Some things I learned from Jas Sandhu and Eric Bentsen at Xantrex

YehoshuaAgapao

If network cables are less than 6.5', echoing can occur, which and can be disruptive to enhanced interactive mode (charge controllers and inverter communicating loads and voltage with each other).

When load shave and sell are both enabled, load shave will not engage until the charge controllers go into night mode. Their explanation was to make sure the batteries are kept full during solar production hours, even when solar production is less than the loads. I'm guessing this is carryover from the Trace Engineering days from the California power rolling blackout and price gouging crisis in 2001 (which will probably repeat when hyperinflation comes to America), and many of these California installations did not have generators (They said I don't need this behavior because I have a backup generator, but apparently they didn't design for that) and the rolling blackouts tended to occur in the mid to late afternoon. From my own experimentation, if sell is disabled, and you can manage to get battery voltage below grid support voltage (which I can't without bumping up the grid support voltage), it will kick into load shave even when solar is still providing for a portion of the loads.

This is my workaround to get load shave working the way I want it to work (with XW-Config over RDP connection from work as I'm not home between 5PM and 6PM):

Set grid support to 50.2V
Press apply (commit the changes)
Uncheck sell
Wait for the sell checkbox to gray out (grays out when battery voltage drops below grid support)
{Do not press apply}
Change grid support to 50.3 or 50.4V (depending on loads)
Press apply (commit both the sell disable and grid support voltage bump-up simultaneously)

Unless you did it to early, it should kick into load shave and stay in load shave. If the loads are too light, battery voltage will bounce up and it will go into grid support mode and generally does not go back into load shave. Have to set grid support to float minus 1V (enhanced interactive mode has grid support at float-1V when charge controllers are on float; enhanced interactive mode works its magic mostly during the absorption phase), wait a bit and start again.

NorthGuy

Re: Some things I learned from Jas Sandhu and Eric Bentsen at Xantrex

YehoshuaAgapao wrote: »

If network cables are less than 6.5', echoing can occur, which and can be disruptive to enhanced interactive mode (charge controllers and inverter communicating loads and voltage with each other).

Very strange. CANBus should work fine even with shortest connections. Although, in all practical application, it's hard to have less than 6.5' of cable.

YehoshuaAgapao wrote: »

This is my workaround to get load shave working the way I want it to work (with XW-Config over RDP connection from work as I'm not home between 5PM and 6PM):

You need to re-program it every day just to make it work?

Looks like they programmed it to do what they think it should do no matter what - probably a safeguard against idiots who change settings without any understanding of what they're doing. As a result, when you try to change something, it has a lot of unintended side effects in other places, which you simply cannot control.

This works fine if you have a generic UPS application, but it's a nightmare when you want to fine-tune it to your needs.

I'm so happy that I don't need any of that weired logic off-grid.

YehoshuaAgapao

Re: Some things I learned from Jas Sandhu and Eric Bentsen at Xantrex

NorthGuy wrote: »

Very strange. CANBus should work fine even with shortest connections. Although, in all practical application, it's hard to have less than 6.5' of cable.

My system is XW-Config<-5'->SCP<-3'->CC4<-3'->CC3<-3'->CC2<-3'->CC1<-5'->Inverter<-~50'->AGS

NorthGuy wrote: »

You need to re-program it every day just to make it work?

Looks like they programmed it to do what they think it should do no matter what - probably a safeguard against idiots who change settings without any understanding of what they're doing. As a result, when you try to change something, it has a lot of unintended side effects in other places, which you simply cannot control.

This works fine if you have a generic UPS application, but it's a nightmare when you want to fine-tune it to your needs.

I'm so happy that I don't need any of that weired logic off-grid.

To get it the way I want it to work and the way I think the manual says it should work yeah I have to change it every day. But it looks like it only works if there is no charge controller input and it was intentional design dating form the trace engineering days and the California power crisis (rolling blackouts and price gouging). Looks like this feature only works as advertised in the manual when there is only a battery bank but no renewable energy source.

NorthGuy

Re: Some things I learned from Jas Sandhu and Eric Bentsen at Xantrex

YehoshuaAgapao wrote: »

My system is XW-Config<-5'->SCP<-3'->CC4<-3'->CC3<-3'->CC2<-3'->CC1<-5'->Inverter<-~50'->AGS

See, that's 72', way longer that 6.5'.

YehoshuaAgapao

Re: Some things I learned from Jas Sandhu and Eric Bentsen at Xantrex

NorthGuy wrote: »

See, that's 72', way longer that 6.5'.

Think he meant more than 6.5' per segment, between devices, not the total length of everything.

BB.

Re: Some things I learned from Jas Sandhu and Eric Bentsen at Xantrex

I don't know about Xanbus--But there can be issues with short cables connecting a group of devices together and then a long run to another device.

Having a bunch of devices right next to each other can lower the "local impedance" of the bus, and cause reflection between the "short" and "long" sections of the bus.

So, for some types of buses, the 3' cable lengths can be "too short" and should be replaced with longer cables. Some of the older Ethernet over hard coax cable (using "vampire" taps) had something like 10-15 minimum between taps. And I had problems with a "too short electrically" disc drive back plane I had to redesign with lots of swooping "S" bends to get the traces long enough that the disc drives did not "locally" load the bus too much.

That being said, we had one person here with similar problems and cable lengths were not the issue.

Here are a couple threads--Not sure that they have any answers:

Config help with Xantrex SCP and 3 XW-MPPT60
Xantrex XW6048 Reliability

-Bill

NorthGuy

Re: Some things I learned from Jas Sandhu and Eric Bentsen at Xantrex

YehoshuaAgapao wrote: »

Think he meant more than 6.5' per segment, between devices, not the total length of everything.

BB. wrote: »

Having a bunch of devices right next to each other can lower the "local impedance" of the bus, and cause reflection between the "short" and "long" sections of the bus.

CANBus has 120 Ohm terminators at the ends (60 Ohm taken together). The resistance of devices is very high compared to this and is supposed to be negligible. The prsence of devices that are only listening shouldn't have any influence on the operations (same as if you would attach a voltmeter). So, the device driving the bus only "sees" the total length of the cable and terminators at the end.

It was designed to be used in cars, tractors, boats where uneven runs of cable of various length are very likely to occur.

I tried short and relatively long cables on mine and it worked all the time end appeared to be very resilient - you can reconnect devices, remove one of the terminators, reconnect back, and it's working without encountering any problems.

BB.

Re: Some things I learned from Jas Sandhu and Eric Bentsen at Xantrex

When are are dealing with "transmission lines" where the wave length of the frequencies involved is on the order of or shorter than the wave length of the cable, then each tap (receive and transmit) has a bit of capacitance/inductance that affects the wave form.

When you have multiple taps "jammed" together in one "local" area, they add together and dramatically affects the transmission line characteristics... A cat 5 cable is around 100 +/- 15 ohms. So, the endpoint termination should be somewhere around 100 ohms.

Again, I have no indication that there have been problems with Xanbus (Canbus?) and minimum length of transmission line impedance issues.

I guess that 50kHz, 125kHz, 250kHz, 500kHz, 800kHz and 1MHz (really kBit or Mbit per second) are common standards for CanBus--Those are not very high data rates so, the cabling should not be affected by minimum connection spacings.

-Bill

NorthGuy

Re: Some things I learned from Jas Sandhu and Eric Bentsen at Xantrex

BB. wrote: »

I guess that 50kHz, 125kHz, 250kHz, 500kHz, 800kHz and 1MHz (really kBit or Mbit per second) are common standards for CanBus--Those are not very high data rates so, the cabling should not be affected by minimum connection spacings.

Xan Bus operates at 250kbps.

YehoshuaAgapao

Re: Some things I learned from Jas Sandhu and Eric Bentsen at Xantrex

BB. wrote: »

When you have multiple taps "jammed" together in one "local" area, they add together and dramatically affects the transmission line characteristics... A cat 5 cable is around 100 +/- 15 ohms. So, the endpoint termination should be somewhere around 100 ohms.

I'm using 24 gauge cat6a cable. It is partially shielded - only an external foil. This is on everything but between the inverter and the AGS. The longer cable is using 26 gauge cat6a fully shielded (SSTP) cable. 24 gauge is 25.67 milliohms per foot while 26 gauge is 40.81 milliohms per foot.

bmet

Re: Some things I learned from Jas Sandhu and Eric Bentsen at Xantrex

YehoshuaAgapao wrote: »

I'm using 24 gauge cat6a cable. It is partially shielded - only an external foil. This is on everything but between the inverter and the AGS. The longer cable is using 26 gauge cat6a fully shielded (SSTP) cable. 24 gauge is 25.67 milliohms per foot while 26 gauge is 40.81 milliohms per foot.

Is CAT 6 UTP not sufficient? I have ran it along large, long runs of stagepin in cable looms (stage lighting) for video conversion transformers without a hint of 60 cycle hum.

niel

Re: Some things I learned from Jas Sandhu and Eric Bentsen at Xantrex

now i have dabbled a bit in digital stuff as a ham many years ago and with baud rates that low i don't see there to be much in the way of problems. i even used straight cables with higher baud rates in lengths about 4-5ft or so with no problems. cat 5 and cat 6 should work just fine imo and may be overkill, but if you have it, go with it as it shouldn't hurt.
but what do i know?

NorthGuy

Re: Some things I learned from Jas Sandhu and Eric Bentsen at Xantrex

niel wrote: »

cat 5 and cat 6 should work just fine imo and may be overkill, but if you have it, go with it as it shouldn't hurt.

I think flat telephone cables would work just fine, but I used CAT 5 because the signal wires are twisted. I don't really know if twisting is that important, but it's better be safe than sorry.

BB.

Re: Some things I learned from Jas Sandhu and Eric Bentsen at Xantrex

Twisting can help--But it can also hurt.

There are different pinouts for "Cat" cabling and non "Cat" applications (i.e., everyone uses the simple "telephone" connectors, but they do not follow the standards for pin outs).

So, check the documentation and see what pin out/off the shelf cabling the may recommend. Twisted pairs when the "wrong" signals are "twisted" can be much worse (coupling different AC signals together).

-Bill

NorthGuy

Re: Some things I learned from Jas Sandhu and Eric Bentsen at Xantrex

BB. wrote: »

So, check the documentation and see what pin out/off the shelf cabling the may recommend. Twisted pairs when the "wrong" signals are "twisted" can be much worse (coupling different AC signals together).

There are only two signal wires. They're in the middle of connector and are connected to green/light-green pair, which is twisted together in all standard Ethernet cables. I think they designed it this way especially for twisting. At least I don't see any benefits of using CAT5 instead of telephone cable except for twisting. I might be wrong here.

BB.

Re: Some things I learned from Jas Sandhu and Eric Bentsen at Xantrex

Twisting is good for high frequency noise suppression (both receiving and transmitting interference). But this is for pretty high frequency stuff (typically around 30-180 MHz) or so.

For an analog voltage meter, probably not a big problem unless there is a big radio transmitter nearby.

-Bill