Mastering the Obvious (a long winded story) 7

Warning: Long winded story, I won't be offended if you skip it. I just had to share this with all of you in the Drives industry with me. No help needed, but I thought some of you would get a kick out of this story.

I got a call this week to go out and help a contractor program a PID loop on a VFD he installed. He has used me for this in the past and we work well together, but usually he also uses me to help him convince his customers to use VFDs in the first place. This time it was different, his end user contracted with him to just supply and install a drive on a pump application that they had already purchased.

The PID loop was a breeze, extremely simple level control loop for a 100HP centrifugal pump emptying a holding tank of waste-water coming from a gravel wash plant. The water came off of the various processes by gravity into a holding tank, then was pumped up hill into a settling pond. The total dynamic head; only 13 feet above the bottom of the tank. The tank was 7 feet deep. After the fact I thought a VFD was overkill but I wasn't hired to consult on this in the beginning, just to make it work. They had an existing pump that was being replaced by a new "more efficient design" that would save them energy and were told the VFD would compound that energy savings by modulating speed. Remember, I did NOT get the consult contract on this marvel of engineering thought. The fact that there was no throttling valve to be replaced by the VFD nor a solid reason to even bother maintaining a level in the tank was apparently never even brought up (until I got there). Still, I went about the job I was hired to do and programmed the PID loop and the ultrasonic level transducer like a good little soldier. As I said, a piece of cake.

The only "minor" problem I encountered was that the tank overflowed! Why? Because at full speed, the pump was not pumping out as much fluid as was coming into the tank! Hello??? Rocket Science here? I think not. Bottom line, the pump was too small for the task. My assessment took all of about 38 seconds, the time it took for the tank to overflow with the new pump running full speed AFTER the old pump was turned off. So they all look at me and start pointing fingers that I had not "tuned the drive and level controller" properly. I explained that this had NOTHING to do with "tuning" anything. The VFD was at FULL SPEED the entire time (after about a 5 second ramp soak from the loop). For some reason, all these other genius mechanical engineers had a difficult time grasping my 38 second conclusion. I saw more graphing calculators getting whipped out in front of me than I have seen in one place in my entire life. All of them attempting to prove to ME that I am wrong about this. Meanwhile, THE TANK IS STILL OVERFLOWING BEHIND THEM!!!! I soooooo wish I had set up a video camera, the ironic humor video would have been a gold mine!

So I investigate, because obviously I have made some kind of grave error in my assessment of the dirty water flowing over the top of the tank. Something MUST be wrong with the VFD, motor etc. Even though I had nothing to do with the selection, I was having a bit of fun with this, so I started poking around. Here is what I found.

Old pump: 300HP 460V, rated capacity; 14,400 GPM
New Pump: 100HP 460V, rated capacity; 4250 GPM at rated speed

Apparently they thought that somehow the VFD was supposed to magically "fix" this disparity. Granted, the outflow pipe had been significantly shortened from the original design that needed the 300HP pump, so that was why the "geniuses" had come up with only needing 100HP. I said, "Why not just put a VFD on the existing pump and modulate the speed, using only as much power as you need? You would get the same net effect and not have to put in another pump." Their answer: "The 300HP VFD would have been way too expensive." Me; "Compared to a 100HP VFD, new service to feed it, 100HP motor and new pump?" No clear answer came to THAT piece of pretzel logic.

Instead they said "Can't the VFD run the pump faster than 60Hz?" Absolutely I said, but that will do you NO GOOD. "Do it anyway!" So I did, restarted the VFD and of course the pump accelerated, filled the outlet pipe, then promptly slowed down until the VFD took it off-line on OL trip. "See, that VFD is too small. THAT's the problem!" they said. Having not sold the VFD I technically shouldn't have cared, but now I was getting angry because they were implying that somehow this was an ELECTRICAL problem, essentially besmirching my chosen profession! I conducted an in-situ class on VFD technology and what you can and cannot expect from them. I finally convinced a few key management types (the crowd was growing bigger every minute) that maybe this WAS a mechanical problem. It was a tough sell, but the WATER CASCADING OVER THE TOP OF THE TANK helped my argument a lot. They finally let me leave, but I have to be on call all next week for when they call a big meeting on this, because the MEs are STILL convince that their calculations are correct, and this IS an electrical problem.

What tickles me most now is, I will get paid handsomely to sit in on that meeting as the MEs attempt to convince everyone in the room that the water was NOT flowing over the top of that tank, I guess! In addition, one of the managers said "From now on, we will hire you to evaluate all vendor proposals that include VFDs."

Bottom line, I win.

JRaef.com
Eng-Tips: Help for your job, not for your homework Read faq731-376

 

[itsmoked]

Funny thing is the VFD would've worked, they'd have never known.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[ozmosis]

until somebody else came along and then said..why don't you fix the return valve?..
In actual fact, we did some calcualtions a fews months later and whilst it is a good application for a drive, the payback time and ROI was very long (4~5 years) as they reduced output at the quarry so were only operating for typically 5 hrs/day 5 days a week. The cost of a 75kW drive in a 'quarry-proof' enclosure was quite high making it an expensive proposal too.
The ROI on fixing the non-return valve was about 3 hrs and then the savings were quite apparant as the pump was not actually running for about 50% of the time.
I think we did supply a softstart eventually so the salesman got a return on having to clean his shoes..

 

[xnuke]

I was a working as a field service engineer for a company that manufactured production tools for the semiconductor industry. I was assigned to copper electroplaters, which were a new product line for our company. Each plater had four plating cells, each with lift and spin servo motors.

We got the tool installed and ran all of the mechanical motions of the process with deionized water to avoid working around harmful chemicals. All four plating cells functioned as they should: they began spinning, dropped into the water-filled cells, then lifted out again and stopped spinning so that the pick-and-place robot could remove the product and advance it to the next stage of the process.

We then dumped the water and replaced it with electrolyte solution. During the first plating run, cells 1, 2, and 3 functioned normally, but cell 4 had an error associated with the lift motor that would not allow it to lift from the cell. We retrieved the product, then tried it again with the same result. We put the motors through the steps under manual control and everything worked fine.

As this occurred right at the end of our 12-hour night shift as well as our work week, we turned it over to the day shift crew and the lead engineer. The lead (who was really into programming, but also trained in electronics) decided at the turnover meeting that it was a software problem. I tried my best to convince him that it was an EMC issue, as the only thing different between the water runs and the plating runs was the addition of plating current. I explained (in vain, of course) that the plating current conductor from the cathode ran parallel with the motor cables for both the lift and spin motors through a metal tube so that the whole head assembly could be swung out for maintenance, and that the collapsing magnetic field when the plating current turned off was wreaking havoc with the lift motor by inducing a voltage spike somewhere, most likely in the encoder cable. I noted that the motors sequenced correctly with water in the cells, which meant it most likely wasn't a software problem or it would have shown up during earlier sequencing.

The lead spent the whole day on the phone with the software people, who couldn't duplicate the problem on the lab prototype, and wouldn't allow the day shift crew to work on the machine until the software folks figured out the problem.

When the new night shift crew came on that night, they read the troubleshooting log. One of the engineers who really trusted my troubleshooting ability and judgment decided to look at the hardware, since there was nothing he could do about testing the software. He found a problem with the encoder cable. At the motor end, the pins for the power supply common wire and the cable shield had been installed reversed inside the connector. Every time the plating current shut off, the EMI from the collapsing magnetic field was inducing a voltage spike into the common of the encoder power supply via the shield drain wire pin. The encoder would freak out, the servomotor would lose track of its position, and the machine would fault. The pins were swapped, and everything worked just fine.

An interesting note: the lead assumed the problem was in his area of expertise and neglected to examine everything else, so he was only guilty of wearing blinders, not of trying to avoid responsibility. I'm much more used to the "It's not a [insert discipline here] problem, it's a [insert some other discipline here] problem" type of finger-pointing.

xnuke
"Do you think you used enough dynamite there, Butch?"
Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[jraef]

OMG sed2, an EPIPHANY!!!
I can't believe it. You may have just described what the probelm is in MY application! It never occurred to me that when they shut down the big pump, there might not be a check valve and the backwash is what is overwhelming the smaller pump! Maybe THAT's why the pump guy's calculations are so far off! I'm going to have to "volunteer" to check that out before the finger pointing session starts this week.

JRaef.com
Eng-Tips: Help for your job, not for your homework Read faq731-376

 

[ozmosis]

Jeff
I just read properly your original post. You did say skip it and I guess I did skim through it without taking in some of the detail, that's probably what prodded the memory of my site adventure a few years ago (you did say you wouldn't be offended if it's skipped! )
It does sound mighty similar. I think I'd check the backwash.

 

[cswilson]

Glad I'm not the only one to see these types of issues! Since everyone seems to be enjoying these, I'll toss in a couple of my war stories:

Under the category of "blame the PID", because we build complex positioning servo controllers used by a lot of people who don't have a good understanding of what's going on, we often get the blame first. Many years ago, a customer using our controller for a large positioning table complained that our PID could not stabilize his application. Phone support was fruitless (and we didn't have the diagnostic and communications tools we have now), so we had to send someone across the country to "fix it". Sent a guy who wasn't a degreed engineer, but with lots of experience and horse sense.

He gets to the facility, takes one look at the operation of the table, wipes down the slides with a rag and sprays them with WD-40. PID problem fixed -- in 5 minutes!

One of my favorites, though, came from a student user (which makes it a little more excusable, I think). He calls up one day, saying he just cannot get his (bare) motor to spin at all with our controller. So we walk him through step by step to break down the problem. First, we have him open the servo loop and just force an open loop 50% output on the DAC. The motor still does not move. We have him read back the number in the DAC register to confirm software operation. It reads back 16K, half of the 32K range, so we conclude that the software is operating properly, and he must have a hardware problem. Next, we have him pull out a voltmeter and read the output voltage. He gets 5V, half of the 10V range.

So we tell him that our controller is working fine in both software and hardware, so the next thing to check is his amplifier. To which his response is, "Amplifier???"

Curt Wilson
Delta Tau Data Systems

 

[itsmoked]

xnuke was that a lead(IC leads not Pb) plater?

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[xnuke]

It was a silicon wafer plater for forming interconnects on microprocessor dies.

xnuke
"Do you think you used enough dynamite there, Butch?"
Please see FAQ731-376 for tips on how to make the best use of Eng-Tips.

 

[itsmoked]

Thanks! Seemed a little too complex for lead plating, a generally linear process.

Keith Cress
Flamin Systems, Inc.-

http://www.flaminsystems.com

 

[rbalex]

"All problems within 20'(and further if over 600V) of a wire (any size or purpose) are electrical problems until proven otherwise."

I learned that well over 35 years ago

 

[controlnovice]

Working in a refinery, I received a call about a 100HP motor that kept on tripping during startup. I went to investigate with the electricians.

When we arrived, before we could investigate, the operator immediately reset the breaker and hit the start button. Sure enough, the breaker tripped.

Electricians locked it out, closed all piping valves, and meggered the motor. No problems.

"It has to be the motor", claims the operations supervisor.

Come to find out that this is the second motor, recently installed.

I asked them to reset the breaker and open the valves.

Once the valves opened, the motor started to run backwards. None of the valves, nor the start/stop station was in line of site of the motor shaft.

I had learned enough in the past to watch and listen, so no one else saw what was going on.

It was a pleasure to tell the operations supervisor on my way out with the electricians that the check valve had failed, causing the pump and motor to spin backwards while off.

______________________________________________________________________________
This is normally the space where people post something insightful.