I don't know what IGBTs or MOSFETs are restricted, but you could attempt purchasing a krytron through ebay.
Hi Dave,
Regarding the motherboard you will need for those 2 new Xeons, I recommend you the Supermicro X9DAi board
http://www.supermicro.nl/products/motherboard/Xeon/C600/X9DAi.cfm.
It has all the features you need - It supports DDR3 non-ECC RAMs, has a sound card, has SATA III/6Gbps, has 3 16x PCI-E slots.
Your power supply is enough to power this thing, you do not need a new one. However, you need a reduction from Molex to 4/8pin CPU power connector. Those boards require 2 4/8pin CPU power connectors, your power supply only has one (like 99% of desktop power supplies). Also check that you can fit the board into your chassis.
TO-BUY list if you choose this board:
- Motherboard (obviously
) - Molex to 4/8 pin CPU (something like this http://www.directron.com/ad202.html)
- Possibly a new chassis. Depends if you can fit the EATX board into your chassis.
- CPU Coolers
Had this Casio FX790P programmable computer/calculator in school more than 20 years ago, bought from Tandy.
Programmed a few little but quite neat programs with it in Basic.
Also own the cute little thermal printer that came with the deal.
Unfortunately it broke down on me some years after (flatcable between the two halves is VERY sensitive) and is now lying in multiple pieces waiting for a, long forgotten, repair.
One of my favourite reads on nuclear weapons history is this:
http://blog.nuclearsecrecy.com/Restricted Data is a blog about nuclear secrecy, past and present, run by Alex Wellerstein, a historian of science at the Stevens Institute of Technology.
A really interesting perspective from a talented researcher.
Re: Vibration resistance of SIP components in the Subaru ECU.
I suspect that the module was made to mount vertically (like on the firewall, etc.) and with the two big tabs on either side, so the giant connector is at the bottom.
In this orientation, note that all the SIP components/modules are vertical, so the majority of vibration will be vertical also (from bumpy roads, etc.)
I suspect this wasn't by accident that they are all vertical. Maybe even those can radial electrolytics have their pins arranged in that orientation as well.
O Level sounds distinctly like a vestige of British colonial rule; we used to have O Levels (at 16) and A Levels (at 18).
I don't understand why you would call someone "not a proper pcb designer", because of old school square routing, Dave. :/
It might even be a software limitation, I have redrawn pcb's for Bosch made in the 80's that where all square routing and they demanded that I kept the square routing, low speed signals, what's the problem with not having 45º corners?
All the electrons will crash in the turn?
Re: Vibration resistance of SIP components in the Subaru ECU.
I suspect that the module was made to mount vertically (like on the firewall, etc.) and with the two big tabs on either side, so the giant connector is at the bottom.
In this orientation, note that all the SIP components/modules are vertical, so the majority of vibration will be vertical also (from bumpy roads, etc.)
I suspect this wasn't by accident that they are all vertical. Maybe even those can radial electrolytics have their pins arranged in that orientation as well.
It's also worth noting that engine controllers from this era, were mounted in the vehicles main cabin, rather than underbonnet, and hence have a much lower requirement for environmental sealing and vibration resistance. (being bolted to a ton of body shell automatically limits the peak operating g the unit can undergo!) Modern units are now underbonnet or even engine mounted to reduce the wiring loom cost and improve the engine integration aspect where one engine family is used across multiple platforms.
Modern devices are entirely SMC and use drivers in packages like SOIC-PowerSO with a large soldered tag underneath for heat sinking and mechanical security:
For example here is an early 2000's EDC17 Bosch ecu:
/Bosch_DCU17HD01-4_41_Engine_Control_Module_(iSi)_-_Main_PCB_Top.jpg)
being bolted to a ton of body shell automatically limits the peak operating g the unit can undergo!
Yes, it wouldn't be about peak G's, it would be about hitting resonant frequencies of free standing packages from broadband vibration.
I've seen free standing TO-220 packages vibrate clean off from just wheeling up and down a production floor for a few weeks on a metal trolley.
Yes, it wouldn't be about peak G's, it would be about hitting resonant frequencies of free standing packages from broadband vibration.
I've seen free standing TO-220 packages vibrate clean off from just wheeling up and down a production floor for a few weeks on a metal trolley.
And that TO-220 device is mounted "horizontally" which would make it MORE vulnerable to vertical vibration.
The pin pattern was in a triangle, but the leads were formed to that everything above the surface of the PC board was parallel.
It would probably be speculative to assume they designed it that way so that the springiness of the leads absorbed the vibration from the mass of the package.....
I was also very surprised it wasn't conformally coated. But if it was designed to go inside the passenger compartment, that might explain it.
Had this Casio FX790P programmable computer/calculator in school more than 20 years ago, bought from Tandy.
GeI was the one who sent it in. It used to belong to my dad, he wanted to throw it away but I told him I knew somebody who collects calculators. I showed him the video; he was convinced that his one was a bit different than the one in the video - until I told him I send Dave the calculator.
The calculator was bought in Singapore by my uncle who was on his way to Australia (or came back). The calculator has been all over the world! By the way, I know that the Australian import/export rules are strict, that is why I didn't include the batteries, just to be sure.
And about the rectifier: yes, each pair of mosfet rectifiers face the same way because I intended use of a heat sink. If a heat sink is actually necessary depends on the current drawn and on the mosfets. The major source of heat will be originate from conduction losses, so the Rds(on) is the most important parameter. Just if someone wants to know!
Regarding the motherboard you will need for those 2 new Xeons, I recommend you the Supermicro X9DAi board http://www.supermicro.nl/products/motherboard/Xeon/C600/X9DAi.cfm.
I'll second that SuperMicro recommendation - we build quite a few systems round that platform, the
X9DR3-LN4F is our preferred board, but we're typically looking at it from an IO standpoint.
Trap for young players - remember to check what warranty you have on the board - the 3 year warranty is an optional extra.
I was hoping that the package from Digilent had a Zybo instead of the Basys 3. I've been wanting to play around with the Zynq 7000 SoC and I've been eyeing the Zybo because it seems to be the cheapest dev board for the Zynq. I've also been looking at the Altera Cyclone V SoC development board from Terasic which seems to be a bit more bang for the buck.
When you do the video on the Digilent board can you consider the question of why you might consider the new chip over the old? From a beginner learning POV. I know the new one would be faster and better for new design work, but if you wanted to buy a second hand one, should you?
If you just want to learn FPGA's, makes no difference at all.
I see my item in the background, at least I know it arrived safely! O0
When you do the video on the Digilent board can you consider the question of why you might consider the new chip over the old? From a beginner learning POV. I know the new one would be faster and better for new design work, but if you wanted to buy a second hand one, should you?
If you just want to learn FPGA's, makes no difference at all.
The only difference for the Digilent boards is that the new one is supported by their new IDE.
The ones with Spartan 6 and older FPGAs are not supported.
You need the Xilinx ISE development environment for them.
It's a good and capable IDE, but active development has ended.
it's still supported by Xilinx.
Regarding the motherboard you will need for those 2 new Xeons, I recommend you the Supermicro X9DAi board
Someone has kindly donated one of these boards ( I can pick):
...
I was going to suggest a cheaper way of getting more cores for video rendering than using Xeons but if people are donating the CPUs
and motherboards then I guess there's not much point.
