Topic: Hacking a BFL Jalapeno to 32GH and beyond....(???) (Read 54341 times)

I will say I have been busy these past few months. Actually I'm going to Defcon this year in Vegas, anyone want to go and meet up for a discussion on Hacking Miners for Fun, Fires, and Profit?

I'll post details and thoughts over here in this thread:

https://bitcointalksearch.org/topic/m.7232230

It could be fun....

Nice to know, I have bookmarked this thread thanks, just a bit difficult to understand for an unexperienced miner tho.

good info,
but I have not been able to understand this long discussion
This seems nice, and maybe one day if I've understood could try this

Thanks very much for the info.  That's pretty much what I thought, but was getting confirmation from the experts.

Thanks again!

Quick answer: No but it is fun. At this point difficulty has caught up, but if you want a kick ass conversation piece you can still buy the chips on ebay and I'll install them.

C

Greetings all!  I know that it may be a bit odd for my first post to be in the hardware section, but that's me.  Straight to it.

I recently have gotten into the mining scene (about 4ish months) by picking up a 10gh Jalapeno and a few Antminer U2s.  20gh total.
Nothing spectacular, but fun nonetheless.
And, yes, I know, late to the party, so to speak.

Now, I started doing some looking and ran across this thread about mod'ing the Jalapenos for moar power, as it were.
This was after I got to thinking about the fact that if this Jalapeno was orig a 5gh (plate on the back says 5), what was done to get it to 10.  I don't have any equip/software to really do any kind of diag on this thing, so I am going to ask the experts.

So, my here's my question, is it worth messing with this unit by adding chips, etc. to get it up to say 20+, or just keep rocking it as is?
It currently is a 3 chip unit.  After taking the cover off, it dropped from 60+C, down to avg'ing 48C.  The error rate stays at a steady 3.8%.  I am not sure if there is room for improvement currently, or not.
Also, when you first kick it off, there is some garbage that shows up in the mining client.  Both, BFG and the Bitminter app.  I've read that may be the FTDI chip, but I could be mistaken.


Thanks y'all!

Indeed, it brings out that little bit of adventure in all of us.

So I got it home and put it through it's paces: Was clocked at the factory at 290mhz, which lead to about a 6% error rate and a speed of 7.3gh. Too high on the errors, backed it to speed 7 with my firmware release and it's down to 274mhz with 0.2% errors and a 7.3gh hashing speed. Less errors=less power.

Added three chips, no problem. However witrh the sides off and the fan on the top it's clocking at 70c temps, which is a little bit warm. I have the copper heat sinks on all the FETs so it's ok, but once again I see a serious difference in temps between 4 chip jallies and 5 chip ones. 5 seems to be the tipping point for a stock heat sink; once you get to that level stuff gets more complex. However it is putting out a very solid 20gh, so the added chips are all hashing at >4gh each.

I also have three jallies in from Sweeden, one had a broken off power connector, the other two hashed at 5gh. Flashing them brought one to 7gh, the other was still at 5 (has a pair of C class chips). The one with the broken power connector was a simple fix, and it also went to 7.

The user enclosed 4 chips, so I boosted each one with two more chips, bringing them to 14.9 and 15.8gh. Not bad actually, and they run fine without extra heat sinking. For the third one I'm going to see how adding two then three chips works, should be able to get it past 16gh.

Never dull. If you want your jally to be boosted, drop me a line. Despite bitcoin price drops I am honoring my standard prices (.1btc to add two user supplied chips). I also have a few chips available that I will sell for install as well (for much less than the Ebay guy who is being a bit irrational).

Never dull :-)

Sounds badass, it brings people together in ways some wouldn't or never would.

This is why I love Bitcoins. I had to get a miner from a local guy, who wanted me to add 3 chips to the unit, so I asked him to drop it off while I and the kids were playing XP Laser sport in MD. This is my wife's point of view:

Oddly though, before all that, my husboo tells me before the game starts, "I'll be getting a package dropped off for me by a man at the front desk while we're playing. I'll need you to pick it up." We agreed that my code word would be "aluminum." I TOTALLY thought he was kidding. He and the kids went in to play, and I went up to observe them from the upper balcony. Another guy was up there too, a friendly, middle-aged-looking guy, watching the laser tag game and laughing. It's dark up there, kind of hard to see. At a certain point, he squints at me, comes over, and asks me "Are you with C?" I agree that I was. I figured this was one of his old laser tag buddies from when he used to regularly game. The guy tells me "I have the package for him."  I tell him "Oh right, I'm supposed to say 'aluminum' for the password." like I actually know what's going on. He laughs and says "Yes, and I said I'd flap my arms like wings." So he does. Then he hands me the box and says, "I'll just pay up front, I just went to the ATM.", and counts me out a hundred mumble dollars. I take it and put it in my pocket like I have some fucking clue what's going on. He says this was like a spy movie, and I agree it was all very cloak and dagger and we chuckle, say good-bye, and he leaves. Hmm.

Of course part of me knows that Lightfoot fixes bit coin miners for people, and I figured at a certain point that this nonsense was related to that, but really. How was I supposed to know he was REALLY going to get a delivery from someone at the laser tag place while he was gaming with the kids?  This is my life.

I simply... love... bitcoin. :-)

I'm wondering can anyone in the UK make two Jalapenos in to one, IE pull the two chips off one board and add it to the other board?

Both use the newer type chips and run the 292 firmware, one gives about 8.1GH and the other does just over 5.5GH.

Any ideas?

Another note: I did some power calculations, and it looks like the 8 chip jally pulls 5.2 watts per gh, while a standard single/30 pulls about 4.0 watts per gh. I'm guessing most of the difference is the cooling pump, but I think the newer boards are more efficient than turbo-charged jallies.

C

Very interesting little update: Last night my 6 miners out in the shed faulted out my Rpi at 4am. Went out at 6am to check on them, brought a laptop and after some analysis found that my 28gh air cooled jalapeno was screwing up the USB bus. Not sure why, but it's out of the loop for now and everything else (8 chip jally, chili, two single/30's and two single/60's) are running normally.

Unit was a factory good 2 chip December Jally that was upgraded to 7 chips with dual heat sinks. It tended to get quite warm but not insanely hot.

Now it is true the temps last night in the shed dropped under 15f, and this was a confirmed good unit, but it's still odd. I'll have to sit down and figure out what's going on.

C

Can you show the difference in pix ?

IGBTs, or single power FETs

What they look like.

All I know is the old NPN sets from old motherboards.

However I will say this: The BFL units are putting a fair bit more than a few watts of heat from the FETs. There's also the issue of imbalance due to thermal conductivity issues at work here.

Personally I prefer IGBTs, or single power FETs.

C

complicated What is so cool tho, is the end result is even more info for a learner like me.

Thank You Mr Teal...

In a year or two I will understand every word.

I gotta get one of these machines..

This is why Mr. Teal builds boards and I fix them. Excellent explanation, and thank you.

C

Interesting note for those that are interested to follow up on what lightfoot said. The highside (top) FET doesn't actually conduct any higher peak current than the lowside (synchronous or bottom) FETs, and the average current is actually quite a bit lower than the lowside FETs because the duty cycle is quite low. That is why you often see designs with the lowside FETs have a higher current rating and lower resistance.

The power dissipated in each of the lowside fets is pretty simple to calculate, it's just (1-Vout/Vin) * Rds(on) * [(Io/n)^2+(Ir*nphase/n)^2/12], so the percentage of time the lowside is on, times the resistance when it's on, times the current. n there is just the number of lowside FETs in total (6 for the Jalapeno and LS), nphase is the number of phases (2 for the Jalapeno and LS), and Ir is the ripple current. That's usually designed to be about 20%-50% of the output current, and depends only on switching frequency, output voltage and inductance. For the BFL design Ir is ~6.8A. For the lowside FETs the power dissipated is directly proportional to the resistance (Rds(on)) of the mosfets, so you often see a lot of big, beefy FETs used since they have lower Rds(on).

The topside mosfets are a little trickier, since there's two things to calculate. The first is the conduction losses, which is the same formula as for the lowside fets except instead of using (1-Vout/Vin) (0.917 for the 1V from 12V that BFL uses) you use Vout/Vin; in this case 0.083. Obviously if you used the same FETs for the top as you do for the bottom the dissipation from conduction would be a lot lower, 11x lower in this case. However, you never use the same FETs in the top because there's another kind of losses that the highside FETs have to deal with, conduction losses.

The conduction losses aren't from the steady state P=I^2R losses you get when the FET is fully turned on and there's very little voltage across it. Real MOSFETs aren't perfect switches though. When you switch on and off the highside FET, the change doesn't happen instantly, so there is a time when you have a lot of voltage across the FET and have current flowing at the same time, so you burn up a bunch of power there.
The formula is P = 2 * f * (Vcc*Io/n) * Rg * (n/nphase) * Ciss. f is the frequency (300kHz for BFL), Vcc is 12V, Rg is the total gate resistance of the driver and FET, and Ciss is the input capacitance of the FET.
What's interesting about this is that the n term cancels out; adding more highside FETs per phase doesn't actually decrease the switching losses per FET. This is because even though if you run 2 or 3 FETs in parallel each of them sees less current, the time it takes the driver to fully turn on the FETs is 2 or 3 times longer due to the input capacitance being higher. That also means that if you run 3 of the same FETs for the highside, your total switching losses actually go up instead of down like they do with the conduction losses. That's why you generally always see designers use lower current rated FETs with higher Rds(on); those ones have a lot less input capacitance so your switching losses are lower. It's also why you see arrangements like this one, with one highside and two lowside FETs.


Pulling up some calculations I did on the BFL PSU at 1V and 100A output using the new "cool" BSC014/BSC0902 pair (which is what's in my Single), the conduction losses per lowside FET were 0.52W, or 3.12W for all 6. The conduction losses in the highside FETs is 84mW each and the switching losses are a relatively massive 2.2W per device or 13.5W in total. At 100W output, the BFL PSU (at least on paper) would dissipate 16.6W in the FETs and 3.25W in the inductors, so about 20W in total, or 83% efficient.

As an interesting note, if they would have used a transistor better suited to the highside with a lower Ciss like the BSC052N03LS, those hot highside FETs might run a lot cooler. Those would have 0.93W of switching losses each, and with 1 the conduction losses would be 1.64W each, with 2 the conduction losses would be 0.41W each, and with 3 the conduction losses would be 0.18W each. If you ran with two highside BSC052's, not only would your per device dissipation go from 2.28W to 1.34W, but since you only have two of them the total highside dissipation would drop to 5.35W from 13.5W. You should give that a shot next time you're replacing the FETs on a Jalapeno.

This isn't really an overclocking thread, the OP started adding chips to his Jalapeno to increase speed.  The Single is already at the maximum number of chips, so there is no way to increase from that.

Is there anybody here who can flash these devices?

I have a faulty one, would like to have it checked/repaired.

send me a PM if interested, European if possible.

Well, I thought about it, but the problem is these situations are so far beyond normal it's not funny. Posting videos would give the impression that this kind of thing can happen normally, when it really is a test bench sort of problem.

The most a normal jally would do is smoke the FETs then fail the power supply. And since it's in a case it wouldn't damage anything else (which is probably what happened). So I'm posting as a warning to other people doing work on stuff; be careful and always take proper precautions.

It's never dull.

C