A home built upgrade as a mixer equivalent.
Illustrates what can be done with basic parts.
No commercial part will do this job.
The build allows proportional motor control to assist the rudders on the tug, which is otherwise impossible to position in high winds or in reverse. No extra controller is needed, neither is the existing PCB modified in any way.
The pictures show the scratchbuilt board under a final test prior to attempting to fit the board into the Southampton tug.
The waveform is one of many. The yellow would be a half speed motor, but the alteration is the rudder setting modifying the motor speed. The red is the timing pulse for the rudder circuitry.
The mod has no effect until the rudder is moved to at least40% deflection in reverse and 60% in forward. In forward, the faster you go, the less effect the mod has, a bit like power steering in a car.
When stopped, with no throttle, just adjusting the rudder drives both motors gently, up to 40% speed, in opposite sense to turn on the spot.
All figures are adjustable, set by component values.
That's the theory anyhow!
I'm now going to try to remove the deck...

Question re a build by Alan Blond. Commercially available?.
This was shown as a veroboard build. I don't known if/ how it was marketed. The info clearly says it picks up the receiver signal, which is probably the 1-2 msec pulses for servos. That's a fairly standard Arduino function and no mystery to those skilled in that art.
I said in an earlier post that a standard mixer that uses the receiver 1-2 msec signal cannot be used in the tug because it uses a proprietary PCB. As far as I can see from the official circuit that signal does not exist on any line that can be picked off. There is a PWM that goes to the smoker that is a function of throttle but could be anything and no direction info. You do get the rudder servo signal. I'll be explaining the logic of mixing these. The guys who make the commercial items delight in telling of million options.
If you can buy a mixer that picks up the forward and back H bridge signals currently on the motor, as I have to do for this device, then yes I'm reinventing the wheel.
Save me!

I don't want to get into rivalry with the Arduino project. This blog runs in parallel, for those who feel more comfortable with a few logic chips. For any system, once you have the input signals digitised, software is a good choice for processing. But it seems to me that the processors do not escape needing a lot of hardware for model boats, albeit that some of it comes on a board that you buy.
You have to make serial pins into parallel outputs and at the end point, every device needs a pin, and wire, if not a buffer as well, even sometimes for a LED. The end difference in size can be very small.
My personal problem is thinking through and debugging the original logic.
Both methods have the same issue. RIRO.
Rubbish In Rubbish Out!
The build I have just done uses 5 simple logic chips, but the thinking behind this was difficult. However, I could see what was going on in real time, glitches and all, so fault finding wasn't too bad. If your programming logic is wrong, then the builder faces the problem of deciding between a circuit malfunction and his her own error. That's partly why programmers beg borrow and use previous work.
My experience at work was that testers with probes caused more problems than they fixed.
But testing can be essential.
A 7 day unsolved software problem was once fixed by me in 5 minutes by prodding with a scope when the 32 channel(!) logic analyser maintained that all was well. And it was a hardware fault.
There's space for both methods to work together, choice often depends on your comfort zone.
Getting back to this ESC build, what you see, barring connectors, is the complete device with 2 H bridges capable of 5A included. The ICs are on top and the surface mount parts and wiring under. It took 2 days to build and another day to fix after I got too ambitious and took a jump too far, causing damage due to a build error. I will be adding motor suppression on the motors themselves.
This build isnot an economy in time, but a necessity to make the boat work as I feel it should. And for me it's fun😆

Just to be clear.... this blog carries on from the Electronize thread and also from Southampton Tug/ smoke generators. A few folk would like an introduction to Electronics and this ESC project will hopefully serve as a foundation for that.
Please go through the last few weeks of those if you want to know where I started out from.
As of now, I have the deck off the boat. The tips given to me elsewhere on this were useful. You remove all davits to get at some buried screws; two smaller davits at the back seem to be glued in and needed a gentle levering.
Looking at the motors gave a scare as it looked as if they were in series!
My mistake though. Phew!
I can see no reason for this boat to lean to the right in the water, but balancing it on the keel it clearly tips quite heavily. The rudder servo must weigh a ton, it's on the right. Sorry- starboard.
That leads to a decision, to move the battery to the back under the rear hatch, and opposite the servo so as to help the balance. It also means that the battery and a new ON switch can be accessed without removing the rather fragile deckhouse under which the wiring was quite tight. Several wires pass through this , sealed with a glue gun rather haphazardly.
The wires are of course too short so this is the next thing to sort out

The plan (and nothing goes to plan) is to break up the circuit into 4 or 5 parts and explain the electronics, thinking, testing and pitfalls of each one. Building it is one thing, sorting any problems in a safe and systematic fashion is another! I learned long ago to design the test method before the circuit is built.
An objective is to show how a person with basic test gear - a cheap DMM costing 4 pounds - can check out functionality.
Bear in mind that this build is applicable to the Southampton Tug and similar boats using the same basic parts. I detailed the options previously. You can spend your way to the same solution but it will cost a fair bit. Other boats which have all servo signals available can get the same functionality commercially for about 50 pounds. You need a mixer, a Y servo connector and 2 ESCs. And you learn very little.
A scratch build with legged components is getting difficult. The reason is progress. I can buy the IC chips in surface mount types for around 20 p each, in 10 s. Double that for ones. But the legged versions on E bay will be 2 pounds each. Some are still sold but the range is decreasing. I can handle such SMD ICs but it needs a small soldering iron, a steady hand and a decent magnifying glass. SMD passive components cost a penny or two each but you have to buy in 10 s 50s and 100s. They easily fit on veroboard, the 0805 size is easy to handle and the smaller 0603 are OK with decent tweezers to hold them.
For that reason I did the design to minimise the different numbers of components. So for instance you buy 10K resistors. Two in series gives 20K and 2 in parallel gives 5K. As many values needed are not critical, a few values go a long way. Most SMD boards use this method because the machines only handle a limited number of values.
I used SMD components on the wiring side because a lot can be squeezed in. But legged components are perfectly feasible. A big help to this would be a PCB, which would reduce the time and risk considerably. For that we really need to order at least 10 to make costs sensible. More folk would need to join in.
I hope I answered the real question!

Nothing goes to plan. The photos will follow. I installed the board; it ran for a while, and when I applied the power for the third time, I got the dreaded grey smoke. That sort of thing is a nightmare because one does not know why the fault happened. It was working and I didn't touch anything sir.
You have to guess, unless you are certain of having made a mistake. I don't believe in failures due to bad luck. So I made some improvements which I will detail later, and also added a switch so as to turn things on with all inputs active and set. I learned a few things about low current H bridges.
Simple Sailor reminded me earlier about logic level MOSFETs. I was tempted and I fell. But it turns out that all the usual precautions on H bridges are required. You don't really save anything, because logic chips do not have enough Oomph to drive the FETs at a reasonable speed. You really have to add a buffer amplifier. And most switching circuits have a tendency to self destruction which must be resisted. I discerned a design rule to share.
If designing an H bridge with only one drive point on each side, the sum of the Vgs of the top and bottom FET when conducting the nominal load current must be less than the rail voltage. That is generally true for logic level FETs with single battery cell devices, but for 2 cells, or anything above 5V or so, a single point drive of an H bridge with logic level devices is liable to have problems, grey smoke being one of them. You should really use the old, bigger higher Vgs rated devices. Especially when you're a modeller rather than a speed merchant.
You have to drive the top and bottom transistors of one 'leg" of an H bridge so that there is a "dead band" when both are nominally off, or at least nearly off. This goes against nature and what FETs want to do, which is to both stay on together in an embrace to destruction. Normally, that means a driver chip is required but I am far too mean for those and found a workaround. I think a 3 cell 12V device would certainly need a driver chip, which all add their own problems, especially as many do not handle flat out full speed. At a future point I will cover H bridges and their foibles and differences. At no time may any transistor gate be uncontrolled, especially at turn on. The normal ESC is a bargain....
Surprisingly, I rethought my ideas on old fashioned Bipolar devices, and provided the current is no more than 5A they probably work out cheaper and simpler than MOSFETs especially at higher voltages. And low voltages below 2V is possible, where a MOSFET will start to struggle. They can be made more tolerant of abuse and the driving of them is actually quite easy.
Anyway, after a few hours soldering and retest....
Success! At least until it goes in the water. I added some LEDs which tell me if a prop is going forward or back, and the brightness indicates the speed. The circuit seems to be behaving properly; all functions work. Whether its all in the right direction only a trip to some water will tell. And to some extent, its a matter of preference or even practicalities.
I examined the Tug electronics and it's a little more non standard than I expected.
The rudder servo pulse is 1to 2 msec, but occurs at a rate of about 8.6 msec, over twice as fast as usual. Servos don't care about this but I do; cue some value changes to readjust gain.
The motor throttle pulses also come at every 8.6 msec. Full speed is seen, but the minimum starts at 2 msec or just under 1/4 speed. I feel that is excessive for a Tug. In time I will add an adjustment for that. When throttle is stopped and I move the rudders, as designed the motors turn in opposite directions for a spot turn. I can make the motors turn very slowly indeed, albeit in air. So it looks like the 2 msec they have used corresponds to a maximum servo pulse.
Provided it works OK in the water I will start to document the circuit in sections and use each section to illustrate the use of common and inexpensive electronic parts.
This has ended at a slightly higher cost than originally envisaged, but still very light for the functionality. It's the bits at the edges that cost.... I have realised from other blogs that every expert among us accumulates numbers of small tools, methods and knick knacks that help us to practise our expertise. When you come to help someone else get started, the real cost suddenly stares you in the eye. What is cheap and easy for me turns into something expensive for another to duplicate because of such start up costs.

You must always have a fuse he said. But I wanted a switch. So I made a pair of wires with a connector that would normally be joined by a switch. But I fit a fuse into the connector and the fuse acts as the switch. Job done sir.

A couple of photos here showing the bits in the boat. The circuit board could be half the size or less on a proper PCB but a first hand built prototype needs spare space.

The next visit is to the bath or a pond. I think everything is going the right way but my mechanical and spacial ability is zilch.

Still considering how to present this build as an educational series and whether to throw away and replace the output stages as I am not 100% happy with the devices used. And I have used the spares....

As I said, I was not happy with the parts I bought for the output stage and sure enough for no apparent reason one device has developed a slight fault just as I got the deck back on. Unusually, not complete destruction. But typical of a small over voltage spike. At a 12V rating these devices were marginal on my normal design rules so I pay the price and scrap them.
So that output stage is coming out and getting replaced. I need a few days to buy in the proper bits but a temporary set of a silly big size will go in tomorrow. That's another thing down to experience....

Well I said in a recent post that in my experience more damage was caused by testing than the testers fixed in the first place. My "no apparent reason" for a fail was possibly the testing itself. After the damage was removed, on initial retest I put my high impedance scope probe onto a power supply pin and saw a flash. Impossible! But true. Now scope probes regularly go open circuit in their wire. But in over forty years I have never ever had one go short circuit.
There a first time for everything. So testing probably shorted a wire out and the rest becomes history. But probably a good thing as I was mistrustful of the circuit anyway. Fix underway.