ideally has a bit of stand off. Looking closer the hybrid Dave is showing might have also had pins but they were cut off. Castellation is probably cheaper in large quantity but its a pain for smaller runs and rapid prototyping due to the increased cost and lead time for a castellated PCB. Would also prefer to have some additional standoff so I don't have to worry about a keep out area on the PCB beneath where the hybrid is being mounted or not being able to use the bottom layer of the hybrid. Not being through hole is also a strong want to avoid penetrating through and disrupting underlying layers.
I found these https://leadframes.co.uk/surface-mount-leadframes/ which are pretty much exactly what I had in mind but they don't seem to be stocked anywhere?
https://www.digikey.de/products/de/connectors-interconnects/contacts-leadframe/416
McBryce.
Thank you!!
Edit: Unfortunately none of the surface-mount type I am searching for.
Which are you looking for? They are called leadframe contacts, you can get them all over the place.
McBryce.
Regarding the hybrid circuit: I have used ceramic hybrids a lot in my telecom circuit design days. Mostly the ones from Mitel (long gone; absorbed into Microchip after several rounds of acquisitions). It saved doing a ton of wheel-re-inventing analog design work and by having the modules mounted vertically they also saved quite a bit of space. I wish I had some pictures...
The reliability of vertical mounting due to vibration in our application is a concern. This device is going in a vehicle. If we really want vertical mounting there are potential options like adding extra bracing...
Also our current layout works out pretty nice and compactly using a parallel mounted PCB/hybrid.
Which are you looking for? They are called leadframe contacts, you can get them all over the place.
McBryce.
Relinking the particular one I was looking for:
https://leadframes.co.uk/surface-mount-leadframes/As I explained in my previous post, looking for
surface mount lead frames in particular to help with routing in underlying layers. This is for a custom gate drive module in a resonant H-bridge power converter. I think we could get some conventional through-hole lead frames for now and manually bend them into surface mount for prototypes for now? then do a full MOQ when ready for production? but the MOQ is still way too high and quite expensive.
If it's for automotive purposes I'd recommend you consider using castelations on the Hybrid PCB and solder it pin-less to the PCB. Most of those leadframes won't be RoHS and will fail compliance when you get to homologation.
McBryce.