Home
JAQForum Ver 19.10.27
Log In or Join  
Active Topics
Local Time 00:09 09 Apr 2020 Privacy Policy
Jump to

Notice. New forum software under development. It's going to miss a few functions and look a bit ugly for a while, but I'm working on it full time now as the old forum was too unstable. Couple days, all good. If you notice any issues, please contact me.

Forum Index : Solar : Simplified Solar Hot Water

     Page 6 of 6    
Author Message
davef
Guru

Joined: 14/05/2006
Location: New Zealand
Posts: 492
Posted: 07:35pm 16 Mar 2020
Copy link to clipboard 
Print this post

Hi Tony,

  Quote  A much better way to build a window detector would be to use each half of the LM358 as a separate voltage comparator, and then use the outputs to toggle the set/reset inputs of a flip flop.


Took me a minute or so, but then I realised that the LM358 with hysteresis is a window detector.

Was the LM555 designed to operate properly with a variable voltage supply (0V to Vsupply) or does it just happen to work properly in this mode?

I might have to dig deep in the junk pile to find a LM555 to try.
 
Warpspeed
Guru

Joined: 09/08/2007
Location: Australia
Posts: 3087
Posted: 08:15pm 16 Mar 2020
Copy link to clipboard 
Print this post

Plover,

There are various options offered in regards to shipping. The most expensive is DHL, but they can deliver in about three days, China to Australia, and they work 24/7.  I have had deliveries arrive here at 7pm on a Sunday evening.
The cost is very roughly about what you are suggesting, three dollars for the boards and thirty to thirty five dollars shipping to the suburbs.

The cheapest option is usually ordinary surface mail (China post) it comes by sea in a container and can take four to six weeks. Cost is only about ten dollars less though.

Still, if it was the other way around, thirty dollars for the boards, and three dollars postage, its still not a great deal of money these days.

Dave,

Yes, an LM358 with hysteresis would work perfectly well, the problem though is setting up the two voltage levels. Its a very difficult thing to work out suitable resistance values to place two voltage thresholds exactly where you want them using hysteresis and standard resistor values.  

Finding two resistors that have an exact ratio between them is fairly easy using standard resistor values.  Finding three resistor values that have exact ratios between all three is far more difficult. It is a lot easier to use two voltage comparators and set up each comparator separately with only two resistors, then use the two logic outputs to set and reset a flip flop.

I took that idea one step further by arranging my circuit to switch in either direction when the current reaches one milliamp through the voltage setting resistors.
The two range setting resistors then become dead easy to work out, one K ohm for each volt.

And by connecting resistors in series, its easy to make up odd resistor values to experiment with.

For experimental purposes, you could even use a couple of potentiometers with scales calibrated in voltage. The scales could be marked in voltage using an ohm meter initially, although fixed resistors would be preferable for final installation.

Something like a 100K pot in series with a 180K resistor would give an adjustment range of roughly 180v to 280v for the high voltage trip point.
A 100K pot and 100K resistor 100v to 200v for the low voltage trip point.

If its cycling fast enough (at least several Hz) a light bulb connected across the heating element might give a very rough indication of power.
Edited 2020-03-17 07:27 by Warpspeed
Cheers, Tony.
 
davef
Guru

Joined: 14/05/2006
Location: New Zealand
Posts: 492
Posted: 01:08am 17 Mar 2020
Copy link to clipboard 
Print this post

Hi Tony,

Taking the circuit on Simple PV Diversion Unit I now see the difference ... the single op-amp is not configured as a window detector.  Yes, setting up a window detector using a single op-amp is a convoluted process.

My op-amp configuration just switches ON the MOSFET when the upper set point is exceeded and turns it OFF when the voltage on the cap sags down to a lower set limit.

What added feature does a window detector offer in the circuit in this thread?

Thanks,
Dave
Edited 2020-03-17 11:24 by davef
 
plover

Senior Member

Joined: 18/04/2013
Location: Australia
Posts: 289
Posted: 01:08am 17 Mar 2020
Copy link to clipboard 
Print this post

Warpspeed
Just an update, I gave up on JCL, the reason I tried them was I could upload your zip file directly.

I decided I would go to Shenzhen2u.com where I have been before. The problem is that I need to make gerber filename changes, so I have looked at that and prepared the change list.

Before doing so I decided to see what happens if I just up load your file, I partly expect to get a message in a day or two that they have rejected processing.  Placing order however took less than 15 minutes, with careful checking of the shipping cost.

As I am prepared to wait I am trying the E-packet delivery cheaper than China post by 30 cents, I am not sure but I think it may be airfreight. I did notice an offer for SF-xxx I assume this means Sea Freight even cheaper.

Anyway 10 boards $29 Australian, delivered.  

  Quote  The two range setting resistors then become dead easy to work out, one K ohm for each volt.


This little trick using 1 mA really is a gem, I can forget my spreadsheet for calculating values. (Only need this for record nowadays, I forget too easy)  
 
Davo99
Guru

Joined: 03/06/2019
Location: Australia
Posts: 428
Posted: 02:09am 17 Mar 2020
Copy link to clipboard 
Print this post

  Warpspeed said  

Something like a 100K pot in series with a 180K resistor would give an adjustment range of roughly 180v to 280v for the high voltage trip point.
A 100K pot and 100K resistor 100v to 200v for the low voltage trip point.


This is exactly what I was thinking the other day looking at the circuit.
I notice you say for testing purposes.  What I looked up seemed to indicate the trimpots won't handle the power which you have set at 1W resistors.
Is this the reason you seem to suggest they would not be useable as a permanent inclusion? If there are trimmers that could handle the load it would make these even more user friendly.



  Quote  If its cycling fast enough (at least several Hz) a light bulb connected across the heating element might give a very rough indication of power.


Was also playing with this very thing this morning with one of my own setups with an arduino driving a Mosfet.  It's a bit dull so I'm using a 500W halogen to try and find the sweet spot of the fixed timing.  Didn't get far before it started raining but fining up now so I'll go have another crack.  It was what I did just bench testing the pulsing idea the other week.

Someone also mentioned to me that this could be good for driving a space heater like the old bar radiators or even an old electric jug.  Anything resistance isn't going to matter.
 
Warpspeed
Guru

Joined: 09/08/2007
Location: Australia
Posts: 3087
Posted: 02:54am 17 Mar 2020
Copy link to clipboard 
Print this post

  Quote  What added feature does a window detector offer in the circuit in this thread?


Your basic voltage comparator with hysteresis would work perfectly well for this. Only disadvantage is the complexity of working out suitable resistor values to do it. That would be totally beyond many people here on the Forum.
Two separate comparators adds a bit of circuitry, but hugely simplifies the whole resistor calculation problem. So the basic reason for the window comparator is to make the circuit much simpler for others to get working.

Plover,

The original concept behind this was to make something really dead simple that people could get going with absolutely minimal knowledge or test equipment, and not run into problems. Its pretty basic, but quite a bit of thought went into it.

Dave,

Its not power dissipation that is the problem, but voltage rating of the resistors. Now the people that make those really tiny metal film resistors say you can put 500v across them no sweat.

I am a dedicated disciple of Murphy and a certified pessimist.

Something a wee bit physically larger is much less likely to break down with a bit of dust, moisture, or crud that may tend to build up over the years in this (probably located outdoors) type of project.

Two smaller resistors in series would achieve the same result, especially if they were spaced off the board a bit, or perhaps sleeved in plastic.
Edited 2020-03-17 13:33 by Warpspeed
Cheers, Tony.
 
Davo99
Guru

Joined: 03/06/2019
Location: Australia
Posts: 428
Posted: 06:26am 17 Mar 2020
Copy link to clipboard 
Print this post

Spent some Time today playing with the pulsing on and off of the panels through a 1000uF cap .

Confused as to what I learned and saw. It was cloudy and I had 4x250W panels in series.

Playing with the light which I ended up with a 275W heat lamp, the best output was straight connection or pulsing the thing really fast like 20/1000ths  of a sec both ways.

With the element I wasn't sure what I was looking for. To keep the panel voltage up or the amps up. The seemed to be in opposition but I think when I was packing up I saw a measuring mistake where I had the multi meter.  

All seemed to be working slowing the pulsing right down showed corresponding action at the output so maybe just one of those days where things lined up as they were or the cap was too small. Ill try again with more caps later in the week.
 
Davo99
Guru

Joined: 03/06/2019
Location: Australia
Posts: 428
Posted: 04:00am 19 Mar 2020
Copy link to clipboard 
Print this post

Tony,

Boards arrived 5 Minutes ago.  
They look brilliant and can't wait to try the built one out.
Thanks so much again for that, very much appreciated!

I got a CPU heatsink and a fan to mount the mosfets on and the fan will run off another of those little power converters which I'll tap into the supply from the solar.

Just again, in case I missed anything, is there any better amount of capacitance to run?  I have 1000uf caps and can do anything from 1 to 10.

I have 4 x 250W panels set up now I was going to initially test on then go up to 8 from there. Is there a power limit the board will take, IE say 4 Kw of panels on a 3.5 Kw element or...?  what about voltage, under 300?

I can stop checking the letter box 5 times a day now. Even though I had the camera setup, every time I went out the back for a while I had to go check again in case I missed him.  Dog loved all the walks down there though.  In the end I heard postie coming so they were handed to me and never even hit the letterbox.  :0)

Thanks Mate!
 
Warpspeed
Guru

Joined: 09/08/2007
Location: Australia
Posts: 3087
Posted: 04:38am 19 Mar 2020
Copy link to clipboard 
Print this post

Great stuff there Dave, now the fun begins            

The fewer caps, the faster it will cycle, and there is no practical limit to how fast that board can cycle. Cycling speed will not change the amount of power transmitted.

If its going pretty fast, a light bulb connected across the element should give a fair indication of relative power with minimal blinking.

I think you will find its all very non critical to set up.

Its typical to find that you can only reach about 80% of the panel rating under perfect conditions anyway.  And a 2.5Kw element is not going to explode with 2.6Kw.

As long as those mosfets do not become too hot to hold your finger on continuously, they should be fine.
Edited 2020-03-19 15:01 by Warpspeed
Cheers, Tony.
 
Davo99
Guru

Joined: 03/06/2019
Location: Australia
Posts: 428
Posted: 08:26am 24 Mar 2020
Copy link to clipboard 
Print this post

I have been doing the mental arithmetic with the board for a while trying to work out the best way to mount it all.
I am taking my time to try and do it best I can as my skill in this area is poor and I'm determined not to stuff anything up.  I want to -try- and make it as least laughable as possible. Wiring is one thing, neat wiring is quite another.

I am going to use a solar inverter cabinet I have stripped out, ( stripped 2 actually but wasn't happy with the first) which has a massive heat sink I can mount the fets on and Connections for the wiring.  Similar to a solar inverter I'll have the 2 MC4 Inputs for the solar and probably a regular power point for the output side so I can plug in lights, heaters or just put a regular female plug on the water heater. Seems hard to find mountable terminations just for cable. I did order one from flea bay and might be here in a month but the outlet I think would be better anyway.

The only thing I'm unsure of, which I'd like your advise on Tony, is what gauge wire to use?

I'm doing everything in 2.5 But it's a bugger to manipulate and make neat in the cabinet.  I'm undecided if the pulsing makes any difference to the load?
I think with the cap discharge the heavier the wire the better to carry the high amp short pulse.
Then again, like my mate, I'd always wiring a 50W mains LED light like it's the main feeder to a Sub station.   :0)
I looked at 1.5 for this and it just didn't seem right especially if I connect up 2 Kw or more of panels to drive a 3.6 Kw heater element.

I'm thinking I'll use 4x 470 Uf 500V caps. I'll solder them up and glue them to the side of the cabinet where it should be the coolest away from the fets.  Heat sink probably weighs 2 Kg so I'm thinking should be Ok.

Had headaches pretty bad last few days but Hopefully I'll be able to get it done and connected tomorrow and the fun really can begin.

If anyone knows any links to tutorials on how to do the wiring and electrical connections for things like this so they are neat and professional looking, that is something that my non existent skills  would benefit from.

I did have a nice little 4 Hole neutral link bus bar thing with a cover but I used it in one of my solar sub boards. they seem hard to get as most suppliers only seem to have 7 hole units and they are a bit big for a lot of things.
 
Warpspeed
Guru

Joined: 09/08/2007
Location: Australia
Posts: 3087
Posted: 09:03pm 24 Mar 2020
Copy link to clipboard 
Print this post

There are two methods of rating wire sizes, voltage drop, and temperature rise.

If running very long lengths, especially at low voltages, voltage drop can become a serious issue, so we may need to use much heavier gauge wire to keep the voltage drop within reason.

Within a piece of equipment, the wiring is usually quite short, so we can get away with using thinner fine stranded wire which is more flexible, neater, and much easier to work with, provided its not going to get unacceptably hot. How hot it can safely get depends on the type of insulation and your sense of adventure.

I believe for normal house wiring 2.5mm square is rated for 20 to 25 amps which is a pretty good compromise between temperature rise and voltage drop at 230v.

The thinner 1.0mm sq or 1.5mm sq is usually rated for 10 amp lighting circuits and is also a pretty good compromise for voltage drop at 230v in lengths found within a home.

For what we are doing, 2.5mm sq from the solar panels would probably be o/k for a 3.5Kw 230v system,  as long as the wiring runs are quite short.  The usual 4mm sq solar cable might be better, especially at voltages lower than 230v, and especially with longer cable runs.

Within equipment where wire lengths might be only a foot or two, you can really push things. Something like good quality 16 gauge automotive wire 1.3mm diameter and 1.3mm sq can be rated at 22 amps!  That is more than I would be happy with personally, but it does show how the situation can change.
Cheers, Tony.
 
Davo99
Guru

Joined: 03/06/2019
Location: Australia
Posts: 428
Posted: 08:45am 25 Mar 2020
Copy link to clipboard 
Print this post

I cut some good quality Fine stranded ( silver) wire out of the inverters and am using that to wire the board.  It's 2.5 and the longest piece is about 30 Cm so I'll go with that.

I'm aware of " Normal" runs and loads but just wondering about how the pulsed DC would affect things. Ac and Dc are of course different as are the voltages but having the power pulsed and from a cap made me wonder what comes into play here.
 
Warpspeed
Guru

Joined: 09/08/2007
Location: Australia
Posts: 3087
Posted: 08:29pm 25 Mar 2020
Copy link to clipboard 
Print this post

It is a bit "different" but not by enough to worry about.
If you run everything in 2.5 then that will be more than sufficient for 3.6Kw.
Cheers, Tony.
 
plover

Senior Member

Joined: 18/04/2013
Location: Australia
Posts: 289
Posted: 12:39am 27 Mar 2020
Copy link to clipboard 
Print this post

Just an update on dropping Warpseed's zip file in the 'letterbox' at ShenZhen2U
Looks like it has worked, ie I need not change file extensions. This seemed to be an obstacle.

WIll be interesting to see when they arrive some time. The parcel has been in transit in China for a couple of days according to the tracker.
Edited 2020-03-27 10:40 by plover
 
hotwater
Senior Member

Joined: 29/08/2017
Location: United States
Posts: 120
Posted: 01:42pm 27 Mar 2020
Copy link to clipboard 
Print this post

Glad to see everyone experimenting.  Many should watch a video on youtube about adiabatic charging of capacitors. It relates to water heating.
 
Warpspeed
Guru

Joined: 09/08/2007
Location: Australia
Posts: 3087
Posted: 10:33pm 27 Mar 2020
Copy link to clipboard 
Print this post

Professor Yaakov is perfectly correct, but the whole problem needs to be taken into proper context.

If you try to charge or discharge a capacitor in zero time, the current must be infinite, and you lose half your energy.
This becomes a serious problem in extremely high energy systems that need to charge and discharge as rapidly as possible, for example radar, sonar, and pulsed laser power supplies.

How do you put one megawatts worth of stored energy into a high voltage capacitor bank in one tenth of a second to fire off your radar/sonar/laser at full power ten times each second ?

At 9:28 on the video, the Professor tells us that if you slow down the charging process (so that the current becomes less than infinite) it all becomes much more efficient. Which is completely true. I used to design high power pulsed laser power supplies for the defense department at one time.

The ideal charging method is charging from a constant current source, as we will be doing from our solar panels, which is a near perfect solution. We discharge through our heating element, which again slows down the process and makes it all more efficient.

We might be in trouble though, if we charged through solar panels up to say 250 volts dc, but then wanted to discharge the whole lot in a few microseconds at thousands of amps. But doing it far more slowly, we see none of those types of problems.

The only potential problem for us, is the ripple current rating of the capacitors. High currents are flowing in and out of the capacitor each cycle, and the very thin aluminium foil in the capacitor has resistance, which generates heat.

As long as the electrolytic does not overheat it will be fine. If overheating is a problem, we could use higher quality low esr capacitors with a sufficiently high ripple current rating, which will be expensive.
Or we can just parallel up a few more of the cheapies so that the current is shared and the heat per capacitor will be less. If the electrolytic is running just slightly warm it will be fine. If its running really hot, something will need to be done.
Cheers, Tony.
 
plover

Senior Member

Joined: 18/04/2013
Location: Australia
Posts: 289
Posted: 11:29pm 29 Mar 2020
Copy link to clipboard 
Print this post

Update on my boards from ShenZhen2U I have looked at the tracking having chosen E-Packet shipping it looks like this may be China-Post as well I think it is coming airmail


 
Solar Mike
Guru

Joined: 08/02/2015
Location: New Zealand
Posts: 582
Posted: 12:46am 30 Mar 2020
Copy link to clipboard 
Print this post

My PCB's from JLCPCB China were posted registered airmail 26 days ago, just arrived in the mail box this morning, the previous PCB order took 10 days by air post. So this virus is slowing things down a lot.


Mike
 
     Page 6 of 6    
Print this page


To reply to this topic, you need to log in.

© JAQ Software 2020