Another nasty realisation looks like it might be the death of the first version of the PRC344 battery warning module!
Yesterday it came to light that a further 100k resistor was required. Not a difficult thing to add in itself. But while shopping for the 4M7 resistors needed for the timing, and realising that they are rather expensive, a consideration of the alternative, that is going to a 10uF capacitor and dropping the resistor down to 470k, led to the discovery that the 1uF ceramic capacitors im using are only rated at 16V!
I really dont know how far I could safely push that rating! A quick test shows that the voltage on the capacitor peaks at about 16V for a 24V supply. But, a freshly charged battery could easily be up to near 28V! And there are also single use 33V Lithium batteries out there! A check at 28V showed the capacitor to have over 19V on it!
There are 10uF 35V SMD tantalum capacitors available at good prices, but these are considerably bigger than the 1uF 0603 ceramic! No chance of a retro-fit mod here!
So, although I have a fair bit of cash tied up in the parts stock for these v.1 modules, it is looking risky to sell them as capable of being reliable in a 24V Clansman system. One option might be to add a potential divider on the supply line to keep the voltages safe of course! I will test that out. The whole module takes only a couple of mA so it should be a safe method, if a little untidy.
But another issue is that a non 50/50 duty cycle is preferred, and to do that requires extra parts anyway. I will test this out on breadboard, but whether I then invest in a v.2 module for sale will have to depend on if there is likely to be enough demand. I am also however going to test out whether or not I can do away with the switching transistor.
So the v.1 warning module looks like it might be a bit of a hybrid SMT/through hole unit, with fixed 3sec on/3sec off warning. This would at least save the project!
Wednesday 20 March 2019
Touchy circuits!
Having received the LM358s, I set about checking the circuit for the PRC344 battery warning would behave as required on the breadboard.
Working now with just a single 8-pin DIL device meant that much of the circuit was scrunched up on the breadboard, making it tricky to see if everything was connected properly, so it was no surprise when it didnt bloody work!
After some time fiddling and testing, I discovered that the battery level detector side of the circuit did work just fine, but the astable multivibrator side didnt!
This now was where it got strange! The astable would work, but only if I had the tip of the multimeter probe touching pin5! Or, even weirder, if I myself touched pin5 or the leg of the resistor connected to it!
I checked everything! I tried different positions on the breadboard, different feedback resistors, battery supply, capacitors bypassing the pin to ground, capacitors bypassing supply to ground, even THREE different ICs! The damn thing just would NOT work unless I touched it!
By now, I was starting to think the worst! Ive got around £20 of PCBs and components tied up in this! What the heck can be wrong? Are the ICs junk?
No. It turns out, with thanks to MrBungle on the UK Vintage Radio Repair and Restoration forum who pointed this out to me, that pin5 in this case requires a 'virtual ground', made up of the two feedback resistors I had, plus another to the supply rail! I nipped out and added another 100k resistor, and lo and behold the bugger is oscillating again!
This means of course that my shiny PCBs are already out of spec and need modding! But, an additional resistor is hardly a big mod and wont make much difference to the space requirements. One aspect of this though is that the circuit now operates correctly, which it seems it wasnt doing with the 741 op-amp, and gives a 50/50 3sec duty cycle. Making it now do the 10sec off/2sec on that I previously had, requires even more modification, with the addition of at least another resistor and a diode, most likely two diodes. I can of course do this on the existing run of PCBs but its going to start getting a bit messy! If the board does prove to have some demand for it, then a mkII version will have to be made with the extra parts.
For now, im just glad that the damn circuit does in fact work! And yes, the big problem of the optocoupler being always slightly on has been solved by using the LM358!
Working now with just a single 8-pin DIL device meant that much of the circuit was scrunched up on the breadboard, making it tricky to see if everything was connected properly, so it was no surprise when it didnt bloody work!
After some time fiddling and testing, I discovered that the battery level detector side of the circuit did work just fine, but the astable multivibrator side didnt!
This now was where it got strange! The astable would work, but only if I had the tip of the multimeter probe touching pin5! Or, even weirder, if I myself touched pin5 or the leg of the resistor connected to it!
I checked everything! I tried different positions on the breadboard, different feedback resistors, battery supply, capacitors bypassing the pin to ground, capacitors bypassing supply to ground, even THREE different ICs! The damn thing just would NOT work unless I touched it!
By now, I was starting to think the worst! Ive got around £20 of PCBs and components tied up in this! What the heck can be wrong? Are the ICs junk?
No. It turns out, with thanks to MrBungle on the UK Vintage Radio Repair and Restoration forum who pointed this out to me, that pin5 in this case requires a 'virtual ground', made up of the two feedback resistors I had, plus another to the supply rail! I nipped out and added another 100k resistor, and lo and behold the bugger is oscillating again!
This means of course that my shiny PCBs are already out of spec and need modding! But, an additional resistor is hardly a big mod and wont make much difference to the space requirements. One aspect of this though is that the circuit now operates correctly, which it seems it wasnt doing with the 741 op-amp, and gives a 50/50 3sec duty cycle. Making it now do the 10sec off/2sec on that I previously had, requires even more modification, with the addition of at least another resistor and a diode, most likely two diodes. I can of course do this on the existing run of PCBs but its going to start getting a bit messy! If the board does prove to have some demand for it, then a mkII version will have to be made with the extra parts.
For now, im just glad that the damn circuit does in fact work! And yes, the big problem of the optocoupler being always slightly on has been solved by using the LM358!
Tuesday 19 March 2019
Dual Dual Op-Amps!
The LM358 dual op-amps finally arrived today - both types! So I now have oodles of both DIL-8 and SOIC-8 dual op-amps to play with!
The first three of the battery warning modules now have their IC mounted. Im still waiting for the higher value 0603 chip resistors though, so cannot complete and test them just yet.
What I can do though now is rejig the breadboard prototype with an LM358 and test the system on that, which will allow me to find the most appropriate values for the sense resistors RA and RB. I do expect the values to be a little different than the LM741 circuit, but the biggest variation will be down to which of the standard values gives the closest threshold to the desired 20V.
Some of you reading this may also be aware of the sad passing of Rev George Dobbs G3RJV recently. The G-QRP club have set up a tribute fund on behalf of the Alzheimers Society, if you wish to make a donation https://george-dobbs.muchloved.com/
The first three of the battery warning modules now have their IC mounted. Im still waiting for the higher value 0603 chip resistors though, so cannot complete and test them just yet.
 |
| Nearly... |
Some of you reading this may also be aware of the sad passing of Rev George Dobbs G3RJV recently. The G-QRP club have set up a tribute fund on behalf of the Alzheimers Society, if you wish to make a donation https://george-dobbs.muchloved.com/
Monday 18 March 2019
Any RF port in a storm
Although I havent been very active on air for a while, I have been running my U3S WSPR beacon on 40m. This had to stop over the weekend when the storm brought my HF doublet down. As I was working, it had to wait for today for repair. Nothing I like more than trying to make a decent Western Union splice stood atop step ladders! I am now repaired and back on air, although I havent yet had time to check the matching to see if the repair has affected anything. WSPR results since the repair seem ok.
With thanks to Alan G8LIT, my Clansman PRC344 is now housed neatly in its own 58 pattern carrier satchel, which cleaned up nicely with a bit of 'chelating mixture' (that is, flour and vinegar) on the metal work, and a spin through the wash.
Ive also acquired a vintage 2m multimode, from near the very start of the era! An FDK Multi-2000! This is a proper beast of a set, and has some rather curious features, such as its dial selected 10kHz steps, but with a +/- 7kHz VXO for FM! It weights a ton, mostly because it is a dual 12V DC/ 240VAC capable set.
It does need some work doing though! The mains switch seems to work but the supply doesnt get from the switch to the regulator! I can power the set by connecting directly to the reservoir capacitor! The VXO doesnt seem to work, its a tad deaf, the output power is down, possibly due to a failed final transistor, a couple of the filament dial bulbs have blown, and the volume pot is noisy. There are also some joints in the internal wiring that have previously been taped and are now falling apart, but these go to the dial bulbs so i'll be replacing those anyway. Its also about 3kHz off frequency. Cosmetically its not bad for its age, theres a crack in the fascia but the knobs and front panel etc are all ok. Minor repair stuff really. Someone in the past has added a homebrew receive pre-amp as well.
One good thing about radios of this vintage is that the manuals came with the alignment details and circuit diagrams. So I have the manual, and will set about restoring it to usable condition. I dont think it will be too hard, most of the problems are very minor. My only real concern is the PA, getting a suitable replacement final transistor might be a challenge. I suspect the receiver can be brought up a bit just from alignment, but its possible that parts will be needed in the front end, or maybe even that homebrew pre-amp! Those big electrolytics might be a bit suspect as well after all these years!
The switching transistors for the battery warning modules came today, so my first three units now have that fitted. I do hope the rest of the parts dont take too long to arrive now!
With thanks to Alan G8LIT, my Clansman PRC344 is now housed neatly in its own 58 pattern carrier satchel, which cleaned up nicely with a bit of 'chelating mixture' (that is, flour and vinegar) on the metal work, and a spin through the wash.
Ive also acquired a vintage 2m multimode, from near the very start of the era! An FDK Multi-2000! This is a proper beast of a set, and has some rather curious features, such as its dial selected 10kHz steps, but with a +/- 7kHz VXO for FM! It weights a ton, mostly because it is a dual 12V DC/ 240VAC capable set.
 |
| Vintage FDK Multi-2000 2m multimode transceiver |
 |
| Top View |
 |
| Bottom View. Note homebrew pre-amp! |
It does need some work doing though! The mains switch seems to work but the supply doesnt get from the switch to the regulator! I can power the set by connecting directly to the reservoir capacitor! The VXO doesnt seem to work, its a tad deaf, the output power is down, possibly due to a failed final transistor, a couple of the filament dial bulbs have blown, and the volume pot is noisy. There are also some joints in the internal wiring that have previously been taped and are now falling apart, but these go to the dial bulbs so i'll be replacing those anyway. Its also about 3kHz off frequency. Cosmetically its not bad for its age, theres a crack in the fascia but the knobs and front panel etc are all ok. Minor repair stuff really. Someone in the past has added a homebrew receive pre-amp as well.
One good thing about radios of this vintage is that the manuals came with the alignment details and circuit diagrams. So I have the manual, and will set about restoring it to usable condition. I dont think it will be too hard, most of the problems are very minor. My only real concern is the PA, getting a suitable replacement final transistor might be a challenge. I suspect the receiver can be brought up a bit just from alignment, but its possible that parts will be needed in the front end, or maybe even that homebrew pre-amp! Those big electrolytics might be a bit suspect as well after all these years!
The switching transistors for the battery warning modules came today, so my first three units now have that fitted. I do hope the rest of the parts dont take too long to arrive now!
Thursday 14 March 2019
Prototype PCBs
Well, I have to say, im rather impressed!
I have been using the online EasyEDA to design the PCB layouts for the various Clansman projects im working on. Now, as much as I dislike the way EDA software uses device package symbols that bear little relation to how I was taught to lay out a schematic, you cant deny that the advantage of being able to let the software design the circuit boards takes a lot of the hard work of prototype design out. But what has impressed me is something that I can be forgiven for being skeptical of - that EasyEDA is 'tied in' with a Far East supplier of PCBs and components.
Having designed the SMD PCB for the PRC344 battery warning circuit, I decided I would take advantage of the offer of 10 PCBs for just $2 (about £1.60), to see how they came out.
I did make a bit of a mistake by not spotting which shipping service was selected, and got an invoice for well over £20! But having corrected that, I decided that the cheaper shipping, which brought the total up to £6, was still worth the risk. The first thing that raised an eyebrow was the near flawless tracking - the entire process was tracked! But even more surprising was being told the boards were etched and ready to ship just 2 days later!
They took about a week to arrive, which really isnt bad at all from China! And, as it turned out, most of the delay was by Royal Mail! And I have to say, they do look pretty good
Due to their very small size, the silkscreen print is rather limited, and given a choice I personally would have made some of the track a little wider ( I could have done that, but had no reference from which to decide how much thicker to go from the default).
Next step of course is to make one up and test it. Unfortunately, I am still awaiting delivery of several components, not least the LM358 op-amp that is at the heart of the circuit!
I gather that the low cost offer for 10x PCBs holds for boards up to 100mm x 100mm! My boards are about 15mm x 25mm! If I could just find a way to layout SMT parts on both sides of the PCB, I could create some amazingly compact circuits!
But, there are difficulties! To make these boards fit in the only safe available space in the PRC344, I had to choose 0603 sized parts. These are damn small! And to attach them to the board im using solder paste out of a tub! This is a new idea to me - in the past whenever ive had to rework SMT, ive had access to solder paste in dispensing syringes and a controlled pump, plus hot air reflow equipment. Here, im having to apply the paste with the tip of a probe, and then melt it with the tip of my soldering iron. Add on the fact that I dont have access to the high power magnification I used to use either, and it all makes for a rather tricky job!
But, that said, ive made a start. Im making three up at the moment, one is for my '344, the other two for further test and evaluation.
The Zener diode is a miniMELF package. These are tricky buggers, as they are little cylinders and apt to roll off! As the only parts that I currently have stocked are the zener, opto and capacitor, these are the only parts ive been able to install just now. The capacitor is just about visible in the photo below!
Each was then tested for continuity to ensure there were no shorts underneath! Once completed and tested, then connecting wires will be measured and attached, and the whole board housed in a length of heatshrink tubing. As nice as it is to be able to see the electronics, putting it into heatshrink will ensure it cannot cause any problems inside the radio!
As this design uses fixed resistors to save complexity and cost, its likely that a little compromise will be needed as to the exact threshold voltage at which the warning is activated. I'll aim to make this as close to 20v as I can, but it depends on where the standard resistor values come out. If I cant make it close enough with standard values, then one option is to parallel standard values by mounting them on top of each other! All that can be determined though using standard through hole parts on the breadboard once the DIL LM358s arrive! Of course, if I decide to offer some of these as kits, then the buyer can experiment as they see fit with the values of RA and RB, likewise, if they are not happy with the duty cycle of the alarm, they can play with other values as they wish.
I have been using the online EasyEDA to design the PCB layouts for the various Clansman projects im working on. Now, as much as I dislike the way EDA software uses device package symbols that bear little relation to how I was taught to lay out a schematic, you cant deny that the advantage of being able to let the software design the circuit boards takes a lot of the hard work of prototype design out. But what has impressed me is something that I can be forgiven for being skeptical of - that EasyEDA is 'tied in' with a Far East supplier of PCBs and components.
Having designed the SMD PCB for the PRC344 battery warning circuit, I decided I would take advantage of the offer of 10 PCBs for just $2 (about £1.60), to see how they came out.
I did make a bit of a mistake by not spotting which shipping service was selected, and got an invoice for well over £20! But having corrected that, I decided that the cheaper shipping, which brought the total up to £6, was still worth the risk. The first thing that raised an eyebrow was the near flawless tracking - the entire process was tracked! But even more surprising was being told the boards were etched and ready to ship just 2 days later!
They took about a week to arrive, which really isnt bad at all from China! And, as it turned out, most of the delay was by Royal Mail! And I have to say, they do look pretty good
Due to their very small size, the silkscreen print is rather limited, and given a choice I personally would have made some of the track a little wider ( I could have done that, but had no reference from which to decide how much thicker to go from the default).
Next step of course is to make one up and test it. Unfortunately, I am still awaiting delivery of several components, not least the LM358 op-amp that is at the heart of the circuit!
I gather that the low cost offer for 10x PCBs holds for boards up to 100mm x 100mm! My boards are about 15mm x 25mm! If I could just find a way to layout SMT parts on both sides of the PCB, I could create some amazingly compact circuits!
But, there are difficulties! To make these boards fit in the only safe available space in the PRC344, I had to choose 0603 sized parts. These are damn small! And to attach them to the board im using solder paste out of a tub! This is a new idea to me - in the past whenever ive had to rework SMT, ive had access to solder paste in dispensing syringes and a controlled pump, plus hot air reflow equipment. Here, im having to apply the paste with the tip of a probe, and then melt it with the tip of my soldering iron. Add on the fact that I dont have access to the high power magnification I used to use either, and it all makes for a rather tricky job!
But, that said, ive made a start. Im making three up at the moment, one is for my '344, the other two for further test and evaluation.
 |
| Fitting the miniMELF zener |
The Zener diode is a miniMELF package. These are tricky buggers, as they are little cylinders and apt to roll off! As the only parts that I currently have stocked are the zener, opto and capacitor, these are the only parts ive been able to install just now. The capacitor is just about visible in the photo below!
 |
| zener, 1uF capacitor and optocoupler installed |
As this design uses fixed resistors to save complexity and cost, its likely that a little compromise will be needed as to the exact threshold voltage at which the warning is activated. I'll aim to make this as close to 20v as I can, but it depends on where the standard resistor values come out. If I cant make it close enough with standard values, then one option is to parallel standard values by mounting them on top of each other! All that can be determined though using standard through hole parts on the breadboard once the DIL LM358s arrive! Of course, if I decide to offer some of these as kits, then the buyer can experiment as they see fit with the values of RA and RB, likewise, if they are not happy with the duty cycle of the alarm, they can play with other values as they wish.
Monday 11 March 2019
Rev. George Dobbs G3RJV - A deep loss to our hobby community
With great sadness I regret to have to forward on the announcement of the death of the Rev. George Dobbs, G3RJV, founder of the G-QRP club. Anyone who has ever met, or had any contact with, George will know that he was a true gentleman, the guiding light of the QRP movement, and the true epitome of the spirit of Amateur Radio.
There is a tribute collection on behalf of the Alzheimers Society https://george-dobbs.muchloved.com/
There is a tribute collection on behalf of the Alzheimers Society https://george-dobbs.muchloved.com/
Message as received from G-QRP club chairman -
Mon Mar 11, 2019 1:59 am (PDT) . Posted by:
"Steve Hartley" g0fuw
It
is with a heavy heart that I have to report that the Reverend George Dobbs,
G3RJV, founder of the GQRP Club passed away in the early hours of this
morning.
George had been unwell for some time but he had been living quite comfortably in a care home.
Unfortunately, his condition deteriorated quite rapidly over the last few days and his life ended at 04:00 this morning.
Jo, his wife, was with him when he passed.
A full obituary will follow and all GQRP members will receive a special G3RJV memorial edition of SPRAT.
RIP George.
73, & 72, Steve, G0FUW
Chairman
GQRP Club
George had been unwell for some time but he had been living quite comfortably in a care home.
Unfortunately, his condition deteriorated quite rapidly over the last few days and his life ended at 04:00 this morning.
Jo, his wife, was with him when he passed.
A full obituary will follow and all GQRP members will receive a special G3RJV memorial edition of SPRAT.
RIP George.
73, & 72, Steve, G0FUW
Chairman
GQRP Club
Sunday 10 March 2019
Bits and Bobs
Had a day out with Bob M1BBV at Grantham rally today. Somewhat disappointed, it was much smaller than expected, and its a long way. But any trip out with Bob is always an experience! Intended to also take a walk out to the targets at Donna Nook, but the weather was awful.
After much deliberation, advice seeking and calculations, the resistors are on their way to create a grounding and static bleed unit in order to be able to safely fly a kite antenna. Annoyingly, there was a small winch at the rally that I could have got that would have made flying an antenna kite much easier!
One positive is that my gas bill is coming down. Seems the mild winter so far has reduced the wholesale price, and my supplier is quite good at tracking this and reducing prices. They are also still doing their £50 credit offer when switching using a customers personal link, so, i'll put that link at the bottom of this post!
I am now in a position to test some of the PRC344 scan controllers code features against the actual radio. This will be done by a mock-up with test clip wires, and will hopefully prove that the three key control inputs - synth lock, mute and over-ride, work as expected. One aspect of the controller that I hadn't appreciated, and that required some rather complex redesign, was the way the selector switches would behave if isolated!
Essentially, my plan was to use Schottky diodes to isolate my control lines from those of the switches, allowing the existing pull-up resistors to hold unused lines high, and to prevent the switches influencing the synthesiser by disconnecting the switches ground. Sounds like a sensible plan, until you remember that these are BCD switches! If I leave the commons of the switches floating - the net result is that any closed switch is pulled high by leakage through from the pull-ups! As best I can work out, this would completely screw up the system! So, ive designed an isolator module based on 4066 ICs to completely disconnect the switches!
I do still need to test this theory, before I build the isolator. If there is any way to do this without the additional electronics it would of course be preferable!
If anyone is interested in switching energy supplier, do your research, and if Bulb turn out to be good for you, then feel free to use my link to switch, and grab £50 credit! -
bulb.co.uk/
After much deliberation, advice seeking and calculations, the resistors are on their way to create a grounding and static bleed unit in order to be able to safely fly a kite antenna. Annoyingly, there was a small winch at the rally that I could have got that would have made flying an antenna kite much easier!
One positive is that my gas bill is coming down. Seems the mild winter so far has reduced the wholesale price, and my supplier is quite good at tracking this and reducing prices. They are also still doing their £50 credit offer when switching using a customers personal link, so, i'll put that link at the bottom of this post!
I am now in a position to test some of the PRC344 scan controllers code features against the actual radio. This will be done by a mock-up with test clip wires, and will hopefully prove that the three key control inputs - synth lock, mute and over-ride, work as expected. One aspect of the controller that I hadn't appreciated, and that required some rather complex redesign, was the way the selector switches would behave if isolated!
Essentially, my plan was to use Schottky diodes to isolate my control lines from those of the switches, allowing the existing pull-up resistors to hold unused lines high, and to prevent the switches influencing the synthesiser by disconnecting the switches ground. Sounds like a sensible plan, until you remember that these are BCD switches! If I leave the commons of the switches floating - the net result is that any closed switch is pulled high by leakage through from the pull-ups! As best I can work out, this would completely screw up the system! So, ive designed an isolator module based on 4066 ICs to completely disconnect the switches!
I do still need to test this theory, before I build the isolator. If there is any way to do this without the additional electronics it would of course be preferable!
If anyone is interested in switching energy supplier, do your research, and if Bulb turn out to be good for you, then feel free to use my link to switch, and grab £50 credit! -
bulb.co.uk/
Sunday 3 March 2019
Design Success?
The rethink of the battery warning design took only a few minutes in the workshop to prove out. First, I took measurements of the actual current draw of each section of the design. The pulse generator, with no load on the output, requires 1.8mA. The low voltage detector requires about 1.6mA. Hardly going to break the bank with a 4Ah battery! So, its not a big issue if the pulse generator is running all the time.
This made life much easier! A quick rearrange of the breadboard and the output of the pulse generator is now only capable of turning on the optocoupler when the detectors output turns on the control transistor!
This did mean a very simple yet fundamental change to the detector - I had to swap the inverting and non-inverting inputs over. Now, the output of the detector is high when the battery is below threshold. This turns on the transistor and grounds the cathode of the optocoupler.
Of course, at present on the breadboard with 741s in use, the transistor is never actually switched off! I measure 0.7v on the base due to the poor swing limits of the 741. With any luck, the LM358s will not only cure that, but the dual device means I can design the PCB much, much smaller! A surface mount NPN transistor takes up much less space than a SOIC-8 package! With the potential divider changed back to individual fixed resistors, to be selected on test (SoT), my current PCB design is down to slightly less than 24mm x 13mm!
I eventually managed to get my turn on the EasyEDA autorouter, and now have a fully designed PCB! I'll check it all over to ensure ive not made any silly design mistakes, and then see about getting some boards etched.
This made life much easier! A quick rearrange of the breadboard and the output of the pulse generator is now only capable of turning on the optocoupler when the detectors output turns on the control transistor!
This did mean a very simple yet fundamental change to the detector - I had to swap the inverting and non-inverting inputs over. Now, the output of the detector is high when the battery is below threshold. This turns on the transistor and grounds the cathode of the optocoupler.
Of course, at present on the breadboard with 741s in use, the transistor is never actually switched off! I measure 0.7v on the base due to the poor swing limits of the 741. With any luck, the LM358s will not only cure that, but the dual device means I can design the PCB much, much smaller! A surface mount NPN transistor takes up much less space than a SOIC-8 package! With the potential divider changed back to individual fixed resistors, to be selected on test (SoT), my current PCB design is down to slightly less than 24mm x 13mm!
I eventually managed to get my turn on the EasyEDA autorouter, and now have a fully designed PCB! I'll check it all over to ensure ive not made any silly design mistakes, and then see about getting some boards etched.
Miniaturization - Back to the drawing board!
Well, I honestly thought i'd cracked it!
The two 741s work ok, but I need devices that can get much closer to ground in order to ensure my optocoupler is turned off. So, move over to the LM358 - gets very close to ground, and so cheap theyre practically giving them away from China! 100 devices ordered, 50/50 DIL and SOIC, for £2.47. And since I found £2 abandoned in the gym lockers, net cost to me 47p!
Being dual rather than single devices, I thought i'd also got it down to an incredibly tiny little circuit...
...only thats not the case. My circuit relies on the output on the 1st op-amp controlling the ground pin of the 2nd. Only trouble is - a dual device only has one set of supply pins!
So at the moment im still stuck using TWO physical devices, leaving me with two unused op-amps on the board. The only other thing that is preventing me getting really small is the preset resistor. I reverted back to this with the idea of selling a few of these devices to make back the development cost, and thought that the end user would probably prefer to have the threshold voltage adjustable. And at least in the PRC344, space is an absolute premium, and the board has to be physically as small as possible! There are of course SMT preset resistors, but ive yet to find them in the EDA library, so at present cant design a board to use them! That would be ideal and i'll ask on the forums for a package file. The other alternative is to go back to fixed resistors and let the user decide their own values.
At present with the standard PCB preset and two SOIC LM358s, ive got the PCB down to 27mm x 17mm. Thats not bad, but is still a tight fit in the only 'safe' places in the radio.
So, making the advance decision to go back to fixed resistors, my limiting factor is the two ICs. How do I get around this and down to a single IC?
Ive done some playing, and theres no easy way to simply use the output of one op-amp to control the other via its inputs. So perhaps thats not the way to go? Rather than trying to control the op-amp pulse generator itself, maybe just control its output getting to the optocoupler? The detector op-amp could switch a transistor in the cathode of the opto. This would mean that the pulse generator was running continuously, but if theres no load on the output, is that a problem? How much current does it draw with no output load?
Time for some measurements and more testing on the breadboards!
The two 741s work ok, but I need devices that can get much closer to ground in order to ensure my optocoupler is turned off. So, move over to the LM358 - gets very close to ground, and so cheap theyre practically giving them away from China! 100 devices ordered, 50/50 DIL and SOIC, for £2.47. And since I found £2 abandoned in the gym lockers, net cost to me 47p!
Being dual rather than single devices, I thought i'd also got it down to an incredibly tiny little circuit...
...only thats not the case. My circuit relies on the output on the 1st op-amp controlling the ground pin of the 2nd. Only trouble is - a dual device only has one set of supply pins!
So at the moment im still stuck using TWO physical devices, leaving me with two unused op-amps on the board. The only other thing that is preventing me getting really small is the preset resistor. I reverted back to this with the idea of selling a few of these devices to make back the development cost, and thought that the end user would probably prefer to have the threshold voltage adjustable. And at least in the PRC344, space is an absolute premium, and the board has to be physically as small as possible! There are of course SMT preset resistors, but ive yet to find them in the EDA library, so at present cant design a board to use them! That would be ideal and i'll ask on the forums for a package file. The other alternative is to go back to fixed resistors and let the user decide their own values.
At present with the standard PCB preset and two SOIC LM358s, ive got the PCB down to 27mm x 17mm. Thats not bad, but is still a tight fit in the only 'safe' places in the radio.
So, making the advance decision to go back to fixed resistors, my limiting factor is the two ICs. How do I get around this and down to a single IC?
Ive done some playing, and theres no easy way to simply use the output of one op-amp to control the other via its inputs. So perhaps thats not the way to go? Rather than trying to control the op-amp pulse generator itself, maybe just control its output getting to the optocoupler? The detector op-amp could switch a transistor in the cathode of the opto. This would mean that the pulse generator was running continuously, but if theres no load on the output, is that a problem? How much current does it draw with no output load?
Time for some measurements and more testing on the breadboards!
Post-op-amp
My previously mentioned battery warning circuit, using two 741 op-amps, is now working on the breadboard. Although the concept is sound, and it works almost as expected, it has thrown up an interesting problem.
One thing I proved to myself while prototyping this, was the effect of reversing the input pins on the op-amp. As built, with the zener diode on pin 2, the output goes low when the input voltage drops below the threshold. Reversing the inputs does what you'd expect, and reverses the output. The output is low so long as the input is above the threshold. I just wanted to prove that to myself!
This meant that I could then try both ways of powering the 2nd op-amp, either by pulling its GND pin low, or by using the 1st op-amp as source to feed its Vcc pin. Both work, but I think from convention pulling the GND pin low is preferable.
The above photo shows the two op-amps working together. In this mock-up ive reverted to the preset for the threshold setting, and the 2nd op-amp is directly driving the LED via a 3k3 series resistor.
So far, so froody. The problem becomes more apparent (it can be seen in the direct drive mock-up but its not as noticeable) when the opto-coupler is added, as can bee seen in this little video
Because the 741 cant get really close to its supply rails, this means that the output lows are always a volt or two above ground. The cumulative effect of this is that the opto-coupler never actually turns full off!
Clearly this is no good, as it means that whatever is controlled from the opto-coupler will not be switched cleanly. It might now make much difference in practice, but it isnt good practice!
Im hoping that this is something that will be cured by moving to the better specified LM358 dual op-amp. If ive read the datasheet correctly, this can get down to within a few mV of its ground, which should mean that the combined effect of perhaps a few dozen mV is not enough to turn the opto-coupler on.
One thing I proved to myself while prototyping this, was the effect of reversing the input pins on the op-amp. As built, with the zener diode on pin 2, the output goes low when the input voltage drops below the threshold. Reversing the inputs does what you'd expect, and reverses the output. The output is low so long as the input is above the threshold. I just wanted to prove that to myself!
This meant that I could then try both ways of powering the 2nd op-amp, either by pulling its GND pin low, or by using the 1st op-amp as source to feed its Vcc pin. Both work, but I think from convention pulling the GND pin low is preferable.
The above photo shows the two op-amps working together. In this mock-up ive reverted to the preset for the threshold setting, and the 2nd op-amp is directly driving the LED via a 3k3 series resistor.
So far, so froody. The problem becomes more apparent (it can be seen in the direct drive mock-up but its not as noticeable) when the opto-coupler is added, as can bee seen in this little video
Because the 741 cant get really close to its supply rails, this means that the output lows are always a volt or two above ground. The cumulative effect of this is that the opto-coupler never actually turns full off!
Clearly this is no good, as it means that whatever is controlled from the opto-coupler will not be switched cleanly. It might now make much difference in practice, but it isnt good practice!
Im hoping that this is something that will be cured by moving to the better specified LM358 dual op-amp. If ive read the datasheet correctly, this can get down to within a few mV of its ground, which should mean that the combined effect of perhaps a few dozen mV is not enough to turn the opto-coupler on.
'344/'320 low battery circuit
Well it didnt take me long to become dissatisfied with my low battery alarm circuit!
I realised almost straight away that the circuit as built is pretty much game over if the battery reaches the threshold. The tone being constant means that no further listening is possible!
The PRC351 does its low battery warning by pulsing the mute circuit open, causing a brief burst of noise in the audio gear, every few seconds or so. Since my circuit uses an optocoupler output, in theory this could be connected to the '344s mute circuit to achieve the same, if the ouput could be pulsed.
So, last night I threw this little beastie onto the breadboard -
With the values shown I found this lit the LED for about 1sec every 3-4sec. Not sure if that duty cycle is adjustable, it would be nice to be able to choose to lengthen the 'off' time. This simple circuit therefore is ideal for giving a pulsed warning that wont disturb actual radio listening too badly, but will still indicate that the user should change the battery.
Next step then, is to couple it to the low battery detector circuit previously built.
As the detector output goes low when the threshold is reached, my thought is that I can put it in the ground connection of the pulse generator, in order to switch that on. Im not sure this will work, but i'll give it a go. If not then i'll need to add a switching transistor, unless I can juggle the inputs of the detector such that the output goes high when the threshold is reached, and I could then just drive the pulse generator supply line. But my thought is that with the detector output high, I will be putting roughly equal positive voltage on both the supply and ground pins, with a net result of no effective supply.
This design does of course add a minimum of four extra parts into an already grossly cramped space. So to fit in the radio this will have to be made surface mount. All the parts can easily be done this way, although some are not very cheap (the opto and the preset resistor if I make the threshold adjustable again) The 741s will be replaced by a LM358 dual op-amp in a SOIC package.
If im going to do this in SMT, then I may as well also design a PCB and take advantage of the rediculously low cost PCB offers available, make a few up and sell them on!
Im pretty sure the same circuit could be used to put a low battery warning on the PRC320 as well, but i'll look into that later.
I realised almost straight away that the circuit as built is pretty much game over if the battery reaches the threshold. The tone being constant means that no further listening is possible!
The PRC351 does its low battery warning by pulsing the mute circuit open, causing a brief burst of noise in the audio gear, every few seconds or so. Since my circuit uses an optocoupler output, in theory this could be connected to the '344s mute circuit to achieve the same, if the ouput could be pulsed.
So, last night I threw this little beastie onto the breadboard -
With the values shown I found this lit the LED for about 1sec every 3-4sec. Not sure if that duty cycle is adjustable, it would be nice to be able to choose to lengthen the 'off' time. This simple circuit therefore is ideal for giving a pulsed warning that wont disturb actual radio listening too badly, but will still indicate that the user should change the battery.
Next step then, is to couple it to the low battery detector circuit previously built.
As the detector output goes low when the threshold is reached, my thought is that I can put it in the ground connection of the pulse generator, in order to switch that on. Im not sure this will work, but i'll give it a go. If not then i'll need to add a switching transistor, unless I can juggle the inputs of the detector such that the output goes high when the threshold is reached, and I could then just drive the pulse generator supply line. But my thought is that with the detector output high, I will be putting roughly equal positive voltage on both the supply and ground pins, with a net result of no effective supply.
This design does of course add a minimum of four extra parts into an already grossly cramped space. So to fit in the radio this will have to be made surface mount. All the parts can easily be done this way, although some are not very cheap (the opto and the preset resistor if I make the threshold adjustable again) The 741s will be replaced by a LM358 dual op-amp in a SOIC package.
If im going to do this in SMT, then I may as well also design a PCB and take advantage of the rediculously low cost PCB offers available, make a few up and sell them on!
Im pretty sure the same circuit could be used to put a low battery warning on the PRC320 as well, but i'll look into that later.
Saturday 2 March 2019
Clansman PRC344 Low Battery Alarm
Ive decided to slightly rename this project as a low battery alarm rather than a warning, as in use, once the battery is too low, the 2kHz CALL tone pretty much makes any further listening impossible!
The circuit is very simple. I make up a potential divider chain rather than leave one of my few stock presets in circuit, but of course if you have a 100k - 200k preset spare then using that will mean you can adjust the threshold to taste. In this case the alarm threshold is about 20v. This is ideal as it means the battery is not discharged lower than 1v/cell. Note in particular that 'battery' on the diagram refers to a switched 24v line inside the radio - not the raw battery connection!
I knocked the prototype up on matrix board. This is a bit too big really even with the edges trimmed off!. Ideally to fit comfortably in the radio SMT parts would be used. But, wrapped in several layers of insulating tape (ive no heatshrink wide enough!), it is now installed in my '344. The red wire is the 24v, which ive connected to a handy pin6 on one of the regulators. Yellow is 3v from pin3 of the 3v regulator. Black is ground. The blue wire connects to module 17 pin 3. You can see how it fits and how it works in the little video below -
 |
| Battery alarm circuit. NB - 'Battery' has to be a switched line! |
The circuit is very simple. I make up a potential divider chain rather than leave one of my few stock presets in circuit, but of course if you have a 100k - 200k preset spare then using that will mean you can adjust the threshold to taste. In this case the alarm threshold is about 20v. This is ideal as it means the battery is not discharged lower than 1v/cell. Note in particular that 'battery' on the diagram refers to a switched 24v line inside the radio - not the raw battery connection!
I knocked the prototype up on matrix board. This is a bit too big really even with the edges trimmed off!. Ideally to fit comfortably in the radio SMT parts would be used. But, wrapped in several layers of insulating tape (ive no heatshrink wide enough!), it is now installed in my '344. The red wire is the 24v, which ive connected to a handy pin6 on one of the regulators. Yellow is 3v from pin3 of the 3v regulator. Black is ground. The blue wire connects to module 17 pin 3. You can see how it fits and how it works in the little video below -
A later revision of this would be to use a double op-amp to pulse the mute open, similar to how the '351 low battery warning works. But that would have more components and would demand SMT parts to fit the radio.
Battery Warning Made Simple
I was going to incorporate a battery monitor function into the PRC344 scan controller project, but in order that the device itself wouldnt drain the battery, this would have involved so many power saving modes, wake-on-interrupt functions and shockingly expensive ultra low quiescent current regulators, unless the scan controller was actually in operation, that I decided against it. No doubt at a much later stage of development I will revisit this function, perhaps when incorporating the external control module, but for now ive decided against putting this function into the MCU.
But one of the flaws of the PRC344, and also the popular PRC320, is that unlike the PRC350, PRC351/2, and the PRC349, they have no warning of a low battery! For amateur use this is a big problem, as it makes it very easy to accidentally leave the set on, and end up potentially destroying a good battery by massively over discharging it.
So, I thought up a little add-on circuit that should be compact enough to fit into either the '344 or the '320, and will give a warning of an excessively low battery. Its very simple, just a cheap 741 op-amp acting as a comparator, a zener diode reference, a fixed potential divider (originally a 200k preset but fixed resistors are smaller and cheaper), and an optocoupler. This is set such that the optocoupler is turned ON when the supply to the circuit falls below around 20v.
Within the '344, there is an easily accessible 2kHz tone oscillator, used to provide the CALL function. As luck would have it, this is part of module 17, which is also the battery input module where the bizarre Schottky diode is! To activate the tone, 3v is connected to pin 3 of this module. All I have to do is connect my little device's optocoupler emitter to Pin 3 of module 17, the collector to 3v, its supply to the switched 24v, and a convenient 0v. When the battery is too low, the CALL tone will be heard in the audio gear.
The battery warning circuit draws just 2mA, rising to 3mA in a low battery condition. Add to this 4.7mA to drive the PC817 optocoupler, for a maximum 7.7mA. At first it seems a bit counter intuitive that the battery low warning system should draw more current when the battery is low than when fresh, but this is of little consequence as the warning signal should cause the operator to either change the battery or turn the set off!
The Scan Controller code is now coming along, with much help from Mikroelectronika forum members Hexreader and Janni. These two kind persons have helped me with many code examples, and, to a greater extent, complete sections of workable code! Its not quite ready to implement in the radio yet, but it is now very close to a working system on breadboard. A few minor snags with interrupt handling need to be smoothed out.
How it will work in the radio itself ive yet to discover! Until I receive the TSSOP break out boards im unable to make use of the HV multiplexer ICs for testing, but I might be able to connect up the Synth Lock and Mute lines to test against the breadboard, plus the CALL switch for the over-ride interrupt.
But one of the flaws of the PRC344, and also the popular PRC320, is that unlike the PRC350, PRC351/2, and the PRC349, they have no warning of a low battery! For amateur use this is a big problem, as it makes it very easy to accidentally leave the set on, and end up potentially destroying a good battery by massively over discharging it.
So, I thought up a little add-on circuit that should be compact enough to fit into either the '344 or the '320, and will give a warning of an excessively low battery. Its very simple, just a cheap 741 op-amp acting as a comparator, a zener diode reference, a fixed potential divider (originally a 200k preset but fixed resistors are smaller and cheaper), and an optocoupler. This is set such that the optocoupler is turned ON when the supply to the circuit falls below around 20v.
Within the '344, there is an easily accessible 2kHz tone oscillator, used to provide the CALL function. As luck would have it, this is part of module 17, which is also the battery input module where the bizarre Schottky diode is! To activate the tone, 3v is connected to pin 3 of this module. All I have to do is connect my little device's optocoupler emitter to Pin 3 of module 17, the collector to 3v, its supply to the switched 24v, and a convenient 0v. When the battery is too low, the CALL tone will be heard in the audio gear.
 |
| Battery Warning circuit mock-up |
The Scan Controller code is now coming along, with much help from Mikroelectronika forum members Hexreader and Janni. These two kind persons have helped me with many code examples, and, to a greater extent, complete sections of workable code! Its not quite ready to implement in the radio yet, but it is now very close to a working system on breadboard. A few minor snags with interrupt handling need to be smoothed out.
How it will work in the radio itself ive yet to discover! Until I receive the TSSOP break out boards im unable to make use of the HV multiplexer ICs for testing, but I might be able to connect up the Synth Lock and Mute lines to test against the breadboard, plus the CALL switch for the over-ride interrupt.
Subscribe to:
Posts (Atom)