Arco Zeus regulator - first impressions and a look inside

[cheesegas]

8 hours ago, dmr said:

I don't really see what the big issue is with float mode and tacho signal, maybe I have not thought hard enough 😀, surely it only needs a volt (or even much less) to deduce speed, so as long as a tiny bit of rotor excitation is applied (well below float voltage) there should be enough voltage. Its a zero crossing thing? Is the residual magnetism not enought?

The problem is just that. The Zeus can apply a minimum field current, no matter the RPM. If you set it to something like 5%, there isn't enough magnetism in the rotor to induce a large enough current in the stator for a meaningful signal to be present on the W terminal, so the tacho reads zero. In my setup, 24% was just enough at idle to get a tacho signal; 23% meant it jumped between zero and 700rpm. This is the threshold of the tiny bit of excitation you mentioned.

 

However, at higher RPMs when the rotor is moving faster past the stator, that same 24% field drive induces a greater current in the stator than at idle. It's not only enough to drive the tacho but it's also enough to push the output of the alternator above the target voltage when the load is low. 

 

A user configurable map of minimum field % against RPM would solve this, so it's dynamic and reduces the minimum field as RPMs increase.

[dmr]

30 minutes ago, cheesegas said:

The problem is just that. The Zeus can apply a minimum field current, no matter the RPM. If you set it to something like 5%, there isn't enough magnetism in the rotor to induce a large enough current in the stator for a meaningful signal to be present on the W terminal, so the tacho reads zero. In my setup, 24% was just enough at idle to get a tacho signal; 23% meant it jumped between zero and 700rpm. This is the threshold of the tiny bit of excitation you mentioned.

 

However, at higher RPMs when the rotor is moving faster past the stator, that same 24% field drive induces a greater current in the stator than at idle. It's not only enough to drive the tacho but it's also enough to push the output of the alternator above the target voltage when the load is low. 

 

A user configurable map of minimum field % against RPM would solve this, so it's dynamic and reduces the minimum field as RPMs increase.

 

ok, I have no first hand experience of this. I wonder how many volts Zeus needs to deduce speed?. A user configurable limit would be a bit of a pain, I would think that an automatic system could be deduced, something like Nick desribes for his alternator control chip.

 

The "tacho required" is a bit usefull in that it alows the start delay and soft ramp to work correctly even if the engine is a bit slow to start for some reason, but in almost all cases this should not really be required.

 

I have been using Zeus to control Alternator temperature. It looks like there is some learning going on and it has got very good at holding a steady temperature (within 1 degree) with only small modulations in duty cycle......BUT, I have to specify a target temperature and a maximum temperature (65 and 75 in my case) and Zeus aims for something mid way between these (69/70) while I think it should aim for the lower temperature. A  problem with this is it constantly gives a "temperature high" status message.

 

More seriously, despite setting a temperature limit and duty cycle limit, I believe I saw Zeus apply 100% field for maybe a minute or too causing a high alternature temperature. It could potentially have given a high voltage if it had been past the bulk phase. Sadly I did not manage to record this. I suspect this happened because I had gone back into the saloon with my phone which was right on the edge of the Bluetooth range and Zeus somehow got sidetracked into trying to reconnect, but this is only a guess.

[cheesegas]

21 minutes ago, dmr said:

ok, I have no first hand experience of this. I wonder how many volts Zeus needs to deduce speed?. A user configurable limit would be a bit of a pain, I would think that an automatic system could be deduced, something like Nick desribes for his alternator control chip.

It would only need a software update (I expect!) to be able to add the feature to create a map. You’d fully charge the batteries, allow the reg to drop into float mode, disconnect all other charge sources and then slowly increase RPMs whilst the regulator hunts for and then sets the minimum field for each RPM setpoint. Once the map is created everything would be automatic. 

I don’t remember the peak-peak voltages when I had the oscilloscope on it, and neither did I put a multimeter on it to get the RMS voltage…but I did have a spare tacho dial attached to the W terminal. I noticed that the dial would jump around at roughly the same point the Zeus begins to lose signal, which suggests it’s an alternator output issue. Further refining the input signal processing on the Zeus may help though.

 

[dmr]

A bit of lateral thinking......does this actually matter in most instalations?

On my engine the main tacho comes from the "starter" alternator. The Zeus controls the domestic alternator and if that goes into float the only consequence is that Zeus cannot display rpm, but as its not really doing anything that does not matter?

The issue is that 0% minimum duty should be the default, and should only be changed in special circumstances?

 

The entire concept of driving the tacho off the alternator is a bit historic, and on modern systems that run something like the Zeus the tacho should ideally come from a cam or crank sensor (though this is still rare).

 

To answer your post, a lot depends on how Zeus handles the tacho signal, it might be possible to do a firmware upgrade but if the tacho signal goes via a comparator or directly to a digital input then its a hardware change that would be required.

 

If you get the scope out again then report back on typical voltages at low field drives???

[cheesegas]

True, on a modern installation with dual alternators it doesn’t matter except for when you have ‘wait for tacho’ enabled. That requires a minimum field current. I think it should also have a warning in the app…

 

There should either be an enable/disable option for minimum field with a warning when you enable it, or it defaults to zero and then when you enter in a number it pops up with the warning. There’s a half page dedicated to this very pitfall in the Wakespeed manual for example, and you’re also less likely to use minimum field on the Wakespeed as it doesn’t have a wait for tacho option. 
 

My 35 year old Isuzu originally had a sender on the camshaft for the tacho, but newer engines don’t, pretty much all mechanically injected engines of the sort we see on inland craft have an alternator driven tacho. The Wakespeed can take engine RPM from CANBUS but I don’t know if the Zeus can. 
 

I was going to try connecting the Zeus’s tacho input to my Isuzu’s sender, but it doesn’t allow for a tacho ratio below 1 - it needs to be 0.5 as the camshaft is half engine speed. I could knock together something to output a second pulse, but at that point you may as well install a sender on the crank pulley or something!

[dmr]

9 minutes ago, cheesegas said:

True, on a modern installation with dual alternators it doesn’t matter except for when you have ‘wait for tacho’ enabled. That requires a minimum field current. I think it should also have a warning in the app…

 

There should either be an enable/disable option for minimum field with a warning when you enable it, or it defaults to zero and then when you enter in a number it pops up with the warning. There’s a half page dedicated to this very pitfall in the Wakespeed manual for example, and you’re also less likely to use minimum field on the Wakespeed as it doesn’t have a wait for tacho option. 
 

My 35 year old Isuzu originally had a sender on the camshaft for the tacho, but newer engines don’t, pretty much all mechanically injected engines of the sort we see on inland craft have an alternator driven tacho. The Wakespeed can take engine RPM from CANBUS but I don’t know if the Zeus can. 
 

I was going to try connecting the Zeus’s tacho input to my Isuzu’s sender, but it doesn’t allow for a tacho ratio below 1 - it needs to be 0.5 as the camshaft is half engine speed. I could knock together something to output a second pulse, but at that point you may as well install a sender on the crank pulley or something!

 

Unless I missed it there is almost zero documentation available for the Zeus, its operation and the the App, just the installation wiring diagrams. Maybe this will turn up soon???

Have just set my minimum duty to 0%.  I want to get 100% confident in the Zeus before I think about replacing the Trojans with Lithium.

Also just reduced my maximum duty to 65% and still got over 80 amps out of the alternator (100amp Iskra) when I first started the engine.  I would like to limit it to about 70 as its running off a single V belt, but if I do that I might not get enough current when the alternator is hot.  I will post some graphs of logged data in the next day or so.

[nicknorman]

11 hours ago, dmr said:

A bit of lateral thinking......does this actually matter in most instalations?

 

It matters if you want have an alternator warning light that would eg come on when the belt snapped. If the field current can fall to zero / no phase signal when things are spinning, that is indistinguishable from a rotation stoppage. And so in order to avoid false warning when the former occurs, the system has to be designed so that neither case brings on the warning light.

10 hours ago, dmr said:

 

Also just reduced my maximum duty to 65% and still got over 80 amps out of the alternator (100amp Iskra) when I first started the engine.  I would like to limit it to about 70 as it’s running off a single V belt, but if I do that I might not get enough current when the alternator is hot.  I will post some graphs of logged data in the next day or so.

Again, this is not great design. As you imply, if you can only specify a max duty cycle then the actual field current will vary according to temperature. But what you want is to specify a max output, which corresponds to a max field current, not a max duty cycle. Which is why I use the max field current setting, not the max duty cycle setting (which is also available).

 

 

The only thing I don’t understand is why so much field current is required to maintain a phase output. If the phase output is close to zero, the diodes don’t conduct and no current flows. In fact no current flows until the diodes start to conduct and that is with a phase voltage swing of battery + 2 diode drops, say 14v at least. Until the diodes conduct, the power output of the alternator is more or less zero. So I would have thought it would just take an extremely small % of field current to get the phase voltage up to the point of diodes almost conducting. I must be missing something if you need 20-something %.

Edited March 26 by nicknorman

[dmr]

5 minutes ago, nicknorman said:

It matters if you want have an alternator warning light that would eg come on when the belt snapped. If the field current can fall to zero / no phase signal when things are spinning, that is indistinguishable from a rotation stoppage. And so in order to avoid false warning when the former occurs, the system has to be designed so that neither case brings on the warning light.

 

The Zeus does not suppport a traditional warning light, maybe this is part of the reason.

It should be relatively easy to sort this out in software, if there is no field drive then a tacho signal is not expected so the light should not/can not light, but when there is a field drive then the light should come on if there is no output current. This is where that alternator shunt really does become useful 😀.

 

I am starting to see that an alternator shunt is actually potentially quite useful. Alternator current does drop off quite a lot as the alternator warms up so if I want to limit current when the alternator first starts (to control belt slip) but still have a good current when the alternator is hot then working off alternator current rather than duty cycle might be a good idea......but Zeus does not do this.

 

 

==============================

 

Here is part of a data log from Zeus.

This shows the current falling off as the alternator warms up. Ive got duty cycle set to 65 maximum and alternator temperature set at 65/75 which is probably a bit low.  This all just about works but I would be happier limiting start up current to 70 amps but this would stop me getting 70 amps when the alternator is hot (or would if I increased the alternator temperature a bit)

 

blue: current

red:  duty

green alternator temperature

 

 

[cheesegas]

25 minutes ago, dmr said:

The Zeus does not suppport a traditional warning light, maybe this is part of the reason.

It should be relatively easy to sort this out in software, if there is no field drive then a tacho signal is not expected so the light should not/can not light, but when there is a field drive then the light should come on if there is no output current. This is where that alternator shunt really does become useful 😀

Unfortunately it's not as simple as that - if there is field drive and the engine is running, there is always output current. You can't have the field on and no current flowing in the stator. This is the problem; to maintain float voltage, almost zero current must flow out the alternator (well ok, maybe 100ma...) which means the field must be sufficiently low for this to happen. This level of field current is not enough for the tacho to work.

 

Your logic system would work fine when in bulk or absorption, but in float with no loads switched on in the boat there's the situation of the field being very low and the output current of the alternator being next to zero; the tolerance of a shunt is probably less than the current needed to maintain a lithium battery in float. 

 

Lithium systems are even more difficult as there's a situation which could last 10 minutes or so when the regulator drops the voltage from say 14.2v to 13.4v; the battery must actually discharge a little in order for the regulator to achieve target voltage. For this to happen, the field must be zero and the output current and thus tacho will be zero also. This looks the same to the regulator as the engine being switched off.

 

A crude alternative would be to force the alternator to always output current by switching in a ballast resistor when in float, but that's wasting energy...

 

Another, better, workaround for the tacho issue is on dual alternator installations, to connect the regulator's tacho wire to the starter alternator so it always gets signal. I'm frankly surprised this isn't recommended in the manual.

 

 

10 hours ago, dmr said:

Unless I missed it there is almost zero documentation available for the Zeus, its operation and the the App, just the installation wiring diagrams. Maybe this will turn up soon???

I agree that documentation is not very good and they seem to rely heavily on the brief blurbs in the app rather than having a detailed manual. There's no explanation of how various settings interact, for example the minimum field. Some descriptions are bad too - the voltage at which the regulator drops the field under a high voltage fault condition is called Maximum Voltage. This sounds like the maximum allowable charge voltage under normal conditions, and the blurb in the dialog box doesn't do much to explain. High Voltage Cutout would be more descriptive.

Edited March 26 by cheesegas
missed an important 'not'!

[dmr]

3 minutes ago, cheesegas said:

Unfortunately it's not as simple as that - if there is field drive, there is always output current. You can't have the field on and no current flowing in the stator. This is the problem; to maintain float voltage, almost zero current must flow out the alternator (well ok, maybe 100ma...) which means the field must be sufficiently low for this to happen. This level of field current is not enough for the tacho to work.

 

Your logic system would work fine when in bulk or absorption, but in float with no loads switched on in the boat there's the situation of the field being very low and the output current of the alternator being next to zero; the tolerance of a shunt is probably less than the current needed to maintain a lithium battery in float. 

 

Lithium systems are even more difficult as there's a situation which could last 10 minutes or so when the regulator drops the voltage from say 14.2v to 13.4v; the battery must actually discharge a little in order for the regulator to achieve target voltage. For this to happen, the field must be zero and the output current and thus tacho will be zero also. This looks the same to the regulator as the engine being switched off.

 

A crude alternative would be to force the alternator to always output current by switching in a ballast resistor when in float, but that's wasting energy...

 

Another, better, workaround for the tacho issue is on dual alternator installations, to connect the regulator's tacho wire to the starter alternator so it always gets signal. I'm frankly surprised this isn't recommended in the manual.

 

 

I agree that documentation is not very good and they seem to rely heavily on the brief blurbs in the app rather than having a detailed manual. There's no explanation of how various settings interact, for example the minimum field. Some descriptions are bad too - the voltage at which the regulator drops the field under a high voltage fault condition is called Maximum Voltage. This sounds like the maximum allowable charge voltage under normal conditions, and the blurb in the dialog box doesn't do much to explain. High Voltage Cutout would be more descriptive.

 

Maybe I didn't explain my thinking too well?? If we want a simulation of a traditional warning light which can detect belt (or brush) failures then this should only come on if Zeus is providing significant field drive and not detecting corresponding output current. The light should not come on if there is no charge, and probably not during float. Something more complicated like Nick describes might be better, but I would really just like to just see a light come on and go off when I start the engine so a simpler system should work.

 

Assuming most Zeus systems are controlling the domestic alternator comprehensive detection of belt failure is a bit less critical as its usually the starter belt that drives the water pump.

 

Taking the tacho signal from the starter alternator sounds like an exellent idea, but I don't think I can be bothered to redo the wiring but will if Arco make suitable changes to their software.  I have disabled the "tacho required" option because as you say its really not that significant.

[nicknorman]

9 minutes ago, cheesegas said:

Unfortunately it's not as simple as that - if there is field drive, there is always output current. You can't have the field on and no current flowing in the stator. This is the problem; to maintain float voltage, almost zero current must flow out the alternator (well ok, maybe 100ma...) which means the field must be sufficiently low for this to happen. This level of field current is not enough for the tacho to work.

 Picking up on this and having answered my previous question about why the field current % has to be so high for the tacho to work at low rpm:

 

I was for some reason thinking an alternator was effectively a current source. But at the heart is Faraday’s Law which describes emf being a function of magnetic field and (effectively) rpm. So the alternator is fundamentally an emf generator, not a current generator.

 

So in order to get a decent phase signal of say 10v pk-pk one might need quite a lot of field current at low rpm. But no current flows because the diodes don’t go into conduction, until the phase voltage gets to battery voltage + diode drops. For a given low rpm, if you ramp up the field, no current flows as the emf ramps up towards (say) 14v. Then just a slight increase in field current is needed to push the voltage beyond the battery voltage + diode drops, and suddenly the alternator is outputting current. So there should be a fairly wide margin of field current at low rpm to keep a modest voltage - but a voltage below diode conduction - on the phase output. The problem arises due to the rpm term in faraday’s equation, such that a wide range of field current is needed to maintain the “phase output live but diodes not conducting”, over a wide range of rpm.

[dmr]

4 minutes ago, nicknorman said:

 Picking up on this and having answered my previous question about why the field current % has to be so high for the tacho to work at low rpm:

 

I was for some reason thinking an alternator was effectively a current source. But at the heart is Faraday’s Law which describes emf being a function of magnetic field and (effectively) rpm. So the alternator is fundamentally an emf generator, not a current generator.

 

So in order to get a decent phase signal of say 10v pk-pk one might need quite a lot of field current at low rpm. But no current flows because the diodes don’t go into conduction, until the phase voltage gets to battery voltage + diode drops. For a given low rpm, if you ramp up the field, no current flows as the emf ramps up towards (say) 14v. Then just a slight increase in field current is needed to push the voltage beyond the battery voltage + diode drops, and suddenly the alternator is outputting current. So there should be a fairly wide margin of field current at low rpm to keep a modest voltage - but a voltage below diode conduction - on the phase output. The problem arises due to the rpm term in faraday’s equation, such that a wide range of field current is needed to maintain the “phase output live but diodes not conducting”, over a wide range of rpm.

 

I agree. So if Zeus is reading the phase/tacho/W signal via an ADC then it could all be done very neatly, but if Zeus is reading the squared up signal as a digital input it can't be done quite as well or safely.

If Zeus is in float or shut down then the Tacho signal signal is really not important (unless its sharing it on the CANbus), as far as I can see its only real purpose is to support the speed dependant duty cycle limit.

As an aside the speed dependant duty cycle limit does not really work for me, there are not enough speed points, so I can't limit it at tickover without also reducing it at lower running speeds.

 

The new Victron DC-DC converters look dead good so an option for lithium would be to connect the domestic alternator to the starter battery and charge the lithiums via the DC-DC. Two converters cost less than a Zeus 😀. I really need the duty cycle and temperature limits so the Zeus is the best option for me.

[nicknorman]

Been down to the boat for a few days, thought I would generate some log data. The voltage and rpm data comes from the regulator chip - and the communication protocol does rather limit the resolution - and from the Mastershunt for the current and SoC.

Graphs can be difficult to read when they have several different parameters but hopefully one gets the idea that the scale for current and SoC is on the left, and for volts and rpm on the right. I factored rpm by 100 to make it work alongside voltage.

 

The batteries are already well charged at around 94%. At the LH side of the graph I flick the target SoC switch to 100% and it begins charging. We are cruising and rpm is up and down for moored boats etc, since the setup is on field current limit to 2A, the charge current varies quite a lot according to rpm. But note that the current remains at 75 to 95A until the SoC is 97% - the joys of Li! In fact you can see that the SoC jumps from 98% to 100% as the Mastershunt synchronises itself when the current falls to 5% capacity for a couple of minutes. So in fact, the SoC where the current starts to fall was around 98.5%!

When the system detects fully charged (current below 5% for a couple of minutes) the regulated voltage changes to 13.3 which is less than battery voltage, so you can see the battery current is negative (discharging) and alternator output is zero. However the rpm readout from the chip is unaffected due to the "phase keep alive" function I mentioned previously. Whether the phase output would also be enough to operate a tacho I am not so sure.

 

Oh and you can see there is a lot of smoothing on the current trace, this is a function of the Mastershunt. I really should use the BMV712 current which is much more instantaneous, but this “currently” only goes to the BMS, it doesn’t get passed on to the regulator. I’ll need to change the code a bit.

 

Edited April 14 by nicknorman

[dmr]

You will note that in my plot a few posts back I was really lucky because current, duty cycle and temperature all fit very nicely on the sames Y axis.😀

 

I note that your speed resolution is low, just like the Zeus. It should be possible to do much better, especially if the controller has a counter/timer, but this will depend on both hardware and software detail.

 

I do not yet know if Zeus is going to handle end of charge detection correctly as a low current could be due to a fully charged battery OR potentially due to the engine speed dropping and not taking account of low speed duty cycle reduction. I suspect our market is very small and most sailing boats will not see as much big speed changes as canal boats, but there is always mooring manouvers so it should be ok.

 

I think I read that the Wakepeed can go into a zero current mode rather than a zero charge/float mode, dunno if the Zeus can do the same.

 

A liitle job later this week is to get a proper ignition light working. I reckon a relay on the old ignition wire will do this. (output from diode trio activates a changeover relay and breaks the indicator light drive. Neater than a big resistor.)

[nicknorman]

1 hour ago, dmr said:

You will note that in my plot a few posts back I was really lucky because current, duty cycle and temperature all fit very nicely on the sames Y axis.😀

 

I note that your speed resolution is low, just like the Zeus. It should be possible to do much better, especially if the controller has a counter/timer, but this will depend on both hardware and software detail.

 

I do not yet know if Zeus is going to handle end of charge detection correctly as a low current could be due to a fully charged battery OR potentially due to the engine speed dropping and not taking account of low speed duty cycle reduction. I suspect our market is very small and most sailing boats will not see as much big speed changes as canal boats, but there is always mooring manouvers so it should be ok.

 

I think I read that the Wakepeed can go into a zero current mode rather than a zero charge/float mode, dunno if the Zeus can do the same.

 

A liitle job later this week is to get a proper ignition light working. I reckon a relay on the old ignition wire will do this. (output from diode trio activates a changeover relay and breaks the indicator light drive. Neater than a big resistor.)

 

On the speed resolution thing, this is just a feature of the LIN interface standard not the internal workings of the chip. The data field is 8 bits but it is somewhat logarithmic so the alternator rpm that I am operating at is way up near the top where each bit is worth a lot of rpm. It is annoying because one can burn some programmable "OTP fuses" on the chip to specify the number of poles, which would bring the reported alternator rpm down to mid-range where the resolution is much better. But unfortunately for some reason I still dont understand, NXP don't include how you do this on the data sheet (only that you can do it, and give you the bits to change) and refused to tell me when I asked nicely.

 

Still, it is of no real consequence as it doesn't affect the operation of the device beyond the logging data.

 

I really only included the graph to make you jealous of the "taking full charge whilst at 98% SoC" thing that lithium batteries do!😁

 

Edited April 14 by nicknorman

[dmr]

55 minutes ago, nicknorman said:

 

On the speed resolution thing, this is just a feature of the LIN interface standard not the internal workings of the chip. The data field is 8 bits but it is somewhat logarithmic so the alternator rpm that I am operating at is way up near the top where each bit is worth a lot of rpm. It is annoying because one can burn some programmable "OTP fuses" on the chip to specify the number of poles, which would bring the reported alternator rpm down to mid-range where the resolution is much better. But unfortunately for some reason I still dont understand, NXP don't include how you do this on the data sheet (only that you can do it, and give you the bits to change) and refused to tell me when I asked nicely.

 

Still, it is of no real consequence as it doesn't affect the operation of the device beyond the logging data.

 

I really only included the graph to make you jealous of the "taking full charge whilst at 98% SoC" thing that lithium batteries do!😁

 

Many many years I had the use of a minicomputer that had ROM chips based on real OTP fuses, and the myth is true, over time the blown fuses really do grow back and make some horrible bugs.

 

The Zeus rpm does bounce between two adjacent values, just like yours, so if I have set a speed dependant duty cycle limit that bounces about too, no real problem but it looks a bit messy on the logged data.

[nicknorman]

41 minutes ago, dmr said:

Many many years I had the use of a minicomputer that had ROM chips based on real OTP fuses, and the myth is true, over time the blown fuses really do grow back and make some horrible bugs.

 

The Zeus rpm does bounce between two adjacent values, just like yours, so if I have set a speed dependant duty cycle limit that bounces about too, no real problem but it looks a bit messy on the logged data.

The duty cycle limit (actually, a field current limit in my case) has some “ramping” (ie low pass filtering) applied on the way up, but not on the way down. So if the rpm goes up the field current limit will ramp up slowly, but if the rpm goes down it will rapidly reduce the field current limit, to avoid a transient high mechanical load. This seems to avoid any issue with noticeable “hunting”. And anyway there are still a lot of counts between idle (850 engine rpm) and 1300rpm - the latter being the point at which field current restriction is lifted - so even without the ramping I don’t think hunting would be noticeable.

[dmr]

11 minutes ago, nicknorman said:

The duty cycle limit (actually, a field current limit in my case) has some “ramping” (ie low pass filtering) applied on the way up, but not on the way down. So if the rpm goes up the field current limit will ramp up slowly, but if the rpm goes down it will rapidly reduce the field current limit, to avoid a transient high mechanical load. This seems to avoid any issue with noticeable “hunting”. And anyway there are still a lot of counts between idle (850 engine rpm) and 1300rpm - the latter being the point at which field current restriction is lifted - so even without the ramping I don’t think hunting would be noticeable.

 

I didn't mean hunting, just the speed jumping between the two adjacent "digitisation" states due to the low speed resolution and some sort of noise, not an issue but it just looks bad.   Likewise Zeus only displays and logs temperature to a 1 degree resolution but it holds that temperature really well so its internal measurement must have more resolution (or averaging). 

I just don't like seeing digitisation steps 😀