Replacing (or eliminating) antique MOVs?

Stud-mount diodes (four in a bridge configuration) are protected by MOVs: one across the AC connections, one across the DC connections.
These MOVs are antique-looking finned affairs (a la selenium rectifiers). Both sport PN S01AAA?AAA (?=illegible) with Westinghouse trademark. The wiring diagram symbol for these looks like a zener diode with an extra arrow head (both arrow points touch, with the familiar "z" shaped line between them).
The AC supply voltage to the bridge is 75 rms. The bridge diodes are originally 100V/1A devices; I'm replacing with 400V/35A ones.
This rectifier drives a DC motor via 2 contactors: one for direction (by reversing polarity) and the other for "GO/STOP". The MOVs are to limit the kick from opening contacts, I presume.
Someone suggested that with such stout diodes available today that maybe the MOVs are not required.
If you think I need to keep the MOVs, how do I select new ones?
What say y'all?
Thanks,
--
DaveC
snipped-for-privacy@bogusdomain.net
  Click to see the full signature.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
DaveC wrote:

You need the MOV's. The coils in the DC motors are wonderful inductors and can put out all kinds of HV spikes. If you have measured the input voltage correctly than the bridge diodes will have 106 volts applied across them even without considering mains voltage fluctuations etc. This makes selecting a suitable MOV impossible (even for the manufacturer)
MOV's are specified with an AC operating voltage which, in your case, should be a little higher than 75 volts and a clamping voltage which should be much lower than the PIV of the diodes (400 volts in your new case). You also need to calculate the amount of stored power in joules - with a moving mass of the motor this could be very high.
I suspect that something is not right with the circuit as you have described it.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
David Eather wrote:

MOVs would not be used to absorb spikes from the motor, if they were, they wouldn't last long.
You should look at Transient Voltage Supressors, aka TransZorbs. Vishay has a nice selection of them
http://www.vishay.com/diodes/protection-tvs-esd/trans-zorb /
So do a number of other manufactures.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

I'll bet that they are not MOVs. They are what is referred to as "Tranzorbs".
MOVs would only be able to handle a certain LOW number of transient events before they undergo operating characteristic changes. A Tranzorb, on the other hand, will ALWAYS kill a transient, unless you breach the tranzorb itself. With an MOV, each event changes the device's characteristics of operation until it finally functions as a completely different circuit element.
Try NOT to follow the lead of other idiots, and refrain from cross-posting your queries.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

To apply MOVs (and other clamp devices) there needs to be a decent difference between the semiconductor voltage rating and the actual peak circuit operating voltage. In this case the circuit operates at 75 Vrms or 106 V peak plus a margin for supply variation, say 117 V max, and the devices were rated at 100 V. There isn't a MOV voltage that would protect the semis and not destroy itself in normal operation.
But luckily it survived (maybe the semis could actually take a lot more than 100 V?), and you're using 400 V diodes, so you can safely choose a MOV that starts to break down around 150-250 Vdc or 100-150 Vac. And since MOVs are so cheap now you can just go with something rated for, say, 100-150 Joules, which should be more than adequate.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

Check the spec on MOV's Mycelium is correct, after a number of uses they are considered GONE.
Hard to believe anyone would design a circuit 'requiring' the presence of MOV's to protect during 'normal' operation catching flyback surges.
Yes, replace with tranzorbs. And try to match the characteristics of your new diodes to the old. Higher current devices tend to be slower and the current going the wrong way for that short time may heat things up.
I have seen the turn off time of diodes matched to a circuit. Diodes selected based upon allowing a slight amount of reverse current to slowly die off during voltage reversal. With the end result the diodes turn off 'quietly'. For example, I've seen linear power supplies with faster diodes [the diodes made very high frequency shut off noise] fail conducted RFI/EMI testing caused by this phenomenon.
Robert
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

Maybe you replied to the wrong post (I didn't see or comment on Mycelium's post)? Anyway, usenet certainly supports both those who answer posters' questions directly and those who point out other potential problems.
Yes MOVs eventually wear out, but they are cheap and readily available with high ratings. In this case a few 10s of ms transient of a few 10s of Watts amounts to under a Joule. A 100+ Joule MOV as suggested would last way longer than the relays that this circuit supplies. Transzorbs would be fine too - depends what's on hand or easy to get. If you think about it, it's not so hard to believe that designers could do such a thing; it's often just practical reality.
The original 100 V diodes on a 75 Vrms AC supply? well that's another matter.
I agree with you about the diode speed IF it's a SMPS. But it sounded to me like an ordinary AC mains application, in which case the 35 A bridge is still quite an overkill, but it's just NOT going to cause the problems you mentioned. I assumed it was a part the OP had on hand, which would be fine.
Cheers, Tony
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

Goggle access to Usenet implies a 'string' of responses that tack on end to end. Thus, the appearance I was replying to you.
Good comments, too
Robert
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

If the device you are replacing looks like a big diode, it is a Tranzorb. If it looks like an old TV varistor, it's an MOV, which is what it is a variant of.
If indeed it looks like a diode, you would be better off sourcing the suitable Transzorb and doing the service only one time. With the MOVs, you only get a hundred or so spike suppressions before they are operationally different than when they were made.
http://search.digikey.com/scripts/DkSearch/dksus.dll?Cate6225&keywords=transzorb
Even better:
http://search.digikey.com/scripts/DkSearch/dksus.dll?Cate5429&keywords=transzorb
That is the purpose these devices were made for.
We used them all the time across our FETs and IGBTs right at the lead frame or real close to it on the PCB. They are great for suppressing polarity switching transients with exacting precision.
The other benefit is that you do not have to go back into the assembly and keep repeating the process of replacing the MOVs, which would fail in this application.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
On Wed, 08 Jul 2009 17:34:45 -0700, Mycelium

Fair enough. But how about apples for apples? the transzorbs in your link are rated at only 500W. Sure the common MOVs degrade when hit with their rated 100 kW or so, but if you hit them with only 500 W (like the transzorbs) they will last virtually forever. The physical volume of active material is simply vastly larger in a MOV than in a transzorb. Transzorbs are great if used within their ratings, but MOVs can have vastly higher ratings, and can also provide good lifetime when de-rated in lower energy circuits, so they shouldn't be dismissed without proper consideration.
Cheers, Tony
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

I saw 5kW devices. There were about five ranges I saw.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

They certainly should if the originally designed in devices are transzorbs, which every evidence says they are.
You also should examine the application. There will be a maximum expected spike potential, and energy. The energy component of which is pretty low. So big, monster MOVs is overkill, if you want to look at it that way. Still points toward Transzorbs being the right choice.
Do you need to see it on a Silver Platter?
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
On Fri, 10 Jul 2009 21:38:51 -0700, Mycelium

Not at all. I don't dispute that Transzorbs are a better choice for high repetition low energy spikes - I use them like that wherever possible, and if I needed to go out buy something, I would have used them for the OP's application as well. I only disagreed (and still do) with the suggestions that (paraphrasing): - the original devices were transzorbs (unlikely, given the disc / fin construction), - MOVs wear out fast even when hit infrequently with a tiny fraction of their rating. - using an adequate but imperfect solution is always stupid and never acceptable, - the concept that ng posters always need to be told they're wrong, even when their proposed (and often already implemented) solutions are quite adequate.
Cheers, Tony
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

The joules that kick back will usually be less than or equal to the energy put in, so any spike must be less than or equal to the what went into the motor. That implies that the designer 'could' have taken into account the MOV's ability to handle lesser energy than its full rating for a much longer number of hits.
But then again I've never seen an MOV with fins. Plus, symbol for MOV I'm used to is a long rectangle [like a consumer] and a diagonal line drawn across it with a line drawn parallel to the rectangle at each end of the diagonal line. Tranzorbs I've always seen 'nose to nose' diodes with diagonal short lines off each end of the cathode line [very open 'z'] which sounded like the original symbol.
Robert
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

The schematic symbol I heard described indicated transzorbs, so I consider it pretty likely that is what was designed into the application.
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload
On Tue, 07 Jul 2009 07:48:36 -0700, Robert Macy wrote:

I concur wholeheartedly here. I once worked at a place that built several different models of HV power supplies, to do experiments with electron beams and stuff. They had MOVs all over the place, and the MOVs were a royal pain in the ass.
One day, one of our engineers saw the press release on TransZorbs, and after evaluating them, the company issued EOs to replace every single MOV with a TransZorb. (i.e., the savings over the long-term was worth all the extra work!)
Cheers! Rich
Add pictures here
<% if( /^image/.test(type) ){ %>
<% } %>
<%-name%>
Add image file
Upload

Polytechforum.com is a website by engineers for engineers. It is not affiliated with any of manufacturers or vendors discussed here. All logos and trade names are the property of their respective owners.