Indramat Drive Capacitor replacement

I have another thread in the DMG forum section about my giant Deckel Maho MC800H restoration ("Contest for the more insane member") but I need some feedback in regards to leakage current and ESR ratings on capacitors and how it may affect the circuit performance if they are outside a certain range of the originals.

I've spent a solid week dismantling and recording the values, types and sizes of every single capacitor on the entire machine. Every... Single... One.

Philips (now Vishay - BC Components) offers these little caps that are all over the Indramat products:



These are Philips 036 Low leakage caps. Almost no capacitor that is stocked with any North American distributors has leakage as low as these units do (.006CV+3µA after 1 min, .001CV+3µA after 5 min). The Nichicon UKL units are very low (.002CV after 1 or 2 min depending on case size), have better ripple current ratings and better life ratings but I'm not an EE and I want to replace all the capacitors with units that are a close a possible to hopefully prevent complications. When I have looked into what others say on capacitor upgrading I keep hearing that increasing the voltage rating is no problem but I'm finding you get an instant increase of your leakage current unless the original wasn't a particularly low leakage unit to begin with and the new ones have less leakage factor. Also I fear if the voltage rating is too much higher than the operating voltage the forming of the cap will not stay as "complete" and may have a negative impact on long term performance. I would love to have all the circuits drawn out and learn how to calculate everything I need but I just don't have time to go to such extents. I have around 144 caps to replace.

This question is for the pros that rebuild drives on at least a semi-regular basis: The Nichicon UKL caps I mentioned before have really low impedance as compared to the originals. If a cap has really low impedance I don't know if this can cause higher inrush current that could damage the circuit, or what values could possibly negatively impact any circuits that are tuned somehow? How critical is low leakage in computer processing type circuits like the controlling side of the drives? I assume leakage will directly affect what voltage the cap will stay at in operation? Does it matter all that much with Electrolytics having such wide tolerances that change with age and temp? I know ESR can go up with age and cause damage but I'm not sure how low I can go without causing problems as well.

I'm going to replace the DC bus caps with the largest capacitance I can fit in the same spot with caps that have higher ripple ratings (sometimes much higher) and lower ESR. I don't know if I go too low on the ESR if it could cause a resonant voltage problem or something in an inverter DC BUS application?

I saw a photo of a Indramat board that appeared to have Nippon Chemi-con caps in the place of the 036 Philips units, but I can't make out what series they are. In my experience Fanucs are quite reliable and they use Chemi-con caps so I like to use them when I can.

Any thoughts on this would be quite appreciated! This is a ton of work but when finished I want all the electrical to be problem free for a couple decades if possible.

I am particularly impressed with the Nippon Chemicon (United Chemi-con) KZN specifications. The seemed to be filled with magic and still have a wide range and great stock locally. The ripple current capability and life expectancy is a whole order of magnitude better than most others I've been going through.

I think you will find that many components, even on premium brand boards, were selected for low cost and high availability rather than absolute specification. I wouldn't worry too much about those little bypass caps, but that's just me. If the circuit is working correctly, they should be doing nothing anyway.

When I rebuild old electronics, I just find a cap with the same farad rating and close to the same voltage. I buy low ESR caps, but most modern caps are very low in ESR compared to the caps that are currently working just fine in the drives.

Some things, like switch mode power supplies, need really good caps to reduce output voltage ripple. As long as that power is nice and clean, the down stream caps have much less work to do.

Just my thoughts.

Thank you, I appreciate feedback based on experience!

This has been a very educational week - I'm going to run with the assumption that if the cap has lower leakage, higher ripple current capability across the entire frequency range and lower ESR than the replacement cap it will work at least as well if not better than the one it is replacing when it was in new condition and much better than it is now that it is 22 years old. I have been learning the appropriate formulas and wonder if the frequency where the cap becomes inductive is critical for the smaller caps or are there elements already into the circuit to reduce the range of frequency the caps are going to see? Can ripple also be getting created on the output side of the control boards as current is being sent to the IGBT stages?

I can check the ripple output of the main power supply after I recap it with my scope, it would be nice to see what the original specification allowed as a max but with the highest quality caps I can find that will fit that application I can only hope it will be lower than it was, and then I hope your statement about the rest having to do less work will apply. I know the electronics were all working fine when this was last powered off years ago, the problems with the machine were elsewhere as I have found them with my extensive dismantling and reassembly of the entire machine ( http://www.practicalmachinist.com/v...ember-they-will-maho-owner-321712/index3.html )

What is your favorite cap for SMPS? A preference on the input vs output stages? Do you have a favorite? If cost was no object, would ceramic or film caps be worth considering?

Also, I really need to find some suitable replacements for the larger 4 pin Siemens series B43306 caps that are used in my KDV 3.1 Drive power supply. Of course they aren't made anymore by Epcos/TDK who bought out that division of Siemens years ago and I can't find anyone that supports that pin pattern. I may have to drill new holes and make new pads for the caps? I don't think that board is a multilayer board...

Thanks again for your feedback. More feedback is also very welcome!!

NO leakage will affect its capacity. But first what makes you think its faulty? What do you think your going to achieve by replacing it? If its working you would do well to leave it alone unless any given capacitors are showing signs of bulging or leakage, thats bonding compound on the board bellow that cap not leakage to give it a mechanical support.

Voltage rating wise, even doubling it will have no effect on the formation - maintenance of the layers that stops a typical capacitor from short circuiting, the thickness of those layers is directly governed by the voltage the capacitor regularly sees, weather it be a capacitor rated for say 12V or 200V if its been ran for more than 3-4 years at say 5V, both would be equally likely to go pop if you suddenly gave them 10 volts! A given voltage builds - maintains a given layer thickness.

As to DC Bus voltage capacitors, IMHO DO NOT INCREASE THERE CAPACITANCE!!!!!!! If you fit lower ESR and - or higher capacitance you greatly increase start up and fault surge current potentials well above what what the control components may be able to handle in the drive. Drives and any capacitance they have - use is important in there design, you would do really well not to mess with that design unless you know a lot more that your post indicates. By all means fit higher voltage ones if you can, but theres no need to go stupid either. The next voltage range up is plenty on any general capacitor upgrade.

IMHO your badly pissing up a rope here, the IC's on thoes boards have tin plated legs, that tin is oxidized over time, that oxide is porous and humidity will be making its way up those legs into the chip. IC's have a finite life time in a typical breathable atmosphere. Based on the age of that board theres only so much of that left and IMHO 2 decades is probably a lot longer than that life span!

Your border line at the point were your trying to resuscitate a 95 year old, yes you may succed in not haveing them die from a heart attack but it then just falls to the next weakest component and on thoes boards theres going to be plenty weaker components, which are now seeing a higher voltage, with larger current inrushes due to the reduced ESR.

Ultra low leakage times - rates really should not be a issue on that kinda drive, think about it, capacitors frequently vary in capacitance by 10's of percent, losing 0.006 or 0.1V after a couple of minutes should be immaterial to a permanently powered drive. Your looking at numbers you don’t understand the meaning of and in a device you don't seam to even grasp how it works let alone why its designed the way it is. Depending on the drives control circuitry design changing things may well have some unintended results!!!

Interesting post adama.

What I agree with in that interesting post:

That IS bonding compound on the bottom of the caps... I have been dealing with that since about 30 years ago, when I started working with electronics.

99+% of the time I have only replaced the electrolytic capacitors on a drive or power supply. Just ordered the same or higher voltage rating and the same capacitance rating and they worked fine. This is on Fanuc, Yasnac, and Misubishi drives and power supplies, not Indramat.

I only recall once replacing several ceramic caps on a Yaskawa motherboard. There was a power supply problem that resulted in the blown up caps on the motherboard.

I think you have too much time on your hands if you are worrying about minuscule difference in capacitor specs.
On old boards I solder only what is truly necessary and not just go changing components for the fun of it.

Old fragile board and it's pretty easy to loose a trace from the board, the caps aren't the only thing that is getting old.

If it's not broken why fix it?

Marko

Thank you very much for all your replies.

"If it isn't broke why fix it?"

I have other machining centers that are currently being used for production. I have had them go down because of capacitor related issues in the past. This machine I'm replacing the caps on is completely dismantled, it is a extremely rare machine which has great value to me in what it will be able to do once finished. It is apart now and I don't have work that depends on the machine functioning... yet. This is a large 5 axis trunnion HMC with a pallet changer, I don't have another one, there were only 3 brought to North America and the other two may not even be the same spec with the trunnion. I heard NASA had one maybe someone can confirm? I'm told it was worth around a million dollars in 1997, in 1997 dollars.

I cannot afford to have production go down while I rip out drives only to find that I damaged something from a capacitor failing. This has happened in the past to members on this forum who have remarked "I would replace them all ahead of time just for peace of mind". They are cheap. Diagnosis can be hard once damage has started.

It is much cheaper and easier now for everything to be changed than to mess around once it is back together and running. I have done many restorations of many types of machines this way, from Automotive to CNC and it always fixes all the problems just by the extent of the preventative health care up front.

A great example is finding a nice classic car under a tarp that has been in a shed for 20 years. Sure it ran fine when it was put there, its not broke why fix it right? Well it might be broken now, mice could have eaten wires, cylinders may be rusted... if you resist the urge to try to start it but instead rip it all apart and fix whatever you find it will cost far less in the long run than if you are foolish and try to start it. You could have a fire, toast a piston, hit a sticking valve (had to completely rebuild an engine for a guy that did this very thing by starting an old bike that sat too long). I need to be clear: The machine is NOT running right now in my shop, it hasn't seen power in a few years but when it did have power last it didn't have errors with the drives at that time. Since then I acquired it and I'm restoring it very carefully while not causing any harm to anything in the process. I put a link to the PM thread earlier, you will see what I mean. Back on topic...

Electrolytic caps aren't a fixed/broken component like almost all other components on a electronic board. They are a ticking time bomb that MUST be changed before anything else will need replacing, typically. They are 99% most likely to fail before anything else and that is very highly likely to happen in the first few seconds this machine sees power again. I can try reforming the caps but typically this doesn't last for long when the caps are old and changing them is an obvious choice. Did I mention it is cheap?

Thinking I have too much time on my hands for making damn sure I don't destroy 6 Indramat drives with it's power supply and a insanely expensive RAC3.1 Spindle drive on power up? You must have far more money than I do... it would cost a LOT of money to fix that damage. Many tens of thousands of dollars.

Luckily, there is one warehouse in the world that does have the really special caps that are on these boards and I'm getting them brought in. The rest of the caps on the machine are nowhere near as unusual in the specifications of them as those caps.

"An ounce of prevention is worth a pound of cure"

All I can say is good luck, and hope you have a good rework station and capacitance meter to test the new caps, and all the anti static gear to complete the job without accidentally killing something from any of the drives you are working on.

It's always a calculated risk changing parts in a working configuration, needless to say verify all the new caps before soldering the on the board. Not all old caps are bad nor are all the new ones good.

Marko

I do have a good rework station, I take anti-static very seriously as it doesn't take much to damage IC's that are intended for 5V TTL signals. I wore a wrist strap tethered to my work station as well as the frame of each drive which is then tethered to the main ground line by my panel. This was just to take the values off the caps so far.

I'm planning to verify the caps prior to install with a quality Fluke ESR meter, just to put them through an initial test in case they decide to fail or not meet specifications. Is there a "dud test" you like to perform beyond this?

I consider this "working configuration" to be more of a Schrodinger's cat in that until I put power to the boards they are both functional and broken at the same time.

99.99 percent of cases the electrolytic capacitor leakage current is irrelevant.

Doubling voltage rating is not a problem, lower than original ESR is also not a problem in 99.99% cases. Just dont select some super-duper low esr caps, they seem to often have less stable electrolyte. Worst case is that you replace old caps that would work fine for another 20 years with chemically unstable junk that expires in 5 years.

Nichico upw is my normal cap choice. According to some EE folklore the -55 temp rating usually means a more stable inorganic electrolyte.
UPW Series - Nichicon - General Purpose | Online Catalog | DigiKey Electronics

V=IR Its the basic rule of electronics, if resistance is lower and voltage is the same you get more current. P=Isquared R hence that slight increase in current puts a large increase into thermal stress in the part's controlling that power flow. You need to pay some serious respect if the drives are as hard to replace as you say to what the capacitors are that are in them were, bigger capacitance and lower resistance in that capacitor on the power side will bump peak current flows, on old drive components that could well push things over the edge. Equally increaseing any capacitance on the output of the drive can also have negative effects to ripple currents and other things. Capacitors in a drive enviroment are doing a lot more than just holding a voltage up, this is especially relevant when your dealing with inductors as well like the motor your drive is controlling is.

You can like, dislike or agree or not, but electronics is the one thing i am qualified in and trust me, your flirting with problems here with the notion more is better when it comes to capacitance, because it really may not be and it really may be the bit that kills things. Same goes for lower ESR, on high power drive electronics like these it needs calculateing not guessing, because lower is not the safe way to go with ESR!!! Higher ESR may just reduce available delivered power, but won't cause the harm Low ESR has the potential too. ESR stands for electronic series resistance, its litrally in effect how much resistance that capacitor shows to slow the rate of charge and discharge, chose caps with half the ESR and you have literally just doubled your inrush surge currents on switch on!!! On a control board that won't matter much, on the power handeling side on a 50hp drive thats a major problem. Please don't ignore this, look into it don't just take my word for it, but really understand how it affects your plans.

If you want true long term reliability rather than try and fix dinosaurs it might be easier to look for suitable off the shelf available replacements.

As far as power supply ripple, the specs I have seen are usually less than 30mV under load, especially for older gear. I've seen guys try to use cheap PC power supplies in old controls and it never seems to work very well. A typical ATX will struggle to produce less than 150mV ripple.

adama said:

......If you want true long term reliability rather than try and fix dinosaurs it might be easier to look for suitable off the shelf available replacements.

Click to expand...

+1 on this!

In the late 80s to mid 90s I worked for the MAHO distributor in my area. Indramat drives were referred to as laundramat drives by all of us in the service department as we felt they were better suited to drive a washing machine than a CNC.

They haven't gotten any better. They wear the "Rexroth" name now. I went to work on a trio of gear hobbing machines once during a factory shut down. Fired each machine up and you could hear the breakers trip instantly. Among 3 machines, I found 4 faulty drives. The machines were just sitting with no power connected.

Im not saying moderns better or more reliable long term, but if its a std interface you at least can buy one off the shelf shipped straight to site and be up and running in well under 24 hours.

adama said:

Im not saying moderns better or more reliable long term, but if its a std interface you at least can buy one off the shelf shipped straight to site and be up and running in well under 24 hours.

Click to expand...

You're thinking like this guy is running a 200 man production facility and that a 30 year old Maho is some kind of keystone machine that could bring the whole place to it's knees.

More likely, the OP is a home shop guy who is happy to spend a month of Sundays with a solder pump and a Digikey catalog hoping to head off those tiny voltage spikes.

I just replaced a 25 year old 15hp Allen Bradley VFD for a customer. The control transformer blew in the old one. New transformer was about $75 from several sources. New Power Flex 755 drive from Allen Bradley was North of $4000 plus a day and a half for me to get it to play nice with the existing PLC. The customer didn't even flinch. Shipped in the new drive UPS red and tossed the old one in the scrap bin.

They told me they had $1/2 million in work waiting to be finished up on that machine. I got the feeling that did that kind of $ about every week...

Thank you everyone for taking the time to share your experiences on this topic.

I'm confident the time I'm spending properly specifying the replacement components of these drives will greatly increase their reliability. There are a few caps in those drives that have really low life ratings.

I wasn't planning on changing capacitance anywhere but in the DC bus itself. Fanuc and Indramat both offer add-on capacitor packs that you can add to the bus so during regenerative braking the bus voltage doesn't go too high. Indramat simply states to "add capacitance" if you are suffering from this problem:

"In servo drive applications in which a high number of acceleration and deceleration operations are common, for example in nibble machines and roll feed units, auxiliary capacitance modules can be connected to the DC bus. This prevents the bleeder resistor from being activated in the power supply module when the drives are braked, which significantly reduces the heat losses generated in the enclosure. The stored energy can be used for acceleration, reducing the energy consumed by the system."

They are even kind enough to share the formulas to calculate the minimum you should add given your specific circumstances. I am tempted to move that bleeding resistor outside the enclosure too, keeping temps low is paramount to long electrolytic capacitor life.

On the topic of raising the voltage rating of the cap if you can't find the capacitance you need at the voltage the cap was at before: Capacitors have decreased in size over the years and my application is not restricted in terms of volume. If I raise the voltage rating of the cap it will likely be much closer in size to the original, and have lower ESR as well as be more robust.

My selection priorities have boiled down to (In order of importance)

Availability
Capacitance
Ripple current rating
ESR/Tan delta/Lead spacing
Voltage
Leakage (Unless it is the special extremely low leakage Philips situation)
Life
Diameter
Length

I do care about the manufacturer but honestly I have a short list of who I will use. I don't even bother looking outside this list. Japanese unless it is the large bus caps. TDK/Epcos still makes replacements for some of my larger caps and the stud mounting arrangement that the old caps have is still easy to get.

Manufacturers have more options for caps than we do. They can buy them in boxes of 10,000 or more so they can get whatever is still being manufactured no matter who is stocking it. If they decide to spec regular grade caps all over the place and then ultra extreme low leakage caps in a few select locations I feel very confident they didn't do this without a good reason. Without going to the extent of drawing up the circuits and starting to figure out exactly why this is the case I think this is the best compromise. I have over 16 boards I'm working on... I don't have that kind of time.

I came on this forum to ask for advice from drive specialists that might know why these specific capacitors would have been selected for this application, not because I am completely ignorant to anything regarding electronics.

Time will tell if this was the best approach. The other option I would consider would not be to find replacement Indramat drives, rather I would just replace the entire control and drives as a package with the servos. This would be so expensive and time consuming to implement that at that point I would be better off finding a new machine. This isn't about me having nothing better to do, it is a compromise of all the options I have.

Hey all, just wanted to say that although I understood only the basics of most of this discussion, it was very interesting and touched on some things that I had been thinking about.

Amazing the depth of knowledge on this site.

NateA2- I'm with you- I'm a happy camper if I can locate a blown fuse and replace it!
Dan