Stump The Band (Dan)?

[matt_grooms]

Disclaimer - this is not a LOR question as much as it is an AC supply question. Turn back if you ave a weak heart.......

I have a LOR configuration (in an automobile) that has been the same since Jan 2010 and has a history of running flawlessly when powered either by local utility (110 VAC wall outlet) OR when powered by my 5,000 watt DC->AC inverter. I have run it for up to ten days 24x7 on public utility and up to 48 hrs on inverter without incident.

That's nine months of flawless function either parked or mobile. The inverter is no cheap piece of junk. It cost $500 and is designed for 100% duty cycle in a marine environment. It is powered by six 12 volt gelled electrolyte electric vehicle batteries. They are each the size of a large car battery.

In the last two weeks a problem has cropped up that has stopped my mobile (inverter powered) show in it's tracks.

When the 'system' (PC, 1602W, 2 16ch AC controllers) is powered (PUBLIC UTILITY) and running a scheduled show via everything works perfectly.

When the 'system' is powered (BATTERY/INVERTER) and runs the same show via the same schedule the LOR controllers lose synch across the RS485 network a few seconds into the first sequence of the show. Once the controllers lose synch there is no way to get them synched again and the serial port driver is hung. You can't kill the process, nor shut down the show, nor shut down Windows. The only way out is to poke the PC in the eye to force a reboot.

This problem is reporduceable ad naseum. It NEVER EVER happens on wall power and it happens EVERY SINGLE time on battery/inverter.

I have swapped out a new RS485 dongle, USB cable and all RJ45 cables. no change.

I have removed the USB serial port via device manager in Windows and have the system find it /install it after a reboot. Nothing changes.

wall power = 100% flawless S2 operation.

battery/inverter = 100% usb seriall port hang 1-2 seconds into first sequence. Sequence type makes no difference.

Running apps on the PC (battery/inverter power) that don't use the USB serial port are fine and will run forever on battery/inverter. Can run Winamp till my ears bleed and no hangs.

Inverter output is 124 volts AC / 59.9 hz as reported by my Kill-o-watt. Don't have an O-scope to look at battery/inverter wave form.

Automobile electrical system is eliminated as a culprit as all testing is done while parked with the battery bank being charged by public utility connected battery charger. If test is repeated with battery charger disconnected or with engine running the results are the same.

Clearly the problem is tied to the AC supplied by the battery inverter pair, but what is it? The PC runs fine on this inverter derrived AC. It's just the PC/usb sp driver/ RS485 dongle/ LOR controllers that have problems.


Dan - any ideas?

PS - the one thing I haven't tried is another computer. I'll try a laptop tonight and see if there is a change.

PSS - the single thing that changed just before this started happeing is I upgraded S2 from 2.7.X (whatever was current last January) to 2.8.X (latest). Possible connection? IF so WHY only when INVERETR powered? It shouldn't know the difference.

[-klb-]

Do you have the ferrite filtered USB cable? One with at least two large ferrites, if not three? If not, that is the first place to start.

If there have been any changes in your USB cable, or you AC power adapter for the computer, or even a loosening of the AC connections, I would think the filtered cable is the best place to start, and that electronic noise from the power supply is the culprit.

I had a highly variable issue with locking up last year while I was trying to upgrade firmware on 40 controllers. I noticed that when one of the volunteers was hanging around, I only averaged about 1 controller between lockups. When he was gone, I would usually get 5 or more.

Turns out that the extension cord was a little flakey on the connection, and that often when somebody moved it walking by, the AC adapter briefly turned off, and back on.. It was quite reproducible that plugging in the AC adapter would lock up the USB-485B. Switching to one of the 3 ferrite USB cables solved the issue...

I think you have something similar where some EMF noise is causing the USB link to get out of state, or locked up...

[matt_grooms]

I don't doubt a possible EMF based problem.

I do use the double ferrite bead USB cable supplied by LOR with the 485 dongle.

This evening, before substituting a laptop for the computer I have been using for nine months, I tried bringing things up on battery/inverter power (engine running even). To my great surprise the scheduled show started withowt issues (as if connected to wall power). I stopped/started the show scheduler several times without incident.

It seemed as though my gremlin had departed, as I was now a good twenty minutes into a show running (without incident) on battery/inverter.

I then simply took a step to the left and a section of rope light turned on (not programmed) and the USB serial lockup had returned ! Subsequent recovery attempts did not return things to normal.

I will test things with a laptop (if succesfull eliminating the PC hardware, USB serial port driver, S2 software, USB cable, and dongle from consideration). I suspect however that the gremlin will remain.

Either way I only have to survive till Novemebr 2nd (Dia de los Muertos event). After that I'm tearing down the Fiesta Mobile completely in order to start building Fiesta Mobile II, based on a much larger truck than Fiesta Mobile I was built around.

Thanks for the input.

[displaysinmotion_will]

im going to say that unless your inverter is a "pure sine" thats more than likley your problem most "cost inverters" and whe i say lower cost, most large pure sine wave inverters range in the 1500 + range are " modified sine" meaning its is not the same type of wave length that comes off of a public utilty grid. and creates problems for lor and othe sensitive electronics. the most common problem is that fades do not work properly

[matt_grooms]

I might buy into that idea if the problem was more random in nature, and if the problem had been present from the beginning (nine months ago).

The problem must exist in one of three places:

1) The inverter

2) The quad outlet hardwired to the output terminals of the inverter

3) The AC delivered by the inverter.

I'm leaning towards the the AC and need to get a scope on the waveform to see just how choppy it is and how much noise is present.

I guess what I'm really after is this:

The PC will run happily forever on the inverter output. Only the USB serial port driver is affected.

IF the LOR controllers are less tolerant on less than ideal AC power and are hanging, how does a controller hanging have an effect on the USB port serial driver.

In a normal LOR setup you can go around turning controllers off/on all night and it has no effect on the overall system.

The output stream from the PC to the controllers is one way for the most part so why is the USB serial port driver hanging?

[matt_grooms]

I might buy into that idea if the problem was more random in nature, and if the problem had been present from the beginning (nine months ago).

The problem must exist in one of three places:

1) The inverter

2) The quad outlet hardwired to the output terminals of the inverter

3) The AC delivered by the inverter.

I'm leaning towards the the AC and need to get a scope on the waveform to see just how choppy it is and how much noise is present.

I guess what I'm really after is this:

The PC will run happily forever on the inverter output. Only the USB serial port driver is affected.

IF the LOR controllers are less tolerant on less than ideal AC power and are hanging, how does a controller hanging have an effect on the USB port serial driver.

In a normal LOR setup you can go around turning controllers off/on all night and it has no effect on the overall system.

The output stream from the PC to the controllers is one way for the most part so why is the USB serial port driver hanging?

[Steve Saunders]

The computers power supply, or the capcitors on the motherboard could also be a source of frustration. We have seen a lot of electorlytic caps failing on mainboards lately.

-Steve

[-klb-]

I don't think anyone was suggesting that the controllers having issues on inverter would cause issues on the USB side.. Rather that especially on older firmware, dimming could produce totally unexpected results. I expect that even now, on a modified sine wave inverter, that there are dead gaps in the dimming curves.

In general, the RS-485 buss is quite unlikely to mess up the show computer, or USB drivers. But the USB link between the ftdichip drivers and the ftdichip hardware in the USB-485 has a track record of being a source of headaches in strong electrical noise fields. Note that everything, from the computer power supply to the strings of lights themselves will emit more electrical noise when running on inverter than on wall power.

As for the software upgrade, I think it was further back when S2 changed how it interacted with the serial port, and I think everything was to make this kind of issue less likely, and easier to recover. It is possible that the new version is sending the commands in a different order than the previous one, resulting in different noise generated. If it seems like it is one point in the sequence, does it still happen at the same time, if a different sequence runs first?

Separately, you may be one of the few cases where a SC485 adapter may work better for you, in that it does not have the USB link that is locking up on you. There have been more USB issues in the past, but they have been pretty rare the last couple of years. The ferrite filtered cords, and improvements in how the software interacts with the port have cleared up the majority of the issues reported here.

[gizmomkr]

Matt:

This is going to start ot sound like a broken record - but in a very general sense :

You allready diagnosed the problem. You know when running on clean AC power, everything is fine, and on inverter power (I never saw you specifically state it was pure sine, or not - only that it was "expensive") it screws up.

You seem to have some knoledge of electronics as you have made refferences to the wave form...

You KNOW nethier your PC or the LOR hardware were designed to be run on anything short of a pure sine wave. you probably understand the conceptual level of PWM dimming and how this changes the wave form.

Despite all this - you seem baffeled why feeding all this precision electronic equipment with far less / different sources than designed is causing some type of undesirable behavior.

I saw the original post, and I almost immediately saw the question of the pure sine inverter raised, and I really thought that would be the end of it. Yet I still see the questions flying - why does the pc lock up, why did it just start doing this - I guess my question is : why should it. You are running equipment in more ways than one under completely different conditions than it was intended to be run.

My advise - use a DC board with DC lights - I would wager that would not crash and burn. But I would also tell you to get that pc on a DC PSU.

All of that asside, I'll share a similar issue I had.

About 8 years ago, I installed a PC in the trunk of my car. I had an inverter feeding a standard PSU - and for almost a year it worked fine. All of a sudden, the video output went to 16 colors. I looked at device manager, and the video card was now an "unknown device" - keeping this story short and to the point: When I fed the PC from AC Power, it worked fine. Reconnected to Inverter: "unknown device"

Why did an inverter cause my video card to become an unknown device ? I really have no idea, but I know there are all kinds of reasons NOT to run my PC from an inverter.

I bought an industrial DC-DC PSU and eliminated the inverter - Everything worked perfect from that day on.

[matt_grooms]

-klb- wrote:

I don't think anyone was suggesting that the controllers having issues on inverter would cause issues on the USB side.. Rather that especially on older firmware, dimming could produce totally unexpected results.

the USB link between the ftdichip drivers and the ftdichip hardware in the USB-485 has a track record of being a source of headaches in strong electrical noise fields.

possible that the new version is sending the commands in a different order than the previous one, resulting in different noise generated.

does it still happen at the same time, if a different sequence runs first?

may be one of the few cases where a SC485 adapter may work better for you, in that it does not have the USB link that is locking up on you.

klb,

THANK YOU! This is the direction I hoped the conversation would eventualy go.

I suggested the LOR controller on modified sine wave inverter affecting the USB connected 485 device / driver since the PC has a 100% track record on inverter.

Dimming is not an issue as in all cases when the 485 adapter/driver hangs it is doing simple on/off/on step commands.

history of ftdichip drivers and the ftdichip hardware problems in strong electrical noise environments. Now this IS interesting. The lights on board the vehicle HAVE changed during the Halloween reconfiguration, AND the USB cable connecting the RS485 adapter has been slightly repositioned.

New version sending commands in different order. OK, but again why an impact on inverter and not public urtility?

When RS485 buss / ftdichip driver hangs? 99% of the time it is .5 to 1 second into the first sequence of a show. Doesn't matter if it's a background, startup, animation, musical, or shutdown sequence it has the same timing / effect.

SC485, that's the serial port adapter? Man I wish I had one of those to test with!

I fully understand the mantra of 'use a sine wave inverter', but this modified wave inverter worked fine for the past nine months and I'm not going to ditch it for a $2,000 sine wave unit. There is a reason this configuration has stopped being reliable and I'd like to figure out why, rather than throwing money at it.

KLB. thanks again. your well response gives me hope I may be able to figure this one out with a little help.

Matt

[-klb-]

Ah.. The lighting change, especially if there are more LED, and even more so if any strings with capacitors in them, you will have increased the electrical noise considerably.

And I think I have said this before, everything radiates more electrical noise when running from modified sine wave inverters. Changing geometry, by changing the USB adapter position could have also allowed the USB cable to pick up more signal.

As for the possibility of different command sequence, I do think that was further back in the release history, but if you had a number of commands starting at the same time, I think the newer version may try to do better at getting more of the channels to actually change state at the same time, possibly generating a larger spike as the inverter tries to adapt to the sudden change in load. But, I think the change in lighting is probably more likely to relate.

Did you ever try with the laptop? How did that compare? One thing that could be tried with the laptop is running the controllers off the existing inverter, and the laptop off a DC adapter. Or even off a different inverter...

[matt_grooms]

gizmomkr wrote:

Matt:

This is going to start ot sound like a broken record - but in a very general sense :

Agreed.

You allready diagnosed the problem. You know when running on clean AC power, everything is fine, and on inverter power (I never saw you specifically state it was pure sine, or not - only that it was "expensive") it screws up.

Modified sine wave, and it has a nine month perfect track record of "not" screwing up. Recent changes (SW version, lights controlled, cable routing, etc) have created the problem. I'd like help finding the problem so I can correct it.

You seem to have some knoledge of electronics as you have made refferences to the wave form...

Yes.

You KNOW nethier your PC or the LOR hardware were designed to be run on anything short of a pure sine wave. you probably understand the conceptual level of PWM dimming and how this changes the wave form.

Modern switching power supplies are very tolerant of out of spec AC input. Worldwide public utility is nothing like we are blessed with in the US. I don't buy your argument at all with regards to PC power supplies. The LOR controllers with their zero cross timing issues are certainly more critical with respect to AC supply. That said, reference my nine month track record of 100% reliable operation on this same modified sine wave inverter, fed from the same six battery bank, charged by the same 140 amp alternator.

Despite all this - you seem baffeled why feeding all this precision electronic equipment with far less / different sources than designed is causing some type of undesirable behavior.

Baffled? Your words not mine. I'm curious what change in my setup has uncovered this weakness, and what aspect of LOR operation is is the culprit.

I saw the original post, and I almost immediately saw the question of the pure sine inverter raised, and I really thought that would be the end of it. Yet I still see the questions flying - why does the pc lock up

PC doesn't lock up; driver hangs, RS485 buss is dead.

., why did it just start doing this

Yes

- I guess my question is : why should it. You are running equipment in more ways than one under completely different conditions than it was intended to be run.

You're getting preachy here.......

My advise - use a DC board with DC lights - I would wager that would not crash and burn. But I would also tell you to get that pc on a DC PSU.

The PC is happy running on it's AC psupply. DC lights aren't an option. I have AC powered devices.

All of that asside, I'll share a similar issue I had.

About 8 years ago, I installed a PC in the trunk of my car. I had an inverter feeding a standard PSU - and for almost a year it worked fine. All of a sudden, the video output went to 16 colors. I looked at device manager, and the video card was now an "unknown device" - keeping this story short and to the point: When I fed the PC from AC Power, it worked fine. Reconnected to Inverter: "unknown device"

Why did an inverter cause my video card to become an unknown device ? I really have no idea, but I know there are all kinds of reasons NOT to run my PC from an inverter.

I bought an industrial DC-DC PSU and eliminated the inverter - Everything worked perfect from that day on.

Glad that fixed your problem. Mine is different.

Thanks for the input.

[matt_grooms]

-klb- wrote:

Ah.. The lighting change, especially if there are more LED,

No LEDs introduced at all. Change consists of exactly four elements. 1) Four ghoul characters, each with a short (20 lamp) AC string. 2) Two strings of AC driveway marker lamps (one string of 50? AC lamps spread across five plastic fixtures in each) 3) Four individual lanterns, each containing a single 'flame' (flicker) bulb 4) Three strings (in series) of 'flame'(flicker) bulbs. These bulbs and the four in number 3 above each use two plates with an air gap between them instead of a filament to create a flame/flicker effect. (I'm hearing the audience gasp and say in unison "electrically noisy, noisy, noisy".

And I think I have said this before, everything radiates more electrical noise when running from modified sine wave inverters. Changing geometry, by changing the USB adapter position could have also allowed the USB cable to pick up more signal.

Certainly possible.

As for the possibility of different command sequence, I do think that was further back in the release history, but if you had a number of commands starting at the same time, I think the newer version may try to do better at getting more of the channels to actually change state at the same time, possibly generating a larger spike as the inverter tries to adapt to the sudden change in load. But, I think the change in lighting is probably more likely to relate.

Did you ever try with the laptop? How did that compare? One thing that could be tried with the laptop is running the controllers off the existing inverter, and the laptop off a DC adapter. Or even off a different inverter...

Haven't tried the laptop, but will. Had a similar thought of running a hybrid test arrangement with laptop (or original PC) on public utility with controllers on inverter and monitoring the results. If that worked could reverse with PC on inverter and LOR gear on public utility to see which combination was same/better/worse.

Thanks for keeping my brain moving in the right direction. I wonder just how noisy are those little arc lamps after all???????????

[-klb-]

The reason that modified sine wave inverters are noisy is that they are a 4 step cycle. Off, positive step, off, negative step. This does offer several improvements over older square wave units, that just simply alternated between postive steps and negative steps. But the modified sine wave unit causes 4 square transitions per cycle instead of 2, so the harmonic spectrum of the square corner is created twice as often. Remember that for a EMF perspective, every waveform is really the sum of a series of sine waves, of increasing frequency, until you build up the square corner. Where the real world implementation runs out of frequency response to generate that corner, you get a ring around the peak frequency, that is both radiated, and a voltage spike in the system.

As for the little arc lamps, I really ought to try some out for my own information. I'm not sure if they really operate in more of an arcing mode, or more like a plasma lamp, like a neon lamp...

But, if you look back at the discovery of radio, the first transmitters were simple arc gap, which wound up transmitting in a whole spectrum of frequencies, simply by creating an arc, with a reasonably high current, the resulting RF could be detected quite a distance away.

[gizmomkr]

I can appreciate wanting to know "why" - but I guess my perspective is the "why" clearly is the presence of the modified sine wave. "why now... does it lock up" the question lingers...

I offer my story and symptoms as it seems very relivant. We both experienced computer problems due to the use of an inverter - we have different equipment, therefore different symptoms - but who is to say that changing your AC PSU for A DC one may remove enough of the noise from your system to aleviate the problem. Maybe it wont - But you would never know if you didnt try.

I do appologize if I came across as "preachy"

I prefer to "Fix the problems source" - To me, changing the USB-rs485 to a Serial-485 is a band aid (even if it works, its just masking the problem) But I would absolutely want to know IF it worked. But I would want to know EVERYTHING else that might work

The inverter I used in my setup had all sorts of OTHER side effects - most of them didnt matter, but I chose not to invest my efforts trying to determin WHY my video card suddenly developed this issue. I went to the SROUCE and got rid of my inverter for powering the PC.

You seem interested in testing with a Serial-to-rs485, yet you dont even want to consider trading your AC Psu for a DC-DC PSU. You talk about how expensive your inverter is - implying because you spent a lot of money on it, surely it must be good.

Im just saying keep an open mind- explore all your options. If it were my setup I would try to find a way to run tests with a DC PSU AND a serial adapter. And then I'd go back to the USB adapter - I would want to know IF it worked with each and every possible combination. When its all done you learned something about your setup, and you are armed with that information should you run into problems in the future for this or other projects.

[matt_grooms]

Thanks. I'm not eager to try a DC power supply since it's not as likely a candidate as the USB<->RS485 adapter.

I'm tied to the inverter because I have AC powered display elements that have no DC equivalent (air compressor, fog machine, AC powered PA amplifiers).

Based on klb's (?) recollection that the USB<->RS485 adapter has a history of vulnerabilty to noise, it seems like the most likely candidate. If a SC485 cures my problem for $30, I'll be able to move forward.

No arguement that a sine wave inverter might make my life better, I can't justify the expense. If I can get things to work with less than perfect AC power, that's what I'll do.

Again, no arguement that a DC<-> power supply would be quieter, I don't feel that's where the problem lies.

Appreciate you input on figuring out my problem.

[-klb-]

One thing I never did figure out was if my issue was due to noise picked up in the USB cable through EMF, or passed through the laptop power to the USB cable. Where I was experiencing my issue, it was just an extension cord from an old outlet in a fire house, a power strip with one controller plugged in right side only, the laptop power supply, laptop, USB cable & adapter, and one section of cat 5. I know I could reproduce the issue by unplugging the laptop power adapter and plugging it back into AC. I never did try and isolate if it would cause the issue without being plugged into the laptop.

One thing that makes life more fun is that I expect you are operating with an entirely floating ground... I was going to suggest that you might try adding some grounded metal shielding around the USB cable, but then it occurred to me that anything you might call a ground could also be an antenna, and that such an attempt might just couple more signal into the cable....

You might also look around, and see how short a USB cable you can find. I know that I have a 1 foot mini USB cable that came with a seagate HDD, but that would only work if you have the USB-485, not the USB-485B... Oh, and I did manage to replicate the issue on either of those adapters. Right now, I can't think of any likely sources for a short standard USB cable.

I'll also comment that I do think the laptop is still worth testing, as it is possible that your issue is noise feeding through the PC, where the external laptop DC power supply, (or even battery power) may provide better isolation from line noise if the path happened to be through the PC, rather than picked up as EMF...

[matt_grooms]

-klb- wrote:

I expect you are operating with an entirely floating ground... I was going to suggest that you might try adding some grounded metal shielding around the USB cable, but then it occurred to me that anything you might call a ground could also be an antenna, and that such an attempt might just couple more signal into the cable....

The negative side of the battery bank is tied to vehicle ground (required for battery bank charging from dedicated (2nd) 140A allternator I installed. The inverter case is gounded via mounting hardware. The inverter output includes a ground but I never verified is tied to the inverter chassis. The inverter ground is not connected to the metal quad outlet box (mounted on lumber) but is of course connected to the ground lugs of the outlets. All electrical devices are connected via three grounded power strips. The computer, PA amplifiers, and two fo three LOR controllers (1602 W, a kit ) all use grounded power cords. The third LOR controller uses scavenged extension cords, ungrounded in a plastic enclosure.

how short a USB cable you can find.

I too have one of those 1 foot USB mini's. No ferrites, and the cable is a bear to get to. might try that AFETR laptop.

I do think the laptop is still worth testing,

Will try tonight with both factory AC wall wart and DC (cig lighter) adapter.

Looks like I better stop Halloween cruising and get to work!

Thanks all.

[matt_grooms]

I rerouted the USB cable away from AC lines feeding and controlled by the two LOR controllers in the immediate area (1602W and a 16ch assemble yourself P/N ????). The USB serial port / RS486 buss hang problem went away. It really does look like emf is the culprit.

So while the show was running merrily along for two hours on battery/inverter power I looked at a couple of other things.

The nominal voltage / frequency comming out of the inverter is 123 volts / 59.5 hz as measure by my Kill-A-Watt device.

The string of c17 'flame' bulbs around the top of the roof look crisp and bright when operated off public utility. When connected directly to the inverter output however, only about half of them light and those that do appear dim and listless. The voltage (124 v) and frequency (59.5) reported by the Kill-A-Watt are right there but the observed light ouyput is sub par.

When the same string is connceted to a LOR controller output channel set to 100%, the observed brightness, etc, is the same as when connected directly to the inverter output.

The weird part is the frequency is about double what is should be (114 hz vice 60 hz)? Suspecting a problem with the 1602W controlelr I moved the lights to a different LOR controller and the same thing is observed. The current consumption and wattage calculated by the Kill-A-Watt is the same as when connected directly to inverter power.

In summary - 'flame' light string on public utility is fine. Same string on inverter power is dim and if connected to a controller output appears to be double the frequency it should be.

Perhaps this is just an anomoly with the Kill-A-Watt and it would look different if viewed on an o-scope.

An additional observation is the performance of a fog machine (basically a heating element and a motor driven pump. On public utility the thing works perfect. On inverter power the pump often will not work.

Unfortunately as odd behavior (on inverter power) continues to stack up it looks more and more like I'm going to have to break the piggy bank and purchae either a sine wave inverter, or more likely, a mega-expensive, super quiet Honda pure sine wave 6,500 watt generator.

Any ideas? Anyone?

[-klb-]

I suspect those flame lamps are rather non linear.

Remember that the RMS voltage measurement is the equivalent to the DC voltage that would have the same heating effect in a restive load as the actual AC waveform observed. So for pure resistive loads, many different waveforms can have the same heating effect.

But for non linear loads, those waveforms, that will all have the same RMS voltage readings may have significantly different results. For example, true sine has voltage going all the way up to a peak of nearly 170 V. But the modified sine wave inverter is only switching between off, positive voltage, off, negative voltage, and back through. Since it spends about equal time in each state, it uses a lower voltage for positive and negative steps, typically about 145V. With the right pulse widths, this results in the exact same readings on a true RMS meter, and exactly the same output from a load like a light bulb.

But, motors often take 20% more power, and don't produce as much torque. That may be what is affecting the fogger pump. And with the flicker lamps, that difference between 145 and 170 is probably quite significant..

And the long zero at zero crossing from the modified sine wave probably makes the killawatt very susceptible to any kind of unexpected behavior causing a false 2x reading..

But I am very glad to hear that just rerouting the USB cable appears to have solved the primary problem for now!