Since we are gasping at straws, this is another idea: http://www.newark.com/3m/3659-26/round-to-flat-cable-26-conductor/dp/03F5946

Try doing your in-house test with a grounded piece of aluminum foil around the ribbon cable or change the length.

 

Since we are gasping at straws, this is another idea: http://www.newark.com/3m/3659-26/round-to-flat-cable-26-conductor/dp/03F5946

Try doing your in-house test with a grounded piece of aluminum foil around the ribbon cable or change the length.

Well, I'm one step ahead of you there. I bought some of that cable and built a cable for the pump. It had very little affect on it. We even tried grounding the shielding foil and again, no appreciable difference. Actually Aluminum foil was my very first try. We did seem to get a lot of help using ferrite beads on the ribbon cable. And, we were really gasping at straws there, because we really didn't know much about them. I purchased the ferrites firstly by the size that would fit on the cable. Everything was listed at 100 MHz while I knew I wanted them as close to the frequency that we were trying to shield. Another point where a professional, not just an EE, could have helped. I put like 10 of these little things on the cable and when we pushed them towards the PC bd connection we got the best results. I added two more I had left over at the Other end. None of us knew much about how to choose the ferrites and it shows. I start having problems with the 60601-1 (Safety) testing when I start using metal shielding inside the machine. I did shorten the cable to around 12" from 18", but an earlier poster said that may have made it worse. I was trying to shorten the antenna, but I don't know near enough about wave lengths to calculate anything like that. I don't have equipment to test anything like that. I appreciate the input.

 

Nobody should mistake my intentions. I have no wish for the TS to fail at his task, I'm just trying to prepare him for one of many possible outcomes. The decision to spend time and resources is probably above his pay grade. He needs to be honest with the decision makers about his prospects for success. He might get lucky, and I sincerely hope for that outcome, but he could also spend a great deal of time and money spinning his wheels, and I think all would agree that he needs to avoid the appearance of doing that. My advice is to put a time and/or dollar limit on your efforts. If you don't have a solution or the prospect of a solution, you own up to the fact and lay out the options for the decision makers. Any other course of action has enormous personal and professional risks in my humble opinion.

 

A stupid page, http://www.wavelengthcalculator.com/ but still. So lengths that are multiples of the wavelength divided by an integer are suspects. So about 25 to 30 inches.

The beads did something, apparently.

Can you "reverse engineer" what the ribbon connects too? At a bare minimum what numbers are on the IC's near the ribbon cable?

Is it easy to shield that small PCB? e.g. A plate with holes mounted with standoffs? Use edge molding around the rough edges to pass the cables through.

Do you have an oscilloscope of any type and, if so, what is the bandwidth?

To really track it down you might need something like http://www.aaronia.com/products/antennas/Near-Field-Probe-Set-PBS2/?gclid=CITx-pbIyc8CFVdahgod1AAKVA and a spectrum analyzer. The latter you can rent.

If you have a source of interference, a possible cheap way might be this http://www.ezprobe.com/specialty/TEXR750.html and a multimeter. The bandwidth of the meter doesn't really matter.

In any event, for the broken analyzer, is the fault permanent? If you short the power supply leads (no power) for like 5 minutes and try again did it work?

 

Nobody should mistake my intentions. I have no wish for the TS to fail at his task, I'm just trying to prepare him for one of many possible outcomes. The decision to spend time and resources is probably above his pay grade. He needs to be honest with the decision makers about his prospects for success. He might get lucky, and I sincerely hope for that outcome, but he could also spend a great deal of time and money spinning his wheels, and I think all would agree that he needs to avoid the appearance of doing that. My advice is to put a time and/or dollar limit on your efforts. If you don't have a solution or the prospect of a solution, you own up to the fact and lay out the options for the decision makers. Any other course of action has enormous personal and professional risks in my humble opinion.

Papa, I took your remark that way. You were just stating what we all here know. I'm in a small family business, my older brother is the CEO and we do talk honestly and realistically. What you said is good advice and while I overrate my abilities a lot, good, sound candor is needed and appreciated. I was just having some fun with Tranzz. I understand people have different views. The guy that designed the new boards for us is Chinese. He's a very dear friend and an elder at the place where I fellowship, but he thinks if different ways as well. Cost and quantity! He doesn't understand why we don't get 1,000 boards because the price is so much better. I'd be dead and in my grave before we could sell 1,000 of these and I'm not sure we could afford to hold that many until we used them all. What we underestimated was the extent the FDA had changed in the last few years. 10 years ago, this would have been a piece of cake. We spent so much money on that Medical Device tax, I could have hired a full time EE. Live and learn. I'm just glad there are people out there that will not only talk to people like me, but will be honest. Honesty is highly Under rated these days! Thanks, Joe

 

Papa, I took your remark that way. You were just stating what we all here know. I'm in a small family business, my older brother is the CEO and we do talk honestly and realistically. What you said is good advice and while I overrate my abilities a lot, good, sound candor is needed and appreciated. I was just having some fun with Tranzz. I understand people have different views. The guy that designed the new boards for us is Chinese. He's a very dear friend and an elder at the place where I fellowship, but he thinks if different ways as well. Cost and quantity! He doesn't understand why we don't get 1,000 boards because the price is so much better. I'd be dead and in my grave before we could sell 1,000 of these and I'm not sure we could afford to hold that many until we used them all. What we underestimated was the extent the FDA had changed in the last few years. 10 years ago, this would have been a piece of cake. We spent so much money on that Medical Device tax, I could have hired a full time EE. Live and learn. I'm just glad there are people out there that will not only talk to people like me, but will be honest. Honesty is highly Under rated these days! Thanks, Joe

I learned my craft in the technology heydays of the late 1960's and 1970's. The biggest sin was not missing a schedule or a commitment -- it was not telling management the nanosecond after realizing the problem. It was easy to get forgiveness and another chance as long as you gave them a reason to trust you. Keeping them informed built mountains of trust. Delivering on commitments made me a mountain range.

 

A stupid page, http://www.wavelengthcalculator.com/ but still. So lengths that are multiples of the wavelength divided by an integer are suspects. So about 25 to 30 inches.

The beads did something, apparently.

Can you "reverse engineer" what the ribbon connects too? At a bare minimum what numbers are on the IC's near the ribbon cable?

Is it easy to shield that small PCB? e.g. A plate with holes mounted with standoffs? Use edge molding around the rough edges to pass the cables through.

Do you have an oscilloscope of any type and, if so, what is the bandwidth?

To really track it down you might need something like http://www.aaronia.com/products/antennas/Near-Field-Probe-Set-PBS2/?gclid=CITx-pbIyc8CFVdahgod1AAKVA and a spectrum analyzer. The latter you can rent.

If you have a source of interference, a possible cheap way might be this http://www.ezprobe.com/specialty/TEXR750.html and a multimeter. The bandwidth of the meter doesn't really matter.

In any event, for the broken analyzer, is the fault permanent? If you short the power supply leads (no power) for like 5 minutes and try again did it work?

You are talking way over my head now. The two boards that the ribbon cable plugs into are the two bds we had made for us. I'm sure we can get the schematics on those, but I wouldn't understand them. We had this same problem before we modified the machine. It failed at these same two frequencies before. It just didn't die like it did this time, and these modifications had passed through these same tests before without dying. We were going through a retest for CE compliance, which is the exact same tests but the machine is plugged into 230 VAC instead. It has a switching PS, like 99% of devices out there now. But it was connected at 110 VAC at the time it died. It did fine at the Higher frequency tests as well. It was just at 385 MHz that we had issues. I don't have any equipment at my disposal. If I knew how to use some of it, I'd get some off eBay and have a small shop. I appreciate the advice.

 

I learned my craft in the technology heydays of the late 1960's and 1970's. The biggest sin was not missing a schedule or a commitment -- it was not telling management the nanosecond after realizing the problem. It was easy to get forgiveness and another chance as long as you gave them a reason to trust you. Keeping them informed built mountains of trust. Delivering on commitments made me a mountain range.

Papa you are a wise man!

 

You are between a rock and a hard place. I design medical devices for a living and have spent a lot of time with the IEC 60601 standards. I have come to think of EMC problems as 'a leaky dam' type issue, you have cork up all of the holes to stop the leaks. From a troubleshooting perspective, you need to hypothesize the coupling mode, make a design changes to address and repeat the process. And just like the leaky dam, do not remove any modifications (corks) until the issue is resolved. Only after the device is working can you remove modifications to determine what was actually required.

 

You are between a rock and a hard place. I design medical devices for a living and have spent a lot of time with the IEC 60601 standards. I have come to think of EMC problems as 'a leaky dam' type issue, you have cork up all of the holes to stop the leaks. From a troubleshooting perspective, you need to hypothesize the coupling mode, make a design changes to address and repeat the process. And just like the leaky dam, do not remove any modifications (corks) until the issue is resolved. Only after the device is working can you remove modifications to determine what was actually required.

I'm sure that was good advice, but too much for me to fathom. If I can get past this odd issue of the machine dying, I'm not going to worry about the 60601-1-2 Table 9 issues, as they are not relevant to this machine and I'm told we can get past FDA with a conditional statement. It actually, passed the 3rd ed. It died when we sent it back to do the CE testing. If we had an inkling of the problem the FDA had become in recent years, we probably would not have started this or at least, hired someone like yourself to help us through. The FDA has gotten so bad lately we had to hire a 3rd party reviewer to help us. They won't even talk to you anymore there.. Maybe we talk again, some time. Thanks

 

OK, then assume it's not those parts. Do you photos of the Chinese portion?

I wonder if the power supply passes? What model do you use?

What can you tell us about the power entry module, if there is one. Have you tried a shielded power cord?
Here's a power entry module http://www.alliedelec.com/schurter-5220-1023-3/70080809/ thats medical approved.
Medical modules have lower leakages.

 

Susceptibility failures like this are usually due to two things. The first is poor bypassing of signal lines that exit or enter the enclosure and the second is a ground system that is too high impedance to accomplish its intended goal.

This can't be solved without a full set of assembly diagrams, schematics and PCB layouts.

 

Hello,

You could also try to replace the ribbon cable with "shielded ribbon cable"
http://www.newark.com/shielded-ribbon-cable-flat-cable/prl/results

Bertus

 

Hello,

You could also try to replace the ribbon cable with "shielded ribbon cable"
http://www.newark.com/shielded-ribbon-cable-flat-cable/prl/results

Bertus

Thanks, Bertus, but that was my first move and it had very little affect on it. Grounding the cable or no, it didn't seem to matter. There was just a very slight reduction in the interference. Actually the Conductive Spay paint on the inside of my Plastic case housing was my first thought, but that turned into a UL nightmare. We aren't trying for a UL listing, but the FDA makes up pass teh 60601-1 tests anyway. The FDA can't grasp teh concept that an Infiltration pump is no where near the same as an Infusion pump. Anyway, they are making a mountain out of a hole hill at my expense, but ultimately the users of health care in this country get to pay for all of that stuff. Thanks for the input. Joe

 

Can you do a board swap and find out which part died?

Like I said, there is a possibility nothing died permanently.

 

Can you do a board swap and find out which part died?

Like I said, there is a possibility nothing died permanently.

Well, I opened it up and everything looked OK. The motor wouldn't run, but everything else worked. I checked voltage and the Power supply was outputting 24 VDC as it should.. I tested the DC leads to the Main PC bd where the motor gets its power from, and there was No power on the pins where the motor plugged in. It appears something on the main PC bd did die. I've got a friend coming over who is an EE to look at it. He designed the LED and stacker boards for me. All of that seems to work. It appears that something on that main PC bd has died. We'll test and get back to everyone. Thanks, Joe

 

Can you post a pic of the main board when you get a chance?

If possible with the leads connected.

 

Can you post a pic of the main board when you get a chance?

If possible with the leads connected.

There are two more. I hope this helps. Thx, Joe

 

Sorry, two more pictures.

 

Any numbers on U1, U4 and U10.
Any numbers on the part above C3. 7805?

What I'd like you do do is with the unit unplugged and/or the 24 V PS disconnected short the Vcc and GND pins on J5. You can short the 24 V PS connector as well (At the PCB connector) if you want too.

Make sure it does not get plugged in until the shorts are removed. If it doesn't fix it, there will be no harm done.

Measure between Vcc and Gnd after the short is removed and the system is powered up. I suspect 5V or 3.3 V.