Yesterday’s post was rather long, but was intended for mostly technical audiences outside of amateur radio.  This post serves as a brain dump of volatile memory before I go to sleep for the night.  (Human conscious memory is more like D-RAM than one might realise.)

So, many in our group use packet radio TNCs already, with a good number using the venerable Kantronics KPC3.  These have a DB9 port that connects to the radio and a second DB25 RS-323 port that connects to the computer.

My proposal: we make an audio interface that either plugs into that DB9 port and re-uses the interface cables we already have, or directly into the radio’s data port.

This should connect to an audio interface on the computer.

For EMI’s sake, I’d recommend a USB sound dongle like this, or these, or this as that audio interface.  I looked on Jaycar and did see this one, which would also work (and burn a hole in your wallet!).

If you walk in and the asking price is more than \$30, I’d seriously consider these other options.  Of those options, U-Mart are here in Brisbane; go to their site, order a dongle then tell the site you’ll come and pick it up.  They’ll send you an email with an order number when it’s ready, you just need to roll up to the store, punch that number into a terminal in the shop, then they’ll call your name out for you to collect and pay for it.

Scorptec are in Melbourne, so you’ll have to have items shipped, but are also worth talking to.  (They helped me source some bits for my server cluster when U-Mart wouldn’t.)

USB works over two copper pairs; one delivers +5V and 0V, the other is a differential pair for data.  In short, the USB link should be pretty immune from EMI issues.

At worst, you should be able to deal with it with judicious application of ferrite beads to knock down the common mode current and using a combination of low-ESR electrolytic and ceramic capacitors across the power rails.

If you then keep the analogue cables as short as absolutely possible, you should have little opportunity for RF to get in.

I don’t recommend the TigerTronics Signalink interfaces, they use cheap and nasty isolation transformers that lead to serious performance issues.

For the receive audio, we feed the audio from the radio and we feed that via potentiometer to a 3.5mm TRS (“phono”) plug tip, with sleeve going to common.  This plugs into the Line-In or Microphone input on the sound device.

### Push to Talk and Transmit audio

I’ve bundled these together for a good reason.  The conventional way for computers to drive PTT is via an RS-232 serial port.

We can do that, but we won’t unless we have to.

Unless you’re running an original SoundBLASTER card, your audio interface is likely stereo.  We can get PTT control via an envelope detector forming a minimal-latency VOX control.

Another 3.5mm TRS plug connects to the “headphone” or “line-out” jack on our sound device and breaks out the left and right channels.

The left and right channels from the sound device should be fed into the “throw” contacts on two single-pole double-throw toggle switches.

The select pin (mechanically operated by the toggle handle) on each switch thus is used to select the left or right channel.

One switch’s select pin feeds into a potentiometer, then to the radio’s input.  We will call that the “modulator” switch; it selects which channel “modulates” our audio.  We can again adjust the gain with the potentiometer.

The other switch first feeds through a small Schottky diode then across a small electrolytic capacitor (to 0V) then through a small resistor before finally into the base of a small NPN signal transistor (e.g. BC547).  The emitter goes to 0V, the collector is our PTT signal.

This is the envelope detector we all know and love from our old experiments with crystal sets.  In theory, we could hook a speaker to the collector up to a power source and listen to AM radio stations, but in this case, we’ll be sending a tone down this channel to turn the transistor, and thus or PTT, on.

The switch feeding this arrangement we’ll call the “PTT” switch.

By using this arrangement, we can use either audio channel for modulation or PTT control, or we can use one channel for both.  1200-baud AFSK, FreeDV, etc, should work fine with both on the one channel.

If we just want to pass through analogue audio, then we probably want modulation separate, so we can hold the PTT open during speech breaks without having an annoying tone superimposed on our signal.

It may be prudent to feed a second resistor into the base of that NPN, running off to the RTS pin on an RS-232 interface.  This will let us use software that relies on RS-232 PTT control, which can be added by way of a USB-RS232 dongle.

The cheap Prolific PL-2303 ones sold by a few places (including Jaycar) will work for this.  (If your software expects a 16550 UART interface on port 0x3f8 or similar, consider running it in a virtual machine.)

Ideally though, this should not be needed, and if added, can be left disconnected without harm.

## Software

There are a few “off-the-shelf” packages that should work fine with this arrangement.

### AX.25 software

AGWPE on Windows provides a software TNC.  On Linux, there’s soundmodem (which I have used, and presently mirror) and Direwolf.

Shouldn’t need a separate PTT channel, it should be sufficient to make the pre-amble long enough to engage PTT and rely on the envelope detector recognising the packet.

### Digital Voice

FreeDV provides an open-source digital voice platform system for Windows, Linux and MacOS X.

This tool also lets us send analogue voice.  Digital voice should be fine, the first frame might get lost but as a frame is 40ms, we just wait before we start talking, like we would for regular analogue radio.

For the analogue side of things, we would want tone-driven PTT.  Not sure if that’s supported, but hey, we’ve got the source code, and yours truly has worked with it, it shouldn’t be hard to add.

### Slow-scan television

The two to watch here would be QSSTV (Linux) and EasyPal (Windows).  QSSTV is open-source, so if we need to make modifications, we can.

Not sure who maintains EasyPal these days, not Eric VK4AES as he’s no longer with us (RIP and thank-you).  Here, we might need an RS-232 PTT interface, which as discussed, is not a hard modification.

Most is covered by FLDigi.  Modes with a fairly consistent duty cycle will work fine with the VOX PTT, and once again, we have the source, we can make others work.

## Custom software ideas

So we can use a few off-the-shelf packages to do basic comms.

• We need auditability of our messaging system.  Analogue FM, we can just use a VOX-like function on the computer to record individual received messages, and to record outgoing traffic.  Text messages and files can be logged.
• Ideally, we should have some digital signing of logs to make them tamper-resistant.  Then we can mathematically prove what was sent.
• In a true  emergency, it may be necessary to encrypt what we transmit.  This is fine, we’re allowed to do this in such cases, and we can always turn over our audited logs for authorities anyway.
• Files will be sent as blocks which are forward-error corrected (or forward-erasure coded).  We can use a block cipher such as AES-256 to encrypt these blocks before FEC.  OpenPGP would work well here rather doing it from scratch; just send the OpenPGP output using FEC blocks.  It should be possible to pick out symmetric key used at the receiving end for auditing, this would be done if asked for by Government.  DIY not necessary, the building blocks are there.
• Digital voice is a stream, we can use block ciphers but this introduces latency and there’s always the issue of bit errors.  Stream ciphers on the other hand, work by generating a key stream, then XOR-ing that with the data.  So long as we can keep sync in the face of bit errors, use of a stream cipher should not impair noise immunity.
• Signal fade is a worse problem, I suggest a cleartext (3-bit, 4-bit?) gray-code sync field for synchronisation.  Receiver can time the length of a fade, estimate the number of lost frames, then use the field to re-sync.
• There’s more than a dozen stream ciphers to choose from.  Some promising ones are ACHTERBAHN-128, Grain 128a, HC-256, Phelix, Py, the Salsa20 family, SNOW 2/3G, SOBER-128, Scream, Turing, MUGI, Panama, ISAAC and Pike.
• Most (all?) stream ciphers are symmetric.  We would have to negotiate/distribute a key somehow, either use Diffie-Hellman or send a generated key as an encrypted file transfer (see above).  The key and both encrypted + decrypted streams could be made available to Government if needed.
• The software should be capable of:
• Real-time digital voice (encrypted and clear; the latter being compatible with FreeDV)
• File transfer (again, clear and encrypted using OpenPGP, and using good FEC, files will be cryptographically signed by sender)
• Voice mail and SSTV, implemented using file transfer.
• Radioteletype modes (perhaps PSK31, Olivia, etc), with logs made.
• Analogue voice pass-through, with recordings made.
• All messages logged and time-stamped, received messages/files hashed, hashes cryptographically signed (OpenPGP signature)
• Operation over packet networks (AX.25, TCP/IP)
• Standard message forms with some basic input validation.
• Ad-hoc routing between interfaces (e.g. SSB to AX.25, AX.25 to TCP/IP, etc) should be possible.
• The above stack should ideally work on low-cost single-board computers that are readily available and are low-power.  Linux support will be highest priority, Windows/MacOS X/BSD is a nice-to-have.
• GNU Radio has building blocks that should let us do most of the above.

So, there’s been a bit of discussion lately about our communications infrastructure. I’ve been doing quite a bit of thinking about the topic.

# The situation today

Here in Australia, a lot of people are being moved over to the National Broadband Network… with the analogue fixed line phone (if it hasn’t disappeared already) being replaced with a digital service.

For many, their cellular “mobile” phone is their only means of contact. More than the over-glorified two-way radios that was pre-cellular car phones used by the social elites in the early 70s, or the slightly more sophisticated and tennis-elbow inducing AMPS hand-held mobile phones that we saw in the 80s, mobile phones today are truly versatile and powerful hand-held computers.

In fact, they are more powerful than the teen-aged computer I am typing this on. (And yes, I have upgraded it; 1GB RAM, 250GB mSATA SSD, Linux kernel 4.0… this 2GHz P4 still runs, and yes I’ll update that kernel in a moment. Now, how’s that iPhone 3G going, still running well?)

All of these devices are able to provide data communications throughput in the order of millions of bits per second, and outside of emergencies, are generally, very reliable.

It is easy to forget just how much needs to work properly in order for you to receive that funny cat picture.

## Mobile networks

One thing that is not clear about the NBN, is what happens when the power is lost. The electricity grid is not infallible, and requires regular maintenance, so while reliability is good, it is not guaranteed.

For FTTP users, battery backup is an optional extra. If you haven’t opted in, then your “land line” goes down when the power goes out.

This is not a fact that people think about. Most will say, “that’s fine, I’ve got my mobile” … but do you? The typical mobile phone cell tower has several hours battery back-up, and can be overwhelmed by traffic even in non-emergencies.They are fundamentally engineered to a cost, thus compromises are made on how long they can run without back-up power, and how much call capacity they carry.

In the 2008 storms that hit The Gap, I had no mobile telephone coverage for 2 days. My Nokia 3310 would occasionally pick up a signal from a tower in a neighbouring suburb such as Keperra, Red Hill or Bardon, and would thus occasionally receive the odd text message… but rarely could muster the effective radiated power to be able to reply back or make calls. (Yes, and Nokia did tell me that internal antennas surpassed the need for external ones. A 850MHz yagi might’ve worked!)

## Emergency Services

Now, you tell yourself, “Well, the emergency services have their own radios…”, and this is correct. They do have their own radio networks. They too are generally quite reliable. They have their problems. The Emergency Alerting System employed in Victoria was having capacity problems as far back as 2006 (emphasis mine):

A high-priority project under the Statewide Integrated Public Safety Communications Strategy was establishing a reliable statewide paging system; the emergency alerting system. The EAS became operational in 2006 at a cost of \$212 million. It provides coverage to about 96 per cent of Victoria through more than 220 remote transmitter sites. The system is managed by the Emergency Services Telecommunications Agency on behalf of the State and is used by the CFA, VICSES and Ambulance Victoria (rural) to alert approximately 37,400 personnel, mostly volunteers, to an incident. It has recently been extended to a small number of DSE and MFB staff.

Under the EAS there are three levels of message priority: emergency, non-emergency, and administrative. Within each category the system sends messages on a first-in, first-out basis. This means queued emergency messages are sent before any other message type and non-emergency messages have priority over administrative messages.

A problem with the transmission speed and coverage of messages was identified in 2006. The CFA expressed concern that areas already experiencing marginal coverage would suffer additional message loss when the system reached its limits during peak events.

To ensure statewide coverage for all pagers, in November 2006 EAS users decided to restrict transmission speed and respond to the capacity problems by upgrading the system. An additional problem with the EAS was caused by linking. The EAS can be configured to link messages by automatically sending a copy of a message to another pager address. If multiple copies of a message are sent the overall load on the system increases.

By February 2008 linking had increased by 25 per cent.

During the 2008 windstorm in Victoria the EAS was significantly short of delivery targets for non-emergency and administrative messages. The Emergency Services Telecommunications Agency subsequently reviewed how different agencies were using the system, including their message type selection and message linking. It recommended that the agencies establish business rules about the use of linking and processes for authorising and monitoring de-linking.

The planned upgrade was designed to ensure the EAS could cope better with more messages without the use of linking.

The upgrade was delayed several times and rescheduled for February 2009; it had not been rolled out by the time of Black Saturday. Unfortunately this affected the system on that day, after which the upgrade was postponed indefinitely.

I can find mention of this upgrade taking place around 2013. From what I gather, it did eventually happen, but it took a roasting from mother nature to make it happen. The lesson here is that even purpose built networks can fall over, and thus particularly in major incidents, it is prudent to have a back-up plan.

# Alternatives

For the lay person, CB radio can be a useful tool for short-range (longer-than-yelling-range) voice communications. UHF CB will cover a few kilometres in urban environments and can achieve quite long distances if good line-of-sight is maintained. They require no apparatus license, and are relatively inexpensive.

It is worth having a couple of cheap ones, a small torch and a packet of AAA batteries (stored separately) in the car or in a bag you take with you. You can’t use them if they’re in a cupboard at home and you’re not there.

The downside with the hand-helds, particularly the low end ones, is effective radiated power. They will have small “rubber ducky” antennas, optimised for size, and will typically have limited transmit power, some can do the 5W limit, but most will be 1W or less.

If you need a bit more grunt, a mobile UHF CB set and magnetic mount antenna could be assembled and fitted to most cars, and will provide 5W transmit power, capable of about 5-10km in good conditions.

HF (27MHz) CB can go further, and with 12W peak envelope power, it is possible to get across town with one, even interstate or overseas when conditions permit. These too, are worth looking at, and many can be had cheaply second-hand. They require a larger antenna however to be effective, and are less common today.

Beware of fakes though… A CB radio must meet “type approval”, just being technically able to transmit in that band doesn’t automatically make it a CB, it must meet all aspects of the Citizens Band Radio Service Class License to be classified a CB.

If it does more than 5W on UHF, it is not a UHF CB. If it mentions a transmit range outside of 476-478MHz, it is not a UHF CB.  Programming it to do UHF channels doesn’t change this.

Similarly, if your HF CB radio can do 26MHz (NZ CB, not Australia), uses FM instead of SSB/AM (UK CB, again not Australia), does more than 12W, or can do 28-30MHz (10m amateur), it doesn’t qualify as being a CB under the class license.

## Amateur radio licensing

If you’ve got a good understanding of high-school mathematics and physics, then a Foundation amateur radio license is well within reach.  In fact, I’d strongly recommend it for anyone doing first year Electrical Engineering … as it will give you a good practical grounding in electrical theories.

Doing so, you get to use up to 10W of power (double what UHF CB gives you; 6dB can matter!) and access to four HF, one VHF and one UHF band using analogue voice or hand-keyed Morse code.

You can then use those “CB radios” that sell on eBay/DealExtreme/BangGood/AliExpress…etc, without issue, as being un-modified “commercial off-the-shelf”, they are acceptable for use under the Foundation license.

# Beyond Voice: amateur radio digital modes

Now, all good and well being able to get voice traffic across a couple of suburban blocks. In a large-scale disaster, it is often necessary to co-ordinate recovery efforts, which often means listings of inventory and requirements, welfare information, etc, needs to be broadcast.

You can broadcast this by voice over radio… very slowly!

You can put a spreadsheet on a USB stick and drive it there. You can deliver photos that way too. During an emergency event, roads may be in-passable, or they may be congested. If the regular communications channels are down, how does one get such files across town quickly?

Amateur radio requires operators who have undergone training and hold current apparatus licenses, but this service does permit the transmission of digital data (for standard and advanced licensees), with encryption if needed (“intercommunications when participating in emergency services operations or related training exercises”).

Amateur radio is by its nature, experimental. Lots of different mechanisms have been developed through experiment for intercommunication over amateur radio bands using digital techniques.

## Morse code

The oldest by far is commonly known as “Morse code”, and while it is slower than voice, it requires simpler transmitting and receiving equipment, and concentrates the transmitted power over a very narrow bandwidth, meaning it can be heard reliably at times when more sophisticated modes cannot. However, not everybody can send or receive it (yours truly included).

I won’t dwell on it here, as there are more practical mechanisms for transmitting lots of data, but have included it here for completeness. I will point out though, due to its simplicity, it has practically no latency, thus it can be faster than SMS.

Okay, there are actually quite a few modes that can be described in this manner, and I’ll use this term to refer to the family of modes. Basically, you can think of it as two dumb terminals linked via a radio channel. When you type text into one, that text appears on the other in near real-time. The latency is thus very low, on par with Morse code.

The earliest of these is the RTTY mode, but more modern incarnations of the same idea include PSK31.

These are normally used as-is. With some manual copying and pasting pieces of text at each end, it is possible to encode other forms of data as short runs of text and send files in short hand-crafted “packets”, which are then hand-deconstructed and decoded at the far end.

This can be automated to remove the human error component.

The method is slow, but these radioteletype modes are known for being able to “punch through” poor signal conditions.

When I was studying web design back in 2001, we were instructed to keep all photos below 30kB in size. At the time, dial-up Internet was common, and loading times were a prime consideration.

Thus instead of posting photos like this, we had to shrink them down, like this. Yes, some detail is lost, but it is good enough to get an “idea” of the situation.

The former photo is 2.8MB, the latter is 28kB. Via the above contrived transmission system, it would take about 20 minutes to transmit.

The method would work well for anything that is text, particularly simple spread sheets, which could be converted to Comma Separated Values to strip all but the most essential information, bringing file sizes down into realms that would allow transmission times in the order of 5 minutes. Text also compresses well, thus in some cases, transmission time can be reduced.

To put this into perspective, a drive from The Gap where that photo was taken, into the Brisbane CBD, takes about 20 minutes in non-peak-hour normal circumstances. It can take an hour at peak times. In cases of natural disaster, the roads available to you may be more congested than usual, thus you can expect peak-hour-like trip times.

## Radio Faximile and Slow Scan Television

This covers a wide variety of modes, ranging from the ancient like Hellschreiber which has its origins in the German Military back in World War II, various analogue slow-scan television modes through to the modern digital slow-scan television.

This allows the transmission of photos and visual information over radio. Some systems like EasyPAL and its elk (based on HamDRM, a variant of Digital Radio Mondiale) are in fact, general purpose modems for transmitting files, and thus can transmit non-graphical data too.

Transmit times can vary, but the analogue modes take between 30 seconds and two minutes depending on quality. For the HamDRM-based systems, transmit speeds vary between 86Bps up to 795kBps depending on the settings used.

Packet radio is the concept of implementing packet-switched networks over radio links. There are various forms of this, the most common in amateur radio being PACTOR, WINMOR, the 1200-baud AFSK and 9600-baud FSK and 300-baud AFSK packet modes.

300-baud AFSK is normally used on HF links, and hails from experiments using surplus Bell 103 modems modified to work with radio. Similarly, on VHF and UHF FM radio, experiments were done with surplus Bell 202 modems, giving rise to the 1200-baud AFSK mode.

The 9600-baud FSK mode was the invention of James Miller G3RUH, and was one of the first packet radio modes actually developed by radio amateur operators for use on radio.

These are all general-purpose data modems, and while they can be used for radioteletype applications, they are designed with computer networking in mind.

The feature facilities like automatic repeating of lost messages, and in some cases support forward error correction. PACTOR/WINMOR is actually used with the Winlink radio network which provides email services.

The 300-baud, 1200-baud and 9600-baud versions generally use a networking protocol called AX.25, and by configuring stations with multiple such “terminal node controllers” (modems) connected and appropriate software, a station can operate as a router, relaying traffic received via one radio channel to a station that’s connected via another, or to non-AX.25 stations on Winlink or the Internet.

It is well suited to automatic stations, operating without human intervention.

AX.25 packet and PACTOR I are open standards, the later PACTOR modems are proprietary devices produced by SCS in Germany.

AX.25 packet is capable of transmit speeds between 15Bps (300 baud) and 1kBps (9600 baud). PACTOR varies between 5Bps and 650Bps.

In theory, it is possible to develop new modems for transmitting AX.25, the HamDRM modem used for slow-scan television and the FDMDV modem used in FreeDV being good starting points as both are proven modems with good performance.

These simply require an analogue interface between the computer sound card and radio, and appropriate software.  Such an interface made to link a 1200-baud TNC to a radio could be converted to link to a low-cost USB audio dongle for connection to a computer.

If someone is set up for 1200-baud packet, setting up for these other modes is not difficult.

## High speed data

Going beyond standard radios, amateur radio also has some very high-speed data links available. D-Star Digital Data operates on the 23cm microwave band and can potentially transmit files at up to 16KBps, which approaches ADSL-lite speeds. Transceivers such as the Icom ID-1 provide this via an Ethernet interface for direct connection to a computer.

General Electric have a similar offering for industrial applications that operates on various commercial bands, some of which can reach amateur frequencies, thus would be usable on amateur bands. These devices offer transmit speeds up to 8KBps.

A recent experiment by amateurs using off-the-shelf 50mW 433MHz FSK modules and Realtek-based digital TV tuner receivers produced a high-speed speed data link capable of delivering data at up to 14KBps using a wideband (~230kHz) radio channel on the 70cm band.  They used it to send high definition photos from a high-altitude balloon.

## The point?

We’ve got a lot of tools at our disposal for getting a message through, and collectively, 140 years of experience at our disposal. In an emergency situation, that means we have a lot of different options, if one doesn’t work, we can try another.

No, a 1200-baud VHF packet link won’t stream 4k HD video, but it has minimal latency and will take less than 20 minutes to transmit a 100kB file over distances of 10km or more.

A 1kB email will be at the other end before you can reach for your car keys.  Further experimentation and development means we can only improve.  Amateur radio is far from obsolete.

Another approach to the selecting a source is to avoid microcontrollers altogether and just rely on non-programmable logic. This is inspired a bit by the Saturday Clock, or rather, my thinking of how it could be done without an MCU.

In selecting a source, we really only care about one thing: is the battery voltage high enough? If no, we need to hunt for one that is.

This question can be answered by a simple analogue comparator such as the LM311, a shift register, and a few other logic gates.

Here, we have such a circuit. Up the top, is our shift register, set up as a ring counter. The buffers there are stand-ins for diodes, if any of them is a 1, the output is a 1 and the NOT gate on the input outputs a 0.

The outputs of the shift register are used to select a battery, which has its own comparator and select logic. The comparator is represented here by the D flip-flop at the extreme left: in essence I’m using this as a switch, Logisim doesn’t provide one, only a momentary button. We need a signal that is high when the battery is above acceptable voltage. We also need its inverse.

The select line from the shift register controls the gate on two tri-state buffers, allowing us to inhibit the comparator’s output. The buffered “good” signal is used to SET the “enable” D-flip-flop that drives the switch turning the battery on. This same (buffered “good”) signal also passes thorough a diode-OR arrangement that indicates whether a source is “available”.

To emulate make-before-break, inverted “select” signal and the “source available” signal pass through an AND gate and into the RESET of the “enable” flip flop, so it gets turned off when another source is turned on.

Finally, the buffered “bad” signal from all modules is fed back on one shared line, inhibiting the clock until a battery drops below the minimum level.

A glitch here is if multiple batteries are initially turned on with none above the minimum voltage, this will cause multiple sources to be selected. This is not too hard to manage in software, and the solution might in fact be to implement this on an ATTiny24A as mentioned in the previous post; this logic circuit can be implemented quite easily in C, with comparators in hardware or using the ADC as a software comparator as I’m doing in the charge controller.

So, further progress on the charge controller.

Thinking about the problem … I realised that I really do not want to be testing for VBNVH when entering the CHARGE_CHECK state, as it’ll prematurely terminate the charge when the battery is being bulk-charged.

Better to wait until the charger decides to stop … which we’ll see due to the battery voltage ceasing to increase. We do want to check we’re not critically high however, so we can swap out VH for VCH.

Next, when we find that VBNVBL, meaning the battery is not charging, there we can check for VBNVH and stop charging at that point.

That change, worked pretty well, but it was still flapping between sources. A little state management helped this. If we declare another state variable, charger_warning, we can flip this to 1 upon first detecting that VBNVBL, then wait a little longer. If after a time-out this is still the case, then we can take action. Thus we define a new timer tCWARN, which delays acting on the not-charging case.

A bit of threshold tweaking, and things are behaving themselves. I’m using an el-cheapo 3-way camping fridge as the stand-in for the cluster. This is an Aldi special bought some years ago that draws about 5-6A… and when running on 12V power, features no thermostat.

We’re finding that its cooling capacity is no match for Brisbane’s early autumn weather anyway, so it’s pretty much would be a constant load even if the thermostat worked on 12V.

The battery we’re using is an old 105Ah AGM battery… which is one of two batteries from the caravan we have here. They were the original batteries, and this battery’s mate had failed when both were replaced. We’ve noted this battery getting warm whilst charging, so we think it might now be on the way out too.

What to replace it with? LiFePO₄ is AU\$1000 for 100Ah, so too expensive. AGM is still the better bet. I can get 300Ah AGM batteries, but they weigh nearly as much as I do. I can just manage the 105Ah, so we’ll stick with those. I will need more than one long-term.

That brings the thorny issue of connecting them. I am not keen to hook batteries in parallel for various reasons. At least not permanently.

Now, the charger I’m using for mains is a 3-channel charger. I can make additional charge controllers (with the caveat that I need heatsinks for the MOSFETs…sigh!) and I can look for 3-channel solar chargers, or just get multiple chargers for the extra batteries. This can be done.

The load is the elephant in the room. I’d ideally like to manage it as a single load, although conceivably, I could put the switch and one storage node on one battery, a second storage node and a compute node on a second, and the final storage and compute nodes on the third. If the switch goes however, my cluster is toast.

I can put batteries in parallel, but this really does need to be done with care, using carefully matched batteries. So the better solution is to have a controller that chooses the battery with the highest voltage.

There might be an analogue means of implementing this, but a microcontroller is a single-chip solution. The ATTiny24As have up to 8 ADC/GPIO pins and three non-ADC GPIO pins. It’s what I’m already using for the charge controller, so is an easy choice.

The cluster will not tolerate a break-before-make switch-over. I thought about using a capacitor bank to keep the cluster alive during a brief (~1sec max) switch-over. Back-of-the-envelope calculations suggested I would need a 10F capacitor bank. I can get a 16V 470mF capacitor for AU\$70 each… and would need 20 of these. Ouch!

A small battery is another option, maybe a 7Ah, but that has its own maintenance issues, and represents a single point of failure.

I can get Schottky diodes capable of 40A, they still present a 0.6V voltage drop. At 30A, that represents 16W! For comparison, a relay with a 225ohm coil resistance will draw ~60mA when the battery is at the maximum of 15V, representing a load of 1W.

Or I can use more MOSFETs like the ones I’m already using, which draw even less power; poor man’s solid-state relays. Latching relays also exist, but they can be rather expensive, more so than a solid-state relay.

I can probably get away with temporary parallel connections, so a make-before-break would let me switch sources. Or, I could place my switch across a Schottky, meaning I put up with that 16W load for a brief moment while I switch sources.

So more to think about, but we are getting close. I can defer this decision until I get a second battery, but I am getting close to the point where the cluster will be running full-time.

So, as promised, the re-design of the charge controller. … now under the the influence of a few glasses of wine, so this should be interesting…

As I mentioned in my last post, it was clear that the old logic just wasn’t playing nice with this controller, and that using this controller to maintain the voltage to the nodes below 13.6V was unrealistic.

The absolute limits I have to work with are 16V and 11.8V.

The 11.8V comes from the combination of regulator voltage drop and ATX PSU power range limits: they don’t operate below about 10.8V, if you add 700mV, you get 11.5V … you want to allow yourself some head room. Plus, cycling the battery that deep does it no good.

As for the 16V limit… this is again a limitation of the LDOs, they don’t operate above 16V. In any case, at 16V, the poor LDOs are dropping over 3V, and if the node is running flat chat, that equates to 15W of power dissipation in the LDO. Again, we want some headroom here.

The Xantrex charger likes pumping ~15.4V in at flat chat, so let’s go 15.7V as our peak.

Those are our “extreme” ranges.

At the lower end, we can’t disconnect the nodes, but something should be visible from the system firmware on the cluster nodes themselves, and we can thus do some proactive load shedding, hibernating virtual instances and preparing nodes for a blackout.

Maybe I can add a small 10Mbps Ethernet module to an AVR that can wake the nodes using WOL packets or IPMI requests. Perhaps we shut down two nodes, since the Ceph cluster will need 2/3 up, and we need at least one compute node.

At the high end, the controller has the ability to disconnect the charger.

So that’s worked out. Now, we really don’t want the battery getting that critically low. Thus the time to bring the charger in will be some voltage above the 11.8V minimum. Maybe about 12V… perhaps a little higher.

We want it at a point that when there’s a high load, there’s time to react before we hit the critical limit.

The charger needs to choose a charging source, switch that on, then wait … after a period check the voltage and see if the situation has improved. If there’s no improvement, then we switch sources and wait a bit longer. Wash, rinse, repeat. When the battery ceases to increase in voltage, we need to see if it’s still in need of a charge, or whether we just call it a day and run off the battery for a bit.

If the battery is around 14.5~15.5V, then that’s probably good enough and we should stop. The charger might decide this for us, and so we should just watch for that: if the battery stops charging, and it is at this higher level, just switch to discharge mode and watch for the battery hitting the low threshold.

Thus we can define four thresholds, subject to experimental adjustment:

 Symbol Description Threshold $V_{CH}$ Critical high voltage 15.7V $V_H$ High voltage 15.5V $V_L$ Low voltage 12.0V $V_{CL}$ Critical low voltage 11.8V

Now, our next problem is the waiting… how long do we wait for the battery to change state? If things are in the critical bands, then we probably want to monitor things very closely, outside of this, we can be more relaxed.

For now, I’ll define two time-out settings… which we’ll use depending on circumstances:

 Symbol Description Period $t_{LF}$ Low-frequency polling period 15 sec $t_{HF}$ High-frequency polling period 5 sec

In order to track the state, I need to define some variables… we shall describe the charger’s state in terms of the following variables:

 Symbol Description Initial value $V_{BL}$ Last-known battery voltage, set at particular points. 0V $V_{BN}$ The current battery voltage, as read by the ADC using an interrupt service routine. 0V $t_d$ Timer delay… a timer used to count down until the next event. $t_{HF}$ $S$ Charging source, an enumeration: 0: No source selected 1: Main charging source (e.g. solar) 2: Back-up charging source (e.g. mains power) 0

The variable names in the actual code will be a little more verbose and I’ll probably use #defines for the enumeration.

Below is the part-state-machine part-flow-chart diagram that I came up with. It took a few iterations to try and describe this accurately, I was going to use a state machine syntax similar to what my workplace uses, but in the end, found the ye olde flow chart shows it best.

In this diagram, a filled in dot represents the entry point, a dot with an X represents an exit point, and a dot in a circle represents a point where the state machine re-enters the state and waits for the main loop to iterate once more.

You’ll note that for this controller, we only care about one voltage, the battery voltage. That said, the controller will still have temperature monitoring duties, so we still need some logic to switch the ADC channel, throw away dummy samples (as per the datasheet) and manage sample storage. The hardware design does not need to change.

We can use quiescent voltages to detect the presence of a charging source, but we do not need to, as we can just watch the battery voltage rise, or not, to decide whether we need to take further action.

So, having knocked the regulation on the LDOs down a few pegs… I am closer to the point where I can leave the whole rig running unattended.

One thing I observed prior to the adjustment of the LDOs was that the controller would switch in the mains charger, see the voltage shoot up quickly to about 15.5V, before going HOLYCRAP and turning the charger off.

I had set a set point at about 13.6V based on two facts:

• The IPMI BMCs complained when the voltage raised above this point
• The battery is nominally 13.8V

As mentioned, I’m used to my very simple, slow charger, that trickle charges at constant voltage with maximum current output of 3A. The Xantrex charger I’m using is quite a bit more grunty than that. So re-visiting the LDOs was necessary, and there, I have some good results, albeit with a trade-off in efficiency.

Ahh well, can’t have it all.

I can run without the little controller, as right now, I have no panels. Well, I’ve got one, a 40W one, which puts out 3A on a good day. A good match for my homebrew charger to charge batteries in the field, but not a good match for a cluster that idles at 5A. I could just plug the charger directly into the battery and be done with it for now, defer this until I get the panels.

But I won’t.

I’ve been doing some thought about two things, the controller and the rack. On the rack, I found I can get a cheap one for \$200. That is cheap enough to be considered disposable, and while sure it’s meant for DJ equipment, two thoughts come to mind:

• AV equipment with all its audio transformers and linear power supplies, is generally not light
• It’s on wheels, meant to be moved around… think roadies and such… not a use case that is gentle on equipment

Thus I figure it’ll be rugged enough to handle what I want, and is open enough to allow good airflow. I should be able to put up to 3 AGM batteries in the bottom, the 3-channel charger bolted to the side, with one charge controller per battery. There are some cheap 30A schottky diodes, which would let me parallel the batteries together to form one redundant power supply.

Downside being that would drop about 20-25W through the diode. Alternatively, I make another controller that just chooses the highest voltage source, with a beefy capacitor bank to handle the switch-over. Or I parallel the batteries together, something I am not keen to do.

I spent some time going back to the drawing board on the controller. The good news, the existing hardware will do the job, so no new board needed. I might even be able to simplify logic, since it’s the battery voltage that matters, not the source voltages. But, right now I need to run. So perhaps tomorrow I’ll go through the changes. 😉