Topic: Klondike - 16 chip ASIC Open Source Board - Preliminary - page 68. (Read 435369 times)

The outer shield is connected to direct to GND. But you gave me an idea.
I haven't seen it happen again after changing the USB cable to the high quality one from my Canon camera that has a ferrite core near the board connector. I was using the other one because it's short and less clutter on my work bench.

I've wired in the inverters now. The signal has very sharp/fast rise times now and it's easy to see the clock delay is 50nS. I tested it manually in ktest and got all GOOD nonces up to 420 MHz doing repeat work filling the queue . At 430 MHz they came back BAD.

Unfortunately on cgminer the error rate hasn't improved. I haven't run that long yet but I'm seeing 3% - 5% error rates. I believe this is now related to noise on the PLL and instability. It can go quite a long time with no errors, and just when I think Ok, it's down to 1% now, but then a bunch pop up and it's back to 3%.

I'm moving onto final board revisions now. Hope to have a new K16 release later today and k1 tomorrow sometime.

I haven't really tested the O/C range yet. I do brief tests up higher but not any long runs. I'm waiting on a set of spare leads for my Ammeter to cut into a PCIe cable for current measurements.

Thanks. The AUS power cords aren't much use here due the curious angled blades. My sister lives near Brisbane but I haven't been to see her in years. I hope to get down there again before too long.

I'm going to order a couple RFI power line filters from Mouser in my next order. There's two kind I looked at but I'll probably cheap out and order the $4 one instead of $22 one.

10VR3 $22.46

16WGF1 $4.40

If there is anything you need that can be sourced in Australia, let me no.. Am flying to BKK on the 21st and can bring it with me...

Reading through burnin's and others designs it looks like these chips probably exceed 2A each if you overclock more than a little bit.  Have you had any results running say, 8 chips on a 16A regulator (or 16 on 2x 16A regulators) and overclocking/overvolting?

Your buck is at 600khz while avalon's is at 300kHz, both spec'd at 2A/chip, so you should have somewhat better OCability, unless the current is the first limit you hit.

How have you connected the outer shield of the usb plug to gnd?
I have often seen at various hubs - that the other shield is also connected via some ferrite beads to gnd.

You might pull 3.3.1 release from github and build with that, or just run with the fix if it works. Maybe it will have to be pulled in anyway if I update and the newer cgminer versions all have the problem. Kano probably knows what's up with this as it's his area.

Presumably the time doesn't matter and it's just that control is released to another thread at that point.

I cloned cgminer from github on July 11, 11:52 EDT(yesterday) on Debian 7 in a Thinkpad T500 laptop.
I repeated this issue today on Debian 7 in VMWare player running on Windows(timeout 999ms).

Very odd. I'm on Ubuntu 12.04 and I don't have this problem.
Or did you pull updates from cgminer github? I used 3.3.1 for mine but have not pulled down any updates since.

Confirmed on Debian 7.

I am on Ubuntu.

Update: I changed the delay to 1 usec only (that is one micro sec), and its still working. Removing the delay again breaks it, and cgminer starts hanging for 1-2 minutes. I am quiet positive there is an issue with threading/locking.

Maybe they have a used server UPS somewhere you could get?  They usually have good power conditioning.

- 'Country selection' tool on APC site:

http://www.apc.com/template/country_selection.cfm

They do offer models adopted for the different power line voltages (i.e. 110 V or 220 - 230 V) and national standards for wall outlets.

Good luck!

I tested it first on Windows several days ago. KTest worked fine but CGMiner always gave timeout(-7) error so I changed the timeout to 999ms and the timeout error's gone but still had the stuck problem and low hashing rate. I moved to a real Linux box and there's no timeout error with 200ms setting but have CGMiner stuck problem until I added the "nmsleep(100);" as terrahash suggested. I tested it on both a laptop with Debian 7 and a desktop with Debian 6, all the same.

Yes, I should pick up something like that. Problem here is that that kind of thing I have to order from USA. What they sell here is either junk or triple price.

Or maybe buy some parts for noise rejection and mount them in the one I have.
Probably a website somewhere with how to do that.

- If such a permanent issue exists, why not to invest in a good power line surge protector?

My choice here always was the top device from the current APC SurgeArrest product line and I honestly must said that I never was disappointed.

Modern APC SurgeArrest PER5, for example, offers 15 - 55 dB EMI / RFI noise rejection in 100 kHz to 10MHz frequency range, amongst the other useful features:

http://www.apcmedia.com/salestools/ASTE-6Z7UZU/ASTE-6Z7UZU_R0_EN.pdf

NOTE, on the K1 revision I will likely use the QFN PIC part# PIC16LF1459-I/ML or E/ML instead of the I/SS. I don't know if anyone has ordered the SS already but due to the small size the QFN is a much better fit, and it's also more readily available. Just a heads up for when I do the revision.

On the K16 fan diode - I will add pads for a SL02 Schottky diode on board, next to the fan, and leave it up to vendors if they want to mount it or not. This diode can handle up to 1A which I'm sure will be adequate for any back-EMF. This is the same one used on the K1 power supply.

The diode doesn't really do anything to protect the fan as it simply shorts back EMF through the fan windings. Rather it protects the board from the voltage spike caused. The NDS351 FET has a schottky diode built in for this purpose. However, the extra diode protects the 12V line from the spikes. If you're assembling boards with good quality fans then they should have a diode in them as part of their internal circuitry. I'm not sure at all how prevalent that is but I did notice on my cheap fans they had a space for a diode on the internal hub and no diode was mounted - cheap bastards, evey half-renminbi counts.

Agreed. +1

Note that  roybitcoin sent me this message about the partial K1 length. Note for those sourcing heat sinks for Nanos.

I vote for adding a diode on board to protect against fan malfunction. There's a lot of crap fans out there, some are outright evil....

Yup. If the code detects no Tach output it will just run open-loop and set a PWM signal for the fan. So 2 wires will work. I noticed that these cheap fans don't get much oomph below about 75%, and stall around 25%. I'd say only the top half is useful, but they're quiet up to about 80-905.

Almost. I just measured a board here. Your 39mm is 36mm. The 25 is ok but when cut from a panel of 2 or 4 would probably be 24mm, depending on cutter blade loss. So that is determined by how you assemble them. Most PnP shops will want to put lots of those in one big panel because they're just too small to do one by one.

edit - important:
Keep in mind that the USB tab is not thick enough for correct insertion. It either needs a thickener as part of the casing/heat sink, or a 2mm board thickness should be used. I believe the USB spec calls for 2mm but often boards are ordered default as 1.6mm.

***

I got the fan working. Turns out all these local *#@$! fans are garbage fakes. They don't actually have a Tach 3rd pin. They just feed back the voltage fed in, which of course is not right for the Tach reader circuit, it applies 12V to the 3.3V circuit when it should be an open collector/drain to GND that pulls down a connected resistor. Did that so many times here I'm lucky something didn't blow. But in the process of figuring this out I also determined that these crap Chinese cheap fans also don't put a Schottky diode in them to prevent back-EMF, which they should have.

So the question arises - should I have a diode on board to protect against connecting a crap fan? The 3 I tried here didn't blow anything but they did unleash about 30V across the FET, which fortunately has it's own built in Schottky diode. Decent CPU fans have a diode in their "hub". One of my fans does have a Tach pin that works and seems to feed back a suitable signal to test with but it's a small 70mm fan that I won't be using except for tests until I buy a real-man's fan. (Which is a bit of a Thai pun... get it?).

BTW my fridge and the small magnifier ring light both cause enough EMI on the power line here to cause a USB disconnect. Probably the source of much random noise. Half the time when the fridge turns on, cgminer re-inits the KLn device. They're on the same outlet circuit.

I did this in my miner with an optocoupler and a power transistor (2955 but that is overkill).   PWM controls the cheapo fans just fine, I'm only use 2 pins per fan, relying upon temperature to set the PWM rate with no feedback.