Diodes Vs. Bridge Rectifiers

[Usman]

I have efficiency concerns on using diodes (set of six for full bridge wave rectification).

Over the years, I have noticed many diode failures due to advent of low quality diodes in the market. Most of these are over-rated and heat-up upon 50% rated amps. Even if these were from approved manufacturers, I seem to have  an uncomfortable feel for their efficiency ratings.

1. -Firstly, I would like to learn from the board of any comparisons data available on these diodes and ready made Bridge Rectifiers, and if Bridge Rectifiers are any better in terms of efficiency and reliability.
   
2. -Secdonly, do these diodes and/or rectifiers, or either one of them, have any problem with delta wired alternators? I have heard someone said that these diodes and/or rectifiers have harmonic disorders of some sort! Not exactly sure what does that mean.
   
   

Any advice would be appreciated.

Thanks.

I think he needs to do his own frigging research and stop sitting around thinking up questions to ask. Then he would know bridges are just combinations of diodes.

[topspeed]

Which of these is better?

[kitestrings]

Hi Usman,

I've always used stud-mounted diode rectifiers.  Jim Sencenbaugh used them on his turbines, and Flux had suggested that they were more robust than potted bridges.  That was enough for me.  The recommendations from this board were to oversize them, and have ample cooling.  That seems to be a consistent theme with other devices, like SS relays as an example.  Heat is the enemy.

I have no experience with running a delta-wired alternator, so I can't speak to that.  There is one other advantage of diodes IMO, and that is that a lightning strike can take out one diode - you replace it and you're back up.  If you smoke a potted bridge, it's all or nothing.

~ks

[Adriaan Kragten]

Which of these is better?

Rectification of a 3-phase current is explained in my public report KD 340. You need minimal six diodes to rectify a 3-phase current. You can use six separate diodes, three blocks with two diodes, or one 3-phase rectifier block which contains six diodes. If you use six separate diodes, three diodes are pressed (or bolted) in an aluminium strip for cooling and three diodes are pressed in another aluminium strip. Both strips must be isolated from each other. Separate diodes are available for two different directions of the current. One needs three diodes of one type and three diodes of the other type. 

The maximum current which can flow through a square 3-phase rectifier block is normally given for a housing temperature of 20 degrees C. At maximum current, the temperature normally becomes much higher, even if a good cooling area is realised. The maximum allowable current is therefore much lower than the nominal current which is specified. The voltage drop over a 3-phase rectifier block with silicon diodes is about 1.4 V so the power dissipation is I * 1.4 W. So if you buy a 25 A block, the heat dissipation in this block is about 35 W at I = 25 A but the temperature at 35 W will become much higher than 20 degrees C. Therefore the maximum practical current is only about 15 A and even for 15 A, you need a heat sink with a rather large cooling area.

One can use a 1-phase rectifier block which contains four diodes and a second 1-phase rectifier block for which only two diodes are used. But two 1-phase blocks of a certain maximum current are normally more expensive that one 3-phase block of the same maximum current. The most heat will be dissipated in the block to which two of the three phases are connected.

One can also use three 1-phase blocks. Every block contains four diodes and it is no good idea to use only two of these four diodes. Therefore the two AC terminals of a 1-phase block are connected to each other. This makes that two diodes are connected in parallel and so the current through one diode is halved. Every phase is connected to a combined AC contact of one rectifier block. All + terminals are connected to each other and all - terminals are connected to each other. The total heat dissipation in the rectifier is the same as if only two of the four diodes are used but this heat is now spread over four diodes and this results in a lower local temperature at each diode. This option is rather expensive if compared to one 3-phase rectifier but if only small square rectifier blocks are available, a much higher maximum current is allowed.

The normal square 3-phase blocks have only a rather low maximum nominal current of about 25 A but one should also choose a type for which the allowable voltage is high enough. If the maximum current is large, one can better use three blocks with two big diodes because these blocks are available for much higher currents.

Different types of separate diodes, large blocks with two diodes, 1-phase blocks and 3-phase blocks can be found in for instance the Farnell catalogue.

[bigrockcandymountain]

I use a chinese no name potted 3 phase rectifier that says 120a on it.  It is bolted to a large heat sink with thermal grease.  At 40a dc output it makes a fair amount of heat.  I usually run max 30a now, so i feel ok about it.

I believe stud mounted diodes would be more robust, but they are a fair bit harder to work with. 

[OperaHouse]

Pretty much any bridge rectifier produces the same heat even if you go for much higher current.  At about 1 1/2 volts for two rectifiers, 40A produces well over 50W.  Higher current rectifiers mean a bigger die which reduces junction temperature.  Junction temperature is much higher than the heat sink temp. This is the reason for using half bridge instead of full bridge as only one forward voltage drop is seen in the package.

[Adriaan Kragten]

This is the reason for using half bridge instead of full bridge as only one forward voltage drop is seen in the package.

To my opinion this isn't correct if you use three 1-phase bridge rectifiers to rectify a 3-phase current in star. If you use only two of the four diodes of a 1-phase bridge rectifier, the whole current I flows through two diodes which are connected in series. The voltage drop over one diode is almost independent of the current and is about 0.7 V for silicon diodes (except for very small currents it is less). So the voltage drop over two diodes connected in series is about 1.4 V. So the power loss over two diodes is about I * U = I * 1.4 W.

If the AC tags are connected to each other, two times two diodes are connected in parallel and so the current through each diode is 1/2 I. But the voltage drop over each diode is the same. So now the total power loss is 2 * 1/2 * I * 1.4 W = I * 1.4 W which is the same. However, now the power loss at each individual diode is half and this results in a lower junction temperature for each diode. So it isn't logic to advice to use only two of the four diodes of a 1-phase bridge rectifier.

The diode which is guiding a current depends on the phases which have the highest and the lowest voltage. Every phase angle of 60 degrees one diode is changing from guiding a current to guiding no current. A coil of the generator winding is guiding a current for only 2/3 of the time. A certain diode is guiding a current for only 1/3 of the time.

The wiring given on the second photo of Topspeed with three rectifier blocks is the best option. You can see that the two AC tags are connected to each other so all four diodes in one rectifier block are used.

[Adriaan Kragten]

I will try to explain why the current flows only during 1/3 of the time through a diode and why it flows during 2/3 of the time through a stator coil. Figure 3 of KD 340 gives a picture of a sinusoidal 3-phase voltage variation. A certain phase has the highest voltage for a phase angle of 120 degrees and another phase has the lowest voltage for a phase angle of 120 degrees but the situation changes every 60 degrees. Figure 5 of KD 340 gives the wire diagram for 3-phase rectification in star.

The rectifier contains six diodes. The diodes D1, D2 and D3 are connected to the plus tag at the outside of the rectifier block. The diodes D4, D5 and D6 are connected to the min tag at the outside of the rectifier block.
D1 is connected in series to D4 and the connecting wire is the AC tag which is connected to phase U.
D2 is connected in series to D5 and the connecting wire is the AC tag which is connected to phase V.
D3 is connected in series to D6 and the connecting wire is the AC tag which is connected to phase W.

There are six paths how the current can flow and for every path the current flows during a phase angle of 60 degrees. These six paths are:

1) D4 - U - W - D3
2) D4 - U - V - D2
3) D5 - V - W - D3
4) D5 - V - U - D1
5) D6 - W - U - D1
6) D6 - W - V - D2

So in this table it can be seen that a certain diode is mentioned only two times which means during a total phase angle of 120 degrees, so during 1/3 of the time. A certain phase is mentioned four times which means during a total phase angle of 240 degrees, so during 2/3 of the time. The reason that a phase is active during the double time as a diode is that the current in a phase can flow in two directions but in a diode it can flow only in one direction. In the table it can also be seen that a current never flows through two diodes which are connected in series, like D4 and D1, at the same time. However, for calculation of the heat losses, it is allowed to assume that the DC current flows always through two diodes which are connected in series.

The DC current which comes finally out of the rectifier has the same value as the current which is flowing through two coils but the DC current flows all the time. The DC current has a small fluctuation which is shown in figure 8 of KD 340.

[Jerrywoo]

Not meaning to hijack the thread but since you are discussing diodes verses a rectifier..I have a turbine I built from a treadmill motor and I used a 50 amp battery isolater off of a motorhome because I didn't have access to blocking diodes..I plan on replacing the isolator with proper blocking diodes.. would a rectifier work better?