Well, the SLAB that was reading 5V has definitely had it! Initially, it looked good, and I was able to transfer it to my Accucel charger. But then, the voltage started to drop, even on charge. It was always unlikely that this one would recover.
The one reading 10V, I was much more conservative with the water (about 2ml per cell), and have left this on the slower charger. Its still only taking a very low current, just a handful of mA, but this is steadily increasing.
So, who knows?
Wednesday, 22 July 2020
Refurbishing dead SLABs?
I use quite a few 12V 7Ah Sealed Lead Acid batteries (SLABs), which I normally obtain as surplus due to routine replacement in UPS equipment. Recently, this source hasn't been available due to the pandemic, and I find myself with one usable battery and two dead ones.
The two dead ones are reading 5V and 10V. Shaking the one reading 5V, it audibly rattled! Clearly, it wasnt as wet in there as it should be!
So, I thought i'd try and refurbish it! The lid of these is lightly glued on, and it was fairly easy to prise it of using a sharp chisel. Under the lid, the cells are closed off using pushed on rubber caps. These easily lifted off, revealing the plates of the cells, all but dry!
As it happens, I had a bottle of de-ionised water on the shelf in the workshop. Using a syringe, ive refilled the cells up to the top, and then gently sucked the excess out of the filler holes. This proved to be a rather messy job, so a roll of paper towel was kept handy to mop up any spills and splashes.
With each cell refilled, the caps replaced, and the lid put loosely back on, it was time to try it on the charger. For this, I am using my 'semi-intelligent' SLAB charger, and monitoring the current into the battery using a multi-meter. This showed the current rapidly ramping up! Ive left it for now as the current was passing 60mA and climbing. A check of the battery voltage showed it to have risen to well over 7V at this point.
Whether the battery will recover enough to be useful, who knows. Its certainly worth a try.
Now, one thing to note - I have added more water to this first test battery than required. There is a reason for this - I dont know yet how easily the glass matting absorbs the water. It may be that it takes several hours to do so. Supposedly, the correct amount of water is just enough for the matting to appear wet. I will top up the second battery more slowly and observe.
The first battery had dropped to about 6.5V over the 7h since I stopped charging. On restarting the charge, the current input was up at 400mA and rising. The charger can supply 1.5A. As it is now daytime and I can keep an eye on it, I will allow it to charge during the rest of the day and see how it does. For safety, it is in a plastic box and covered with paper towels.
The two dead ones are reading 5V and 10V. Shaking the one reading 5V, it audibly rattled! Clearly, it wasnt as wet in there as it should be!
So, I thought i'd try and refurbish it! The lid of these is lightly glued on, and it was fairly easy to prise it of using a sharp chisel. Under the lid, the cells are closed off using pushed on rubber caps. These easily lifted off, revealing the plates of the cells, all but dry!
As it happens, I had a bottle of de-ionised water on the shelf in the workshop. Using a syringe, ive refilled the cells up to the top, and then gently sucked the excess out of the filler holes. This proved to be a rather messy job, so a roll of paper towel was kept handy to mop up any spills and splashes.
With each cell refilled, the caps replaced, and the lid put loosely back on, it was time to try it on the charger. For this, I am using my 'semi-intelligent' SLAB charger, and monitoring the current into the battery using a multi-meter. This showed the current rapidly ramping up! Ive left it for now as the current was passing 60mA and climbing. A check of the battery voltage showed it to have risen to well over 7V at this point.
Whether the battery will recover enough to be useful, who knows. Its certainly worth a try.
Now, one thing to note - I have added more water to this first test battery than required. There is a reason for this - I dont know yet how easily the glass matting absorbs the water. It may be that it takes several hours to do so. Supposedly, the correct amount of water is just enough for the matting to appear wet. I will top up the second battery more slowly and observe.
The first battery had dropped to about 6.5V over the 7h since I stopped charging. On restarting the charge, the current input was up at 400mA and rising. The charger can supply 1.5A. As it is now daytime and I can keep an eye on it, I will allow it to charge during the rest of the day and see how it does. For safety, it is in a plastic box and covered with paper towels.
Tuesday, 21 July 2020
Off-Grid BOINC research computer proof of concept
Quite some time ago, after a mobile phone upgrade, I hard wired a Buck converter module to the battery terminals of a Samsung Galaxy S7, removed the SIM card, and left it crunching numbers as part of the BOINC project, sat in the corner of the shack powered by the main station PSU. In its steady operating state with no display lit, it consumes about 130mA, which is probably less than the operating current consumed by the big linear PSU powering it!
My plan has always been that my main BOINC contribution should be independent of mains supply, so, after a slight debacle involving a fraudulent sale on ebay of a 100W solar panel with 30A charge controller, which when tested was shown to be a 20W panel and the controller hardly built for 10A, I have a solar panel and charge controller with which to set up a proof of concept... and half my money refunded!
The concept is simple. The BOINC phone and its Buck converter connect to the charge controller load terminals, a 12V 7Ah SLAB to the battery terminals, and the 20W panel to, well, the panel terminals of course! Unfortunately, the only battery of this size I have remaining in an operational state is the one used to power the under stairs lighting! This has been borrowed temporarily. It will be replaced for the time being with a 25Ah beast that just happens to be doing nothing.
The whole lot is in a position where it will receive a decent amount of south facing sun through the day - on my eldest boys windowsill! He wont mind - its already covered in IT and computer parts!
Some tests on the phone showed a peak current at start up of about 450mA, and a steady operational current (computing but display off) of about 130mA. This with a fully charged and 100% capacity 7Ah battery should last several days. At full sun output, the 20W panel should be capable of charging a 7Ah battery in around 5h. So in theory, this set-up should be able to operate 24/7. Of course, this battery is not new and far from peak condition!
If it operates without shutting down in the middle of the night using this battery, then I can have confidence in it operating well with a better 2nd hand unit, which I will have to scrounge up another time!
My plan has always been that my main BOINC contribution should be independent of mains supply, so, after a slight debacle involving a fraudulent sale on ebay of a 100W solar panel with 30A charge controller, which when tested was shown to be a 20W panel and the controller hardly built for 10A, I have a solar panel and charge controller with which to set up a proof of concept... and half my money refunded!
The concept is simple. The BOINC phone and its Buck converter connect to the charge controller load terminals, a 12V 7Ah SLAB to the battery terminals, and the 20W panel to, well, the panel terminals of course! Unfortunately, the only battery of this size I have remaining in an operational state is the one used to power the under stairs lighting! This has been borrowed temporarily. It will be replaced for the time being with a 25Ah beast that just happens to be doing nothing.
The whole lot is in a position where it will receive a decent amount of south facing sun through the day - on my eldest boys windowsill! He wont mind - its already covered in IT and computer parts!
Some tests on the phone showed a peak current at start up of about 450mA, and a steady operational current (computing but display off) of about 130mA. This with a fully charged and 100% capacity 7Ah battery should last several days. At full sun output, the 20W panel should be capable of charging a 7Ah battery in around 5h. So in theory, this set-up should be able to operate 24/7. Of course, this battery is not new and far from peak condition!
If it operates without shutting down in the middle of the night using this battery, then I can have confidence in it operating well with a better 2nd hand unit, which I will have to scrounge up another time!
Thursday, 16 July 2020
NAVTEX Antenna repair - bypassed
The bypassing is now in place. A 100nF cap from pin 1 to pin 11 of the mixer ICs pads (remember that the IC isnt there!), this effectively connects the secondary of L3 to the supply rail. Since this system uses a combined power/signal line, that should now be feeding the receiver. A temporary link takes the other side of the secondary to ground. One leg of a 680 ohm resistor has been lifted to isolate the DC from the secondary, and a final temporary link jumps the missing L4 secondary to re-establish the DC supply to the front end transistors.
With the NAVTEX receiver in spectrum test mode, I can now see 'noise' where-as before this was totally blank, unless I put a finger on the supply wire. I can get a peak to appear on the display if I crank up a 518kHz signal from the Marconi 2955 test set. This is air coupled from a whip antenna, and had to be cranked up to about -20dBm output, but moving the antenna and its feeder around near the whip does suggest that the pick-up is by the antenna and not via the feeder or any other means. So far so good, but the proof of course will be whether it actually receives a NAVTEX transmission!
I have the station reserve receiver tuned to 518kHz as well to compare. Sadly, at the time of writing, were in a quiet part of the transmission schedules!
With the NAVTEX receiver in spectrum test mode, I can now see 'noise' where-as before this was totally blank, unless I put a finger on the supply wire. I can get a peak to appear on the display if I crank up a 518kHz signal from the Marconi 2955 test set. This is air coupled from a whip antenna, and had to be cranked up to about -20dBm output, but moving the antenna and its feeder around near the whip does suggest that the pick-up is by the antenna and not via the feeder or any other means. So far so good, but the proof of course will be whether it actually receives a NAVTEX transmission!
I have the station reserve receiver tuned to 518kHz as well to compare. Sadly, at the time of writing, were in a quiet part of the transmission schedules!
NASA Active Dual Channel NAVTEX Antenna Repair
Despite what the title says, its unlikely this will ever be a dual channel antenna again!
For a start, the PIC 12C508 IC, which provides the switching control and the local oscillator for the channel conversion, is not working. Whether it is just physically damaged but the code can be read, will have to be discovered. Even if the code will read, chances are that its been protected and cannot be used.
The mixer IC MC1496 has also been removed from the board. I have no easy way of testing this, so have to assume for now that it is faulty. Ive removed it to prevent any damage from causing problems with the rest of the circuit.
And so we come to L4 (as I have designated it). This is the output tuned transformer. And its knackered! I had noticed some corrosion under it, as there was with the PIC and the previously rotted off crystal. On removing it from the board I found that the fine wires of the primary had completely corroded away.
This seems at first a major problem, but in reality probably not. The frequency its tuned to is known, 518kHz, and its series capacitance also known, 150pF, so its a simple matter to calculate the secondary inductance needed, which is 630uH. Replacing it is just a matter of finding a suitable 10mm can transformer with that impedance secondary and the same winding configuration.
But for now, I have the trouble of bypassing all this to try and get the direct 'pass-through' 518kHz channel working. All the transistors tested out OK off board on their DC parameters, so hopefully its just a case of getting the supply to them and the signal from them!
The secondary of L3 is the problem. This would normally feed the mixer inputs, but also has a DC level on it for one of those input pins. So, I need to isolate that DC level from the transformer, provide a DC block from the transformer to the output/supply wire to allow the signal out, and temporarily ground the other side of the winding. I 'think' that I can isolate the DC by lifting one leg of one of the 680 ohm resistors, and a 100nF cap should work for coupling the signal.
For a start, the PIC 12C508 IC, which provides the switching control and the local oscillator for the channel conversion, is not working. Whether it is just physically damaged but the code can be read, will have to be discovered. Even if the code will read, chances are that its been protected and cannot be used.
The mixer IC MC1496 has also been removed from the board. I have no easy way of testing this, so have to assume for now that it is faulty. Ive removed it to prevent any damage from causing problems with the rest of the circuit.
And so we come to L4 (as I have designated it). This is the output tuned transformer. And its knackered! I had noticed some corrosion under it, as there was with the PIC and the previously rotted off crystal. On removing it from the board I found that the fine wires of the primary had completely corroded away.
This seems at first a major problem, but in reality probably not. The frequency its tuned to is known, 518kHz, and its series capacitance also known, 150pF, so its a simple matter to calculate the secondary inductance needed, which is 630uH. Replacing it is just a matter of finding a suitable 10mm can transformer with that impedance secondary and the same winding configuration.
But for now, I have the trouble of bypassing all this to try and get the direct 'pass-through' 518kHz channel working. All the transistors tested out OK off board on their DC parameters, so hopefully its just a case of getting the supply to them and the signal from them!
The secondary of L3 is the problem. This would normally feed the mixer inputs, but also has a DC level on it for one of those input pins. So, I need to isolate that DC level from the transformer, provide a DC block from the transformer to the output/supply wire to allow the signal out, and temporarily ground the other side of the winding. I 'think' that I can isolate the DC by lifting one leg of one of the 680 ohm resistors, and a 100nF cap should work for coupling the signal.
Saturday, 11 July 2020
NAVTEX Active Antenna
Ive finally come back to the antenna for the NASA Target NAVTEX receiver.
Ive mapped out the circuit diagram, and can now start working out how to fix it.
My suspicion of the lack of 518kHz was with the transistors, but ive had them off the board today and tested, and they check out ok.
Now, I know the PIC oscillator is not working. The original crystal was corroded, so has been replaced by one close but not exactly right. I would have expected it to work but give perhaps slightly off frequencies. I suspect that at some time the antenna feed has been reverse polarised, and this has killed the chip. It might well have killed the MC1496 as well.
So, the plan now is to remove the MC1496 and the PIC 12C508 from the board, plus lift any associated components that might affect the signal path, and see if I can get it working just on 518kHz, by bypassing the conversion circuitry. Im also going to remove the transformers and test them off-board to ensure their windings are intact, as some corrosion/dirt can be seen under one of them. Testing them on board is pointless as the DC circuitry is routed through them/across their pins.
If I can get that working, then I can work on the conversion side another time.
Ive mapped out the circuit diagram, and can now start working out how to fix it.
My suspicion of the lack of 518kHz was with the transistors, but ive had them off the board today and tested, and they check out ok.
Now, I know the PIC oscillator is not working. The original crystal was corroded, so has been replaced by one close but not exactly right. I would have expected it to work but give perhaps slightly off frequencies. I suspect that at some time the antenna feed has been reverse polarised, and this has killed the chip. It might well have killed the MC1496 as well.
So, the plan now is to remove the MC1496 and the PIC 12C508 from the board, plus lift any associated components that might affect the signal path, and see if I can get it working just on 518kHz, by bypassing the conversion circuitry. Im also going to remove the transformers and test them off-board to ensure their windings are intact, as some corrosion/dirt can be seen under one of them. Testing them on board is pointless as the DC circuitry is routed through them/across their pins.
If I can get that working, then I can work on the conversion side another time.
Subscribe to:
Posts (Atom)