New PCBs from JLCPCB - Too Much Current?
ฝัง
- เผยแพร่เมื่อ 25 ก.ย. 2024
- Design & Order on EasyEDA, PCB+SMT $10 Off: easyeda.com/ed...
Get coupon & Join JLC&EDA Group: jlcpcb.com/EDA
How much current can you put through a PCB?
www.7pcb.com/t...
Subscribed to my 2nd channel? Watch all my 3 minute videos on Julian's Shorts: / @juliansshorts
Interested in my new garden workshop? Follow the entire build on Julian's Shednanigans:
/ @juliansshednanigans
Join me on Odysee/LBRY: lbry.tv/$/invi...
Octopus Energy referral link: share.octopus.energy/aqua-birch-918
(We each get £50)
#ElectronicsCreators
The old irons had a thermal mass, that's why they take ages to warm up. They are best for large connections.
ive recently rediscovered the merits of my Weller portasol butane powered soldering iron. it gets so hot it quickly destroys components and traces and thus was put away for years. i've discovered recently however that it's wonderful for heavy jobs for example soldering xt90's to 10awg silicone jacketed wire. and it heats up insanely fast to boot!
The old irons have their place in the world. I can never do a delicate connection with them.
I was most impressed with how little that old iron smoked on power up. My old 60w would set of modern sensitive smoke alarms if I used it indoors.
I still have a massive iron that apparently was used in plumbing. Takes about 15mins to heat up and gets used very rarely. It has its place though. Occasionally it is the only tool for the job when there is a massive ground plane or heat sink.
To me this looks like a case where cheap copper bus bars would maybe have been a better option. You could buy a couple of lengths for about a fiver. Drill them at each end for mounting with long nylon bolts and nylon separators to give two parallel bars. Then drill holes in the bars for your spade terminals. Tap and screw them on. Mount the connectors so pos on on bar and the neg on the bar below. You then have a row of connectors al nicely distributed along the length with plenty of capacity for current.
Well for this use case ... Why not an Y cable ... must be to easy.
One downside of using PCBs with nuts and bolts as external connections, is crush forces.
In your design, the single 3mm plated thru hole is the only thing stopping the fibreglass substrate from crushing.
An improved version (seen in many other designs) is to add via stitching around the central hole.
That is a ring of smaller holes around the main hole, to distribute the forces.
Yes, I saw that on the Ideal Diode PCB. I might give it a try :)
You should add some lock washers to prevent the nuts from coming lose due to thermal expansion.
Why don't you put the light on the normally open connector of relay to not waste the half an Amp?
Another option for increasing the current would be to go to a 4-layer board. The cost differential from JLCPCB is very small...
Good point!
Not liking the current carrying ability of the nuts and washers. How about some heavy copper wire soldered into the XT90 and thru the holes on the board. If you want to stack boards then leave the tails on the wires long and solder them to the second board (could use little plastic washers to maintain a gap, or just hold the gap while you solder).
PCBWay ad before this video, fitting.
PCBs _can_ look neat and tidy but so could a "Y" wire splitter. With the benefit of not having to wait for PCBs. You could still keep it flexible. Attach your XT90 to wires going to ring terminals. Join those in a "Y" (bolt) and cover with your large heat shrink tubing. Done and protected. Cut the heat shrink to reconfigure.
pcbs are more fun tho, which is the whole point
Have you considered using the XT90-PW series connectors which can be directly soldered to the PCB?
For high currents, use copper nuts and bolts or better soldering directly on the XT90 connector....
I don't know for sure what is the maximum current on the copper line on the PCB? If the PCB is not protected by purple insulator paint, you can add tin on the line
Seems like you could have the ground ring terminal screwed onto the bottom directly to the surface, and the VCC directly onto the surface of the top, with them mounted into the XT90 180 to each other so that the bent part of the ring terminal would work with the board thickness almost enough. So the XT90 would be fully in-line with the board, rather than offset on top of it. Steel nuts aren't nearly as conductive as copper right?
Literature on the solar panels typically reads "Do not disconnect under load." Another relay like you plan would be doing just that.
Best not to disconnect a string of 10 solar panels in series (360 volts ish) under load, but a single solar panel (36V) removed from an 8-cell battery (28V) is an 8 volt disconnect.
I thought the reason was because the high voltage from a string of panels can cause an arc in the disconnect switch, which will be sustained due to the nature of DC current. Still not an issue in this application. Although the poor quality contacts on these cheap relays may throw a spaniel in the works over time.
@@UpLateGeek woof!! 🐶
I'll give it a week before you drop something on the new PCBs and there'll be a bang. Make sure to record it for us to see.
Another PCB to cover the bolts and some hot glue will fix that :)
I have 20V cordless powertools. BMS battery and modul have small PCB circuit.
Why PCB not burn when my drill output 20 ampere current?
How much PCB can take current?
There's no issues with the BCP taking the current. You'll find those tracks would take 20A + with relative ease.
The connection with the bolts and nuts will be the point of highest resistance.
Could get quite warm being steel, brass bolts and nuts would be lower resistance. Solid copper wire from large cross-section twin&earth cable soldered between connector and PCB would be even better.
Good idea on the brass nuts and bolts :)
fantastic
Have you considered setting a Raspberry pi up as an MQTT broker so you can control and monitor the shed remotely, using Node Red ?
The Shelly 1 switches will run no 240v or 12 to 60v DC and give you a pair of volt free 16A contacts. The Wemos D1 mini can be used to measure voltages and temperatures.
Both work with mqtt when flashed with the Tasmota software.
I'm using a D1 mini and relay to turn the active balancer on, to top, balance above a set voltage.
Sounds like a big learning curve on that lot!
@@JulianIlett Yes, but think of the number of videos...
MQTT on the pi, easy lots of videos on Mosquito install. Tasmota software install, easy using tasmotizer, it does all the work for you. Getting the msg.topic set up is a little challenging but not too bad.
Yes node red is a challenge but worth the effort and there are lots of flows on the web you can copy. Once you get it running for one sensor the rest falls into place. Start by flashing the tasmota software and operate relays using the web interface.
ACS712 is really hard to find. All the Allegro current sensors are low in stock.
I had to use ACS716 with an op amp lever shifter to get 0 to 1 volt for 0 to 10 amps.
That explains why my PCBs were so expensive. I've had 10 cell connect PCBs made, with ACS712-05 pre-assembled by JLCPCB. The ACS712 ICs were $3 each!
Don't you need to put some kind of cover over the board so you don't short out the terminals when you drop a spanner or screwdriver or water on it?
Current Carrying Capability. Classic!! :-)
If you just got a PCB with no holes, and only spots were the masking was removed you could flip the rings over and hit both sides of the board directly. You COULD leave the holes to add bolts while soldering, and then remove the bolts after.
that is actually a very clever little bit of board.and looks way better than the bolt stack...... now just ad another one of those relays to solar side to monitor the battery so it can cut out the solar before the bms does and you'll have a perfect setup
That's the plan: Fit the 8-cell balancer, then put another relay module on the solar panel (which acts like a simple charge controller).
Given that the board has ground on bottom and B+ on top,, it sounds like it would act as a capacitor . Two plates, separated by an insulator.
Does the stray capacitance of this board cause any issues?
A few picofarads. That might affect high frequency signals, but will have a negligible effect on DC.
You could also just print the current shunt resistor on your PCB buss bars, by selecting the appropriate width traces.
Yep I have a 100w iron for those task.
That BMS will eventually balance the cells! Are against Solar charge controllers?
I think you may be overinvested in XT90's.
Somewhat :)
Those little red ring terminals are not rated for any real current. Why not make up a cable with multiple xt90 in parrellel for your project?
Thought: if you could isolate the ground and use a synchronous buck (with cc/cv), you could charge your cells in parallel (2 leads per battery) and still have it connected as you do now. It sure if it would get to the same efficiency as an active Bms but it would guarantee balancing and you’d eliminate the partial charge/discharge of the active balancing.
Also, why not make the solar relay on the xt90 boards instead of as a separate one? (Or build it all together)
How much capacitance that PCB has?
A few pF :)
I'd use copper tube myself
How many racing drone builders thought he was making an ESC power distribution board?
wait, so whats the point of the giant lightbulb?
It's a resistor - to limit current coming from the solar panel. It's a temporary measure.
When the Bms cuts out the solar is still feeding the miner, it could keep starting and stopping depending on sunlight until the Bms cuts in again.
Yes, that needs looking at. Another one of those voltage controlled relays could fix that - acting like a primitive charge controller. Eventually I want the BMS to protect the battery in a worst case scenario, but otherwise do nothing at all.
Love the Ant Miners. Your neighbors must hate them.
How are the earnings on them?
I would've called it the XTribution 90 board. Then the temptation would be to make it X shaped, which would really be form over function, so probably not a good idea.
Honestly I'm a little surprised you don't get any warming from those little ring terminals. There's not a huge amount of meat between the ring and the wire entry. Although I suppose if you think about the volume of wire that would be used with it, squishing that to the same thickness as the terminal would probably be smaller than the width of the terminal at that point.
That A.C. soldering iron is probably a "thermally stabilized" temperature rather than temperature controlled. A T12 clone from china (at about 70 watt and temperature controlled) does a much better job and heats in seconds.
I've been really tempted by the Pinecil, which is a TS100/80 clone that Adrian Black featured in a recent mail call video. It's got both USB type C and 5.5mm barrel jack power connections, and it takes the same T12 clone tips that the TS100/80 takes. But the best part is that it's a lot cheaper than a TS100/80, even after you add a USB-C wire and power supply or 20V power brick.
Your regulator looks to be a linear type. A switching regulator would give you a longer run time.
still got soeyetus then.
Have you ever heard of a solar charge controller, maybe put a pwm5 on it! "thing never bloody worked"
I really like those boards. So clean. Ya, bus bars could replace them but these are thin and Fun!! :-) Thanks, Julian, for years of great video adventures and learning!
Did you ever test miners on a 4S LiFePO4 battery ( instead of 8S plus buck)? Hard to imagine there would be a deal breaking voltage mismatch there! Also did you consider the various user suggestions for a better solution than all these relays? It's trivially easy to program a Raspberry Pi to watch some GPIO pin voltages and run bash or python script to do unattended ssh into the miners and send stop command, to r the linux 'reboot' (or 'shutdown') command when it's time to restart . If these ideas are successful the efficiency, reliability and flexibility of the solution will be dramatically better!!
I want to try that. But: I'll need a buck-type charge controller between the solar panel and 12V battery. Antminer voltage range is 11.6V to 13V - that's 2.9V to 3.25V per cell (if balanced). But it's worth a try :)
@@JulianIlett Methinks that voltage range suits LiFePO4 like a glove. They won't stay above 3.25V loaded unless grossly overcharged . And LiFePO4 has very flat discharge curve at ~3.2v once reach 2.9 to 3.0 they're done and extending lower will stress the cells . In any case, one of the BMS or the ant miner will cut first, and either will be more or less an appropriate/safe level to do so.
Bummer you have to get a new charge controller. Id hope you can do better with efficiency vs the Chinese asynchronous buck at circa 80%. I wonder if this might be one of those situations where a PWM version may actually stand to be more reliable and efficient ?
In any case , seems like at 4S your balancing situation will be dramatically improved , in practice, and you stand to gain a significant amount of effective battery capacity as a result.
finally , a software controlled Raspberry Pi alternative to the relays seems like it will net you maybe another 5-10 watts in saved holding current, and allow you to programmatically tackle all the strange hysteresis phenomena via a combination of voltage sensing and timers. A bit of work to set up, but a neat project in itself, using hardware you already own, and will be dependable once tuned.
good luck - neat project - thanks for sharing!
I didn't know the Raspberry Pi had an ADC in it. you can do all of that with an esp8266 at 1/10th the price. Also the charge controller will switch on my grid tie when enough voltage and switch off when batteries are empty. Can also send commands over UDP or Telnet connection to have it switch and see voltage and current
@@JulianIlett - IMHO a lot can go wrong running the miners directly off of batteries whereas the buck converter can provide great flexibility with different battery voltages, etc. I don't recall Julian mentioning the efficiency of the converter but 80% would be really low. That converter would be getting pretty warm if it was dissipating 50W continuously. A quick eBay scan showed 90-92%. Perhaps Julian can do an efficiency test.
What could increase run time would be to run the miner at the minimum voltage that keeps them stable. That might mean having a buck converter that can be adjusted BUT even if the buck converter outputs exactly 12.0V that could be a ~8% savings over 13.0V.
In a previous video it was mentioned about using SSH (etc.) to politely shutdown the miner instead of just cutting its power. Many reasons and many possible improvement could be made by using SSH. That would be ideal but it would be even better if it didn't need shutting down because of a bigger battery.
@@PetRatty ahh thank you for pointing out- my mistake- i forgot that you'd need an ADC board with the pi. but yes there are several other boards one could directly use, eg esp, particle, etc
"risk the pcb might be shorted?" dont worry, it wont be shorted for long.