Corrosion Pattern on Press-Fit SCR Pole Faces

Looks like electrical polarization of the grain surface....

They are power SCRs, not Triacs. I had used them in the 70s for electronic power and industrial applications such as battery forming rectifiers, electroplating machines, UPS, etc. and other high amperage devices. The device are normally are sandwiched, both sides pressed to a certain torque required to activate. The pittings that I had observed were due to the unevenness between the aluminum heat sinks mounting plates and the device's sides which correspond to the voids filled by the conductive thermal compound used. These voids may incur some arcing when the thermal compound dries up and during high energy conduction.

Re dissimilar metals: dissimilar + humidity + large DC current --> electrolytic corrosion?

I must say that the original description is one of the best that I have seen on CR4, no guessing games!

I am wondering if some sort of conductive paste needs to be between the two faces to take up the "slack" so to say......maybe one of those with silver in them? Something like this maybe?:-

Silver Paste

Has the OP tried something like that?

Also, is he certain that the ones he is at present using have been correctly mounted/clamped/tightened? As even temperature changes could promote looseness....which could maybe lead to the problem he is seeing....

Is the OP certain that he has the original maker quality and not some cheap Chinese copy?

Dear colleagues,

I am extremely well impressed for the quality of your answers, which reflect a highly professional attitude from all of you. I feel proud to belong to this group.

Trying to resume all of your questions, let me try to make a list:

1- Solar Eagle, thanks for your photos. They really show a very similar pattern, if only on a smaller scale. (I read the attached article, up to where I could)

2- Europium, what if I tell you this very week I got a microscope from a friend that had one idling in his home and thought I could make a better use of it? Yes there are lots of crystals and some(few) whiskers. If I am correct, material has transferred from the SCR copper face to the aluminum heat sink, and the traces on the SCr face are actual channels. Will share the photos as soon as I figure out how to get an adequate camera (not so easy here, but this goes outside the scope of the thread). I am open to any questions on personal mail.

3- The SCR installation method is as professional as possible. SCRs are installed under hydraulic pressure and mounting force is calculated from the oil pressure, and set per the SCR spec sheet. For these ones, mounting force is 5500 lb-f or 24.5KN. The system is installed and the mounting springs compressed to the proper force, then the bolts are hand adjusted plus ¼ turn.

4- Mounting surfaces on SCRs are not touched. On the heatsinks, mounting surface is first machined , then stone rectified, then honed to an almost mirror finish. Water-based (not silicone) thermal compound is then spread on both mounting surfaces, then removed with a clean cloth, in order to leave compound only where necessary.

5- After mounting, SCRs are measured for Vdrm, Vrrm, Vgt and Igt, then a report is written and the mounted pieces delivered to the customer.

6- I consider that after all this procedure there should not be any gap between mating surfaces. However I am trying to import a pressure-reading tape to check the surface fitting

7- Corrosion, or whatever it may be, takes its time to develop. Yesterday I opened two assemblies that were on duty for less than a year. No worm at all.

8- The diode in the photo was so stuck to the heatsink it was necessary to use a rubber mallet to set it free.

9- Vsar, you are right. They are large SCRs, not triacs. Your ref to the 70s is right as well. This is about the time I begun to work with power SCRs, and here in Venezuela most industrial equipment is quite old. I am familiar with vacuum tubes, mercury-arc rectifiers, ignitrons, magamps, all the way to these large SCRs, very popular in paper mills and oil drilling rigs.

10- Andy, I have not tried but have always asked everyone and his dog for a conductive paste, and got no answer until now. Let me get a sample, try it and will come back to you! Vielen Dank!

11- Yes I am sure pieces and spares are originals. In this country, I can tell you long stories of not-so-cheap Chinese copies of even complete DC drives, but I risk going off-topic.

Great post, thanks.

Have you been to a shop where PCs are built? They often use a quality of thermal paste loaded with silver between the main chip and the cooling fins, more to remove heat, but also allowing electrical contact as well...

There are several types/makes of paste available, some with silver, some not.....

Like here:-

silver-filled-grease

Maybe that will help you further?

What is the conductivity and the current-carrying capacity of the grease? My concern is that it may not be adequate in the OP's case where the contact area carries thousands of amps.

Actually, if both faces are flat, its almost unimportant. Think about it!! it will only "fill-in" where gaps are and as silver is a really good conductor, will actually improve (reduce) contact resistance, probably only in the micro Ohms range.

eg. Difficult to quantify with normal meters....but a bad contact will get hot quickly under load!!

If you had opened the link I posted and read it fully, you could have probably answered your own question also!

Here is the second paragraph, can you pick out the important attributes?

High electrical conductivity
Excellent thermal conductivity
Provides protection against wear
Remains stable in a wide temperature range; -70 to 485°F (-57 to 252°C)
Protects against moisture and corrosion
Very low viscosity vs. temperature change

Now large temperature changes under load indicate a bad connection with resistance, which would need to be corrected, but probably due to some mechanical error, the problem would not be due to this paste.

The paste would most probably improve the quality of the connection, not make it worse....

Now that wasn't difficult was it?

"Now that wasn't difficult, was it?"

You have no idea.

I see several tell-tales in your fine description of the situation and process.

In item 3 you mention copper SCR terminals and aluminum heat sinks. It is generally known in HV equipment not to mate those metals due to the galvanic reaction that can occur (which you are seeing). The usual remedy is an oxide inhibiting compound such as Burndy's Penetrox A-13, which is a synthetic base with suspended zinc particles, and is rated for all voltages:

(http://ecat.burndy.com/Comergent/en/US/adirect/burndy?cmd=catProductDetail&showAddButton=true&productID=PENA134&_bcs_=-1%08%23%23%08%23%23%08http%3A%2F%2Fecat.burndy.com%3A80%2FComergent%2Fen%2FUS%2Fadirect%2Fburndy%3Fcmd%3DcatLanding%08%23%23%08true%08%070%08602382%2F602919%2F602945*%08Oxide+Inhibitor%08http%3A%2F%2Fecat.burndy.com%3A80%2FComergent%2Fen%2FUS%2Fadirect%2Fburndy%3Fcmd%3DcatDisplayStyle%26DispStyle%3DSingleLevelCategory%26catKey%3D602945%08%23%23%08false%08)

and possibly additionally a bimetallic plate such as Sefcor's ATP: (http://www.sefcor.com/products/view/tees-couplers-and-misc/aluminium-misc/ATP)

These are commonly used in MV and HV applications where aluminum and copper/bronze need to be bolted together.

In item 4 you say that the thermal compound you use is water-based. That is the other half of the equation. If it were silicone or other synthetic-based rather than aqueous, you may avoid the corrosion with the current materials.

(Sorry for the long links! For some reason CR4 is not letting me edit the hyperlinks today to give them a shorter title.)

Hi rilgran,

Why were these diodes and SCRs pulled from the equipment? Apart from the corrosion were they defective?

Hi Europium and the whole group,

Yes they were removed because they were defective. When dismounted, the corrosion showed up. I have at least 10 more SCRs with the same type defect. The one shown was the only diode.

Will respond at length in a couple of days.

Best regards to everybody

Just an add on during your repair process..

Be sure each SCR device has its own properly matching protective di/dt circuitry.

Heartfelt thanks indeed, to all the participants in this thread. I am still working on the issues raised, and appreciated very much all your answers. I have to proceed slowly, since the present working conditions on my country force me to do so. However, I said slowly, not stopping, and i am making subtle changes to the way I install the SCRs. Then I have to wait, and see what happens when they come back. I got a microscope and was able to confirm that copper goes out from the poleface (hence the channels) and gets firmly embedded in the aluminum heatsink. I have made some changes, and now I am waiting for new results.

Thanks very much to all of the participants. You have teached me something new. I appreciate and respect that, and feel extremely thankful.

Best regards to all of you,

Ricardo.