Diode Problem

[IASMike]

we are using a Powerex CS241250 diode as a half-wave rectifier to energize a DC Brake. The brake pulls about 9 amps. The diode is rated for 50 amps. The cicuit consists of the diode feed by 120VAC. The DC side has several contacts to control the 53VDC. When the contacts are open we read 53VDC. When we close the contacts the voltage drops to 2.2VDC @ .5Amps. From what we can see the diode should be capable of handling the load. We thought we might of had a circuit problem so we took everything out of the circuit to try it with the same results. Any ideas on what is happening ?

 

[itsmoked]

IASMike,

The halfwave rectifier circuit you describe creates a very strange waveform that typical meters can make no sense of. You must use a "true RMS" meter to get a reading with much significance. So, your readings could be totally bogus.

You mention once contacts close you get a much lower voltage. Can you jumper one contact at a time,(presuming they run different things), and find which load drags the voltage down the most.

Often magnetic devices use variable inductance to reduce the load by up to ten times once they are pulled in. If your brake can't make it all the way in it could present a very large load that drags your voltage down.

 

[Skogsgurra]

Is there a transformer involved? It usually doesn't like the DC load that a single diode represents. Core gets DC magnetized.

 

[OperaHouse]

While readings can be strange 2.2V is a little low. I'd try the circuit with a 100W light bulb ib place of the brake. Try it before the diode too. Even though this is a 1200V diode, if you do not have any snubbing on this brake the kick back voltage could kill the diode. I'd have a second diode across the coil at a minimum and add a RC snubber too.

 

[VEBill]

Putting half-wave AC into a 'DC Brake' (presumably a big solenoid) might not be acceptable. Has this aspect been checked ?

 

[buzzp]

Sounds like the ripple cap is not of sufficient size or has failed.
Typically with so much power requirements a full wave bridge is used to allow more power consumption without pulling down the bus.
So check your ripple cap size and also see if it failed. And I would use a full wave configuration as even if ripple cap is good and of the right size you will probably see the voltage drop when a load is applied simply because you are using a half wave rectifier.

 

[VEBill]

To me, it seemed like he wasn't using a 'ripple cap'.

Using a half-wave rectifier with a 9-Amp load would require one heck of a 'ripple cap'.

 

[itsmoked]

VE1BLL you are correct the cap would be a doozy. They usually don't use caps in these circuits. This is very common, the one diode "gimmick". I see this a lot and usually the diode is smaller 5A kinda sizes.
Opera the Light bulb is a great idea! You use that a lot don't you?!

 

[OperaHouse]

With a lamp, what you see is what you get. I have a friend at Lamp division and he always smacked me in the head when I called it a bulb. I got a suspicion that there is a source resistance problem. That would be an unusual diode failure mode.

 

[buzzp]

Then it definately requires a full wave rectifier. Your not going to pull 9A from a half wave rectified signal without dragging down the bus some, sorry just won't happen, especially since its a solenoid which can pull up to 20 times continuous current rating upon energization. Hope the solenoid coil is not rated for 53VDC cause it will never see it with a half wave on the front end and a 9A load.

Is this an existing circuit or something you have designed and are testing?

 

[buzzp]

If you have not missed the obvious such as short circuits and such and this is a new design, better put in a full wave recitfier circuit (assumed no shorts cause you said everything was disconnected and still pulled down but then how would anything get pulled down when everything is disconnected?).

Also where was the voltage and current measured? Since you have a current measurement in the mix, I assume this is the current through the diode? I hope the 2.2 is not the voltage across the diode because this is not so abnormal (lil high but dont know the material your diode is made from).

 

[elfgriper]

IASMike,
I think looking a this signal (at brake) with a oscope will show a large negative voltage as the diode switches off and the Magnetic field collaspes. I don't think you can just connect a std meter.
-elf

 

[itsmoked]

Opera; I like those "lamp" solutions because it does show high impedance problems. So many troubleshooting escapades have gone wrong because of the new high impedance DMM we all use. "There's voltage here!!" Yeah, at 1 uA.

I would go for much smaller diodes. I would use the following diode bridge which will do the job fine, scratch that, BETTER and cost FARRRRRR less.

http://rocky.digikey.com/WebLib/General Semi/Web Data/GBPC12, 15, 25 and 35.pdf

Then add one of these diodes across the DC posts on the bridge pointed backwards, that is, not conducting normally.

http://rocky.digikey.com/WebLib/General Semi/Web Data/1N5400 - 1N5408.pdf

Good luck.

 

[Warpspeed]

That brake solenoid was probably originally designed for smooth dc operation, probably with a solid steel or iron core.

A half wave rectifier is going to be putting 170 volt peaks into that coil. Maybe it is saturating, or maybe the eddy current in the core are causing some fairly dramatic current spikes at peak voltage.

That is a horrible circuit. As others have said, first use a transformer to reduce the voltage, and use a full wave rectifier. If it still will not behave, add some series inductance after the rectifier in the form of a properly designed choke with laminated core and air-gap, it need not be large.

It will work much better, and everything will be under far less electrical stress.

 

[buzzp]

Thanks, Warpspeed, someone who finally sees the light! Although, I dont think the spikes are harming a thing, just simply not enough power available.

 

[IASMike]

Thanks Everyone!

Our biggest problem is we need to conserve power. So we
decided to go with two SCRs to release the brake and stop firing one SCR after the brake is released. The voltages that I gave you were taken with a True RMS meter referenced to ground and we did look at it with a scope. The scope said the voltage was being pulled down. One thing we still do not quite understand is that we we put in a 40A SCR that someone had blown the gate on so now it acts like a diode, is that it would energize the brake.

 

[Warpspeed]

On rather intriguing thought has just occurred to me Mike.

If you have one diode and a resistive load, you then get the classic text book half wave rectified waveform.

BUT with a purely inductive (no loss)load, there will be the half sinusoid forward voltage sure enough, followed during the other half cycle with a back EMF across the inductive load. And quite possibly no net measurable dc!

To fix this you need to place a second flywheel diode directly across the load. It will still be just a half wave rectifier, but the flywheel diode would be required with an inductive load for the rectifier to work properly.

If diodes are fairly foolproof, SCRs most certainly are not. I think that looking at it on an oscilloscope may at some stage be well worth doing, especially if SCRs are planned with an inductive load.

It always seem to be the simplest circuits that can cause the most grief. At least it is like that for me.

 

[itsmoked]

[It always seem to be the simplest circuits that can cause the most grief.]

A-men brother!