Blocking Diodes

[WastefulMiser]

Is it advised to put blocking diodes on both the postive and negative in lieu just on just the postive for power supplies (24VDC)?

 

[cadpoint]

Can you please describe what's going on, why you need these diodes?

I this for visual signaling on circuit or that the 24 volts is still toastie ? :grin:

If you put the diode in parallel than you'd only need one ((right outside the line side of the trannie - as I recall)), as you stated it you'd be in series per pole thus two would be required...

 

[cadpoint]

Frankly the load side is where you want to put one or two...

 

[gar]

100926-1830 EST

WastefulMiser:

Please provide a clear description of your circuit and what these diodes are supposed to do.

You have a 24 VDC supply. Is either + or minus - defined as common and which if either one is connected to the metal enclosure of the equipment, and/or earth (ground)?

What does the power supply supply?

Is the output of the power supply switched with a switch or relay contacts?

Is the load inductive?

Are you trying to prevent or reduce transient noise?

Are you trying to protect the power supply in some way?

.

 

[dereckbc]

When I hear the term BLOCKING DIODE makes me think of solar PV systems where they are used to prevent batteries from back feeding solar panels at night causing the batteries to discharge through the panels. However with charge controllers the Blocking Diode is antiquated.

So yeah we need a system description and function served. Are you maybe referring to Steering Diodes?

 

[ron]

This is the way we do it for generator starting batteries.
http://www.sens-usa.com/upload/fmproduct_right_menua/Appn_Note_111028A_397.pdf

 

[WastefulMiser]

Thank you all for the replies.

Sorry for the delay. Hopefully the following clears things up:

100926-1830 EST

WastefulMiser:

Please provide a clear description of your circuit and what these diodes are supposed to do.

Parallaling (4) 500W, 48-24VDC, ungrounded power supplies.

You have a 24 VDC supply. Is either + or minus - defined as common and which if either one is connected to the metal enclosure of the equipment, and/or earth (ground)?

Neither the + or - are grounded. I do not know if the manufacture uses "conventional" electron flow: + to -; or if they use actual electron flow: - to +. I hate to assume, but I believe most products use conventional electron flow theory as that's what is taught.

What does the power supply supply?

PLC inputs and outputs.

Is the output of the power supply switched with a switch or relay contacts?

I imagine relay contacts -- the PLC is probably controlling VFD's for some process. The PLC is not in my scope actually nor is the redundant power supplies, I am just curious. Mayhap you could clarify on the difference between switch and relay contacts.

Is the load inductive?

Is there a concern with induction for DC? Nevertheless, I would say no.

Are you trying to prevent or reduce transient noise?

I do not know the actual intent the end-user has for the blocking diodes. The spec isn't in front of me, but it says blocking diodes on the secondary of the power supplies. I guess they are worried about back-feeding the 48VDC circuit. I assume most power supplies have blocking diodes on the secondary internally.

Are you trying to protect the power supply in some way?

Yes. I think that is the intent.

I guess my two questions I have are:

1. Is it recommended to put a diode on both the + and -? Or is that a waste of a diode? Any benefit to it?
2. Load sharing. How critical is it for such small loads? Would the power supplies balance themselves out evenly?

 

[gar]

100926-2229 EST

WastefulMiser:

Since the application is PLC inputs I would expect that moderately well filtered DC would be desired. In this application 10% ripple would probably cause no problem. But you might actually have regulated supplies, and there may not be any good reason for precision regulation.

What is a 48-24 V supply mean? Is it two 24 V supplies in series with the center point grounded and thus providing +/- 24 V relative to common (ground)? Probably not really important whether these have some point grounded. However, I would be inclined to provide a ground reference.

You do not need a diode on both + and - of each supply for isolation unless this is a +/- supply relative to common. I would think it is unlikely a that +/- supply would be required for PLC inputs. I would put the diodes on the non-commoned side of the supplies.

If the supply type is an unregulated rectifier -- capacitor input filter, then there is no need for the isolating diodes unless the intent is to provide automatic isolation if one or more of the rectifying diodes shorted. But instead this might be handled by a fuse. The current distribution from each supply will depend upon each supply's internal voltage and impedance and the load current.

If the supplies are regulated, then there might be more reason for isolation diodes.

Are these supplies also supplying outputs of the PLC? Seems like a lot of power just for inputs.

Based on these comments what new questions can you ask to further clarify the needs of the circuit?

Virtually all electrical engineers use positive current flow as the convention. When you look at the symbol for a diode the arrow effect of the symbol points in the direction of positive current flow for low voltage drop (the forward biased direction). The primary type device where electrons actually flow is in a vacuum tube, but here one still uses the positive current flow convention for circuit analysis. Only when you are concerned with the inside of the tube do you talk about electron flow.

.

 

[dbuckley]

I'm reading this as 48VDC in, 24VDC out PSUs, such as you would find in a telecom facility where 48VDC is pervasive.

Assuming that the outputs of the power supplies are truly floating, then just a single diode per PSU will do fine.

Please do not assume that the presence of steering diods will make these power supplies loadshare, giving you 2KW (or n+1 1.5KW). The combined safe rating of the four units is still going to be 500W, but you will have a lot of resilience (n+3). The reason they won't loadshare is that the PSUs will all have slightly different output voltages, and thus one steering diode will conduct, and the other three will be reverse biased and thus not conduct.

 

[dereckbc]

I'm reading this as 48VDC in, 24VDC out PSUs, such as you would find in a telecom facility where 48VDC is pervasive.

Where are you picking that up at? I still have figured out what the OP has, he has not clarified.

 

[rcwilson]

This is a common application for critical equipment like plant control systems that require reliable power or for UPS logic power supplies.

Two 24V power supples, each rated for 100%+ of the load are fed from separate sources. Diodes on the output of each PS isolate the common DC load bus from the power supply. The power supplies might or might not share the load, but when one fails or has an internal short, the other one can still supply control power.

Diodes are needed on both positive and negative for reliable isolation from a faulted power supply.

 

[hurk27]

Not sure if this will apply, but back when I messed with DC powered RF amps being used on line voltage AC some of these amps would exceed the available ratings of power supply's we could get at a decent price, so we would gang (parallel) several together to reach our input ratting of the amp. we did this by free wheeling diodes and biasing resistors the diodes prevented voltage feed back into the other power supply's which would cause the voltage regulators to go nuts on us other wise, and the bias resistors would limit each power supply to its rated capacity, one 2kw 10 meter amp I remember, took 10, 25 amp 13.6 volt power supply's to run, trying to find a 13.6 volt x 250 amp power supply would have cost a fortune, I eventually started making my own from buck boost transformers and, old battery chargers, using TO-66 transistors for voltage regulation finals, but still costly, this still had its own problems of getting the voltages balanced equal across all power supply's, but with the biasing resistors it wasn't too bad, the resistors were made out of stock tank coil (Ni-chrome wire) I had in stock for making my own output tank coils.

So the only thing I can see the use of the free wheeling diodes is for is to prevent the voltage back feed into the other power supply's which would send these other power supply trying to regulate the voltages of the other power supply's. which like I said above will make them go nuts lol.

 

[dbuckley]

Where are you picking that up at? I still have figured out what the OP has, he has not clarified.

From this statement:

Parallaling (4) 500W, 48-24VDC, ungrounded power supplies

I deduce that the power supplies are 48V in, 24V out, and that the outputs are floating. I may be wrong, but from whats written, that's the closest sane information I can extract.

Diodes are needed on both positive and negative for reliable isolation from a faulted power supply.

For floating outputs, although diodes in both +ve and -ve wont do much harm (other than dissipating 15W of extra heat over just one diode), they shouldn't be necessary, as shorts to ground are not common power supply failure conditions.

 

[adelle]

Note too that each silicon diode will produce a 0.7v drop. And yes, with enough current that will make some heat so heatsinks may be needed.

Schottkey (sp?) diodes only have a 0.3v drop but may not be available with the current ratings you need.

 

[WastefulMiser]

I'm reading this as 48VDC in, 24VDC out PSUs, such as you would find in a telecom facility where 48VDC is pervasive.

Assuming that the outputs of the power supplies are truly floating, then just a single diode per PSU will do fine.

Please do not assume that the presence of steering diods will make these power supplies loadshare, giving you 2KW (or n+1 1.5KW). The combined safe rating of the four units is still going to be 500W, but you will have a lot of resilience (n+3). The reason they won't loadshare is that the PSUs will all have slightly different output voltages, and thus one steering diode will conduct, and the other three will be reverse biased and thus not conduct.

I must not be that incoherent, but you are correct. 48VDC IN ---> 24VDC OUT.

I tried uploading a .PDF yesterday, now I'll try to attach a .JPEG. Nope, can't, it has to be a URL. And trying to upload something repeatedly fails. Must be a firewall or some such.

I didn't think adding steering diodes meant load sharing. I was just wondering if selecting a power supply with load sharing communication interconnect between the power supplies was applicable.

I don't understand why/how slight different voltages of the power supplies would impact the additive property of the power supplies when paralleled. I would agree there could be an imbalance in loading of the power supplies if load sharing is not used.

 

[WastefulMiser]

This is a common application for critical equipment like plant control systems that require reliable power or for UPS logic power supplies.

Two 24V power supples, each rated for 100%+ of the load are fed from separate sources. Diodes on the output of each PS isolate the common DC load bus from the power supply. The power supplies might or might not share the load, but when one fails or has an internal short, the other one can still supply control power.

Diodes are needed on both positive and negative for reliable isolation from a faulted power supply.

Would the diode point toward the -ve or away from the -ve for isolation? Ah, toward the -ve terminal of the power supply.

As mentioned in a below post, how do you compensate for the 1.4VDC voltage drop across the diodes?

 

[nollij]

As mentioned in a below post, how do you compensate for the 1.4VDC voltage drop across the diodes?

The ones used at our plant are adjustable within a range. Others may or may not be.

Diodes are needed on both positive and negative for reliable isolation from a faulted power supply.

The redundant DC supplies used here have diodes only on the positive terminals. Downstream of the diodes the positives of the two supplies and the negatives are connected together when they reach the equipment. These systems are ungrounded. How would adding the diode to the negative terminals affect the isolation between the supplies?

 

[rcwilson]

Making the negative bus common is only an issue if one power supply has an internal ground fault and another ground occurs on the system. Isolation diodes on just the positive outputs will protect against 98% of the problems. We do both + & - because the cost is low compared to the $1M + cost of our control systems. Most systems don't need that level of reliability.

 

[nollij]

Making the negative bus common is only an issue if one power supply has an internal ground fault and another ground occurs on the system. Isolation diodes on just the positive outputs will protect against 98% of the problems. We do both + & - because the cost is low compared to the $1M + cost of our control systems. Most systems don't need that level of reliability.

I just made a doodle of the ground fault path and I see what you are saying.

I believe our control systems are in the same category. If this system loses power then a very vital part of the plant trips and everything comes tumbling down.

 

[dbuckley]

I didn't think adding steering diodes meant load sharing. I was just wondering if selecting a power supply with load sharing communication interconnect between the power supplies was applicable.

If you have load-share capable power supplies, and then use steering diodes on the outputs, then (in general, there may be exceptions, but I've never seen one) the power supplies can no longer load-share. For load share capable PSUs to load share, their outputs must be directly connected together.

I don't understand why/how slight different voltages of the power supplies would impact the additive property of the power supplies when paralleled. I would agree there could be an imbalance in loading of the power supplies if load sharing is not used.

This is because of the properties of a diode. A diode will only conduct when it is "forward biassed", which means that the anode is more positive than the cathode, and that voltage differential exceeds (silicon) 0.7v.

Which ever diode has the highest voltage source will reverse bias all the other diodes, so only one conducts.

This mechanism ensures that if you have two sources of power and a diode steering mechanism, the source with the higher voltage wins. Its a very neat and reliable answer to high availability power, fire and burglar alarm panels use steering diodes for just this purpose; if the mains is on, the power comes from the mains supply. If not, then the next source of power is the standby battery.