Ok. Back again, for my 2nd try at this (first one was wiped by the board software after 15 minutes typing it out!).
I picked up a number of things but the highly desired PIC chips had not yet arrived. Disappointed. I know they're overdue and should be there any day now. So I'll be calling in and going again as soon as they are there. I had to go this time due to other demands and was hoping they arrived before me. No luck.
But the good news is from forum member
chaoztc who, just before I left, bread boarded up a couple of the same PICs and tested out the data signalling. He had good results with prelim code that transfers bursts of 104 bytes from one PIC to the other using my planned methods (software output, NOR gate, delay and UART input). Here's a few photos from him:
Breadboard showing NOR gates and PICs.
Burst of data between PICs.
Clock (yellow) and data (green) after NOR gate and RC delay going into UART.
This is great advance work and he sent me the sample code to work from going forward. I believe he's also using some info I sent his way to optimize for faster speeds.
The IR Thermometer arrived and works great. Thank You, Bicknellski.
No heat sinks qrrived but I did check a few shops in Bkk and got some useful stuff there. I have some small heat sinks for the K1 now, and a larger double size 115mm x 254mm x 30mm heat sink for the K16. I'll try mounting 2x K16 on this but it could be cut down for 2 units. This would work using the method where PCB mounts against heat sink as support member. They're going to get me a price on 400mm long for a 4x bar. The way these places work - I know they are buying from another local source for about half the price. I paid $6 for this double heat sink, so that's $3 each. I bet some wholesaler in Bkk has them for $3 each, $1.50 per K16. Hard to find out where. Will try to post some photos later but the ones I got look much the same as other posted here.
I picked up several tools and fans, and various bits which I'll be testing out soon. I expect to order the scope today thanks to KS who arranged a bitcoin sale for me and made a large donation in US$ to get this going. For as much as I love Bitcoin it's still a hassle to transfer in/out of $$.
How big of a performance hit would there be if a software clocking algorithm took 500uS to 600uS to load the Avalon chips?
Thanks
A 300uS loss on 0.9 S work cycle (assuming 16 chips) would be 0.03%, so even if I can't get it fastest possible choaztc has already shown a speed that will make other friction much greater than not being able to push data at full speed. I'm confident we'll get quite close and if the ASIC is super picky about input timing I have a back up plan that may get us bang on 125nS timing, but is a bit tricky software wise.
Air? Any regular heatsink with a regular, slow spinning fan should be more than enough. A 1K/W heatsink isn't anything special, and that's with no fan! So, basically, as long as you have any aluminum piece with fins on it and have a little air blowing at it too, everything is sweet! :-)
In hindsight the whole calculation isn't really necessary. I didn't expect it to work out that nicely.
Ente
Spot on Ente. I think from your calcs and the Avalon photos we can be pretty sure that a heat sink is needed, and air flow is needed, but after that it's not going to be critical just how much or how big the fins are. Good board -> heat sink contact is our main concern but I think we'll be fine there.
